fix: endpoints for single local disks must retain port info (#15585 )

upgrade golang-lint to the latest (#15600 )
fix: .golangci.yml key 'linters-settings' duplicate entry (#15589 )
2026-08-05 15:12:01 +03:00 · 2022-08-26 12:53:15 -07:00 · 2022-08-26 12:52:29 -07:00 · 2022-08-25 23:06:06 -07:00 · 2022-08-25 13:07:15 -07:00 · 2022-08-25 11:30:52 -07:00
357 changed files with 13281 additions and 6596 deletions
@@ -0,0 +1,14 @@
+name: 'Dependency Review'
+on: [pull_request]
+
+permissions:
+  contents: read
+
+jobs:
+  dependency-review:
+    runs-on: ubuntu-latest
+    steps:
+      - name: 'Checkout Repository'
+        uses: actions/checkout@v3
+      - name: 'Dependency Review'
+        uses: actions/dependency-review-action@v1
@@ -20,7 +20,7 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        go-version: [1.17.x, 1.18.x]
+        go-version: [1.18.x, 1.19.x]
        os: [ubuntu-latest]
    steps:
      - uses: actions/checkout@629c2de402a417ea7690ca6ce3f33229e27606a5 # v2
@@ -20,7 +20,7 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        go-version: [1.17.11b7, 1.18.3b7]
+        go-version: [1.18.5b7]
        os: [ubuntu-latest]
    steps:
      - uses: actions/checkout@v2
@@ -20,7 +20,7 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        go-version: [1.17.x]
+        go-version: [1.18.x, 1.19.x]
        os: [ubuntu-latest]
    steps:
      - uses: actions/checkout@v2
@@ -48,3 +48,5 @@ jobs:
          sudo sysctl net.ipv6.conf.default.disable_ipv6=0
          make verify-healing
          make verify-healing-inconsistent-versions
+          make verify-healing-with-root-disks
+          make verify-healing-with-rewrite
@@ -20,7 +20,7 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        go-version: [1.17.x, 1.18.x]
+        go-version: [1.18.x, 1.19.x]
        os: [ubuntu-latest, windows-latest]
    steps:
      - uses: actions/checkout@v2
@@ -20,7 +20,7 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        go-version: [1.17.x, 1.18.x]
+        go-version: [1.18.x, 1.19.x]
        os: [ubuntu-latest]
    steps:
      - uses: actions/checkout@v2
@@ -61,7 +61,7 @@ jobs:
      # are turned off - i.e. if ldap="", then ldap server is not enabled for
      # the tests.
      matrix:
-        go-version: [1.17.x]
+        go-version: [1.18.x]
        ldap: ["", "localhost:389"]
        etcd: ["", "http://localhost:2379"]
        openid: ["", "http://127.0.0.1:5556/dex"]
@@ -27,4 +27,4 @@ jobs:
        npm install -g markdownlint-cli
        markdownlint --fix '**/*.md' \
          --config /home/runner/work/minio/minio/.github/markdown-lint-cfg.yaml \
-          --disable MD013 MD040
+          --disable MD013 MD040 MD051
@@ -21,7 +21,7 @@ jobs:

    strategy:
      matrix:
-        go-version: [1.17.x]
+        go-version: [1.18.x, 1.19.x]

    steps:
      - uses: actions/checkout@v2
@@ -20,7 +20,7 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        go-version: [1.17.x]
+        go-version: [1.18.x, 1.19.x]
        os: [ubuntu-latest]

    steps:
@@ -2,6 +2,13 @@ linters-settings:
  golint:
    min-confidence: 0

+  gofumpt:
+    lang-version: "1.18"
+
+    # Choose whether or not to use the extra rules that are disabled
+    # by default
+    extra-rules: false
+
  misspell:
    locale: US

@@ -14,28 +21,15 @@ linters:
    - govet
    - revive
    - ineffassign
-    - gosimple
-    - deadcode
-    - structcheck
    - gomodguard
    - gofmt
-    - unused
-    - structcheck
    - unconvert
-    - varcheck
+    - unused
    - gocritic
    - gofumpt
    - tenv
    - durationcheck

-linters-settings:
-  gofumpt:
-    lang-version: "1.17"
-
-    # Choose whether or not to use the extra rules that are disabled
-    # by default
-    extra-rules: false
-
 issues:
  exclude-use-default: false
  exclude:
@@ -2,3 +2,5 @@ CVE-2020-26160
 CVE-2020-15136
 CVE-2020-15115
 CVE-2020-15114
+CVE-2022-2835
+CVE-2022-2837
@@ -2741,186 +2741,6 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

 ================================================================

-github.com/briandowns/spinner
-https://github.com/briandowns/spinner
----------------------------------------------------------------
-                                 Apache License
-                           Version 2.0, January 2004
-                        http://www.apache.org/licenses/
-
-   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
-
-   1. Definitions.
-
-      "License" shall mean the terms and conditions for use, reproduction,
-      and distribution as defined by Sections 1 through 9 of this document.
-
-      "Licensor" shall mean the copyright owner or entity authorized by
-      the copyright owner that is granting the License.
-
-      "Legal Entity" shall mean the union of the acting entity and all
-      other entities that control, are controlled by, or are under common
-      control with that entity. For the purposes of this definition,
-      "control" means (i) the power, direct or indirect, to cause the
-      direction or management of such entity, whether by contract or
-      otherwise, or (ii) ownership of fifty percent (50%) or more of the
-      outstanding shares, or (iii) beneficial ownership of such entity.
-
-      "You" (or "Your") shall mean an individual or Legal Entity
-      exercising permissions granted by this License.
-
-      "Source" form shall mean the preferred form for making modifications,
-      including but not limited to software source code, documentation
-      source, and configuration files.
-
-      "Object" form shall mean any form resulting from mechanical
-      transformation or translation of a Source form, including but
-      not limited to compiled object code, generated documentation,
-      and conversions to other media types.
-
-      "Work" shall mean the work of authorship, whether in Source or
-      Object form, made available under the License, as indicated by a
-      copyright notice that is included in or attached to the work
-      (an example is provided in the Appendix below).
-
-      "Derivative Works" shall mean any work, whether in Source or Object
-      form, that is based on (or derived from) the Work and for which the
-      editorial revisions, annotations, elaborations, or other modifications
-      represent, as a whole, an original work of authorship. For the purposes
-      of this License, Derivative Works shall not include works that remain
-      separable from, or merely link (or bind by name) to the interfaces of,
-      the Work and Derivative Works thereof.
-
-      "Contribution" shall mean any work of authorship, including
-      the original version of the Work and any modifications or additions
-      to that Work or Derivative Works thereof, that is intentionally
-      submitted to Licensor for inclusion in the Work by the copyright owner
-      or by an individual or Legal Entity authorized to submit on behalf of
-      the copyright owner. For the purposes of this definition, "submitted"
-      means any form of electronic, verbal, or written communication sent
-      to the Licensor or its representatives, including but not limited to
-      communication on electronic mailing lists, source code control systems,
-      and issue tracking systems that are managed by, or on behalf of, the
-      Licensor for the purpose of discussing and improving the Work, but
-      excluding communication that is conspicuously marked or otherwise
-      designated in writing by the copyright owner as "Not a Contribution."
-
-      "Contributor" shall mean Licensor and any individual or Legal Entity
-      on behalf of whom a Contribution has been received by Licensor and
-      subsequently incorporated within the Work.
-
-   2. Grant of Copyright License. Subject to the terms and conditions of
-      this License, each Contributor hereby grants to You a perpetual,
-      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
-      copyright license to reproduce, prepare Derivative Works of,
-      publicly display, publicly perform, sublicense, and distribute the
-      Work and such Derivative Works in Source or Object form.
-
-   3. Grant of Patent License. Subject to the terms and conditions of
-      this License, each Contributor hereby grants to You a perpetual,
-      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
-      (except as stated in this section) patent license to make, have made,
-      use, offer to sell, sell, import, and otherwise transfer the Work,
-      where such license applies only to those patent claims licensable
-      by such Contributor that are necessarily infringed by their
-      Contribution(s) alone or by combination of their Contribution(s)
-      with the Work to which such Contribution(s) was submitted. If You
-      institute patent litigation against any entity (including a
-      cross-claim or counterclaim in a lawsuit) alleging that the Work
-      or a Contribution incorporated within the Work constitutes direct
-      or contributory patent infringement, then any patent licenses
-      granted to You under this License for that Work shall terminate
-      as of the date such litigation is filed.
-
-   4. Redistribution. You may reproduce and distribute copies of the
-      Work or Derivative Works thereof in any medium, with or without
-      modifications, and in Source or Object form, provided that You
-      meet the following conditions:
-
-      (a) You must give any other recipients of the Work or
-          Derivative Works a copy of this License; and
-
-      (b) You must cause any modified files to carry prominent notices
-          stating that You changed the files; and
-
-      (c) You must retain, in the Source form of any Derivative Works
-          that You distribute, all copyright, patent, trademark, and
-          attribution notices from the Source form of the Work,
-          excluding those notices that do not pertain to any part of
-          the Derivative Works; and
-
-      (d) If the Work includes a "NOTICE" text file as part of its
-          distribution, then any Derivative Works that You distribute must
-          include a readable copy of the attribution notices contained
-          within such NOTICE file, excluding those notices that do not
-          pertain to any part of the Derivative Works, in at least one
-          of the following places: within a NOTICE text file distributed
-          as part of the Derivative Works; within the Source form or
-          documentation, if provided along with the Derivative Works; or,
-          within a display generated by the Derivative Works, if and
-          wherever such third-party notices normally appear. The contents
-          of the NOTICE file are for informational purposes only and
-          do not modify the License. You may add Your own attribution
-          notices within Derivative Works that You distribute, alongside
-          or as an addendum to the NOTICE text from the Work, provided
-          that such additional attribution notices cannot be construed
-          as modifying the License.
-
-      You may add Your own copyright statement to Your modifications and
-      may provide additional or different license terms and conditions
-      for use, reproduction, or distribution of Your modifications, or
-      for any such Derivative Works as a whole, provided Your use,
-      reproduction, and distribution of the Work otherwise complies with
-      the conditions stated in this License.
-
-   5. Submission of Contributions. Unless You explicitly state otherwise,
-      any Contribution intentionally submitted for inclusion in the Work
-      by You to the Licensor shall be under the terms and conditions of
-      this License, without any additional terms or conditions.
-      Notwithstanding the above, nothing herein shall supersede or modify
-      the terms of any separate license agreement you may have executed
-      with Licensor regarding such Contributions.
-
-   6. Trademarks. This License does not grant permission to use the trade
-      names, trademarks, service marks, or product names of the Licensor,
-      except as required for reasonable and customary use in describing the
-      origin of the Work and reproducing the content of the NOTICE file.
-
-   7. Disclaimer of Warranty. Unless required by applicable law or
-      agreed to in writing, Licensor provides the Work (and each
-      Contributor provides its Contributions) on an "AS IS" BASIS,
-      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
-      implied, including, without limitation, any warranties or conditions
-      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
-      PARTICULAR PURPOSE. You are solely responsible for determining the
-      appropriateness of using or redistributing the Work and assume any
-      risks associated with Your exercise of permissions under this License.
-
-   8. Limitation of Liability. In no event and under no legal theory,
-      whether in tort (including negligence), contract, or otherwise,
-      unless required by applicable law (such as deliberate and grossly
-      negligent acts) or agreed to in writing, shall any Contributor be
-      liable to You for damages, including any direct, indirect, special,
-      incidental, or consequential damages of any character arising as a
-      result of this License or out of the use or inability to use the
-      Work (including but not limited to damages for loss of goodwill,
-      work stoppage, computer failure or malfunction, or any and all
-      other commercial damages or losses), even if such Contributor
-      has been advised of the possibility of such damages.
-
-   9. Accepting Warranty or Additional Liability. While redistributing
-      the Work or Derivative Works thereof, You may choose to offer,
-      and charge a fee for, acceptance of support, warranty, indemnity,
-      or other liability obligations and/or rights consistent with this
-      License. However, in accepting such obligations, You may act only
-      on Your own behalf and on Your sole responsibility, not on behalf
-      of any other Contributor, and only if You agree to indemnify,
-      defend, and hold each Contributor harmless for any liability
-      incurred by, or claims asserted against, such Contributor by reason
-      of your accepting any such warranty or additional liability.
-
-================================================================
-
 github.com/buger/jsonparser
 https://github.com/buger/jsonparser
 ----------------------------------------------------------------
@@ -4885,26 +4705,300 @@ SOFTWARE.
 github.com/eclipse/paho.mqtt.golang
 https://github.com/eclipse/paho.mqtt.golang
 ----------------------------------------------------------------
-This project is dual licensed under the Eclipse Public License 1.0 and the
-Eclipse Distribution License 1.0 as described in the epl-v10 and edl-v10 files.
+Eclipse Public License - v 2.0 (EPL-2.0)

-The EDL is copied below in order to pass the pkg.go.dev license check (https://pkg.go.dev/license-policy).
+This program and the accompanying materials
+are made available under the terms of the Eclipse Public License v2.0
+and Eclipse Distribution License v1.0 which accompany this distribution.
+
+The Eclipse Public License is available at
+  https://www.eclipse.org/legal/epl-2.0/
+and the Eclipse Distribution License is available at
+  http://www.eclipse.org/org/documents/edl-v10.php.
+
+For an explanation of what dual-licensing means to you, see:
+https://www.eclipse.org/legal/eplfaq.php#DUALLIC

 ****
-Eclipse Distribution License - v 1.0
+The epl-2.0 is copied below in order to pass the pkg.go.dev license check (https://pkg.go.dev/license-policy).
+****
+Eclipse Public License - v 2.0

-Copyright (c) 2007, Eclipse Foundation, Inc. and its licensors.
+    THE ACCOMPANYING PROGRAM IS PROVIDED UNDER THE TERMS OF THIS ECLIPSE
+    PUBLIC LICENSE ("AGREEMENT"). ANY USE, REPRODUCTION OR DISTRIBUTION
+    OF THE PROGRAM CONSTITUTES RECIPIENT'S ACCEPTANCE OF THIS AGREEMENT.

-All rights reserved.
+1. DEFINITIONS

-Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
+"Contribution" means:

-    Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
-    Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
-    Neither the name of the Eclipse Foundation, Inc. nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
+  a) in the case of the initial Contributor, the initial content
+     Distributed under this Agreement, and

-THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+  b) in the case of each subsequent Contributor:
+     i) changes to the Program, and
+     ii) additions to the Program;
+  where such changes and/or additions to the Program originate from
+  and are Distributed by that particular Contributor. A Contribution
+  "originates" from a Contributor if it was added to the Program by
+  such Contributor itself or anyone acting on such Contributor's behalf.
+  Contributions do not include changes or additions to the Program that
+  are not Modified Works.

+"Contributor" means any person or entity that Distributes the Program.
+
+"Licensed Patents" mean patent claims licensable by a Contributor which
+are necessarily infringed by the use or sale of its Contribution alone
+or when combined with the Program.
+
+"Program" means the Contributions Distributed in accordance with this
+Agreement.
+
+"Recipient" means anyone who receives the Program under this Agreement
+or any Secondary License (as applicable), including Contributors.
+
+"Derivative Works" shall mean any work, whether in Source Code or other
+form, that is based on (or derived from) the Program and for which the
+editorial revisions, annotations, elaborations, or other modifications
+represent, as a whole, an original work of authorship.
+
+"Modified Works" shall mean any work in Source Code or other form that
+results from an addition to, deletion from, or modification of the
+contents of the Program, including, for purposes of clarity any new file
+in Source Code form that contains any contents of the Program. Modified
+Works shall not include works that contain only declarations,
+interfaces, types, classes, structures, or files of the Program solely
+in each case in order to link to, bind by name, or subclass the Program
+or Modified Works thereof.
+
+"Distribute" means the acts of a) distributing or b) making available
+in any manner that enables the transfer of a copy.
+
+"Source Code" means the form of a Program preferred for making
+modifications, including but not limited to software source code,
+documentation source, and configuration files.
+
+"Secondary License" means either the GNU General Public License,
+Version 2.0, or any later versions of that license, including any
+exceptions or additional permissions as identified by the initial
+Contributor.
+
+2. GRANT OF RIGHTS
+
+  a) Subject to the terms of this Agreement, each Contributor hereby
+  grants Recipient a non-exclusive, worldwide, royalty-free copyright
+  license to reproduce, prepare Derivative Works of, publicly display,
+  publicly perform, Distribute and sublicense the Contribution of such
+  Contributor, if any, and such Derivative Works.
+
+  b) Subject to the terms of this Agreement, each Contributor hereby
+  grants Recipient a non-exclusive, worldwide, royalty-free patent
+  license under Licensed Patents to make, use, sell, offer to sell,
+  import and otherwise transfer the Contribution of such Contributor,
+  if any, in Source Code or other form. This patent license shall
+  apply to the combination of the Contribution and the Program if, at
+  the time the Contribution is added by the Contributor, such addition
+  of the Contribution causes such combination to be covered by the
+  Licensed Patents. The patent license shall not apply to any other
+  combinations which include the Contribution. No hardware per se is
+  licensed hereunder.
+
+  c) Recipient understands that although each Contributor grants the
+  licenses to its Contributions set forth herein, no assurances are
+  provided by any Contributor that the Program does not infringe the
+  patent or other intellectual property rights of any other entity.
+  Each Contributor disclaims any liability to Recipient for claims
+  brought by any other entity based on infringement of intellectual
+  property rights or otherwise. As a condition to exercising the
+  rights and licenses granted hereunder, each Recipient hereby
+  assumes sole responsibility to secure any other intellectual
+  property rights needed, if any. For example, if a third party
+  patent license is required to allow Recipient to Distribute the
+  Program, it is Recipient's responsibility to acquire that license
+  before distributing the Program.
+
+  d) Each Contributor represents that to its knowledge it has
+  sufficient copyright rights in its Contribution, if any, to grant
+  the copyright license set forth in this Agreement.
+
+  e) Notwithstanding the terms of any Secondary License, no
+  Contributor makes additional grants to any Recipient (other than
+  those set forth in this Agreement) as a result of such Recipient's
+  receipt of the Program under the terms of a Secondary License
+  (if permitted under the terms of Section 3).
+
+3. REQUIREMENTS
+
+3.1 If a Contributor Distributes the Program in any form, then:
+
+  a) the Program must also be made available as Source Code, in
+  accordance with section 3.2, and the Contributor must accompany
+  the Program with a statement that the Source Code for the Program
+  is available under this Agreement, and informs Recipients how to
+  obtain it in a reasonable manner on or through a medium customarily
+  used for software exchange; and
+
+  b) the Contributor may Distribute the Program under a license
+  different than this Agreement, provided that such license:
+     i) effectively disclaims on behalf of all other Contributors all
+     warranties and conditions, express and implied, including
+     warranties or conditions of title and non-infringement, and
+     implied warranties or conditions of merchantability and fitness
+     for a particular purpose;
+
+     ii) effectively excludes on behalf of all other Contributors all
+     liability for damages, including direct, indirect, special,
+     incidental and consequential damages, such as lost profits;
+
+     iii) does not attempt to limit or alter the recipients' rights
+     in the Source Code under section 3.2; and
+
+     iv) requires any subsequent distribution of the Program by any
+     party to be under a license that satisfies the requirements
+     of this section 3.
+
+3.2 When the Program is Distributed as Source Code:
+
+  a) it must be made available under this Agreement, or if the
+  Program (i) is combined with other material in a separate file or
+  files made available under a Secondary License, and (ii) the initial
+  Contributor attached to the Source Code the notice described in
+  Exhibit A of this Agreement, then the Program may be made available
+  under the terms of such Secondary Licenses, and
+
+  b) a copy of this Agreement must be included with each copy of
+  the Program.
+
+3.3 Contributors may not remove or alter any copyright, patent,
+trademark, attribution notices, disclaimers of warranty, or limitations
+of liability ("notices") contained within the Program from any copy of
+the Program which they Distribute, provided that Contributors may add
+their own appropriate notices.
+
+4. COMMERCIAL DISTRIBUTION
+
+Commercial distributors of software may accept certain responsibilities
+with respect to end users, business partners and the like. While this
+license is intended to facilitate the commercial use of the Program,
+the Contributor who includes the Program in a commercial product
+offering should do so in a manner which does not create potential
+liability for other Contributors. Therefore, if a Contributor includes
+the Program in a commercial product offering, such Contributor
+("Commercial Contributor") hereby agrees to defend and indemnify every
+other Contributor ("Indemnified Contributor") against any losses,
+damages and costs (collectively "Losses") arising from claims, lawsuits
+and other legal actions brought by a third party against the Indemnified
+Contributor to the extent caused by the acts or omissions of such
+Commercial Contributor in connection with its distribution of the Program
+in a commercial product offering. The obligations in this section do not
+apply to any claims or Losses relating to any actual or alleged
+intellectual property infringement. In order to qualify, an Indemnified
+Contributor must: a) promptly notify the Commercial Contributor in
+writing of such claim, and b) allow the Commercial Contributor to control,
+and cooperate with the Commercial Contributor in, the defense and any
+related settlement negotiations. The Indemnified Contributor may
+participate in any such claim at its own expense.
+
+For example, a Contributor might include the Program in a commercial
+product offering, Product X. That Contributor is then a Commercial
+Contributor. If that Commercial Contributor then makes performance
+claims, or offers warranties related to Product X, those performance
+claims and warranties are such Commercial Contributor's responsibility
+alone. Under this section, the Commercial Contributor would have to
+defend claims against the other Contributors related to those performance
+claims and warranties, and if a court requires any other Contributor to
+pay any damages as a result, the Commercial Contributor must pay
+those damages.
+
+5. NO WARRANTY
+
+EXCEPT AS EXPRESSLY SET FORTH IN THIS AGREEMENT, AND TO THE EXTENT
+PERMITTED BY APPLICABLE LAW, THE PROGRAM IS PROVIDED ON AN "AS IS"
+BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, EITHER EXPRESS OR
+IMPLIED INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OR CONDITIONS OF
+TITLE, NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR
+PURPOSE. Each Recipient is solely responsible for determining the
+appropriateness of using and distributing the Program and assumes all
+risks associated with its exercise of rights under this Agreement,
+including but not limited to the risks and costs of program errors,
+compliance with applicable laws, damage to or loss of data, programs
+or equipment, and unavailability or interruption of operations.
+
+6. DISCLAIMER OF LIABILITY
+
+EXCEPT AS EXPRESSLY SET FORTH IN THIS AGREEMENT, AND TO THE EXTENT
+PERMITTED BY APPLICABLE LAW, NEITHER RECIPIENT NOR ANY CONTRIBUTORS
+SHALL HAVE ANY LIABILITY FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING WITHOUT LIMITATION LOST
+PROFITS), HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ARISING IN ANY WAY OUT OF THE USE OR DISTRIBUTION OF THE PROGRAM OR THE
+EXERCISE OF ANY RIGHTS GRANTED HEREUNDER, EVEN IF ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+7. GENERAL
+
+If any provision of this Agreement is invalid or unenforceable under
+applicable law, it shall not affect the validity or enforceability of
+the remainder of the terms of this Agreement, and without further
+action by the parties hereto, such provision shall be reformed to the
+minimum extent necessary to make such provision valid and enforceable.
+
+If Recipient institutes patent litigation against any entity
+(including a cross-claim or counterclaim in a lawsuit) alleging that the
+Program itself (excluding combinations of the Program with other software
+or hardware) infringes such Recipient's patent(s), then such Recipient's
+rights granted under Section 2(b) shall terminate as of the date such
+litigation is filed.
+
+All Recipient's rights under this Agreement shall terminate if it
+fails to comply with any of the material terms or conditions of this
+Agreement and does not cure such failure in a reasonable period of
+time after becoming aware of such noncompliance. If all Recipient's
+rights under this Agreement terminate, Recipient agrees to cease use
+and distribution of the Program as soon as reasonably practicable.
+However, Recipient's obligations under this Agreement and any licenses
+granted by Recipient relating to the Program shall continue and survive.
+
+Everyone is permitted to copy and distribute copies of this Agreement,
+but in order to avoid inconsistency the Agreement is copyrighted and
+may only be modified in the following manner. The Agreement Steward
+reserves the right to publish new versions (including revisions) of
+this Agreement from time to time. No one other than the Agreement
+Steward has the right to modify this Agreement. The Eclipse Foundation
+is the initial Agreement Steward. The Eclipse Foundation may assign the
+responsibility to serve as the Agreement Steward to a suitable separate
+entity. Each new version of the Agreement will be given a distinguishing
+version number. The Program (including Contributions) may always be
+Distributed subject to the version of the Agreement under which it was
+received. In addition, after a new version of the Agreement is published,
+Contributor may elect to Distribute the Program (including its
+Contributions) under the new version.
+
+Except as expressly stated in Sections 2(a) and 2(b) above, Recipient
+receives no rights or licenses to the intellectual property of any
+Contributor under this Agreement, whether expressly, by implication,
+estoppel or otherwise. All rights in the Program not expressly granted
+under this Agreement are reserved. Nothing in this Agreement is intended
+to be enforceable by any entity that is not a Contributor or Recipient.
+No third-party beneficiary rights are created under this Agreement.
+
+Exhibit A - Form of Secondary Licenses Notice
+
+"This Source Code may also be made available under the following
+Secondary Licenses when the conditions for such availability set forth
+in the Eclipse Public License, v. 2.0 are satisfied: {name license(s),
+version(s), and exceptions or additional permissions here}."
+
+  Simply including a copy of this Agreement, including this Exhibit A
+  is not sufficient to license the Source Code under Secondary Licenses.
+
+  If it is not possible or desirable to put the notice in a particular
+  file, then You may include the notice in a location (such as a LICENSE
+  file in a relevant directory) where a recipient would be likely to
+  look for such a notice.
+
+  You may add additional accurate notices of copyright ownership.

 ================================================================

@@ -22029,6 +22123,40 @@ OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.


+================================================================
+
+github.com/pierrec/lz4/v4
+https://github.com/pierrec/lz4/v4
+----------------------------------------------------------------
+Copyright (c) 2015, Pierre Curto
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+* Redistributions of source code must retain the above copyright notice, this
+  list of conditions and the following disclaimer.
+
+* Redistributions in binary form must reproduce the above copyright notice,
+  this list of conditions and the following disclaimer in the documentation
+  and/or other materials provided with the distribution.
+
+* Neither the name of xxHash nor the names of its
+  contributors may be used to endorse or promote products derived from
+  this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+
 ================================================================

 github.com/pkg/errors
@@ -23215,6 +23343,213 @@ https://github.com/prometheus/procfs

 ================================================================

+github.com/prometheus/prom2json
+https://github.com/prometheus/prom2json
+----------------------------------------------------------------
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
+
+================================================================
+
 github.com/rcrowley/go-metrics
 https://github.com/rcrowley/go-metrics
 ----------------------------------------------------------------
@@ -27415,17 +27750,3 @@ limitations under the License.

 ================================================================

-maze.io/x/duration
-https://maze.io/x/duration
----------------------------------------------------------------
-MIT License
-Copyright (c) <year> <copyright holders>
-
-Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
-
-================================================================
-
@@ -15,11 +15,11 @@ checks: ## check dependencies
 	@(env bash $(PWD)/buildscripts/checkdeps.sh)

 help: ## print this help
-	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' Makefile | awk 'BEGIN {FS = ":.*?## "}; {printf "\033[36m%-30s\033[0m %s\n", $$1, $$2}'
+	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' Makefile | awk 'BEGIN {FS = ":.*?## "}; {printf "\033[36m%-40s\033[0m %s\n", $$1, $$2}'

 getdeps: ## fetch necessary dependencies
 	@mkdir -p ${GOPATH}/bin
-	@echo "Installing golangci-lint" && curl -sSfL https://raw.githubusercontent.com/golangci/golangci-lint/master/install.sh | sh -s -- -b $(GOPATH)/bin v1.45.2
+	@echo "Installing golangci-lint" && curl -sSfL https://raw.githubusercontent.com/golangci/golangci-lint/master/install.sh | sh -s -- -b $(GOPATH)/bin
 	@echo "Installing msgp" && go install -v github.com/tinylib/msgp@v1.1.7-0.20211026165309-e818a1881b0e
 	@echo "Installing stringer" && go install -v golang.org/x/tools/cmd/stringer@latest

@@ -39,11 +39,14 @@ lint: ## runs golangci-lint suite of linters
 check: test
 test: verifiers build ## builds minio, runs linters, tests
 	@echo "Running unit tests"
-	@CGO_ENABLED=0 go test -tags kqueue ./...
+	@MINIO_API_REQUESTS_MAX=10000 CGO_ENABLED=0 go test -tags kqueue ./...

 test-decom: install
 	@echo "Running minio decom tests"
 	@env bash $(PWD)/docs/distributed/decom.sh
+	@env bash $(PWD)/docs/distributed/decom-encrypted.sh
+	@env bash $(PWD)/docs/distributed/decom-encrypted-sse-s3.sh
+	@env bash $(PWD)/docs/distributed/decom-compressed-sse-s3.sh

 test-upgrade: build
 	@echo "Running minio upgrade tests"
@@ -55,9 +58,9 @@ test-race: verifiers build ## builds minio, runs linters, tests (race)

 test-iam: build ## verify IAM (external IDP, etcd backends)
 	@echo "Running tests for IAM (external IDP, etcd backends)"
-	@CGO_ENABLED=0 go test -tags kqueue -v -run TestIAM* ./cmd
+	@MINIO_API_REQUESTS_MAX=10000 CGO_ENABLED=0 go test -tags kqueue -v -run TestIAM* ./cmd
 	@echo "Running tests for IAM (external IDP, etcd backends) with -race"
-	@GORACE=history_size=7 CGO_ENABLED=1 go test -race -tags kqueue -v -run TestIAM* ./cmd
+	@MINIO_API_REQUESTS_MAX=10000 GORACE=history_size=7 CGO_ENABLED=1 go test -race -tags kqueue -v -run TestIAM* ./cmd

 test-replication: install ## verify multi site replication
 	@echo "Running tests for replicating three sites"
@@ -87,6 +90,16 @@ verify-healing: ## verify healing and replacing disks with minio binary
 	@(env bash $(PWD)/buildscripts/verify-healing.sh)
 	@(env bash $(PWD)/buildscripts/unaligned-healing.sh)

+verify-healing-with-root-disks: ## verify healing root disks
+	@echo "Verify healing with root drives"
+	@GORACE=history_size=7 CGO_ENABLED=1 go build -race -tags kqueue -trimpath --ldflags "$(LDFLAGS)" -o $(PWD)/minio 1>/dev/null
+	@(env bash $(PWD)/buildscripts/verify-healing-with-root-disks.sh)
+
+verify-healing-with-rewrite: ## verify healing to rewrite old xl.meta -> new xl.meta
+	@echo "Verify healing with rewrite"
+	@GORACE=history_size=7 CGO_ENABLED=1 go build -race -tags kqueue -trimpath --ldflags "$(LDFLAGS)" -o $(PWD)/minio 1>/dev/null
+	@(env bash $(PWD)/buildscripts/rewrite-old-new.sh)
+
 verify-healing-inconsistent-versions: ## verify resolving inconsistent versions
 	@echo "Verify resolving inconsistent versions build with race"
 	@GORACE=history_size=7 CGO_ENABLED=1 go build -race -tags kqueue -trimpath --ldflags "$(LDFLAGS)" -o $(PWD)/minio 1>/dev/null
@@ -132,6 +145,8 @@ clean: ## cleanup all generated assets
 	@echo "Cleaning up all the generated files"
 	@find . -name '*.test' | xargs rm -fv
 	@find . -name '*~' | xargs rm -fv
+	@find . -name '.#*#' | xargs rm -fv
+	@find . -name '#*#' | xargs rm -fv
 	@rm -rvf minio
 	@rm -rvf build
 	@rm -rvf release
@@ -2,6 +2,6 @@

 MinIO creates FIPS builds using a patched version of the Go compiler (that uses BoringCrypto, from BoringSSL, which is [FIPS 140-2 validated](https://csrc.nist.gov/csrc/media/projects/cryptographic-module-validation-program/documents/security-policies/140sp2964.pdf)) published by the Golang Team [here](https://github.com/golang/go/tree/dev.boringcrypto/misc/boring).

-MinIO FIPS executables are available at http://dl.min.io - they are only published for `linux-amd64` architecture as binary files with the suffix `.fips`. We also publish corresponding container images to our official image repositories.
+MinIO FIPS executables are available at <http://dl.min.io> - they are only published for `linux-amd64` architecture as binary files with the suffix `.fips`. We also publish corresponding container images to our official image repositories.

 We are not making any statements or representations about the suitability of this code or build in relation to the FIPS 140-2 standard. Interested users will have to evaluate for themselves whether this is useful for their own purposes.
@@ -125,7 +125,7 @@ You can also connect using any S3-compatible tool, such as the MinIO Client `mc`

 ## Install from Source

-Use the following commands to compile and run a standalone MinIO server from source. Source installation is only intended for developers and advanced users. If you do not have a working Golang environment, please follow [How to install Golang](https://golang.org/doc/install). Minimum version required is [go1.17](https://golang.org/dl/#stable)
+Use the following commands to compile and run a standalone MinIO server from source. Source installation is only intended for developers and advanced users. If you do not have a working Golang environment, please follow [How to install Golang](https://golang.org/doc/install). Minimum version required is [go1.18](https://golang.org/dl/#stable)

 ```sh
 GO111MODULE=on go install github.com/minio/minio@latest
@@ -0,0 +1,87 @@
+//go:build ignore
+// +build ignore
+
+//
+// MinIO Object Storage (c) 2022 MinIO, Inc.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//
+
+package main
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"log"
+	"os"
+	"time"
+
+	"github.com/minio/madmin-go"
+)
+
+func main() {
+	// Note: YOUR-ACCESSKEYID, YOUR-SECRETACCESSKEY are
+	// dummy values, please replace them with original values.
+
+	// API requests are secure (HTTPS) if secure=true and insecure (HTTP) otherwise.
+	// New returns an MinIO Admin client object.
+	madmClnt, err := madmin.New(os.Args[1], os.Args[2], os.Args[3], false)
+	if err != nil {
+		log.Fatalln(err)
+	}
+
+	opts := madmin.HealOpts{
+		Recursive: true,                  // recursively heal all objects at 'prefix'
+		Remove:    true,                  // remove content that has lost quorum and not recoverable
+		Recreate:  true,                  // rewrite all old non-inlined xl.meta to new xl.meta
+		ScanMode:  madmin.HealNormalScan, // by default do not do 'deep' scanning
+	}
+
+	start, _, err := madmClnt.Heal(context.Background(), "healing-rewrite-bucket", "", opts, "", false, false)
+	if err != nil {
+		log.Fatalln(err)
+	}
+	fmt.Println("Healstart sequence ===")
+	enc := json.NewEncoder(os.Stdout)
+	if err = enc.Encode(&start); err != nil {
+		log.Fatalln(err)
+	}
+
+	fmt.Println()
+	for {
+		_, status, err := madmClnt.Heal(context.Background(), "healing-rewrite-bucket", "", opts, start.ClientToken, false, false)
+		if status.Summary == "finished" {
+			fmt.Println("Healstatus on items ===")
+			for _, item := range status.Items {
+				if err = enc.Encode(&item); err != nil {
+					log.Fatalln(err)
+				}
+			}
+			break
+		}
+		if status.Summary == "stopped" {
+			fmt.Println("Healstatus on items ===")
+			fmt.Println("Heal failed with", status.FailureDetail)
+			break
+		}
+
+		for _, item := range status.Items {
+			if err = enc.Encode(&item); err != nil {
+				log.Fatalln(err)
+			}
+		}
+
+		time.Sleep(time.Second)
+	}
+}
@@ -3,6 +3,8 @@
 set -e

 export GORACE="history_size=7"
+export MINIO_API_REQUESTS_MAX=10000
+
 ## TODO remove `dsync` from race detector once this is merged and released https://go-review.googlesource.com/c/go/+/333529/
 for d in $(go list ./... | grep -v dsync); do
    CGO_ENABLED=1 go test -v -race --timeout 100m "$d"
@@ -39,7 +39,7 @@ function start_minio_5drive() {
    "${MINIO[@]}" --address ":$start_port" "${WORK_DIR}/cicd-corpus/disk{1...5}" > "${WORK_DIR}/server1.log" 2>&1 &
    pid=$!
    disown $pid
-    sleep 30
+    sleep 5

    if ! ps -p ${pid} 1>&2 >/dev/null; then
 	echo "server1 log:"
@@ -0,0 +1,151 @@
+#!/bin/bash -e
+
+set -E
+set -o pipefail
+set -x
+
+WORK_DIR="$PWD/.verify-$RANDOM"
+MINIO_CONFIG_DIR="$WORK_DIR/.minio"
+MINIO_OLD=( "$PWD/minio.RELEASE.2020-10-28T08-16-50Z" --config-dir "$MINIO_CONFIG_DIR" server )
+MINIO=( "$PWD/minio" --config-dir "$MINIO_CONFIG_DIR" server )
+
+if [ ! -x "$PWD/minio" ]; then
+    echo "minio executable binary not found in current directory"
+    exit 1
+fi
+
+function download_old_release() {
+    if [ ! -f minio.RELEASE.2020-10-28T08-16-50Z ]; then
+	curl --silent -O https://dl.minio.io/server/minio/release/linux-amd64/archive/minio.RELEASE.2020-10-28T08-16-50Z
+	chmod a+x minio.RELEASE.2020-10-28T08-16-50Z
+    fi
+}
+
+function verify_rewrite() {
+    start_port=$1
+
+    export MINIO_ACCESS_KEY=minio
+    export MINIO_SECRET_KEY=minio123
+    export MC_HOST_minio="http://minio:minio123@127.0.0.1:${start_port}/"
+    unset MINIO_KMS_AUTO_ENCRYPTION # do not auto-encrypt objects
+    export MINIO_CI_CD=1
+
+    MC_BUILD_DIR="mc-$RANDOM"
+    if ! git clone --quiet https://github.com/minio/mc "$MC_BUILD_DIR"; then
+	echo "failed to download https://github.com/minio/mc"
+	purge "${MC_BUILD_DIR}"
+	exit 1
+    fi
+
+    (cd "${MC_BUILD_DIR}" && go build -o "$WORK_DIR/mc")
+
+    # remove mc source.
+    purge "${MC_BUILD_DIR}"
+
+    "${MINIO_OLD[@]}" --address ":$start_port" "${WORK_DIR}/xl{1...16}" > "${WORK_DIR}/server1.log" 2>&1 &
+    pid=$!
+    disown $pid
+    sleep 10
+
+    if ! ps -p ${pid} 1>&2 >/dev/null; then
+	echo "server1 log:"
+	cat "${WORK_DIR}/server1.log"
+	echo "FAILED"
+	purge "$WORK_DIR"
+	exit 1
+    fi
+
+    "${WORK_DIR}/mc" mb minio/healing-rewrite-bucket --quiet --with-lock
+    "${WORK_DIR}/mc" cp \
+		     buildscripts/verify-build.sh \
+		     minio/healing-rewrite-bucket/ \
+		     --disable-multipart --quiet
+
+    "${WORK_DIR}/mc" cp \
+		     buildscripts/verify-build.sh \
+		     minio/healing-rewrite-bucket/ \
+		     --disable-multipart --quiet
+
+    "${WORK_DIR}/mc" cp \
+		     buildscripts/verify-build.sh \
+		     minio/healing-rewrite-bucket/ \
+		     --disable-multipart --quiet
+
+    kill ${pid}
+    sleep 3
+
+    "${MINIO[@]}" --address ":$start_port" "${WORK_DIR}/xl{1...16}" > "${WORK_DIR}/server1.log" 2>&1 &
+    pid=$!
+    disown $pid
+    sleep 10
+
+    if ! ps -p ${pid} 1>&2 >/dev/null; then
+	echo "server1 log:"
+	cat "${WORK_DIR}/server1.log"
+	echo "FAILED"
+	purge "$WORK_DIR"
+	exit 1
+    fi
+
+    go build ./docs/debugging/s3-check-md5/
+    if ! ./s3-check-md5 \
+	 -debug \
+	 -versions \
+	 -access-key minio \
+	 -secret-key minio123 \
+	 -endpoint http://127.0.0.1:${start_port}/ 2>&1 | grep INTACT; then
+	echo "server1 log:"
+	cat "${WORK_DIR}/server1.log"
+	echo "FAILED"
+	mkdir -p inspects
+	(cd inspects; "${WORK_DIR}/mc" admin inspect minio/healing-rewrite-bucket/verify-build.sh/**)
+
+	"${WORK_DIR}/mc" mb play/inspects
+	"${WORK_DIR}/mc" mirror inspects play/inspects
+
+	purge "$WORK_DIR"
+	exit 1
+    fi
+
+    go run ./buildscripts/heal-manual.go "127.0.0.1:${start_port}" "minio" "minio123"
+    sleep 1
+
+    if ! ./s3-check-md5 \
+	 -debug \
+	 -versions \
+	 -access-key minio \
+	 -secret-key minio123 \
+	 -endpoint http://127.0.0.1:${start_port}/ 2>&1 | grep INTACT; then
+	echo "server1 log:"
+	cat "${WORK_DIR}/server1.log"
+	echo "FAILED"
+	mkdir -p inspects
+	(cd inspects; "${WORK_DIR}/mc" admin inspect minio/healing-rewrite-bucket/verify-build.sh/**)
+
+	"${WORK_DIR}/mc" mb play/inspects
+	"${WORK_DIR}/mc" mirror inspects play/inspects
+
+	purge "$WORK_DIR"
+	exit 1
+    fi
+
+    kill ${pid}
+}
+
+function main() {
+    download_old_release
+
+    start_port=$(shuf -i 10000-65000 -n 1)
+
+    verify_rewrite ${start_port}
+}
+
+function purge()
+{
+    rm -rf "$1"
+}
+
+( main "$@" )
+rv=$?
+purge "$WORK_DIR"
+exit "$rv"
@@ -120,7 +120,7 @@ function start_minio_16drive() {
 	cat "${WORK_DIR}/server1.log"
 	echo "FAILED"
 	mkdir -p inspects
-	(cd inspects; "${WORK_DIR}/mc" admin inspect minio/healing-shard-bucket/unaligned/**)
+	(cd inspects; "${WORK_DIR}/mc" support inspect minio/healing-shard-bucket/unaligned/**)

 	"${WORK_DIR}/mc" mb play/inspects
 	"${WORK_DIR}/mc" mirror inspects play/inspects
@@ -140,7 +140,7 @@ function start_minio_16drive() {
 	cat "${WORK_DIR}/server1.log"
 	echo "FAILED"
 	mkdir -p inspects
-	(cd inspects; "${WORK_DIR}/mc" admin inspect minio/healing-shard-bucket/unaligned/**)
+	(cd inspects; "${WORK_DIR}/mc" support inspect minio/healing-shard-bucket/unaligned/**)

 	"${WORK_DIR}/mc" mb play/inspects
 	"${WORK_DIR}/mc" mirror inspects play/inspects
@@ -0,0 +1,97 @@
+#!/bin/bash -e
+
+set -E
+set -o pipefail
+set -x
+
+
+if [ ! -x "$PWD/minio" ]; then
+    echo "minio executable binary not found in current directory"
+    exit 1
+fi
+
+WORK_DIR="$(mktemp -d)"
+MINIO_CONFIG_DIR="$WORK_DIR/.minio"
+MINIO=( "$PWD/minio" --config-dir "$MINIO_CONFIG_DIR" server )
+
+
+function start_minio() {
+    start_port=$1
+
+    export MINIO_ROOT_USER=minio
+    export MINIO_ROOT_PASSWORD=minio123
+    unset MINIO_KMS_AUTO_ENCRYPTION # do not auto-encrypt objects
+    unset MINIO_CI_CD
+    unset CI
+
+    args=()
+    for i in $(seq 1 4); do
+            args+=("http://localhost:$[${start_port}+$i]${WORK_DIR}/mnt/disk$i/ ")
+    done
+
+    for i in $(seq 1 4); do
+            "${MINIO[@]}" --address ":$[$start_port+$i]" ${args[@]} 2>&1 >"${WORK_DIR}/server$i.log" &
+    done
+
+    # Wait until all nodes return 403
+    for i in $(seq 1 4); do
+	    while [ "$(curl -m 1 -s -o /dev/null -w "%{http_code}" http://localhost:$[$start_port+$i])" -ne "403" ]; do
+		    echo -n ".";
+                    sleep 1;
+            done
+    done
+
+ }
+
+# Prepare fake disks with losetup
+function prepare_block_devices() {
+        mkdir -p ${WORK_DIR}/disks/ ${WORK_DIR}/mnt/
+        for i in 1 2 3 4; do
+                dd if=/dev/zero of=${WORK_DIR}/disks/img.$i bs=1M count=2048
+                mkfs.ext4 -F ${WORK_DIR}/disks/img.$i
+                sudo mknod /dev/minio-loopdisk$i b 7 $[256-$i]
+                sudo losetup /dev/minio-loopdisk$i ${WORK_DIR}/disks/img.$i
+                mkdir -p ${WORK_DIR}/mnt/disk$i/
+                sudo mount /dev/minio-loopdisk$i ${WORK_DIR}/mnt/disk$i/
+                sudo chown "$(id -u):$(id -g)" /dev/minio-loopdisk$i ${WORK_DIR}/mnt/disk$i/
+        done
+}
+
+# Start a distributed MinIO setup, unmount one disk and check if it is formatted
+function main() {
+    start_port=$(shuf -i 10000-65000 -n 1)
+    start_minio ${start_port}
+
+    # Unmount the disk, after the unmount the device id
+    # /tmp/xxx/mnt/disk4 will be the same as '/' and it
+    # will be detected as root disk
+    while [ "$u" != "0" ]; do
+	    sudo umount ${WORK_DIR}/mnt/disk4/
+	    u=$?
+	    sleep 1
+    done
+
+    # Wait until MinIO self heal kicks in
+    sleep 60
+
+    if [ -f ${WORK_DIR}/mnt/disk4/.minio.sys/format.json ]; then
+	    echo "A root disk is formatted unexpectedely"
+            cat "${WORK_DIR}/server4.log"
+            exit -1
+    fi
+}
+
+function cleanup() {
+    pkill minio
+    sudo umount ${WORK_DIR}/mnt/disk{1..3}/
+    sudo rm /dev/minio-loopdisk*
+    rm -rf "$WORK_DIR"
+}
+
+( prepare_block_devices )
+( main "$@" )
+rv=$?
+
+cleanup
+exit "$rv"
+
@@ -80,7 +80,7 @@ func (api objectAPIHandlers) PutBucketACLHandler(w http.ResponseWriter, r *http.
 	}

 	// Before proceeding validate if bucket exists.
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -142,7 +142,7 @@ func (api objectAPIHandlers) GetBucketACLHandler(w http.ResponseWriter, r *http.
 	}

 	// Before proceeding validate if bucket exists.
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -63,14 +63,13 @@ func (a adminAPIHandlers) PutBucketQuotaConfigHandler(w http.ResponseWriter, r *

 	objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.SetBucketQuotaAdminAction)
 	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
 		return
 	}

 	vars := mux.Vars(r)
 	bucket := pathClean(vars["bucket"])

-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -126,14 +125,13 @@ func (a adminAPIHandlers) GetBucketQuotaConfigHandler(w http.ResponseWriter, r *

 	objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.GetBucketQuotaAdminAction)
 	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
 		return
 	}

 	vars := mux.Vars(r)
 	bucket := pathClean(vars["bucket"])

-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -171,12 +169,11 @@ func (a adminAPIHandlers) SetRemoteTargetHandler(w http.ResponseWriter, r *http.
 	// Get current object layer instance.
 	objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.SetBucketTargetAction)
 	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
 		return
 	}

 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -252,7 +249,7 @@ func (a adminAPIHandlers) SetRemoteTargetHandler(w http.ResponseWriter, r *http.
 	}
 	if err = globalBucketTargetSys.SetTarget(ctx, bucket, &target, update); err != nil {
 		switch err.(type) {
-		case BucketRemoteConnectionErr:
+		case RemoteTargetConnectionErr:
 			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrReplicationRemoteConnectionError, err), r.URL)
 		default:
 			writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
@@ -300,12 +297,11 @@ func (a adminAPIHandlers) ListRemoteTargetsHandler(w http.ResponseWriter, r *htt
 	// Get current object layer instance.
 	objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.GetBucketTargetAction)
 	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
 		return
 	}
 	if bucket != "" {
 		// Check if bucket exists.
-		if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+		if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 			writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 			return
 		}
@@ -340,12 +336,11 @@ func (a adminAPIHandlers) RemoveRemoteTargetHandler(w http.ResponseWriter, r *ht
 	// Get current object layer instance.
 	objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.SetBucketTargetAction)
 	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
 		return
 	}

 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -379,29 +374,29 @@ func (a adminAPIHandlers) ExportBucketMetadataHandler(w http.ResponseWriter, r *
 	defer logger.AuditLog(ctx, w, r, mustGetClaimsFromToken(r))

 	bucket := pathClean(r.Form.Get("bucket"))
-	if !globalIsErasure {
+	if globalIsGateway {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
 		return
 	}
 	// Get current object layer instance.
 	objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.ExportBucketMetadataAction)
 	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
 		return
 	}
+
 	var (
 		buckets []BucketInfo
 		err     error
 	)
 	if bucket != "" {
 		// Check if bucket exists.
-		if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+		if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 			writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 			return
 		}
 		buckets = append(buckets, BucketInfo{Name: bucket})
 	} else {
-		buckets, err = objectAPI.ListBuckets(ctx)
+		buckets, err = objectAPI.ListBuckets(ctx, BucketOptions{})
 		if err != nil {
 			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 			return
@@ -624,6 +619,39 @@ func (a adminAPIHandlers) ExportBucketMetadataHandler(w http.ResponseWriter, r *
 	}
 }

+type importMetaReport struct {
+	madmin.BucketMetaImportErrs
+}
+
+func (i *importMetaReport) SetStatus(bucket, fname string, err error) {
+	st := i.Buckets[bucket]
+	var errMsg string
+	if err != nil {
+		errMsg = err.Error()
+	}
+	switch fname {
+	case bucketPolicyConfig:
+		st.Policy = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketNotificationConfig:
+		st.Notification = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketLifecycleConfig:
+		st.Lifecycle = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketSSEConfig:
+		st.SSEConfig = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketTaggingConfig:
+		st.Tagging = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketQuotaConfigFile:
+		st.Quota = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case objectLockConfig:
+		st.ObjectLock = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketVersioningConfig:
+		st.Versioning = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	default:
+		st.Err = errMsg
+	}
+	i.Buckets[bucket] = st
+}
+
 // ImportBucketMetadataHandler - imports all bucket metadata from a zipped file and overwrite bucket metadata config
 // There are some caveats regarding the following:
 // 1. object lock config - object lock should have been specified at time of bucket creation. Only default retention settings are imported here.
@@ -633,16 +661,14 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 	ctx := newContext(r, w, "ImportBucketMetadata")

 	defer logger.AuditLog(ctx, w, r, mustGetClaimsFromToken(r))
-	bucket := pathClean(r.Form.Get("bucket"))

-	if !globalIsErasure {
+	if globalIsGateway {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
 		return
 	}
 	// Get current object layer instance.
 	objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.ImportBucketMetadataAction)
 	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
 		return
 	}
 	data, err := ioutil.ReadAll(r.Body)
@@ -657,47 +683,46 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 		return
 	}
 	bucketMap := make(map[string]struct{}, 1)
-
+	rpt := importMetaReport{
+		madmin.BucketMetaImportErrs{
+			Buckets: make(map[string]madmin.BucketStatus, len(zr.File)),
+		},
+	}
 	// import object lock config if any - order of import matters here.
 	for _, file := range zr.File {
 		slc := strings.Split(file.Name, slashSeparator)
 		if len(slc) != 2 { // expecting bucket/configfile in the zipfile
-			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-			return
-		}
-		b, fileName := slc[0], slc[1]
-		if bucket == "" { // use bucket requested in query parameters if specified. Otherwise default bucket name to directory name within zip
-			bucket = b
+			rpt.SetStatus(file.Name, "", fmt.Errorf("malformed zip - expecting format bucket/<config.json>"))
+			continue
 		}
+		bucket, fileName := slc[0], slc[1]
 		switch fileName {
 		case objectLockConfig:
 			reader, err := file.Open()
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			config, err := objectlock.ParseObjectLockConfig(reader)
 			if err != nil {
-				apiErr := errorCodes.ToAPIErr(ErrMalformedXML)
-				apiErr.Description = err.Error()
-				writeErrorResponse(ctx, w, apiErr, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}

 			configData, err := xml.Marshal(config)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			if _, ok := bucketMap[bucket]; !ok {
-				opts := BucketOptions{
+				opts := MakeBucketOptions{
 					LockEnabled: config.ObjectLockEnabled == "Enabled",
 				}
 				err = objectAPI.MakeBucketWithLocation(ctx, bucket, opts)
 				if err != nil {
 					if _, ok := err.(BucketExists); !ok {
-						writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-						return
+						rpt.SetStatus(bucket, fileName, err)
+						continue
 					}
 				}
 				bucketMap[bucket] = struct{}{}
@@ -705,15 +730,16 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *

 			// Deny object locking configuration settings on existing buckets without object lock enabled.
 			if _, _, err = globalBucketMetadataSys.GetObjectLockConfig(bucket); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, objectLockConfig, configData)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)

 			// Call site replication hook.
 			//
@@ -726,8 +752,8 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				ObjectLockConfig: &cfgStr,
 				UpdatedAt:        updatedAt,
 			}); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 		}
 	}
@@ -736,162 +762,145 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 	for _, file := range zr.File {
 		slc := strings.Split(file.Name, slashSeparator)
 		if len(slc) != 2 { // expecting bucket/configfile in the zipfile
-			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-			return
-		}
-		b, fileName := slc[0], slc[1]
-		if bucket == "" { // use bucket requested in query parameters if specified. Otherwise default bucket name to directory name within zip
-			bucket = b
+			rpt.SetStatus(file.Name, "", fmt.Errorf("malformed zip - expecting format bucket/<config.json>"))
+			continue
 		}
+		bucket, fileName := slc[0], slc[1]
 		switch fileName {
 		case bucketVersioningConfig:
 			reader, err := file.Open()
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			v, err := versioning.ParseConfig(io.LimitReader(reader, maxBucketVersioningConfigSize))
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			if _, ok := bucketMap[bucket]; !ok {
-				err = objectAPI.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
-				if err != nil {
+				if err = objectAPI.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{}); err != nil {
 					if _, ok := err.(BucketExists); !ok {
-						writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-						return
+						rpt.SetStatus(bucket, fileName, err)
+						continue
 					}
 				}
 				bucketMap[bucket] = struct{}{}
 			}

 			if globalSiteReplicationSys.isEnabled() && v.Suspended() {
-				writeErrorResponse(ctx, w, APIError{
-					Code:           "InvalidBucketState",
-					Description:    "Cluster replication is enabled for this site, so the versioning state cannot be changed.",
-					HTTPStatusCode: http.StatusConflict,
-				}, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("Cluster replication is enabled for this site, so the versioning state cannot be suspended."))
+				continue
 			}

 			if rcfg, _ := globalBucketObjectLockSys.Get(bucket); rcfg.LockEnabled && v.Suspended() {
-				writeErrorResponse(ctx, w, APIError{
-					Code:           "InvalidBucketState",
-					Description:    "An Object Lock configuration is present on this bucket, so the versioning state cannot be changed.",
-					HTTPStatusCode: http.StatusConflict,
-				}, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("An Object Lock configuration is present on this bucket, so the versioning state cannot be suspended."))
+				continue
 			}
 			if _, err := getReplicationConfig(ctx, bucket); err == nil && v.Suspended() {
-				writeErrorResponse(ctx, w, APIError{
-					Code:           "InvalidBucketState",
-					Description:    "A replication configuration is present on this bucket, so the versioning state cannot be changed.",
-					HTTPStatusCode: http.StatusConflict,
-				}, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("A replication configuration is present on this bucket, so the versioning state cannot be suspended."))
+				continue
 			}

 			configData, err := xml.Marshal(v)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}

 			if _, err = globalBucketMetadataSys.Update(ctx, bucket, bucketVersioningConfig, configData); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
 		}
 	}

 	for _, file := range zr.File {
 		reader, err := file.Open()
 		if err != nil {
-			writeErrorResponse(ctx, w, importError(ctx, err, file.Name, ""), r.URL)
-			return
+			rpt.SetStatus(file.Name, "", err)
+			continue
 		}
 		sz := file.FileInfo().Size()
 		slc := strings.Split(file.Name, slashSeparator)
 		if len(slc) != 2 { // expecting bucket/configfile in the zipfile
-			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-			return
-		}
-		b, fileName := slc[0], slc[1]
-		if bucket == "" { // use bucket requested in query parameters if specified. Otherwise default bucket name to directory name within zip
-			bucket = b
+			rpt.SetStatus(file.Name, "", fmt.Errorf("malformed zip - expecting format bucket/<config.json>"))
+			continue
 		}
+		bucket, fileName := slc[0], slc[1]
 		// create bucket if it does not exist yet.
 		if _, ok := bucketMap[bucket]; !ok {
-			err = objectAPI.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+			err = objectAPI.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 			if err != nil {
 				if _, ok := err.(BucketExists); !ok {
-					writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-					return
+					rpt.SetStatus(bucket, "", err)
+					continue
 				}
 			}
 			bucketMap[bucket] = struct{}{}
 		}
+		if _, ok := bucketMap[bucket]; !ok {
+			continue
+		}
 		switch fileName {
 		case bucketNotificationConfig:
-			config, err := event.ParseConfig(io.LimitReader(reader, sz), globalSite.Region, globalNotificationSys.targetList)
+			config, err := event.ParseConfig(io.LimitReader(reader, sz), globalSite.Region, globalEventNotifier.targetList)
 			if err != nil {
-				apiErr := errorCodes.ToAPIErr(ErrMalformedXML)
-				if event.IsEventError(err) {
-					apiErr = importError(ctx, err, file.Name, bucket)
-				}
-				writeErrorResponse(ctx, w, apiErr, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}

 			configData, err := xml.Marshal(config)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			if _, err = globalBucketMetadataSys.Update(ctx, bucket, bucketNotificationConfig, configData); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			rulesMap := config.ToRulesMap()
-			globalNotificationSys.AddRulesMap(bucket, rulesMap)
+			globalEventNotifier.AddRulesMap(bucket, rulesMap)
+			rpt.SetStatus(bucket, fileName, nil)
 		case bucketPolicyConfig:
 			// Error out if Content-Length is beyond allowed size.
 			if sz > maxBucketPolicySize {
-				writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrPolicyTooLarge), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf(ErrPolicyTooLarge.String()))
+				continue
 			}

 			bucketPolicyBytes, err := ioutil.ReadAll(io.LimitReader(reader, sz))
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			bucketPolicy, err := policy.ParseConfig(bytes.NewReader(bucketPolicyBytes), bucket)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			// Version in policy must not be empty
 			if bucketPolicy.Version == "" {
-				writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrMalformedPolicy), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf(ErrMalformedPolicy.String()))
+				continue
 			}

 			configData, err := json.Marshal(bucketPolicy)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, bucketPolicyConfig, configData)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
 			// Call site replication hook.
 			if err = globalSiteReplicationSys.BucketMetaHook(ctx, madmin.SRBucketMeta{
 				Type:      madmin.SRBucketMetaTypePolicy,
@@ -899,83 +908,79 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				Policy:    bucketPolicyBytes,
 				UpdatedAt: updatedAt,
 			}); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
-
 		case bucketLifecycleConfig:
 			bucketLifecycle, err := lifecycle.ParseLifecycleConfig(io.LimitReader(reader, sz))
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			// Validate the received bucket policy document
 			if err = bucketLifecycle.Validate(); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			// Validate the transition storage ARNs
 			if err = validateTransitionTier(bucketLifecycle); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			configData, err := xml.Marshal(bucketLifecycle)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			if _, err = globalBucketMetadataSys.Update(ctx, bucket, bucketLifecycleConfig, configData); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
 		case bucketSSEConfig:
 			// Parse bucket encryption xml
 			encConfig, err := validateBucketSSEConfig(io.LimitReader(reader, maxBucketSSEConfigSize))
 			if err != nil {
-				apiErr := APIError{
-					Code:           "MalformedXML",
-					Description:    fmt.Sprintf("%s (%s)", errorCodes[ErrMalformedXML].Description, err),
-					HTTPStatusCode: errorCodes[ErrMalformedXML].HTTPStatusCode,
-				}
-				writeErrorResponse(ctx, w, apiErr, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}

 			// Return error if KMS is not initialized
 			if GlobalKMS == nil {
-				writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrKMSNotConfigured), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s", errorCodes[ErrKMSNotConfigured].Description))
+				continue
 			}
 			kmsKey := encConfig.KeyID()
 			if kmsKey != "" {
 				kmsContext := kms.Context{"MinIO admin API": "ServerInfoHandler"} // Context for a test key operation
-				_, err := GlobalKMS.GenerateKey(kmsKey, kmsContext)
+				_, err := GlobalKMS.GenerateKey(ctx, kmsKey, kmsContext)
 				if err != nil {
 					if errors.Is(err, kes.ErrKeyNotFound) {
-						writeErrorResponse(ctx, w, importError(ctx, errKMSKeyNotFound, file.Name, bucket), r.URL)
-						return
+						rpt.SetStatus(bucket, fileName, errKMSKeyNotFound)
+						continue
 					}
-					writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-					return
+					rpt.SetStatus(bucket, fileName, err)
+					continue
 				}
 			}

 			configData, err := xml.Marshal(encConfig)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			// Store the bucket encryption configuration in the object layer
 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, bucketSSEConfig, configData)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)

 			// Call site replication hook.
 			//
@@ -988,29 +993,30 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				SSEConfig: &cfgStr,
 				UpdatedAt: updatedAt,
 			}); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 		case bucketTaggingConfig:
 			tags, err := tags.ParseBucketXML(io.LimitReader(reader, sz))
 			if err != nil {
-				apiErr := errorCodes.ToAPIErrWithErr(ErrMalformedXML, fmt.Errorf("error importing %s with %w", file.Name, err))
-				writeErrorResponse(ctx, w, apiErr, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}

 			configData, err := xml.Marshal(tags)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, bucketTaggingConfig, configData)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
+
 			// Call site replication hook.
 			//
 			// We encode the xml bytes as base64 to ensure there are no encoding
@@ -1022,32 +1028,33 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				Tags:      &cfgStr,
 				UpdatedAt: updatedAt,
 			}); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 		case bucketQuotaConfigFile:
 			data, err := ioutil.ReadAll(reader)
 			if err != nil {
-				writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			quotaConfig, err := parseBucketQuota(bucket, data)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}

 			if quotaConfig.Type == "fifo" {
-				writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("Detected older 'fifo' quota config, 'fifo' feature is removed and not supported anymore"))
+				continue
 			}

 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, bucketQuotaConfigFile, data)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)

 			bucketMeta := madmin.SRBucketMeta{
 				Type:      madmin.SRBucketMetaTypeQuotaConfig,
@@ -1061,9 +1068,93 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *

 			// Call site replication hook.
 			if err = globalSiteReplicationSys.BucketMetaHook(ctx, bucketMeta); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 		}
 	}
+
+	rptData, err := json.Marshal(rpt.BucketMetaImportErrs)
+	if err != nil {
+		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
+		return
+	}
+
+	writeSuccessResponseJSON(w, rptData)
+}
+
+// ReplicationDiffHandler - POST returns info on unreplicated versions for a remote target ARN
+// to the connected HTTP client. This is a MinIO only extension
+func (a adminAPIHandlers) ReplicationDiffHandler(w http.ResponseWriter, r *http.Request) {
+	ctx := newContext(r, w, "ReplicationDiff")
+	defer logger.AuditLog(ctx, w, r, mustGetClaimsFromToken(r))
+
+	vars := mux.Vars(r)
+	bucket := vars["bucket"]
+
+	if globalIsGateway {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
+		return
+	}
+
+	// Get current object layer instance.
+	objectAPI := newObjectLayerFn()
+	if objectAPI == nil {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
+		return
+	}
+	// check if user has permissions to perform this operation
+	if s3Error := checkRequestAuthType(ctx, r, policy.ListBucketVersionsAction, bucket, ""); s3Error != ErrNone {
+		writeErrorResponse(ctx, w, errorCodes.ToAPIErr(s3Error), r.URL)
+		return
+	}
+
+	// Check if bucket exists.
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
+		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
+		return
+	}
+	opts := extractReplicateDiffOpts(r.Form)
+	if opts.ARN != "" {
+		tgt := globalBucketTargetSys.GetRemoteBucketTargetByArn(ctx, bucket, opts.ARN)
+		if tgt.Empty() {
+			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrInvalidRequest, fmt.Errorf("invalid arn : '%s'", opts.ARN)), r.URL)
+			return
+		}
+	}
+
+	keepAliveTicker := time.NewTicker(500 * time.Millisecond)
+	defer keepAliveTicker.Stop()
+
+	diffCh, err := getReplicationDiff(ctx, objectAPI, bucket, opts)
+	if err != nil {
+		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
+		return
+	}
+	enc := json.NewEncoder(w)
+	for {
+		select {
+		case entry, ok := <-diffCh:
+			if !ok {
+				return
+			}
+			if err := enc.Encode(entry); err != nil {
+				return
+			}
+			if len(diffCh) == 0 {
+				// Flush if nothing is queued
+				w.(http.Flusher).Flush()
+			}
+		case <-keepAliveTicker.C:
+			if len(diffCh) > 0 {
+				continue
+			}
+			if _, err := w.Write([]byte(" ")); err != nil {
+				return
+			}
+			w.(http.Flusher).Flush()
+		case <-ctx.Done():
+			return
+		}
+	}
 }
@@ -30,6 +30,10 @@ import (
 	iampolicy "github.com/minio/pkg/iam/policy"
 )

+// validateAdminReq will validate request against and return whether it is allowed.
+// If any of the supplied actions are allowed it will be successful.
+// If nil ObjectLayer is returned, the operation is not permitted.
+// When nil ObjectLayer has been returned an error has always been sent to w.
 func validateAdminReq(ctx context.Context, w http.ResponseWriter, r *http.Request, actions ...iampolicy.AdminAction) (ObjectLayer, auth.Credentials) {
 	// Get current object layer instance.
 	objectAPI := newObjectLayerFn()
@@ -41,11 +45,16 @@ func validateAdminReq(ctx context.Context, w http.ResponseWriter, r *http.Reques
 	for _, action := range actions {
 		// Validate request signature.
 		cred, adminAPIErr := checkAdminRequestAuth(ctx, r, action, "")
-		if adminAPIErr != ErrNone {
+		switch adminAPIErr {
+		case ErrNone:
+			return objectAPI, cred
+		case ErrAccessDenied:
+			// Try another
+			continue
+		default:
 			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(adminAPIErr), r.URL)
 			return nil, cred
 		}
-		return objectAPI, cred
 	}
 	writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAccessDenied), r.URL)
 	return nil, auth.Credentials{}
@@ -177,6 +177,11 @@ func (a adminAPIHandlers) SetConfigKVHandler(w http.ResponseWriter, r *http.Requ
 }

 // GetConfigKVHandler - GET /minio/admin/v3/get-config-kv?key={key}
+//
+// `key` can be one of three forms:
+// 1. `subsys:target` -> request for config of a single subsystem and target pair.
+// 2. `subsys:` -> request for config of a single subsystem and the default target.
+// 3. `subsys` -> request for config of all targets for the given subsystem.
 func (a adminAPIHandlers) GetConfigKVHandler(w http.ResponseWriter, r *http.Request) {
 	ctx := newContext(r, w, "GetConfigKV")

@@ -189,15 +194,34 @@ func (a adminAPIHandlers) GetConfigKVHandler(w http.ResponseWriter, r *http.Requ

 	cfg := globalServerConfig.Clone()
 	vars := mux.Vars(r)
-	buf := &bytes.Buffer{}
-	cw := config.NewConfigWriteTo(cfg, vars["key"])
-	if _, err := cw.WriteTo(buf); err != nil {
+	key := vars["key"]
+
+	var subSys, target string
+	{
+		ws := strings.SplitN(key, madmin.SubSystemSeparator, 2)
+		subSys = ws[0]
+		if len(ws) == 2 {
+			if ws[1] == "" {
+				target = madmin.Default
+			} else {
+				target = ws[1]
+			}
+		}
+	}
+
+	subSysConfigs, err := cfg.GetSubsysInfo(subSys, target)
+	if err != nil {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
 	}

+	var s strings.Builder
+	for _, subSysConfig := range subSysConfigs {
+		subSysConfig.AddString(&s, false)
+	}
+
 	password := cred.SecretKey
-	econfigData, err := madmin.EncryptData(password, buf.Bytes())
+	econfigData, err := madmin.EncryptData(password, []byte(s.String()))
 	if err != nil {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
@@ -427,42 +451,30 @@ func (a adminAPIHandlers) GetConfigHandler(w http.ResponseWriter, r *http.Reques
 		count += len(cfg[hkv.Key])
 	}
 	for _, hkv := range hkvs {
-		v := cfg[hkv.Key]
-		for target, kv := range v {
-			off := kv.Get(config.Enable) == config.EnableOff
+		// We ignore the error below, as we cannot get one.
+		cfgSubsysItems, _ := cfg.GetSubsysInfo(hkv.Key, "")
+
+		for _, item := range cfgSubsysItems {
+			off := item.Config.Get(config.Enable) == config.EnableOff
 			switch hkv.Key {
 			case config.EtcdSubSys:
-				off = !etcd.Enabled(kv)
+				off = !etcd.Enabled(item.Config)
 			case config.CacheSubSys:
-				off = !cache.Enabled(kv)
+				off = !cache.Enabled(item.Config)
 			case config.StorageClassSubSys:
-				off = !storageclass.Enabled(kv)
+				off = !storageclass.Enabled(item.Config)
 			case config.PolicyPluginSubSys:
-				off = !polplugin.Enabled(kv)
+				off = !polplugin.Enabled(item.Config)
 			case config.IdentityOpenIDSubSys:
-				off = !openid.Enabled(kv)
+				off = !openid.Enabled(item.Config)
 			case config.IdentityLDAPSubSys:
-				off = !xldap.Enabled(kv)
+				off = !xldap.Enabled(item.Config)
 			case config.IdentityTLSSubSys:
 				off = !globalSTSTLSConfig.Enabled
 			case config.IdentityPluginSubSys:
-				off = !idplugin.Enabled(kv)
-			}
-			if off {
-				s.WriteString(config.KvComment)
-				s.WriteString(config.KvSpaceSeparator)
-			}
-			s.WriteString(hkv.Key)
-			if target != config.Default {
-				s.WriteString(config.SubSystemSeparator)
-				s.WriteString(target)
-			}
-			s.WriteString(config.KvSpaceSeparator)
-			s.WriteString(kv.String())
-			count--
-			if count > 0 {
-				s.WriteString(config.KvNewline)
+				off = !idplugin.Enabled(item.Config)
 			}
+			item.AddString(&s, off)
 		}
 	}

@@ -24,6 +24,8 @@ import (
 	"io"
 	"io/ioutil"
 	"net/http"
+	"strings"
+	"time"

 	"github.com/gorilla/mux"
 	"github.com/minio/madmin-go"
@@ -116,19 +118,38 @@ func (a adminAPIHandlers) SRPeerBucketOps(w http.ResponseWriter, r *http.Request
 		_, isLockEnabled := r.Form["lockEnabled"]
 		_, isVersioningEnabled := r.Form["versioningEnabled"]
 		_, isForceCreate := r.Form["forceCreate"]
-		opts := BucketOptions{
+		createdAtStr := strings.TrimSpace(r.Form.Get("createdAt"))
+		createdAt, cerr := time.Parse(time.RFC3339Nano, createdAtStr)
+		if cerr != nil {
+			createdAt = timeSentinel
+		}
+
+		opts := MakeBucketOptions{
 			Location:          r.Form.Get("location"),
 			LockEnabled:       isLockEnabled,
 			VersioningEnabled: isVersioningEnabled,
 			ForceCreate:       isForceCreate,
+			CreatedAt:         createdAt,
 		}
 		err = globalSiteReplicationSys.PeerBucketMakeWithVersioningHandler(ctx, bucket, opts)
 	case madmin.ConfigureReplBktOp:
 		err = globalSiteReplicationSys.PeerBucketConfigureReplHandler(ctx, bucket)
 	case madmin.DeleteBucketBktOp:
-		err = globalSiteReplicationSys.PeerBucketDeleteHandler(ctx, bucket, false)
+		_, noRecreate := r.Form["noRecreate"]
+		err = globalSiteReplicationSys.PeerBucketDeleteHandler(ctx, bucket, DeleteBucketOptions{
+			Force:      false,
+			NoRecreate: noRecreate,
+			SRDeleteOp: getSRBucketDeleteOp(true),
+		})
 	case madmin.ForceDeleteBucketBktOp:
-		err = globalSiteReplicationSys.PeerBucketDeleteHandler(ctx, bucket, true)
+		_, noRecreate := r.Form["noRecreate"]
+		err = globalSiteReplicationSys.PeerBucketDeleteHandler(ctx, bucket, DeleteBucketOptions{
+			Force:      true,
+			NoRecreate: noRecreate,
+			SRDeleteOp: getSRBucketDeleteOp(true),
+		})
+	case madmin.PurgeDeletedBucketOp:
+		globalSiteReplicationSys.purgeDeletedBucket(ctx, objectAPI, bucket)
 	}
 	if err != nil {
 		logger.LogIf(ctx, err)
@@ -436,6 +457,7 @@ func getSRStatusOptions(r *http.Request) (opts madmin.SRStatusOptions) {
 	opts.Users = q.Get("users") == "true"
 	opts.Entity = madmin.GetSREntityType(q.Get("entity"))
 	opts.EntityValue = q.Get("entityvalue")
+	opts.ShowDeleted = q.Get("showDeleted") == "true"
 	return
 }

@@ -1177,7 +1177,7 @@ func (a adminAPIHandlers) AccountInfoHandler(w http.ResponseWriter, r *http.Requ
 			return buckets[i].Name < buckets[j].Name
 		})
 	} else {
-		buckets, err = objectAPI.ListBuckets(ctx)
+		buckets, err = objectAPI.ListBuckets(ctx, BucketOptions{})
 		if err != nil {
 			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 			return
@@ -1189,17 +1189,32 @@ func (a adminAPIHandlers) AccountInfoHandler(w http.ResponseWriter, r *http.Requ
 		// For derived credentials, check the parent user's permissions.
 		accountName = cred.ParentUser
 	}
-	policies, err := globalIAMSys.PolicyDBGet(accountName, false, cred.Groups...)
-	if err != nil {
-		logger.LogIf(ctx, err)
-		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
-		return
-	}

-	buf, err := json.MarshalIndent(globalIAMSys.GetCombinedPolicy(policies...), "", " ")
-	if err != nil {
-		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
-		return
+	var buf []byte
+	if accountName == globalActiveCred.AccessKey {
+		for _, policy := range iampolicy.DefaultPolicies {
+			if policy.Name == "consoleAdmin" {
+				buf, err = json.MarshalIndent(policy.Definition, "", " ")
+				if err != nil {
+					writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+					return
+				}
+				break
+			}
+		}
+	} else {
+		policies, err := globalIAMSys.PolicyDBGet(accountName, false, cred.Groups...)
+		if err != nil {
+			logger.LogIf(ctx, err)
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+
+		buf, err = json.MarshalIndent(globalIAMSys.GetCombinedPolicy(policies...), "", " ")
+		if err != nil {
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
 	}

 	acctInfo := madmin.AccountInfo{
@@ -1514,7 +1529,29 @@ func (a adminAPIHandlers) SetPolicyForUserOrGroup(w http.ResponseWriter, r *http
 		}
 	}

-	updatedAt, err := globalIAMSys.PolicyDBSet(ctx, entityName, policyName, isGroup)
+	// Validate that user or group exists.
+	if !isGroup {
+		if globalIAMSys.GetUsersSysType() == MinIOUsersSysType {
+			_, ok := globalIAMSys.GetUser(ctx, entityName)
+			if !ok {
+				writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errNoSuchUser), r.URL)
+				return
+			}
+		}
+	} else {
+		_, err := globalIAMSys.GetGroupDescription(entityName)
+		if err != nil {
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+	}
+
+	userType := regUser
+	if globalIAMSys.GetUsersSysType() == LDAPUsersSysType {
+		userType = stsUser
+	}
+
+	updatedAt, err := globalIAMSys.PolicyDBSet(ctx, entityName, policyName, userType, isGroup)
 	if err != nil {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
@@ -1524,6 +1561,7 @@ func (a adminAPIHandlers) SetPolicyForUserOrGroup(w http.ResponseWriter, r *http
 		Type: madmin.SRIAMItemPolicyMapping,
 		PolicyMapping: &madmin.SRPolicyMapping{
 			UserOrGroup: entityName,
+			UserType:    int(userType),
 			IsGroup:     isGroup,
 			Policy:      policyName,
 		},
@@ -1543,6 +1581,7 @@ const (
 	groupPolicyMappingsFile    = "group_mappings.json"
 	stsUserPolicyMappingsFile  = "stsuser_mappings.json"
 	stsGroupPolicyMappingsFile = "stsgroup_mappings.json"
+	iamAssetsDir               = "iam-assets"
 )

 // ExportIAMHandler - exports all iam info as a zipped file
@@ -1553,7 +1592,6 @@ func (a adminAPIHandlers) ExportIAM(w http.ResponseWriter, r *http.Request) {
 	// Get current object layer instance.
 	objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.ExportIAMAction)
 	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
 		return
 	}
 	// Initialize a zip writer which will provide a zipped content
@@ -1595,8 +1633,9 @@ func (a adminAPIHandlers) ExportIAM(w http.ResponseWriter, r *http.Request) {
 		stsUserPolicyMappingsFile,
 		stsGroupPolicyMappingsFile,
 	}
-	for _, iamFile := range iamFiles {
-		switch iamFile {
+	for _, f := range iamFiles {
+		iamFile := pathJoin(iamAssetsDir, f)
+		switch f {
 		case allPoliciesFile:
 			allPolicies, err := globalIAMSys.ListPolicies(ctx, "")
 			if err != nil {
@@ -1674,11 +1713,6 @@ func (a adminAPIHandlers) ExportIAM(w http.ResponseWriter, r *http.Request) {
 			}
 			svcAccts := make(map[string]madmin.SRSvcAccCreate)
 			for user, acc := range serviceAccounts {
-				if user == siteReplicatorSvcAcc {
-					// skip the site replicate svc account as it should be created automatically if
-					// site replication is enabled.
-					continue
-				}
 				claims, err := globalIAMSys.GetClaimsForSvcAcc(ctx, acc.Credentials.AccessKey)
 				if err != nil {
 					writeErrorResponse(ctx, w, exportError(ctx, err, iamFile, ""), r.URL)
@@ -1827,7 +1861,8 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {
 	}
 	// import policies first
 	{
-		f, err := zr.Open(allPoliciesFile)
+
+		f, err := zr.Open(pathJoin(iamAssetsDir, allPoliciesFile))
 		switch {
 		case errors.Is(err, os.ErrNotExist):
 		case err != nil:
@@ -1862,7 +1897,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {

 	// import users
 	{
-		f, err := zr.Open(allUsersFile)
+		f, err := zr.Open(pathJoin(iamAssetsDir, allUsersFile))
 		switch {
 		case errors.Is(err, os.ErrNotExist):
 		case err != nil:
@@ -1939,7 +1974,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {

 	// import groups
 	{
-		f, err := zr.Open(allGroupsFile)
+		f, err := zr.Open(pathJoin(iamAssetsDir, allGroupsFile))
 		switch {
 		case errors.Is(err, os.ErrNotExist):
 		case err != nil:
@@ -1977,7 +2012,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {

 	// import service accounts
 	{
-		f, err := zr.Open(allSvcAcctsFile)
+		f, err := zr.Open(pathJoin(iamAssetsDir, allSvcAcctsFile))
 		switch {
 		case errors.Is(err, os.ErrNotExist):
 		case err != nil:
@@ -2073,7 +2108,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {

 	// import user policy mappings
 	{
-		f, err := zr.Open(userPolicyMappingsFile)
+		f, err := zr.Open(pathJoin(iamAssetsDir, userPolicyMappingsFile))
 		switch {
 		case errors.Is(err, os.ErrNotExist):
 		case err != nil:
@@ -2102,7 +2137,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {
 					writeErrorResponseJSON(ctx, w, importError(ctx, errIAMActionNotAllowed, userPolicyMappingsFile, u), r.URL)
 					return
 				}
-				if _, err := globalIAMSys.PolicyDBSet(ctx, u, pm.Policies, false); err != nil {
+				if _, err := globalIAMSys.PolicyDBSet(ctx, u, pm.Policies, regUser, false); err != nil {
 					writeErrorResponseJSON(ctx, w, importError(ctx, err, userPolicyMappingsFile, u), r.URL)
 					return
 				}
@@ -2112,7 +2147,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {

 	// import group policy mappings
 	{
-		f, err := zr.Open(groupPolicyMappingsFile)
+		f, err := zr.Open(pathJoin(iamAssetsDir, groupPolicyMappingsFile))
 		switch {
 		case errors.Is(err, os.ErrNotExist):
 		case err != nil:
@@ -2131,7 +2166,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {
 				return
 			}
 			for g, pm := range grpPolicyMap {
-				if _, err := globalIAMSys.PolicyDBSet(ctx, g, pm.Policies, true); err != nil {
+				if _, err := globalIAMSys.PolicyDBSet(ctx, g, pm.Policies, unknownIAMUserType, true); err != nil {
 					writeErrorResponseJSON(ctx, w, importError(ctx, err, groupPolicyMappingsFile, g), r.URL)
 					return
 				}
@@ -2141,7 +2176,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {

 	// import sts user policy mappings
 	{
-		f, err := zr.Open(stsUserPolicyMappingsFile)
+		f, err := zr.Open(pathJoin(iamAssetsDir, stsUserPolicyMappingsFile))
 		switch {
 		case errors.Is(err, os.ErrNotExist):
 		case err != nil:
@@ -2170,7 +2205,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {
 					writeErrorResponseJSON(ctx, w, importError(ctx, errIAMActionNotAllowed, stsUserPolicyMappingsFile, u), r.URL)
 					return
 				}
-				if _, err := globalIAMSys.PolicyDBSet(ctx, u, pm.Policies, false); err != nil {
+				if _, err := globalIAMSys.PolicyDBSet(ctx, u, pm.Policies, stsUser, false); err != nil {
 					writeErrorResponseJSON(ctx, w, importError(ctx, err, stsUserPolicyMappingsFile, u), r.URL)
 					return
 				}
@@ -2180,7 +2215,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {

 	// import sts group policy mappings
 	{
-		f, err := zr.Open(stsGroupPolicyMappingsFile)
+		f, err := zr.Open(pathJoin(iamAssetsDir, stsGroupPolicyMappingsFile))
 		switch {
 		case errors.Is(err, os.ErrNotExist):
 		case err != nil:
@@ -2199,7 +2234,7 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {
 				return
 			}
 			for g, pm := range grpPolicyMap {
-				if _, err := globalIAMSys.PolicyDBSet(ctx, g, pm.Policies, true); err != nil {
+				if _, err := globalIAMSys.PolicyDBSet(ctx, g, pm.Policies, unknownIAMUserType, true); err != nil {
 					writeErrorResponseJSON(ctx, w, importError(ctx, err, stsGroupPolicyMappingsFile, g), r.URL)
 					return
 				}
@@ -71,16 +71,6 @@ const (
 	mgmtForceStop   = "forceStop"
 )

-func updateServer(u *url.URL, sha256Sum []byte, lrTime time.Time, releaseInfo string, mode string) (us madmin.ServerUpdateStatus, err error) {
-	if err = doUpdate(u, lrTime, sha256Sum, releaseInfo, mode); err != nil {
-		return us, err
-	}
-
-	us.CurrentVersion = Version
-	us.UpdatedVersion = lrTime.Format(minioReleaseTagTimeLayout)
-	return us, nil
-}
-
 // ServerUpdateHandler - POST /minio/admin/v3/update?updateURL={updateURL}
 // ----------
 // updates all minio servers and restarts them gracefully.
@@ -152,7 +142,16 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 		return
 	}

-	for _, nerr := range globalNotificationSys.ServerUpdate(ctx, u, sha256Sum, lrTime, releaseInfo) {
+	// Download Binary Once
+	reader, err := downloadBinary(u, mode)
+	if err != nil {
+		logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	// Push binary to other servers
+	for _, nerr := range globalNotificationSys.VerifyBinary(ctx, u, sha256Sum, releaseInfo, reader) {
 		if nerr.Err != nil {
 			err := AdminError{
 				Code:       AdminUpdateApplyFailure,
@@ -166,13 +165,39 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 		}
 	}

-	updateStatus, err := updateServer(u, sha256Sum, lrTime, releaseInfo, mode)
+	err = verifyBinary(u, sha256Sum, releaseInfo, mode, reader)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
 	}

+	for _, nerr := range globalNotificationSys.CommitBinary(ctx) {
+		if nerr.Err != nil {
+			err := AdminError{
+				Code:       AdminUpdateApplyFailure,
+				Message:    nerr.Err.Error(),
+				StatusCode: http.StatusInternalServerError,
+			}
+			logger.GetReqInfo(ctx).SetTags("peerAddress", nerr.Host.String())
+			logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+	}
+
+	err = commitBinary()
+	if err != nil {
+		logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	updateStatus := madmin.ServerUpdateStatus{
+		CurrentVersion: Version,
+		UpdatedVersion: lrTime.Format(minioReleaseTagTimeLayout),
+	}
+
 	// Marshal API response
 	jsonBytes, err := json.Marshal(updateStatus)
 	if err != nil {
@@ -373,8 +398,21 @@ func (a adminAPIHandlers) MetricsHandler(w http.ResponseWriter, r *http.Request)
 	} else {
 		types = madmin.MetricsAll
 	}
+
+	disks := strings.Split(r.Form.Get("disks"), ",")
+	byDisk := strings.EqualFold(r.Form.Get("by-disk"), "true")
+	var diskMap map[string]struct{}
+	if len(disks) > 0 && disks[0] != "" {
+		diskMap = make(map[string]struct{}, len(disks))
+		for _, k := range disks {
+			if k != "" {
+				diskMap[k] = struct{}{}
+			}
+		}
+	}
+
 	hosts := strings.Split(r.Form.Get("hosts"), ",")
-	byhost := strings.EqualFold(r.Form.Get("by-host"), "true")
+	byHost := strings.EqualFold(r.Form.Get("by-host"), "true")
 	var hostMap map[string]struct{}
 	if len(hosts) > 0 && hosts[0] != "" {
 		hostMap = make(map[string]struct{}, len(hosts))
@@ -384,6 +422,7 @@ func (a adminAPIHandlers) MetricsHandler(w http.ResponseWriter, r *http.Request)
 			}
 		}
 	}
+
 	done := ctx.Done()
 	ticker := time.NewTicker(interval)
 	defer ticker.Stop()
@@ -391,17 +430,20 @@ func (a adminAPIHandlers) MetricsHandler(w http.ResponseWriter, r *http.Request)

 	for n > 0 {
 		var m madmin.RealtimeMetrics
-		mLocal := collectLocalMetrics(types, hostMap)
+		mLocal := collectLocalMetrics(types, hostMap, diskMap)
 		m.Merge(&mLocal)

 		// Allow half the interval for collecting remote...
 		cctx, cancel := context.WithTimeout(ctx, interval/2)
-		mRemote := collectRemoteMetrics(cctx, types, hostMap)
+		mRemote := collectRemoteMetrics(cctx, types, hostMap, diskMap)
 		cancel()
 		m.Merge(&mRemote)
-		if !byhost {
+		if !byHost {
 			m.ByHost = nil
 		}
+		if !byDisk {
+			m.ByDisk = nil
+		}

 		m.Final = n <= 1
 		// Marshal API response
@@ -1149,17 +1191,17 @@ func (a adminAPIHandlers) NetperfHandler(w http.ResponseWriter, r *http.Request)
 	}
 }

-// SpeedtestHandler - Deprecated. See ObjectSpeedtestHandler
-func (a adminAPIHandlers) SpeedtestHandler(w http.ResponseWriter, r *http.Request) {
-	a.ObjectSpeedtestHandler(w, r)
+// SpeedtestHandler - Deprecated. See ObjectSpeedTestHandler
+func (a adminAPIHandlers) SpeedTestHandler(w http.ResponseWriter, r *http.Request) {
+	a.ObjectSpeedTestHandler(w, r)
 }

-// ObjectSpeedtestHandler - reports maximum speed of a cluster by performing PUT and
+// ObjectSpeedTestHandler - reports maximum speed of a cluster by performing PUT and
 // GET operations on the server, supports auto tuning by default by automatically
 // increasing concurrency and stopping when we have reached the limits on the
 // system.
-func (a adminAPIHandlers) ObjectSpeedtestHandler(w http.ResponseWriter, r *http.Request) {
-	ctx := newContext(r, w, "ObjectSpeedtestHandler")
+func (a adminAPIHandlers) ObjectSpeedTestHandler(w http.ResponseWriter, r *http.Request) {
+	ctx := newContext(r, w, "ObjectSpeedTestHandler")

 	defer logger.AuditLog(ctx, w, r, mustGetClaimsFromToken(r))

@@ -1176,8 +1218,9 @@ func (a adminAPIHandlers) ObjectSpeedtestHandler(w http.ResponseWriter, r *http.
 	sizeStr := r.Form.Get(peerRESTSize)
 	durationStr := r.Form.Get(peerRESTDuration)
 	concurrentStr := r.Form.Get(peerRESTConcurrent)
+	storageClass := strings.TrimSpace(r.Form.Get(peerRESTStorageClass))
+	customBucket := strings.TrimSpace(r.Form.Get(peerRESTBucket))
 	autotune := r.Form.Get("autotune") == "true"
-	storageClass := r.Form.Get("storage-class")

 	size, err := strconv.Atoi(sizeStr)
 	if err != nil {
@@ -1189,16 +1232,14 @@ func (a adminAPIHandlers) ObjectSpeedtestHandler(w http.ResponseWriter, r *http.
 		concurrent = 32
 	}

-	if runtime.GOMAXPROCS(0) < concurrent {
-		concurrent = runtime.GOMAXPROCS(0)
-	}
-
 	duration, err := time.ParseDuration(durationStr)
 	if err != nil {
 		duration = time.Second * 10
 	}

-	sufficientCapacity, canAutotune, capacityErrMsg := validateObjPerfOptions(ctx, objectAPI, concurrent, size, autotune)
+	storageInfo, _ := objectAPI.StorageInfo(ctx)
+
+	sufficientCapacity, canAutotune, capacityErrMsg := validateObjPerfOptions(ctx, storageInfo, concurrent, size, autotune)
 	if !sufficientCapacity {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, AdminError{
 			Code:       "XMinioSpeedtestInsufficientCapacity",
@@ -1212,15 +1253,23 @@ func (a adminAPIHandlers) ObjectSpeedtestHandler(w http.ResponseWriter, r *http.
 		autotune = false
 	}

-	bucketExists, err := makeObjectPerfBucket(ctx, objectAPI)
-	if err != nil {
-		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
-		return
+	if customBucket == "" {
+		customBucket = globalObjectPerfBucket
+
+		bucketExists, err := makeObjectPerfBucket(ctx, objectAPI, customBucket)
+		if err != nil {
+			writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
+			return
+		}
+
+		if !bucketExists {
+			defer deleteObjectPerfBucket(objectAPI)
+		}
 	}

-	if !bucketExists {
-		defer deleteObjectPerfBucket(objectAPI)
-	}
+	defer objectAPI.DeleteObject(ctx, customBucket, speedTest+SlashSeparator, ObjectOptions{
+		DeletePrefix: true,
+	})

 	// Freeze all incoming S3 API calls before running speedtest.
 	globalNotificationSys.ServiceFreeze(ctx, true)
@@ -1232,15 +1281,32 @@ func (a adminAPIHandlers) ObjectSpeedtestHandler(w http.ResponseWriter, r *http.
 	defer keepAliveTicker.Stop()

 	enc := json.NewEncoder(w)
-	ch := objectSpeedTest(ctx, speedTestOpts{size, concurrent, duration, autotune, storageClass})
+	ch := objectSpeedTest(ctx, speedTestOpts{
+		objectSize:       size,
+		concurrencyStart: concurrent,
+		duration:         duration,
+		autotune:         autotune,
+		storageClass:     storageClass,
+		bucketName:       customBucket,
+	})
+	var prevResult madmin.SpeedTestResult
 	for {
 		select {
 		case <-ctx.Done():
 			return
 		case <-keepAliveTicker.C:
-			// Write a blank entry to prevent client from disconnecting
-			if err := enc.Encode(madmin.SpeedTestResult{}); err != nil {
-				return
+			// if previous result is set keep writing the
+			// previous result back to the client
+			if prevResult.Version != "" {
+				if err := enc.Encode(prevResult); err != nil {
+					return
+				}
+			} else {
+				// first result is not yet obtained, keep writing
+				// empty entry to prevent client from disconnecting.
+				if err := enc.Encode(madmin.SpeedTestResult{}); err != nil {
+					return
+				}
 			}
 			w.(http.Flusher).Flush()
 		case result, ok := <-ch:
@@ -1250,14 +1316,14 @@ func (a adminAPIHandlers) ObjectSpeedtestHandler(w http.ResponseWriter, r *http.
 			if err := enc.Encode(result); err != nil {
 				return
 			}
+			prevResult = result
 			w.(http.Flusher).Flush()
 		}
 	}
 }

-func makeObjectPerfBucket(ctx context.Context, objectAPI ObjectLayer) (bucketExists bool, err error) {
-	err = objectAPI.MakeBucketWithLocation(ctx, globalObjectPerfBucket, BucketOptions{})
-	if err != nil {
+func makeObjectPerfBucket(ctx context.Context, objectAPI ObjectLayer, bucketName string) (bucketExists bool, err error) {
+	if err = objectAPI.MakeBucketWithLocation(ctx, bucketName, MakeBucketOptions{}); err != nil {
 		if _, ok := err.(BucketExists); !ok {
 			// Only BucketExists error can be ignored.
 			return false, err
@@ -1271,11 +1337,11 @@ func deleteObjectPerfBucket(objectAPI ObjectLayer) {
 	objectAPI.DeleteBucket(context.Background(), globalObjectPerfBucket, DeleteBucketOptions{
 		Force:      true,
 		NoRecreate: true,
+		SRDeleteOp: getSRBucketDeleteOp(globalSiteReplicationSys.isEnabled()),
 	})
 }

-func validateObjPerfOptions(ctx context.Context, objectAPI ObjectLayer, concurrent int, size int, autotune bool) (sufficientCapacity bool, canAutotune bool, capacityErrMsg string) {
-	storageInfo, _ := objectAPI.StorageInfo(ctx)
+func validateObjPerfOptions(ctx context.Context, storageInfo madmin.StorageInfo, concurrent int, size int, autotune bool) (bool, bool, string) {
 	capacityNeeded := uint64(concurrent * size)
 	capacity := GetTotalUsableCapacityFree(storageInfo.Disks, storageInfo)

@@ -1600,7 +1666,7 @@ func (a adminAPIHandlers) KMSCreateKeyHandler(w http.ResponseWriter, r *http.Req
 		return
 	}

-	if err := GlobalKMS.CreateKey(r.Form.Get("key-id")); err != nil {
+	if err := GlobalKMS.CreateKey(ctx, r.Form.Get("key-id")); err != nil {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
 	}
@@ -1622,7 +1688,7 @@ func (a adminAPIHandlers) KMSStatusHandler(w http.ResponseWriter, r *http.Reques
 		return
 	}

-	stat, err := GlobalKMS.Stat()
+	stat, err := GlobalKMS.Stat(ctx)
 	if err != nil {
 		writeCustomErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInternalError), err.Error(), r.URL)
 		return
@@ -1661,7 +1727,7 @@ func (a adminAPIHandlers) KMSKeyStatusHandler(w http.ResponseWriter, r *http.Req
 		return
 	}

-	stat, err := GlobalKMS.Stat()
+	stat, err := GlobalKMS.Stat(ctx)
 	if err != nil {
 		writeCustomErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInternalError), err.Error(), r.URL)
 		return
@@ -1677,7 +1743,7 @@ func (a adminAPIHandlers) KMSKeyStatusHandler(w http.ResponseWriter, r *http.Req

 	kmsContext := kms.Context{"MinIO admin API": "KMSKeyStatusHandler"} // Context for a test key operation
 	// 1. Generate a new key using the KMS.
-	key, err := GlobalKMS.GenerateKey(keyID, kmsContext)
+	key, err := GlobalKMS.GenerateKey(ctx, keyID, kmsContext)
 	if err != nil {
 		response.EncryptionErr = err.Error()
 		resp, err := json.Marshal(response)
@@ -1813,7 +1879,7 @@ func getServerInfo(ctx context.Context, r *http.Request) madmin.InfoMessage {
 		Mode:         string(mode),
 		Domain:       domain,
 		Region:       globalSite.Region,
-		SQSARN:       globalNotificationSys.GetARNList(false),
+		SQSARN:       globalEventNotifier.GetARNList(false),
 		DeploymentID: globalDeploymentID,
 		Buckets:      buckets,
 		Objects:      objects,
@@ -2184,15 +2250,13 @@ func (a adminAPIHandlers) HealthInfoHandler(w http.ResponseWriter, r *http.Reque
 	getAndWriteObjPerfInfo := func() {
 		if query.Get(string(madmin.HealthDataTypePerfObj)) == "true" {
 			concurrent := 32
-			if runtime.GOMAXPROCS(0) < concurrent {
-				concurrent = runtime.GOMAXPROCS(0)
-			}
+
+			storageInfo, _ := objectAPI.StorageInfo(ctx)

 			size := 64 * humanize.MiByte
 			autotune := true

-			sufficientCapacity, canAutotune, capacityErrMsg := validateObjPerfOptions(ctx, objectAPI, concurrent, size, autotune)
-
+			sufficientCapacity, canAutotune, capacityErrMsg := validateObjPerfOptions(ctx, storageInfo, concurrent, size, autotune)
 			if !sufficientCapacity {
 				healthInfo.Perf.Error = capacityErrMsg
 				partialWrite(healthInfo)
@@ -2203,9 +2267,9 @@ func (a adminAPIHandlers) HealthInfoHandler(w http.ResponseWriter, r *http.Reque
 				autotune = false
 			}

-			bucketExists, err := makeObjectPerfBucket(ctx, objectAPI)
+			bucketExists, err := makeObjectPerfBucket(ctx, objectAPI, globalObjectPerfBucket)
 			if err != nil {
-				healthInfo.Perf.Error = "Could not make object perf bucket: " + err.Error()
+				healthInfo.Perf.Error = "Unable to create bucket: " + err.Error()
 				partialWrite(healthInfo)
 				return
 			}
@@ -2215,7 +2279,7 @@ func (a adminAPIHandlers) HealthInfoHandler(w http.ResponseWriter, r *http.Reque
 			}

 			opts := speedTestOpts{
-				throughputSize:   size,
+				objectSize:       size,
 				concurrencyStart: concurrent,
 				duration:         10 * time.Second,
 				autotune:         autotune,
@@ -2527,7 +2591,7 @@ func fetchKMSStatus() madmin.KMS {
 		return kmsStat
 	}

-	stat, err := GlobalKMS.Stat()
+	stat, err := GlobalKMS.Stat(context.Background())
 	if err != nil {
 		kmsStat.Status = string(madmin.ItemOffline)
 		return kmsStat
@@ -2540,7 +2604,7 @@ func fetchKMSStatus() madmin.KMS {

 	kmsContext := kms.Context{"MinIO admin API": "ServerInfoHandler"} // Context for a test key operation
 	// 1. Generate a new key using the KMS.
-	key, err := GlobalKMS.GenerateKey("", kmsContext)
+	key, err := GlobalKMS.GenerateKey(context.Background(), "", kmsContext)
 	if err != nil {
 		kmsStat.Encrypt = fmt.Sprintf("Encryption failed: %v", err)
 	} else {
@@ -2621,6 +2685,100 @@ func checkConnection(endpointStr string, timeout time.Duration) error {
 	return nil
 }

+type clusterInfo struct {
+	DeploymentID string `json:"deployment_id"`
+	ClusterName  string `json:"cluster_name"`
+	UsedCapacity uint64 `json:"used_capacity"`
+	Info         struct {
+		MinioVersion    string `json:"minio_version"`
+		PoolsCount      int    `json:"pools_count"`
+		ServersCount    int    `json:"servers_count"`
+		DrivesCount     int    `json:"drives_count"`
+		BucketsCount    uint64 `json:"buckets_count"`
+		ObjectsCount    uint64 `json:"objects_count"`
+		TotalDriveSpace uint64 `json:"total_drive_space"`
+		UsedDriveSpace  uint64 `json:"used_drive_space"`
+	} `json:"info"`
+}
+
+func embedFileInZip(zipWriter *zip.Writer, name string, data []byte) error {
+	// Send profiling data to zip as file
+	header, zerr := zip.FileInfoHeader(dummyFileInfo{
+		name:    name,
+		size:    int64(len(data)),
+		mode:    0o600,
+		modTime: UTCNow(),
+		isDir:   false,
+		sys:     nil,
+	})
+	if zerr != nil {
+		return zerr
+	}
+	header.Method = zip.Deflate
+	zwriter, zerr := zipWriter.CreateHeader(header)
+	if zerr != nil {
+		return zerr
+	}
+	_, err := io.Copy(zwriter, bytes.NewReader(data))
+	return err
+}
+
+// appendClusterMetaInfoToZip gets information of the current cluster and embedded
+// it in the passed zipwriter, This is not a critical function and it is allowed
+// to fail with a ten seconds timeout.
+func appendClusterMetaInfoToZip(ctx context.Context, zipWriter *zip.Writer) {
+	objectAPI := newObjectLayerFn()
+	if objectAPI == nil {
+		return
+	}
+
+	// Add a ten seconds timeout because getting profiling data
+	// is critical for debugging, in contrary to getting cluster info
+	ctx, cancel := context.WithTimeout(ctx, 10*time.Second)
+	defer cancel()
+
+	resultCh := make(chan clusterInfo)
+	go func() {
+		ci := clusterInfo{}
+		ci.Info.PoolsCount = len(globalEndpoints)
+		ci.Info.ServersCount = len(globalEndpoints.Hostnames())
+		ci.Info.MinioVersion = Version
+
+		si, _ := objectAPI.StorageInfo(ctx)
+
+		ci.Info.DrivesCount = len(si.Disks)
+		for _, disk := range si.Disks {
+			ci.Info.TotalDriveSpace += disk.TotalSpace
+			ci.Info.UsedDriveSpace += disk.UsedSpace
+		}
+
+		dataUsageInfo, _ := loadDataUsageFromBackend(ctx, objectAPI)
+
+		ci.UsedCapacity = dataUsageInfo.ObjectsTotalSize
+		ci.Info.BucketsCount = dataUsageInfo.BucketsCount
+		ci.Info.ObjectsCount = dataUsageInfo.ObjectsTotalCount
+
+		ci.DeploymentID = globalDeploymentID
+		ci.ClusterName = fmt.Sprintf("%d-servers-%d-disks-%s", ci.Info.ServersCount, ci.Info.DrivesCount, ci.Info.MinioVersion)
+		resultCh <- ci
+	}()
+
+	select {
+	case <-ctx.Done():
+		return
+	case ci := <-resultCh:
+		out, err := json.MarshalIndent(ci, "", "   ")
+		if err != nil {
+			logger.LogIf(ctx, err)
+			return
+		}
+		err = embedFileInZip(zipWriter, "cluster.info", out)
+		if err != nil {
+			logger.LogIf(ctx, err)
+		}
+	}
+}
+
 // getRawDataer provides an interface for getting raw FS files.
 type getRawDataer interface {
 	GetRawData(ctx context.Context, volume, file string, fn func(r io.Reader, host string, disk string, filename string, info StatInfo) error) error
@@ -2640,7 +2798,8 @@ func (a adminAPIHandlers) InspectDataHandler(w http.ResponseWriter, r *http.Requ
 	}
 	defer logger.AuditLog(ctx, w, r, mustGetClaimsFromToken(r))

-	o, ok := newObjectLayerFn().(getRawDataer)
+	objLayer := newObjectLayerFn()
+	o, ok := objLayer.(getRawDataer)
 	if !ok {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
 		return
@@ -2763,6 +2922,8 @@ func (a adminAPIHandlers) InspectDataHandler(w http.ResponseWriter, r *http.Requ
 	}); err != nil {
 		logger.LogIf(ctx, err)
 	}
+
+	appendClusterMetaInfoToZip(ctx, zipWriter)
 }

 func createHostAnonymizerForFSMode() map[string]string {
@@ -822,7 +822,7 @@ func (h *healSequence) healBuckets(objAPI ObjectLayer, bucketsOnly bool) error {
 		return h.healBucket(objAPI, h.bucket, bucketsOnly)
 	}

-	buckets, err := objAPI.ListBuckets(h.ctx)
+	buckets, err := objAPI.ListBuckets(h.ctx, BucketOptions{})
 	if err != nil {
 		return errFnHealFromAPIErr(h.ctx, err)
 	}
@@ -202,6 +202,9 @@ func registerAdminRouter(router *mux.Router, enableConfigOps bool) {
 		// RemoveRemoteTargetHandler
 		adminRouter.Methods(http.MethodDelete).Path(adminVersion+"/remove-remote-target").HandlerFunc(
 			gz(httpTraceHdrs(adminAPI.RemoveRemoteTargetHandler))).Queries("bucket", "{bucket:.*}", "arn", "{arn:.*}")
+		// ReplicationDiff - MinIO extension API
+		adminRouter.Methods(http.MethodPost).Path(adminVersion+"/replication/diff").HandlerFunc(
+			gz(httpTraceHdrs(adminAPI.ReplicationDiffHandler))).Queries("bucket", "{bucket:.*}")

 		// Bucket migration operations
 		// ExportBucketMetaHandler
@@ -244,8 +247,8 @@ func registerAdminRouter(router *mux.Router, enableConfigOps bool) {
 				Queries("paths", "{paths:.*}").HandlerFunc(gz(httpTraceHdrs(adminAPI.ForceUnlockHandler)))
 		}

-		adminRouter.Methods(http.MethodPost).Path(adminVersion + "/speedtest").HandlerFunc(httpTraceHdrs(adminAPI.SpeedtestHandler))
-		adminRouter.Methods(http.MethodPost).Path(adminVersion + "/speedtest/object").HandlerFunc(httpTraceHdrs(adminAPI.ObjectSpeedtestHandler))
+		adminRouter.Methods(http.MethodPost).Path(adminVersion + "/speedtest").HandlerFunc(httpTraceHdrs(adminAPI.SpeedTestHandler))
+		adminRouter.Methods(http.MethodPost).Path(adminVersion + "/speedtest/object").HandlerFunc(httpTraceHdrs(adminAPI.ObjectSpeedTestHandler))
 		adminRouter.Methods(http.MethodPost).Path(adminVersion + "/speedtest/drive").HandlerFunc(httpTraceHdrs(adminAPI.DriveSpeedtestHandler))
 		adminRouter.Methods(http.MethodPost).Path(adminVersion + "/speedtest/net").HandlerFunc(httpTraceHdrs(adminAPI.NetperfHandler))

@@ -196,8 +196,9 @@ const (
 	ErrInvalidTagDirective
 	// Add new error codes here.

-	// SSE-S3 related API errors
+	// SSE-S3/SSE-KMS related API errors
 	ErrInvalidEncryptionMethod
+	ErrInvalidEncryptionKeyID

 	// Server-Side-Encryption (with Customer provided key) related API errors.
 	ErrInsecureSSECustomerRequest
@@ -887,7 +888,7 @@ var errorCodes = errorCodeMap{
 	},
 	ErrReplicationRemoteConnectionError: {
 		Code:           "XMinioAdminReplicationRemoteConnectionError",
-		Description:    "Remote service connection error - please check remote service credentials and target bucket",
+		Description:    "Remote service connection error",
 		HTTPStatusCode: http.StatusNotFound,
 	},
 	ErrReplicationBandwidthLimitError: {
@@ -1072,6 +1073,11 @@ var errorCodes = errorCodeMap{
 		Description:    "The encryption method specified is not supported",
 		HTTPStatusCode: http.StatusBadRequest,
 	},
+	ErrInvalidEncryptionKeyID: {
+		Code:           "InvalidRequest",
+		Description:    "The specified KMS KeyID contains unsupported characters",
+		HTTPStatusCode: http.StatusBadRequest,
+	},
 	ErrInsecureSSECustomerRequest: {
 		Code:           "InvalidRequest",
 		Description:    "Requests specifying Server Side Encryption with Customer provided keys must be made over a secure connection.",
@@ -1158,7 +1164,7 @@ var errorCodes = errorCodeMap{
 	// MinIO extensions.
 	ErrStorageFull: {
 		Code:           "XMinioStorageFull",
-		Description:    "Storage backend has reached its minimum free disk threshold. Please delete a few objects to proceed.",
+		Description:    "Storage backend has reached its minimum free drive threshold. Please delete a few objects to proceed.",
 		HTTPStatusCode: http.StatusInsufficientStorage,
 	},
 	ErrRequestBodyParse: {
@@ -1921,6 +1927,8 @@ func toAPIErrorCode(ctx context.Context, err error) (apiErr APIErrorCode) {
 		apiErr = ErrInvalidEncryptionParameters
 	case crypto.ErrInvalidEncryptionMethod:
 		apiErr = ErrInvalidEncryptionMethod
+	case crypto.ErrInvalidEncryptionKeyID:
+		apiErr = ErrInvalidEncryptionKeyID
 	case crypto.ErrInvalidCustomerAlgorithm:
 		apiErr = ErrInvalidSSECustomerAlgorithm
 	case crypto.ErrMissingCustomerKey:
@@ -2066,7 +2074,7 @@ func toAPIErrorCode(ctx context.Context, err error) (apiErr APIErrorCode) {
 		apiErr = ErrRemoteDestinationNotFoundError
 	case BucketRemoteTargetNotFound:
 		apiErr = ErrRemoteTargetNotFoundError
-	case BucketRemoteConnectionErr:
+	case RemoteTargetConnectionErr:
 		apiErr = ErrReplicationRemoteConnectionError
 	case BucketRemoteAlreadyExists:
 		apiErr = ErrBucketRemoteAlreadyExists
@@ -20,7 +20,6 @@ package cmd
 import (
 	"bytes"
 	"encoding/json"
-	"encoding/xml"
 	"fmt"
 	"net/http"
 	"net/url"
@@ -30,6 +29,8 @@ import (

 	"github.com/minio/minio/internal/crypto"
 	xhttp "github.com/minio/minio/internal/http"
+	"github.com/minio/minio/internal/logger"
+	xxml "github.com/minio/xxml"
 )

 // Returns a hexadecimal representation of time at the
@@ -64,9 +65,13 @@ func setCommonHeaders(w http.ResponseWriter) {
 // Encodes the response headers into XML format.
 func encodeResponse(response interface{}) []byte {
 	var bytesBuffer bytes.Buffer
-	bytesBuffer.WriteString(xml.Header)
-	e := xml.NewEncoder(&bytesBuffer)
-	e.Encode(response)
+	bytesBuffer.WriteString(xxml.Header)
+	buf, err := xxml.Marshal(response)
+	if err != nil {
+		logger.LogIf(GlobalContext, err)
+		return nil
+	}
+	bytesBuffer.Write(buf)
 	return bytesBuffer.Bytes()
 }

@@ -199,5 +204,12 @@ func setObjectHeaders(w http.ResponseWriter, objInfo ObjectInfo, rs *HTTPRangeSp
 		lc.SetPredictionHeaders(w, objInfo.ToLifecycleOpts())
 	}

+	if v, ok := objInfo.UserDefined[ReservedMetadataPrefix+"compression"]; ok {
+		if i := strings.LastIndexByte(v, '/'); i >= 0 {
+			v = v[i+1:]
+		}
+		w.Header()[xhttp.MinIOCompressed] = []string{v}
+	}
+
 	return nil
 }
@@ -89,7 +89,8 @@ type ListVersionsResponse struct {
 	IsTruncated bool

 	CommonPrefixes []CommonPrefix
-	Versions       []ObjectVersion
+	DeleteMarkers  []DeleteMarkerVersion `xml:"DeleteMarker,omitempty"`
+	Versions       []ObjectVersion       `xml:"Version,omitempty"`

 	// Encoding type used to encode object keys in the response.
 	EncodingType string `xml:"EncodingType,omitempty"`
@@ -247,50 +248,80 @@ type ObjectVersion struct {
 	Object
 	IsLatest  bool
 	VersionID string `xml:"VersionId"`
-
-	isDeleteMarker bool
 }

-// MarshalXML - marshal ObjectVersion
-func (o ObjectVersion) MarshalXML(e *xml.Encoder, start xml.StartElement) error {
-	if o.isDeleteMarker {
-		start.Name.Local = "DeleteMarker"
-	} else {
-		start.Name.Local = "Version"
+// DeleteMarkerVersion container for delete marker metadata
+type DeleteMarkerVersion struct {
+	Key          string
+	LastModified string // time string of format "2006-01-02T15:04:05.000Z"
+
+	// Owner of the object.
+	Owner Owner
+
+	IsLatest  bool
+	VersionID string `xml:"VersionId"`
+}
+
+// Metadata metadata items implemented to ensure XML marshaling works.
+type Metadata struct {
+	Items []struct {
+		Key   string
+		Value string
 	}
-
-	type objectVersionWrapper ObjectVersion
-	return e.EncodeElement(objectVersionWrapper(o), start)
 }

-// StringMap is a map[string]string
-type StringMap map[string]string
+// Set add items, duplicate items get replaced.
+func (s *Metadata) Set(k, v string) {
+	for i, item := range s.Items {
+		if item.Key == k {
+			s.Items[i] = struct {
+				Key   string
+				Value string
+			}{
+				Key:   k,
+				Value: v,
+			}
+			return
+		}
+	}
+	s.Items = append(s.Items, struct {
+		Key   string
+		Value string
+	}{
+		Key:   k,
+		Value: v,
+	})
+}
+
+type xmlKeyEntry struct {
+	XMLName xml.Name
+	Value   string `xml:",chardata"`
+}

 // MarshalXML - StringMap marshals into XML.
-func (s StringMap) MarshalXML(e *xml.Encoder, start xml.StartElement) error {
-	tokens := []xml.Token{start}
-
-	for key, value := range s {
-		t := xml.StartElement{}
-		t.Name = xml.Name{
-			Space: "",
-			Local: key,
-		}
-		tokens = append(tokens, t, xml.CharData(value), xml.EndElement{Name: t.Name})
+func (s *Metadata) MarshalXML(e *xml.Encoder, start xml.StartElement) error {
+	if s == nil {
+		return nil
 	}

-	tokens = append(tokens, xml.EndElement{
-		Name: start.Name,
-	})
+	if len(s.Items) == 0 {
+		return nil
+	}

-	for _, t := range tokens {
-		if err := e.EncodeToken(t); err != nil {
+	if err := e.EncodeToken(start); err != nil {
+		return err
+	}
+
+	for _, item := range s.Items {
+		if err := e.Encode(xmlKeyEntry{
+			XMLName: xml.Name{Local: item.Key},
+			Value:   item.Value,
+		}); err != nil {
 			return err
 		}
 	}

-	// flush to ensure tokens are written
-	return e.Flush()
+	return e.EncodeToken(start.End())
 }

 // Object container for object metadata
@@ -307,7 +338,7 @@ type Object struct {
 	StorageClass string

 	// UserMetadata user-defined metadata
-	UserMetadata StringMap `xml:"UserMetadata,omitempty"`
+	UserMetadata *Metadata `xml:"UserMetadata,omitempty"`
 }

 // CopyObjectResponse container returns ETag and LastModified of the successfully copied object
@@ -438,6 +469,8 @@ func generateListBucketsResponse(buckets []BucketInfo) ListBucketsResponse {
 // generates an ListBucketVersions response for the said bucket with other enumerated options.
 func generateListVersionsResponse(bucket, prefix, marker, versionIDMarker, delimiter, encodingType string, maxKeys int, resp ListObjectVersionsInfo) ListVersionsResponse {
 	versions := make([]ObjectVersion, 0, len(resp.Objects))
+	deleteMarkers := make([]DeleteMarkerVersion, 0, len(resp.Objects))
+
 	owner := Owner{
 		ID:          globalMinioDefaultOwnerID,
 		DisplayName: "minio",
@@ -445,10 +478,25 @@ func generateListVersionsResponse(bucket, prefix, marker, versionIDMarker, delim
 	data := ListVersionsResponse{}

 	for _, object := range resp.Objects {
-		content := ObjectVersion{}
 		if object.Name == "" {
 			continue
 		}
+
+		if object.DeleteMarker {
+			deleteMarker := DeleteMarkerVersion{
+				Key:          s3EncodeName(object.Name, encodingType),
+				LastModified: object.ModTime.UTC().Format(iso8601TimeFormat),
+				Owner:        owner,
+				VersionID:    object.VersionID,
+			}
+			if deleteMarker.VersionID == "" {
+				deleteMarker.VersionID = nullVersionID
+			}
+			deleteMarker.IsLatest = object.IsLatest
+			deleteMarkers = append(deleteMarkers, deleteMarker)
+			continue
+		}
+		content := ObjectVersion{}
 		content.Key = s3EncodeName(object.Name, encodingType)
 		content.LastModified = object.ModTime.UTC().Format(iso8601TimeFormat)
 		if object.ETag != "" {
@@ -466,12 +514,12 @@ func generateListVersionsResponse(bucket, prefix, marker, versionIDMarker, delim
 			content.VersionID = nullVersionID
 		}
 		content.IsLatest = object.IsLatest
-		content.isDeleteMarker = object.DeleteMarker
 		versions = append(versions, content)
 	}

 	data.Name = bucket
 	data.Versions = versions
+	data.DeleteMarkers = deleteMarkers
 	data.EncodingType = encodingType
 	data.Prefix = s3EncodeName(prefix, encodingType)
 	data.KeyMarker = s3EncodeName(marker, encodingType)
@@ -569,14 +617,14 @@ func generateListObjectsV2Response(bucket, prefix, token, nextToken, startAfter,
 		}
 		content.Owner = owner
 		if metadata {
-			content.UserMetadata = make(StringMap)
+			content.UserMetadata = &Metadata{}
 			switch kind, _ := crypto.IsEncrypted(object.UserDefined); kind {
 			case crypto.S3:
-				content.UserMetadata[xhttp.AmzServerSideEncryption] = xhttp.AmzEncryptionAES
+				content.UserMetadata.Set(xhttp.AmzServerSideEncryption, xhttp.AmzEncryptionAES)
 			case crypto.S3KMS:
-				content.UserMetadata[xhttp.AmzServerSideEncryption] = xhttp.AmzEncryptionKMS
+				content.UserMetadata.Set(xhttp.AmzServerSideEncryption, xhttp.AmzEncryptionKMS)
 			case crypto.SSEC:
-				content.UserMetadata[xhttp.AmzServerSideEncryptionCustomerAlgorithm] = xhttp.AmzEncryptionAES
+				content.UserMetadata.Set(xhttp.AmzServerSideEncryptionCustomerAlgorithm, xhttp.AmzEncryptionAES)
 			}
 			for k, v := range CleanMinioInternalMetadataKeys(object.UserDefined) {
 				if strings.HasPrefix(strings.ToLower(k), ReservedMetadataPrefixLower) {
@@ -588,7 +636,7 @@ func generateListObjectsV2Response(bucket, prefix, token, nextToken, startAfter,
 				if equals(k, xhttp.AmzMetaUnencryptedContentLength, xhttp.AmzMetaUnencryptedContentMD5) {
 					continue
 				}
-				content.UserMetadata[k] = v
+				content.UserMetadata.Set(k, v)
 			}
 		}
 		contents = append(contents, content)
@@ -35,8 +35,8 @@ func shouldEscape(c byte) bool {

 // s3URLEncode is based on Golang's url.QueryEscape() code,
 // while considering some S3 exceptions:
-//	- Avoid encoding '/' and '*'
-//	- Force encoding of '~'
+//   - Avoid encoding '/' and '*'
+//   - Force encoding of '~'
 func s3URLEncode(s string) string {
 	spaceCount, hexCount := 0, 0
 	for i := 0; i < len(s); i++ {
@@ -288,9 +288,10 @@ func checkClaimsFromToken(r *http.Request, cred auth.Credentials) (map[string]in
 }

 // Check request auth type verifies the incoming http request
-// - validates the request signature
-// - validates the policy action if anonymous tests bucket policies if any,
-//   for authenticated requests validates IAM policies.
+//   - validates the request signature
+//   - validates the policy action if anonymous tests bucket policies if any,
+//     for authenticated requests validates IAM policies.
+//
 // returns APIErrorCode if any to be replied to the client.
 func checkRequestAuthType(ctx context.Context, r *http.Request, action policy.Action, bucketName, objectName string) (s3Err APIErrorCode) {
 	_, _, s3Err = checkRequestAuthTypeCredential(ctx, r, action, bucketName, objectName)
@@ -298,9 +299,10 @@ func checkRequestAuthType(ctx context.Context, r *http.Request, action policy.Ac
 }

 // Check request auth type verifies the incoming http request
-// - validates the request signature
-// - validates the policy action if anonymous tests bucket policies if any,
-//   for authenticated requests validates IAM policies.
+//   - validates the request signature
+//   - validates the policy action if anonymous tests bucket policies if any,
+//     for authenticated requests validates IAM policies.
+//
 // returns APIErrorCode if any to be replied to the client.
 // Additionally returns the accessKey used in the request, and if this request is by an admin.
 func checkRequestAuthTypeCredential(ctx context.Context, r *http.Request, action policy.Action, bucketName, objectName string) (cred auth.Credentials, owner bool, s3Err APIErrorCode) {
@@ -26,9 +26,10 @@ import (
 )

 // healTask represents what to heal along with options
-//   path: '/' =>  Heal disk formats along with metadata
-//   path: 'bucket/' or '/bucket/' => Heal bucket
-//   path: 'bucket/object' => Heal object
+//
+//	path: '/' =>  Heal disk formats along with metadata
+//	path: 'bucket/' or '/bucket/' => Heal bucket
+//	path: 'bucket/object' => Heal object
 type healTask struct {
 	bucket    string
 	object    string
@@ -87,14 +87,14 @@ type healingTracker struct {
 // The disk ID will be validated against the loaded one.
 func loadHealingTracker(ctx context.Context, disk StorageAPI) (*healingTracker, error) {
 	if disk == nil {
-		return nil, errors.New("loadHealingTracker: nil disk given")
+		return nil, errors.New("loadHealingTracker: nil drive given")
 	}
 	diskID, err := disk.GetDiskID()
 	if err != nil {
 		return nil, err
 	}
 	b, err := disk.ReadAll(ctx, minioMetaBucket,
-		pathJoin(bucketMetaPrefix, slashSeparator, healingTrackerFilename))
+		pathJoin(bucketMetaPrefix, healingTrackerFilename))
 	if err != nil {
 		return nil, err
 	}
@@ -104,7 +104,7 @@ func loadHealingTracker(ctx context.Context, disk StorageAPI) (*healingTracker,
 		return nil, err
 	}
 	if h.ID != diskID && h.ID != "" {
-		return nil, fmt.Errorf("loadHealingTracker: disk id mismatch expected %s, got %s", h.ID, diskID)
+		return nil, fmt.Errorf("loadHealingTracker: drive id mismatch expected %s, got %s", h.ID, diskID)
 	}
 	h.disk = disk
 	h.ID = diskID
@@ -129,7 +129,7 @@ func newHealingTracker(disk StorageAPI) *healingTracker {
 // If the tracker has been deleted an error is returned.
 func (h *healingTracker) update(ctx context.Context) error {
 	if h.disk.Healing() == nil {
-		return fmt.Errorf("healingTracker: disk %q is not marked as healing", h.ID)
+		return fmt.Errorf("healingTracker: drive %q is not marked as healing", h.ID)
 	}
 	if h.ID == "" || h.PoolIndex < 0 || h.SetIndex < 0 || h.DiskIndex < 0 {
 		h.ID, _ = h.disk.GetDiskID()
@@ -155,15 +155,19 @@ func (h *healingTracker) save(ctx context.Context) error {
 	}
 	globalBackgroundHealState.updateHealStatus(h)
 	return h.disk.WriteAll(ctx, minioMetaBucket,
-		pathJoin(bucketMetaPrefix, slashSeparator, healingTrackerFilename),
+		pathJoin(bucketMetaPrefix, healingTrackerFilename),
 		htrackerBytes)
 }

 // delete the tracker on disk.
 func (h *healingTracker) delete(ctx context.Context) error {
 	return h.disk.Delete(ctx, minioMetaBucket,
-		pathJoin(bucketMetaPrefix, slashSeparator, healingTrackerFilename),
-		false)
+		pathJoin(bucketMetaPrefix, healingTrackerFilename),
+		DeleteOptions{
+			Recursive: false,
+			Force:     false,
+		},
+	)
 }

 func (h *healingTracker) isHealed(bucket string) bool {
@@ -306,7 +310,7 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	}

 	// Prevent parallel erasure set healing
-	locker := z.NewNSLock(minioMetaBucket, fmt.Sprintf("new-disk-healing/%s/%d/%d", endpoint, poolIdx, setIdx))
+	locker := z.NewNSLock(minioMetaBucket, fmt.Sprintf("new-drive-healing/%s/%d/%d", endpoint, poolIdx, setIdx))
 	lkctx, err := locker.GetLock(ctx, newDiskHealingTimeout)
 	if err != nil {
 		return err
@@ -314,8 +318,7 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	ctx = lkctx.Context()
 	defer locker.Unlock(lkctx.Cancel)

-	buckets, _ := z.ListBuckets(ctx)
-
+	buckets, _ := z.ListBuckets(ctx, BucketOptions{})
 	// Buckets data are dispersed in multiple zones/sets, make
 	// sure to heal all bucket metadata configuration.
 	buckets = append(buckets, BucketInfo{
@@ -334,14 +337,14 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	})

 	if serverDebugLog {
-		logger.Info("Healing disk '%v' on %s pool", disk, humanize.Ordinal(poolIdx+1))
+		logger.Info("Healing drive '%v' on %s pool", disk, humanize.Ordinal(poolIdx+1))
 	}

 	// Load healing tracker in this disk
 	tracker, err := loadHealingTracker(ctx, disk)
 	if err != nil {
 		// So someone changed the drives underneath, healing tracker missing.
-		logger.LogIf(ctx, fmt.Errorf("Healing tracker missing on '%s', disk was swapped again on %s pool: %w",
+		logger.LogIf(ctx, fmt.Errorf("Healing tracker missing on '%s', drive was swapped again on %s pool: %w",
 			disk, humanize.Ordinal(poolIdx+1), err))
 		tracker = newHealingTracker(disk)
 	}
@@ -361,12 +364,15 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	}

 	// Start or resume healing of this erasure set
-	err = z.serverPools[poolIdx].sets[setIdx].healErasureSet(ctx, tracker.QueuedBuckets, tracker)
-	if err != nil {
+	if err = z.serverPools[poolIdx].sets[setIdx].healErasureSet(ctx, tracker.QueuedBuckets, tracker); err != nil {
 		return err
 	}

-	logger.Info("Healing disk '%s' is complete (healed: %d, failed: %d).", disk, tracker.ItemsHealed, tracker.ItemsFailed)
+	if tracker.ItemsFailed > 0 {
+		logger.Info("Healing drive '%s' failed (healed: %d, failed: %d).", disk, tracker.ItemsHealed, tracker.ItemsFailed)
+	} else {
+		logger.Info("Healing drive '%s' complete (healed: %d, failed: %d).", disk, tracker.ItemsHealed, tracker.ItemsFailed)
+	}

 	if serverDebugLog {
 		tracker.printTo(os.Stdout)
@@ -374,6 +380,7 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	}

 	logger.LogIf(ctx, tracker.delete(ctx))
+
 	return nil
 }

@@ -390,12 +397,11 @@ func monitorLocalDisksAndHeal(ctx context.Context, z *erasureServerPools) {
 		case <-ctx.Done():
 			return
 		case <-diskCheckTimer.C:
-
 			healDisks := globalBackgroundHealState.getHealLocalDiskEndpoints()
 			if len(healDisks) == 0 {
 				// Reset for next interval.
 				diskCheckTimer.Reset(defaultMonitorNewDiskInterval)
-				break
+				continue
 			}

 			// Reformat disks immediately
@@ -404,7 +410,7 @@ func monitorLocalDisksAndHeal(ctx context.Context, z *erasureServerPools) {
 				logger.LogIf(ctx, err)
 				// Reset for next interval.
 				diskCheckTimer.Reset(defaultMonitorNewDiskInterval)
-				break
+				continue
 			}

 			for _, disk := range healDisks {
@@ -35,7 +35,7 @@ func runPutObjectBenchmark(b *testing.B, obj ObjectLayer, objSize int) {
 	// obtains random bucket name.
 	bucket := getRandomBucketName()
 	// create bucket.
-	err = obj.MakeBucketWithLocation(context.Background(), bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(context.Background(), bucket, MakeBucketOptions{})
 	if err != nil {
 		b.Fatal(err)
 	}
@@ -76,7 +76,7 @@ func runPutObjectPartBenchmark(b *testing.B, obj ObjectLayer, partSize int) {
 	object := getRandomObjectName()

 	// create bucket.
-	err = obj.MakeBucketWithLocation(context.Background(), bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(context.Background(), bucket, MakeBucketOptions{})
 	if err != nil {
 		b.Fatal(err)
 	}
@@ -196,7 +196,7 @@ func runPutObjectBenchmarkParallel(b *testing.B, obj ObjectLayer, objSize int) {
 	// obtains random bucket name.
 	bucket := getRandomBucketName()
 	// create bucket.
-	err := obj.MakeBucketWithLocation(context.Background(), bucket, BucketOptions{})
+	err := obj.MakeBucketWithLocation(context.Background(), bucket, MakeBucketOptions{})
 	if err != nil {
 		b.Fatal(err)
 	}
@@ -152,10 +152,10 @@ func (b *streamingBitrotReader) ReadAt(buf []byte, offset int64) (int, error) {
 		streamOffset := (offset/b.shardSize)*int64(b.h.Size()) + offset
 		if len(b.data) == 0 && b.tillOffset != streamOffset {
 			b.rc, err = b.disk.ReadFileStream(context.TODO(), b.volume, b.filePath, streamOffset, b.tillOffset-streamOffset)
-			if err != nil {
+			if err != nil && err != errDiskNotFound {
 				logger.LogIf(GlobalContext,
-					fmt.Errorf("Error(%w) reading erasure shards at (%s: %s/%s), will attempt to reconstruct if we have quorum",
-						err, b.disk, b.volume, b.filePath))
+					fmt.Errorf("Reading erasure shards at (%s: %s/%s) returned '%w', will attempt to reconstruct if we have quorum",
+						b.disk, b.volume, b.filePath, err))
 			}
 		} else {
 			b.rc = io.NewSectionReader(bytes.NewReader(b.data), streamOffset, b.tillOffset-streamOffset)
@@ -180,7 +180,7 @@ func (b *streamingBitrotReader) ReadAt(buf []byte, offset int64) (int, error) {
 	b.h.Write(buf)

 	if !bytes.Equal(b.h.Sum(nil), b.hashBytes) {
-		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s  -> %s/%s - content hash does not match - expected %s, got %s",
+		logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s  -> %s/%s - content hash does not match - expected %s, got %s",
 			b.disk, b.volume, b.filePath, hex.EncodeToString(b.hashBytes), hex.EncodeToString(b.h.Sum(nil))))
 		return 0, errFileCorrupt
 	}
@@ -38,12 +38,12 @@ type wholeBitrotWriter struct {
 func (b *wholeBitrotWriter) Write(p []byte) (int, error) {
 	err := b.disk.AppendFile(context.TODO(), b.volume, b.filePath, p)
 	if err != nil {
-		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s returned %w", b.disk, err))
+		logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s returned %w", b.disk, err))
 		return 0, err
 	}
 	_, err = b.Hash.Write(p)
 	if err != nil {
-		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s returned %w", b.disk, err))
+		logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s returned %w", b.disk, err))
 		return 0, err
 	}
 	return len(p), nil
@@ -72,12 +72,12 @@ func (b *wholeBitrotReader) ReadAt(buf []byte, offset int64) (n int, err error)
 	if b.buf == nil {
 		b.buf = make([]byte, b.tillOffset-offset)
 		if _, err := b.disk.ReadFile(context.TODO(), b.volume, b.filePath, offset, b.buf, b.verifier); err != nil {
-			logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s -> %s/%s returned %w", b.disk, b.volume, b.filePath, err))
+			logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s -> %s/%s returned %w", b.disk, b.volume, b.filePath, err))
 			return 0, err
 		}
 	}
 	if len(b.buf) < len(buf) {
-		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s -> %s/%s returned %w", b.disk, b.volume, b.filePath, errLessData))
+		logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s -> %s/%s returned %w", b.disk, b.volume, b.filePath, errLessData))
 		return 0, errLessData
 	}
 	n = copy(buf, b.buf)
@@ -20,17 +20,11 @@ package cmd
 import (
 	"context"
 	"io"
-	"io/ioutil"
-	"os"
 	"testing"
 )

 func testBitrotReaderWriterAlgo(t *testing.T, bitrotAlgo BitrotAlgorithm) {
-	tmpDir, err := ioutil.TempDir("", "")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(tmpDir)
+	tmpDir := t.TempDir()

 	volume := "testvol"
 	filePath := "testfile"
@@ -60,7 +54,9 @@ func testBitrotReaderWriterAlgo(t *testing.T, bitrotAlgo BitrotAlgorithm) {
 	if err != nil {
 		t.Fatal(err)
 	}
-	writer.(io.Closer).Close()
+	if bw, ok := writer.(io.Closer); ok {
+		bw.Close()
+	}

 	reader := newBitrotReader(disk, nil, volume, filePath, 35, bitrotAlgo, bitrotWriterSum(writer), 10)
 	b := make([]byte, 10)
@@ -76,6 +72,9 @@ func testBitrotReaderWriterAlgo(t *testing.T, bitrotAlgo BitrotAlgorithm) {
 	if _, err = reader.ReadAt(b[:5], 30); err != nil {
 		t.Fatal(err)
 	}
+	if br, ok := reader.(io.Closer); ok {
+		br.Close()
+	}
 }

 func TestAllBitrotAlgorithms(t *testing.T) {
@@ -65,7 +65,7 @@ func (api objectAPIHandlers) PutBucketEncryptionHandler(w http.ResponseWriter, r
 	}

 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -90,7 +90,7 @@ func (api objectAPIHandlers) PutBucketEncryptionHandler(w http.ResponseWriter, r
 	kmsKey := encConfig.KeyID()
 	if kmsKey != "" {
 		kmsContext := kms.Context{"MinIO admin API": "ServerInfoHandler"} // Context for a test key operation
-		_, err := GlobalKMS.GenerateKey(kmsKey, kmsContext)
+		_, err := GlobalKMS.GenerateKey(ctx, kmsKey, kmsContext)
 		if err != nil {
 			if errors.Is(err, kes.ErrKeyNotFound) {
 				writeErrorResponse(ctx, w, toAPIError(ctx, errKMSKeyNotFound), r.URL)
@@ -155,7 +155,7 @@ func (api objectAPIHandlers) GetBucketEncryptionHandler(w http.ResponseWriter, r

 	// Check if bucket exists
 	var err error
-	if _, err = objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err = objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -198,7 +198,7 @@ func (api objectAPIHandlers) DeleteBucketEncryptionHandler(w http.ResponseWriter

 	// Check if bucket exists
 	var err error
-	if _, err = objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err = objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -67,11 +67,16 @@ const (
 // - Check if a bucket has an entry in etcd backend
 // -- If no, make an entry
 // -- If yes, check if the entry matches local IP check if we
-//    need to update the entry then proceed to update
+//
+//	need to update the entry then proceed to update
+//
 // -- If yes, check if the IP of entry matches local IP.
-//    This means entry is for this instance.
+//
+//	This means entry is for this instance.
+//
 // -- If IP of the entry doesn't match, this means entry is
-//    for another instance. Log an error to console.
+//
+//	for another instance. Log an error to console.
 func initFederatorBackend(buckets []BucketInfo, objLayer ObjectLayer) {
 	if len(buckets) == 0 {
 		return
@@ -201,7 +206,7 @@ func (api objectAPIHandlers) GetBucketLocationHandler(w http.ResponseWriter, r *

 	getBucketInfo := objectAPI.GetBucketInfo

-	if _, err := getBucketInfo(ctx, bucket); err != nil {
+	if _, err := getBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -227,7 +232,6 @@ func (api objectAPIHandlers) GetBucketLocationHandler(w http.ResponseWriter, r *
 // using the Initiate Multipart Upload request, but has not yet been
 // completed or aborted. This operation returns at most 1,000 multipart
 // uploads in the response.
-//
 func (api objectAPIHandlers) ListMultipartUploadsHandler(w http.ResponseWriter, r *http.Request) {
 	ctx := newContext(r, w, "ListMultipartUploads")

@@ -332,7 +336,7 @@ func (api objectAPIHandlers) ListBucketsHandler(w http.ResponseWriter, r *http.R
 	} else {
 		// Invoke the list buckets.
 		var err error
-		bucketsInfo, err = listBuckets(ctx)
+		bucketsInfo, err = listBuckets(ctx, BucketOptions{})
 		if err != nil {
 			writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 			return
@@ -447,7 +451,7 @@ func (api objectAPIHandlers) DeleteMultipleObjectsHandler(w http.ResponseWriter,
 	checkRequestAuthType(ctx, r, policy.DeleteObjectAction, bucket, "")

 	// Before proceeding validate if bucket exists.
-	_, err := objectAPI.GetBucketInfo(ctx, bucket)
+	_, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -649,6 +653,7 @@ func (api objectAPIHandlers) DeleteMultipleObjectsHandler(w http.ResponseWriter,
 			dv := DeletedObjectReplicationInfo{
 				DeletedObject: dobj,
 				Bucket:        bucket,
+				EventType:     ReplicateIncomingDelete,
 			}
 			scheduleReplicationDelete(ctx, dv, objectAPI)
 		}
@@ -726,11 +731,30 @@ func (api objectAPIHandlers) PutBucketHandler(w http.ResponseWriter, r *http.Req
 		}
 	}

-	if s3Error := checkRequestAuthType(ctx, r, policy.CreateBucketAction, bucket, ""); s3Error != ErrNone {
+	cred, owner, s3Error := checkRequestAuthTypeCredential(ctx, r, policy.CreateBucketAction, bucket, "")
+	if s3Error != ErrNone {
 		writeErrorResponse(ctx, w, errorCodes.ToAPIErr(s3Error), r.URL)
 		return
 	}

+	if objectLockEnabled {
+		// Creating a bucket with locking requires the user having more permissions
+		for _, action := range []iampolicy.Action{iampolicy.PutBucketObjectLockConfigurationAction, iampolicy.PutBucketVersioningAction} {
+			if !globalIAMSys.IsAllowed(iampolicy.Args{
+				AccountName:     cred.AccessKey,
+				Groups:          cred.Groups,
+				Action:          action,
+				ConditionValues: getConditionValues(r, "", cred.AccessKey, cred.Claims),
+				BucketName:      bucket,
+				IsOwner:         owner,
+				Claims:          cred.Claims,
+			}) {
+				writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrAccessDenied), r.URL)
+				return
+			}
+		}
+	}
+
 	// Parse incoming location constraint.
 	location, s3Error := parseLocationConstraint(r)
 	if s3Error != ErrNone {
@@ -745,7 +769,7 @@ func (api objectAPIHandlers) PutBucketHandler(w http.ResponseWriter, r *http.Req
 		return
 	}

-	opts := BucketOptions{
+	opts := MakeBucketOptions{
 		Location:    location,
 		LockEnabled: objectLockEnabled,
 		ForceCreate: forceCreate,
@@ -764,7 +788,11 @@ func (api objectAPIHandlers) PutBucketHandler(w http.ResponseWriter, r *http.Req
 				}

 				if err = globalDNSConfig.Put(bucket); err != nil {
-					objectAPI.DeleteBucket(context.Background(), bucket, DeleteBucketOptions{Force: false, NoRecreate: true})
+					objectAPI.DeleteBucket(context.Background(), bucket, DeleteBucketOptions{
+						Force:      false,
+						NoRecreate: true,
+						SRDeleteOp: getSRBucketDeleteOp(globalSiteReplicationSys.isEnabled()),
+					})
 					writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 					return
 				}
@@ -864,7 +892,7 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 		return
 	}

-	if _, ok := crypto.IsRequested(r.Header); !objectAPI.IsEncryptionSupported() && ok {
+	if crypto.Requested(r.Header) && !objectAPI.IsEncryptionSupported() {
 		writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
 		return
 	}
@@ -1040,7 +1068,7 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 		return
 	}
 	if objectAPI.IsEncryptionSupported() {
-		if _, ok := crypto.IsRequested(formValues); ok && !HasSuffix(object, SlashSeparator) { // handle SSE requests
+		if crypto.Requested(formValues) && !HasSuffix(object, SlashSeparator) { // handle SSE requests
 			if crypto.SSECopy.IsRequested(r.Header) {
 				writeErrorResponse(ctx, w, toAPIError(ctx, errInvalidEncryptionParameters), r.URL)
 				return
@@ -1066,7 +1094,7 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 					return
 				}
 			}
-			reader, objectEncryptionKey, err = newEncryptReader(hashReader, kind, keyID, key, bucket, object, metadata, kmsCtx)
+			reader, objectEncryptionKey, err = newEncryptReader(ctx, hashReader, kind, keyID, key, bucket, object, metadata, kmsCtx)
 			if err != nil {
 				writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 				return
@@ -1115,9 +1143,12 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 		Host:         handlers.GetSourceIP(r),
 	})

-	if successRedirect != "" {
-		// Replace raw query params..
-		redirectURL.RawQuery = getRedirectPostRawQuery(objInfo)
+	if redirectURL != nil { // success_action_redirect is valid and set.
+		v := redirectURL.Query()
+		v.Add("bucket", objInfo.Bucket)
+		v.Add("key", objInfo.Name)
+		v.Add("etag", "\""+objInfo.ETag+"\"")
+		redirectURL.RawQuery = v.Encode()
 		writeRedirectSeeOther(w, redirectURL.String())
 		return
 	}
@@ -1161,7 +1192,7 @@ func (api objectAPIHandlers) GetBucketPolicyStatusHandler(w http.ResponseWriter,
 	}

 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -1224,7 +1255,7 @@ func (api objectAPIHandlers) HeadBucketHandler(w http.ResponseWriter, r *http.Re

 	getBucketInfo := objectAPI.GetBucketInfo

-	if _, err := getBucketInfo(ctx, bucket); err != nil {
+	if _, err := getBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseHeadersOnly(w, toAPIError(ctx, err))
 		return
 	}
@@ -1301,7 +1332,10 @@ func (api objectAPIHandlers) DeleteBucketHandler(w http.ResponseWriter, r *http.
 	deleteBucket := objectAPI.DeleteBucket

 	// Attempt to delete bucket.
-	if err := deleteBucket(ctx, bucket, DeleteBucketOptions{Force: forceDelete}); err != nil {
+	if err := deleteBucket(ctx, bucket, DeleteBucketOptions{
+		Force:      forceDelete,
+		SRDeleteOp: getSRBucketDeleteOp(globalSiteReplicationSys.isEnabled()),
+	}); err != nil {
 		apiErr := toAPIError(ctx, err)
 		if _, ok := err.(BucketNotEmpty); ok {
 			if globalBucketVersioningSys.Enabled(bucket) || globalBucketVersioningSys.Suspended(bucket) {
@@ -1468,7 +1502,7 @@ func (api objectAPIHandlers) PutBucketTaggingHandler(w http.ResponseWriter, r *h
 	}

 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -62,7 +62,7 @@ func (api objectAPIHandlers) PutBucketLifecycleHandler(w http.ResponseWriter, r
 	}

 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -121,7 +121,7 @@ func (api objectAPIHandlers) GetBucketLifecycleHandler(w http.ResponseWriter, r
 	}

 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -163,7 +163,7 @@ func (api objectAPIHandlers) DeleteBucketLifecycleHandler(w http.ResponseWriter,
 	}

 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -560,7 +560,7 @@ func (r *RestoreObjectRequest) validate(ctx context.Context, objAPI ObjectLayer)
 	}
 	// Check if bucket exists.
 	if !r.OutputLocation.IsEmpty() {
-		if _, err := objAPI.GetBucketInfo(ctx, r.OutputLocation.S3.BucketName); err != nil {
+		if _, err := objAPI.GetBucketInfo(ctx, r.OutputLocation.S3.BucketName, BucketOptions{}); err != nil {
 			return err
 		}
 		if r.OutputLocation.S3.Prefix == "" {
@@ -32,9 +32,9 @@ import (
 // Validate all the ListObjects query arguments, returns an APIErrorCode
 // if one of the args do not meet the required conditions.
 // Special conditions required by MinIO server are as below
-// - delimiter if set should be equal to '/', otherwise the request is rejected.
-// - marker if set should have a common prefix with 'prefix' param, otherwise
-//   the request is rejected.
+//   - delimiter if set should be equal to '/', otherwise the request is rejected.
+//   - marker if set should have a common prefix with 'prefix' param, otherwise
+//     the request is rejected.
 func validateListObjectsArgs(marker, delimiter, encodingType string, maxKeys int) APIErrorCode {
 	// Max keys cannot be negative.
 	if maxKeys < 0 {
@@ -298,7 +298,6 @@ func proxyRequestByNodeIndex(ctx context.Context, w http.ResponseWriter, r *http
 // This implementation of the GET operation returns some or all (up to 1000)
 // of the objects in a bucket. You can use the request parameters as selection
 // criteria to return a subset of the objects in a bucket.
-//
 func (api objectAPIHandlers) ListObjectsV1Handler(w http.ResponseWriter, r *http.Request) {
 	ctx := newContext(r, w, "ListObjectsV1")

@@ -136,7 +136,7 @@ func (sys *BucketMetadataSys) Update(ctx context.Context, bucket string, configF
 		meta.ReplicationConfigXML = configData
 		meta.ReplicationConfigUpdatedAt = updatedAt
 	case bucketTargetsFile:
-		meta.BucketTargetsConfigJSON, meta.BucketTargetsConfigMetaJSON, err = encryptBucketMetadata(meta.Name, configData, kms.Context{
+		meta.BucketTargetsConfigJSON, meta.BucketTargetsConfigMetaJSON, err = encryptBucketMetadata(ctx, meta.Name, configData, kms.Context{
 			bucket:            meta.Name,
 			bucketTargetsFile: bucketTargetsFile,
 		})
@@ -452,7 +452,7 @@ func (sys *BucketMetadataSys) concurrentLoad(ctx context.Context, buckets []Buck
 			sys.metadataMap[buckets[index].Name] = meta
 			sys.Unlock()

-			globalNotificationSys.set(buckets[index], meta) // set notification targets
+			globalEventNotifier.set(buckets[index], meta) // set notification targets

 			globalBucketTargetSys.set(buckets[index], meta) // set remote replication targets

@@ -105,10 +105,8 @@ type BucketMetadata struct {

 // newBucketMetadata creates BucketMetadata with the supplied name and Created to Now.
 func newBucketMetadata(name string) BucketMetadata {
-	now := UTCNow()
 	return BucketMetadata{
-		Name:    name,
-		Created: now,
+		Name: name,
 		notificationConfig: &event.Config{
 			XMLNS: "http://s3.amazonaws.com/doc/2006-03-01/",
 		},
@@ -121,6 +119,17 @@ func newBucketMetadata(name string) BucketMetadata {
 	}
 }

+// SetCreatedAt preserves the CreatedAt time for bucket across sites in site replication. It defaults to
+// creation time of bucket on this cluster in all other cases.
+func (b *BucketMetadata) SetCreatedAt(createdAt time.Time) {
+	if b.Created.IsZero() {
+		b.Created = UTCNow()
+	}
+	if !createdAt.IsZero() {
+		b.Created = createdAt.UTC()
+	}
+}
+
 // Load - loads the metadata of bucket by name from ObjectLayer api.
 // If an error is returned the returned metadata will be default initialized.
 func (b *BucketMetadata) Load(ctx context.Context, api ObjectLayer, name string) error {
@@ -160,17 +169,19 @@ func loadBucketMetadata(ctx context.Context, objectAPI ObjectLayer, bucket strin
 	if err != nil && !errors.Is(err, errConfigNotFound) {
 		return b, err
 	}
-
+	if err == nil {
+		b.defaultTimestamps()
+	}
 	// Old bucket without bucket metadata. Hence we migrate existing settings.
 	if err := b.convertLegacyConfigs(ctx, objectAPI); err != nil {
 		return b, err
 	}

 	// migrate unencrypted remote targets
-	if err = b.migrateTargetConfig(ctx, objectAPI); err != nil {
+	if err := b.migrateTargetConfig(ctx, objectAPI); err != nil {
 		return b, err
 	}
-	b.defaultTimestamps()
+
 	return b, nil
 }

@@ -345,6 +356,7 @@ func (b *BucketMetadata) convertLegacyConfigs(ctx context.Context, objectAPI Obj
 			b.BucketTargetsConfigJSON = configData
 		}
 	}
+	b.defaultTimestamps()

 	if err := b.Save(ctx, objectAPI); err != nil {
 		return err
@@ -438,7 +450,7 @@ func (b *BucketMetadata) migrateTargetConfig(ctx context.Context, objectAPI Obje
 		return nil
 	}

-	encBytes, metaBytes, err := encryptBucketMetadata(b.Name, b.BucketTargetsConfigJSON, kms.Context{b.Name: b.Name, bucketTargetsFile: bucketTargetsFile})
+	encBytes, metaBytes, err := encryptBucketMetadata(ctx, b.Name, b.BucketTargetsConfigJSON, kms.Context{b.Name: b.Name, bucketTargetsFile: bucketTargetsFile})
 	if err != nil {
 		return err
 	}
@@ -449,14 +461,14 @@ func (b *BucketMetadata) migrateTargetConfig(ctx context.Context, objectAPI Obje
 }

 // encrypt bucket metadata if kms is configured.
-func encryptBucketMetadata(bucket string, input []byte, kmsContext kms.Context) (output, metabytes []byte, err error) {
+func encryptBucketMetadata(ctx context.Context, bucket string, input []byte, kmsContext kms.Context) (output, metabytes []byte, err error) {
 	if GlobalKMS == nil {
 		output = input
 		return
 	}

 	metadata := make(map[string]string)
-	key, err := GlobalKMS.GenerateKey("", kmsContext)
+	key, err := GlobalKMS.GenerateKey(ctx, "", kmsContext)
 	if err != nil {
 		return
 	}
@@ -60,7 +60,7 @@ func (api objectAPIHandlers) GetBucketNotificationHandler(w http.ResponseWriter,
 		return
 	}

-	_, err := objAPI.GetBucketInfo(ctx, bucketName)
+	_, err := objAPI.GetBucketInfo(ctx, bucketName, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -72,7 +72,7 @@ func (api objectAPIHandlers) GetBucketNotificationHandler(w http.ResponseWriter,
 		return
 	}
 	config.SetRegion(globalSite.Region)
-	if err = config.Validate(globalSite.Region, globalNotificationSys.targetList); err != nil {
+	if err = config.Validate(globalSite.Region, globalEventNotifier.targetList); err != nil {
 		arnErr, ok := err.(*event.ErrARNNotFound)
 		if ok {
 			for i, queue := range config.QueueList {
@@ -132,7 +132,7 @@ func (api objectAPIHandlers) PutBucketNotificationHandler(w http.ResponseWriter,
 		return
 	}

-	_, err := objectAPI.GetBucketInfo(ctx, bucketName)
+	_, err := objectAPI.GetBucketInfo(ctx, bucketName, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -144,7 +144,7 @@ func (api objectAPIHandlers) PutBucketNotificationHandler(w http.ResponseWriter,
 		return
 	}

-	config, err := event.ParseConfig(io.LimitReader(r.Body, r.ContentLength), globalSite.Region, globalNotificationSys.targetList)
+	config, err := event.ParseConfig(io.LimitReader(r.Body, r.ContentLength), globalSite.Region, globalEventNotifier.targetList)
 	if err != nil {
 		apiErr := errorCodes.ToAPIErr(ErrMalformedXML)
 		if event.IsEventError(err) {
@@ -166,7 +166,7 @@ func (api objectAPIHandlers) PutBucketNotificationHandler(w http.ResponseWriter,
 	}

 	rulesMap := config.ToRulesMap()
-	globalNotificationSys.AddRulesMap(bucketName, rulesMap)
+	globalEventNotifier.AddRulesMap(bucketName, rulesMap)

 	writeSuccessResponseHeadersOnly(w)
 }
@@ -61,7 +61,7 @@ func (api objectAPIHandlers) PutBucketPolicyHandler(w http.ResponseWriter, r *ht
 	}

 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -145,7 +145,7 @@ func (api objectAPIHandlers) DeleteBucketPolicyHandler(w http.ResponseWriter, r
 	}

 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -191,7 +191,7 @@ func (api objectAPIHandlers) GetBucketPolicyHandler(w http.ResponseWriter, r *ht
 	}

 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -128,7 +128,7 @@ func testCreateBucket(obj ObjectLayer, instanceType, bucketName string, apiRoute
 			defer wg.Done()
 			// Sync start.
 			<-start
-			if err := obj.MakeBucketWithLocation(GlobalContext, bucketName1, BucketOptions{}); err != nil {
+			if err := obj.MakeBucketWithLocation(GlobalContext, bucketName1, MakeBucketOptions{}); err != nil {
 				if _, ok := err.(BucketExists); !ok {
 					t.Logf("unexpected error: %T: %v", err, err)
 					return
@@ -163,7 +163,7 @@ func testPutBucketPolicyHandler(obj ObjectLayer, instanceType, bucketName string
 	credentials auth.Credentials, t *testing.T,
 ) {
 	bucketName1 := fmt.Sprintf("%s-1", bucketName)
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName1, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName1, MakeBucketOptions{}); err != nil {
 		t.Fatal(err)
 	}

@@ -55,7 +55,7 @@ func (api objectAPIHandlers) PutBucketReplicationConfigHandler(w http.ResponseWr
 		return
 	}
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -118,7 +118,7 @@ func (api objectAPIHandlers) GetBucketReplicationConfigHandler(w http.ResponseWr
 		return
 	}
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -157,7 +157,7 @@ func (api objectAPIHandlers) DeleteBucketReplicationConfigHandler(w http.Respons
 		return
 	}
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -198,7 +198,7 @@ func (api objectAPIHandlers) GetBucketReplicationMetricsHandler(w http.ResponseW
 	}

 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -267,7 +267,7 @@ func (api objectAPIHandlers) ResetBucketReplicationStartHandler(w http.ResponseW
 	}

 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -310,7 +310,7 @@ func (api objectAPIHandlers) ResetBucketReplicationStartHandler(w http.ResponseW
 	rinfo.Targets = append(rinfo.Targets, ResyncTarget{Arn: tgtArns[0], ResetID: target.ResetID})
 	if err = globalBucketTargetSys.SetTarget(ctx, bucket, &target, true); err != nil {
 		switch err.(type) {
-		case BucketRemoteConnectionErr:
+		case RemoteTargetConnectionErr:
 			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrReplicationRemoteConnectionError, err), r.URL)
 		default:
 			writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
@@ -356,7 +356,7 @@ func (api objectAPIHandlers) ResetBucketReplicationStatusHandler(w http.Response
 	}

 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -21,12 +21,14 @@ import (
 	"bytes"
 	"fmt"
 	"net/http"
+	"net/url"
 	"regexp"
 	"strconv"
 	"strings"
 	"sync"
 	"time"

+	"github.com/minio/madmin-go"
 	"github.com/minio/minio/internal/bucket/replication"
 	xhttp "github.com/minio/minio/internal/http"
 )
@@ -503,6 +505,8 @@ func getHealReplicateObjectInfo(objInfo ObjectInfo, rcfg replicationConfig) Repl
 	}
 	var dsc ReplicateDecision
 	var tgtStatuses map[string]replication.StatusType
+	var purgeStatuses map[string]VersionPurgeStatusType
+
 	if oi.DeleteMarker || !oi.VersionPurgeStatus.Empty() {
 		dsc = checkReplicateDelete(GlobalContext, oi.Bucket, ObjectToDelete{
 			ObjectV: ObjectV{
@@ -516,15 +520,17 @@ func getHealReplicateObjectInfo(objInfo ObjectInfo, rcfg replicationConfig) Repl
 		}, replication.HealReplicationType, ObjectOptions{}))
 	}
 	tgtStatuses = replicationStatusesMap(oi.ReplicationStatusInternal)
-
+	purgeStatuses = versionPurgeStatusesMap(oi.VersionPurgeStatusInternal)
 	existingObjResync := rcfg.Resync(GlobalContext, oi, &dsc, tgtStatuses)
-
+	tm, _ := time.Parse(time.RFC3339Nano, oi.UserDefined[ReservedMetadataPrefixLower+ReplicationTimestamp])
 	return ReplicateObjectInfo{
-		ObjectInfo:        oi,
-		OpType:            replication.HealReplicationType,
-		Dsc:               dsc,
-		ExistingObjResync: existingObjResync,
-		TargetStatuses:    tgtStatuses,
+		ObjectInfo:           oi,
+		OpType:               replication.HealReplicationType,
+		Dsc:                  dsc,
+		ExistingObjResync:    existingObjResync,
+		TargetStatuses:       tgtStatuses,
+		TargetPurgeStatuses:  purgeStatuses,
+		ReplicationTimestamp: tm,
 	}
 }

@@ -724,3 +730,39 @@ func parseSizeFromContentRange(h http.Header) (sz int64, err error) {
 	}
 	return int64(usz), nil
 }
+
+func extractReplicateDiffOpts(q url.Values) (opts madmin.ReplDiffOpts) {
+	opts.Verbose = q.Get("verbose") == "true"
+	opts.ARN = q.Get("arn")
+	opts.Prefix = q.Get("prefix")
+	return
+}
+
+const (
+	replicationMRFDir = bucketMetaPrefix + SlashSeparator + replicationDir + SlashSeparator + "mrf"
+	mrfMetaFormat     = 1
+	mrfMetaVersionV1  = 1
+	mrfMetaVersion    = mrfMetaVersionV1
+)
+
+// MRFReplicateEntry mrf entry to save to disk
+type MRFReplicateEntry struct {
+	Bucket    string `json:"bucket" msg:"b"`
+	Object    string `json:"object" msg:"o"`
+	versionID string `json:"-"`
+}
+
+// MRFReplicateEntries has the map of MRF entries to save to disk
+type MRFReplicateEntries struct {
+	Entries map[string]MRFReplicateEntry `json:"entries" msg:"e"`
+	Version int                          `json:"version" msg:"v"`
+}
+
+// ToMRFEntry returns the relevant info needed by MRF
+func (ri ReplicateObjectInfo) ToMRFEntry() MRFReplicateEntry {
+	return MRFReplicateEntry{
+		Bucket:    ri.Bucket,
+		Object:    ri.Name,
+		versionID: ri.VersionID,
+	}
+}
@@ -260,6 +260,413 @@ func (z *BucketReplicationResyncStatus) Msgsize() (s int) {
 	return
 }

+// DecodeMsg implements msgp.Decodable
+func (z *MRFReplicateEntries) DecodeMsg(dc *msgp.Reader) (err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, err = dc.ReadMapHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, err = dc.ReadMapKeyPtr()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "e":
+			var zb0002 uint32
+			zb0002, err = dc.ReadMapHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "Entries")
+				return
+			}
+			if z.Entries == nil {
+				z.Entries = make(map[string]MRFReplicateEntry, zb0002)
+			} else if len(z.Entries) > 0 {
+				for key := range z.Entries {
+					delete(z.Entries, key)
+				}
+			}
+			for zb0002 > 0 {
+				zb0002--
+				var za0001 string
+				var za0002 MRFReplicateEntry
+				za0001, err = dc.ReadString()
+				if err != nil {
+					err = msgp.WrapError(err, "Entries")
+					return
+				}
+				var zb0003 uint32
+				zb0003, err = dc.ReadMapHeader()
+				if err != nil {
+					err = msgp.WrapError(err, "Entries", za0001)
+					return
+				}
+				for zb0003 > 0 {
+					zb0003--
+					field, err = dc.ReadMapKeyPtr()
+					if err != nil {
+						err = msgp.WrapError(err, "Entries", za0001)
+						return
+					}
+					switch msgp.UnsafeString(field) {
+					case "b":
+						za0002.Bucket, err = dc.ReadString()
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001, "Bucket")
+							return
+						}
+					case "o":
+						za0002.Object, err = dc.ReadString()
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001, "Object")
+							return
+						}
+					default:
+						err = dc.Skip()
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001)
+							return
+						}
+					}
+				}
+				z.Entries[za0001] = za0002
+			}
+		case "v":
+			z.Version, err = dc.ReadInt()
+			if err != nil {
+				err = msgp.WrapError(err, "Version")
+				return
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z *MRFReplicateEntries) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 2
+	// write "e"
+	err = en.Append(0x82, 0xa1, 0x65)
+	if err != nil {
+		return
+	}
+	err = en.WriteMapHeader(uint32(len(z.Entries)))
+	if err != nil {
+		err = msgp.WrapError(err, "Entries")
+		return
+	}
+	for za0001, za0002 := range z.Entries {
+		err = en.WriteString(za0001)
+		if err != nil {
+			err = msgp.WrapError(err, "Entries")
+			return
+		}
+		// map header, size 2
+		// write "b"
+		err = en.Append(0x82, 0xa1, 0x62)
+		if err != nil {
+			return
+		}
+		err = en.WriteString(za0002.Bucket)
+		if err != nil {
+			err = msgp.WrapError(err, "Entries", za0001, "Bucket")
+			return
+		}
+		// write "o"
+		err = en.Append(0xa1, 0x6f)
+		if err != nil {
+			return
+		}
+		err = en.WriteString(za0002.Object)
+		if err != nil {
+			err = msgp.WrapError(err, "Entries", za0001, "Object")
+			return
+		}
+	}
+	// write "v"
+	err = en.Append(0xa1, 0x76)
+	if err != nil {
+		return
+	}
+	err = en.WriteInt(z.Version)
+	if err != nil {
+		err = msgp.WrapError(err, "Version")
+		return
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z *MRFReplicateEntries) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 2
+	// string "e"
+	o = append(o, 0x82, 0xa1, 0x65)
+	o = msgp.AppendMapHeader(o, uint32(len(z.Entries)))
+	for za0001, za0002 := range z.Entries {
+		o = msgp.AppendString(o, za0001)
+		// map header, size 2
+		// string "b"
+		o = append(o, 0x82, 0xa1, 0x62)
+		o = msgp.AppendString(o, za0002.Bucket)
+		// string "o"
+		o = append(o, 0xa1, 0x6f)
+		o = msgp.AppendString(o, za0002.Object)
+	}
+	// string "v"
+	o = append(o, 0xa1, 0x76)
+	o = msgp.AppendInt(o, z.Version)
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *MRFReplicateEntries) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadMapHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, bts, err = msgp.ReadMapKeyZC(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "e":
+			var zb0002 uint32
+			zb0002, bts, err = msgp.ReadMapHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Entries")
+				return
+			}
+			if z.Entries == nil {
+				z.Entries = make(map[string]MRFReplicateEntry, zb0002)
+			} else if len(z.Entries) > 0 {
+				for key := range z.Entries {
+					delete(z.Entries, key)
+				}
+			}
+			for zb0002 > 0 {
+				var za0001 string
+				var za0002 MRFReplicateEntry
+				zb0002--
+				za0001, bts, err = msgp.ReadStringBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Entries")
+					return
+				}
+				var zb0003 uint32
+				zb0003, bts, err = msgp.ReadMapHeaderBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Entries", za0001)
+					return
+				}
+				for zb0003 > 0 {
+					zb0003--
+					field, bts, err = msgp.ReadMapKeyZC(bts)
+					if err != nil {
+						err = msgp.WrapError(err, "Entries", za0001)
+						return
+					}
+					switch msgp.UnsafeString(field) {
+					case "b":
+						za0002.Bucket, bts, err = msgp.ReadStringBytes(bts)
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001, "Bucket")
+							return
+						}
+					case "o":
+						za0002.Object, bts, err = msgp.ReadStringBytes(bts)
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001, "Object")
+							return
+						}
+					default:
+						bts, err = msgp.Skip(bts)
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001)
+							return
+						}
+					}
+				}
+				z.Entries[za0001] = za0002
+			}
+		case "v":
+			z.Version, bts, err = msgp.ReadIntBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Version")
+				return
+			}
+		default:
+			bts, err = msgp.Skip(bts)
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z *MRFReplicateEntries) Msgsize() (s int) {
+	s = 1 + 2 + msgp.MapHeaderSize
+	if z.Entries != nil {
+		for za0001, za0002 := range z.Entries {
+			_ = za0002
+			s += msgp.StringPrefixSize + len(za0001) + 1 + 2 + msgp.StringPrefixSize + len(za0002.Bucket) + 2 + msgp.StringPrefixSize + len(za0002.Object)
+		}
+	}
+	s += 2 + msgp.IntSize
+	return
+}
+
+// DecodeMsg implements msgp.Decodable
+func (z *MRFReplicateEntry) DecodeMsg(dc *msgp.Reader) (err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, err = dc.ReadMapHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, err = dc.ReadMapKeyPtr()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "b":
+			z.Bucket, err = dc.ReadString()
+			if err != nil {
+				err = msgp.WrapError(err, "Bucket")
+				return
+			}
+		case "o":
+			z.Object, err = dc.ReadString()
+			if err != nil {
+				err = msgp.WrapError(err, "Object")
+				return
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z MRFReplicateEntry) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 2
+	// write "b"
+	err = en.Append(0x82, 0xa1, 0x62)
+	if err != nil {
+		return
+	}
+	err = en.WriteString(z.Bucket)
+	if err != nil {
+		err = msgp.WrapError(err, "Bucket")
+		return
+	}
+	// write "o"
+	err = en.Append(0xa1, 0x6f)
+	if err != nil {
+		return
+	}
+	err = en.WriteString(z.Object)
+	if err != nil {
+		err = msgp.WrapError(err, "Object")
+		return
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z MRFReplicateEntry) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 2
+	// string "b"
+	o = append(o, 0x82, 0xa1, 0x62)
+	o = msgp.AppendString(o, z.Bucket)
+	// string "o"
+	o = append(o, 0xa1, 0x6f)
+	o = msgp.AppendString(o, z.Object)
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *MRFReplicateEntry) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadMapHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, bts, err = msgp.ReadMapKeyZC(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "b":
+			z.Bucket, bts, err = msgp.ReadStringBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Bucket")
+				return
+			}
+		case "o":
+			z.Object, bts, err = msgp.ReadStringBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Object")
+				return
+			}
+		default:
+			bts, err = msgp.Skip(bts)
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z MRFReplicateEntry) Msgsize() (s int) {
+	s = 1 + 2 + msgp.StringPrefixSize + len(z.Bucket) + 2 + msgp.StringPrefixSize + len(z.Object)
+	return
+}
+
 // DecodeMsg implements msgp.Decodable
 func (z *ReplicateDecision) DecodeMsg(dc *msgp.Reader) (err error) {
 	var field []byte
@@ -122,6 +122,232 @@ func BenchmarkDecodeBucketReplicationResyncStatus(b *testing.B) {
 	}
 }

+func TestMarshalUnmarshalMRFReplicateEntries(t *testing.T) {
+	v := MRFReplicateEntries{}
+	bts, err := v.MarshalMsg(nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	left, err := v.UnmarshalMsg(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after UnmarshalMsg(): %q", len(left), left)
+	}
+
+	left, err = msgp.Skip(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after Skip(): %q", len(left), left)
+	}
+}
+
+func BenchmarkMarshalMsgMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	bts := make([]byte, 0, v.Msgsize())
+	bts, _ = v.MarshalMsg(bts[0:0])
+	b.SetBytes(int64(len(bts)))
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		bts, _ = v.MarshalMsg(bts[0:0])
+	}
+}
+
+func BenchmarkUnmarshalMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	bts, _ := v.MarshalMsg(nil)
+	b.ReportAllocs()
+	b.SetBytes(int64(len(bts)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		_, err := v.UnmarshalMsg(bts)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestEncodeDecodeMRFReplicateEntries(t *testing.T) {
+	v := MRFReplicateEntries{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeMRFReplicateEntries Msgsize() is inaccurate")
+	}
+
+	vn := MRFReplicateEntries{}
+	err := msgp.Decode(&buf, &vn)
+	if err != nil {
+		t.Error(err)
+	}
+
+	buf.Reset()
+	msgp.Encode(&buf, &v)
+	err = msgp.NewReader(&buf).Skip()
+	if err != nil {
+		t.Error(err)
+	}
+}
+
+func BenchmarkEncodeMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	en := msgp.NewWriter(msgp.Nowhere)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.EncodeMsg(en)
+	}
+	en.Flush()
+}
+
+func BenchmarkDecodeMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	rd := msgp.NewEndlessReader(buf.Bytes(), b)
+	dc := msgp.NewReader(rd)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		err := v.DecodeMsg(dc)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestMarshalUnmarshalMRFReplicateEntry(t *testing.T) {
+	v := MRFReplicateEntry{}
+	bts, err := v.MarshalMsg(nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	left, err := v.UnmarshalMsg(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after UnmarshalMsg(): %q", len(left), left)
+	}
+
+	left, err = msgp.Skip(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after Skip(): %q", len(left), left)
+	}
+}
+
+func BenchmarkMarshalMsgMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	bts := make([]byte, 0, v.Msgsize())
+	bts, _ = v.MarshalMsg(bts[0:0])
+	b.SetBytes(int64(len(bts)))
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		bts, _ = v.MarshalMsg(bts[0:0])
+	}
+}
+
+func BenchmarkUnmarshalMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	bts, _ := v.MarshalMsg(nil)
+	b.ReportAllocs()
+	b.SetBytes(int64(len(bts)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		_, err := v.UnmarshalMsg(bts)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestEncodeDecodeMRFReplicateEntry(t *testing.T) {
+	v := MRFReplicateEntry{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeMRFReplicateEntry Msgsize() is inaccurate")
+	}
+
+	vn := MRFReplicateEntry{}
+	err := msgp.Decode(&buf, &vn)
+	if err != nil {
+		t.Error(err)
+	}
+
+	buf.Reset()
+	msgp.Encode(&buf, &v)
+	err = msgp.NewReader(&buf).Skip()
+	if err != nil {
+		t.Error(err)
+	}
+}
+
+func BenchmarkEncodeMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	en := msgp.NewWriter(msgp.Nowhere)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.EncodeMsg(en)
+	}
+	en.Flush()
+}
+
+func BenchmarkDecodeMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	rd := msgp.NewEndlessReader(buf.Bytes(), b)
+	dc := msgp.NewReader(rd)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		err := v.DecodeMsg(dc)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
 func TestMarshalUnmarshalReplicateDecision(t *testing.T) {
 	v := ReplicateDecision{}
 	bts, err := v.MarshalMsg(nil)
@@ -197,7 +197,7 @@ func mustReplicate(ctx context.Context, bucket, object string, mopts mustReplica

 	// Disable server-side replication on object prefixes which are excluded
 	// from versioning via the MinIO bucket versioning extension.
-	if globalBucketVersioningSys.PrefixSuspended(bucket, object) {
+	if !globalBucketVersioningSys.PrefixEnabled(bucket, object) {
 		return
 	}

@@ -347,7 +347,7 @@ func checkReplicateDelete(ctx context.Context, bucket string, dobj ObjectToDelet
 // target cluster, the object version is marked deleted on the source and hidden from listing. It is permanently
 // deleted from the source when the VersionPurgeStatus changes to "Complete", i.e after replication succeeds
 // on target.
-func replicateDelete(ctx context.Context, dobj DeletedObjectReplicationInfo, objectAPI ObjectLayer, trigger string) {
+func replicateDelete(ctx context.Context, dobj DeletedObjectReplicationInfo, objectAPI ObjectLayer) {
 	var replicationStatus replication.StatusType
 	bucket := dobj.Bucket
 	versionID := dobj.DeleteMarkerVersionID
@@ -358,7 +358,7 @@ func replicateDelete(ctx context.Context, dobj DeletedObjectReplicationInfo, obj
 	defer func() {
 		replStatus := string(replicationStatus)
 		auditLogInternal(context.Background(), bucket, dobj.ObjectName, AuditLogOptions{
-			Trigger:   trigger,
+			Event:     dobj.EventType,
 			APIName:   ReplicateDeleteAPI,
 			VersionID: versionID,
 			Status:    replStatus,
@@ -403,6 +403,7 @@ func replicateDelete(ctx context.Context, dobj DeletedObjectReplicationInfo, obj
 	lk := objectAPI.NewNSLock(bucket, "/[replicate]/"+dobj.ObjectName)
 	lkctx, err := lk.GetLock(ctx, globalOperationTimeout)
 	if err != nil {
+		globalReplicationPool.queueMRFSave(dobj.ToMRFEntry())
 		logger.LogIf(ctx, fmt.Errorf("failed to get lock for object: %s bucket:%s arn:%s", dobj.ObjectName, bucket, rcfg.RoleArn))
 		sendEvent(eventArgs{
 			BucketName: bucket,
@@ -477,14 +478,23 @@ func replicateDelete(ctx context.Context, dobj DeletedObjectReplicationInfo, obj
 	eventName := event.ObjectReplicationComplete
 	if replicationStatus == replication.Failed {
 		eventName = event.ObjectReplicationFailed
+		globalReplicationPool.queueMRFSave(dobj.ToMRFEntry())
 	}
 	drs := getReplicationState(rinfos, dobj.ReplicationState, dobj.VersionID)
+	if replicationStatus != prevStatus {
+		drs.ReplicationTimeStamp = UTCNow()
+	}
+
 	dobjInfo, err := objectAPI.DeleteObject(ctx, bucket, dobj.ObjectName, ObjectOptions{
 		VersionID:         versionID,
 		MTime:             dobj.DeleteMarkerMTime.Time,
 		DeleteReplication: drs,
 		Versioned:         globalBucketVersioningSys.PrefixEnabled(bucket, dobj.ObjectName),
-		VersionSuspended:  globalBucketVersioningSys.PrefixSuspended(bucket, dobj.ObjectName),
+		// Objects matching prefixes should not leave delete markers,
+		// dramatically reduces namespace pollution while keeping the
+		// benefits of replication, make sure to apply version suspension
+		// only at bucket level instead.
+		VersionSuspended: globalBucketVersioningSys.Suspended(bucket),
 	})
 	if err != nil && !isErrVersionNotFound(err) { // VersionNotFound would be reported by pool that object version is missing on.
 		logger.LogIf(ctx, fmt.Errorf("Unable to update replication metadata for %s/%s(%s): %s", bucket, dobj.ObjectName, versionID, err))
@@ -530,7 +540,7 @@ func replicateDeleteToTarget(ctx context.Context, dobj DeletedObjectReplicationI
 	if dobj.VersionID != "" && rinfo.VersionPurgeStatus == Complete {
 		return
 	}
-	if tgt.IsOffline() {
+	if globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
 		logger.LogIf(ctx, fmt.Errorf("remote target is offline for bucket:%s arn:%s", dobj.Bucket, tgt.ARN))
 		sendEvent(eventArgs{
 			BucketName: dobj.Bucket,
@@ -857,7 +867,7 @@ func getReplicationAction(oi1 ObjectInfo, oi2 minio.ObjectInfo, opType replicati

 // replicateObject replicates the specified version of the object to destination bucket
 // The source object is then updated to reflect the replication status.
-func replicateObject(ctx context.Context, ri ReplicateObjectInfo, objectAPI ObjectLayer, trigger string) {
+func replicateObject(ctx context.Context, ri ReplicateObjectInfo, objectAPI ObjectLayer) {
 	var replicationStatus replication.StatusType
 	defer func() {
 		if replicationStatus.Empty() {
@@ -868,7 +878,7 @@ func replicateObject(ctx context.Context, ri ReplicateObjectInfo, objectAPI Obje
 			replicationStatus = ri.ReplicationStatus
 		}
 		auditLogInternal(ctx, ri.Bucket, ri.Name, AuditLogOptions{
-			Trigger:   trigger,
+			Event:     ri.EventType,
 			APIName:   ReplicateObjectAPI,
 			VersionID: ri.VersionID,
 			Status:    replicationStatus.String(),
@@ -906,6 +916,7 @@ func replicateObject(ctx context.Context, ri ReplicateObjectInfo, objectAPI Obje
 			Object:     objInfo,
 			Host:       "Internal: [Replication]",
 		})
+		globalReplicationPool.queueMRFSave(ri.ToMRFEntry())
 		logger.LogIf(ctx, fmt.Errorf("failed to get lock for object: %s bucket:%s arn:%s", object, bucket, cfg.RoleArn))
 		return
 	}
@@ -988,11 +999,12 @@ func replicateObject(ctx context.Context, ri ReplicateObjectInfo, objectAPI Obje

 	// re-queue failures once more - keep a retry count to avoid flooding the queue if
 	// the target site is down. Leave it to scanner to catch up instead.
-	if rinfos.ReplicationStatus() != replication.Completed && ri.RetryCount < 1 {
+	if rinfos.ReplicationStatus() != replication.Completed {
 		ri.OpType = replication.HealReplicationType
+		ri.EventType = ReplicateMRF
 		ri.ReplicationStatusInternal = rinfos.ReplicationStatusInternal()
 		ri.RetryCount++
-		globalReplicationPool.queueReplicaFailedTask(ri)
+		globalReplicationPool.queueMRFSave(ri.ToMRFEntry())
 	}
 }

@@ -1034,7 +1046,7 @@ func replicateObjectToTarget(ctx context.Context, ri ReplicateObjectInfo, object
 		rinfo.ReplicationResynced = true
 		return
 	}
-	if tgt.IsOffline() {
+	if globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
 		logger.LogIf(ctx, fmt.Errorf("remote target is offline for bucket:%s arn:%s", bucket, tgt.ARN))
 		sendEvent(eventArgs{
 			EventName:  event.ObjectReplicationNotTracked,
@@ -1054,13 +1066,15 @@ func replicateObjectToTarget(ctx context.Context, ri ReplicateObjectInfo, object
 		VersionSuspended: versionSuspended,
 	})
 	if err != nil {
-		sendEvent(eventArgs{
-			EventName:  event.ObjectReplicationNotTracked,
-			BucketName: bucket,
-			Object:     objInfo,
-			Host:       "Internal: [Replication]",
-		})
-		logger.LogIf(ctx, fmt.Errorf("Unable to update replicate for %s/%s(%s): %w", bucket, object, objInfo.VersionID, err))
+		if !isErrObjectNotFound(err) {
+			sendEvent(eventArgs{
+				EventName:  event.ObjectReplicationNotTracked,
+				BucketName: bucket,
+				Object:     objInfo,
+				Host:       "Internal: [Replication]",
+			})
+			logger.LogIf(ctx, fmt.Errorf("Unable to update replicate metadata for %s/%s(%s): %w", bucket, object, objInfo.VersionID, err))
+		}
 		return
 	}
 	defer func() {
@@ -1291,11 +1305,25 @@ func filterReplicationStatusMetadata(metadata map[string]string) map[string]stri
 type DeletedObjectReplicationInfo struct {
 	DeletedObject
 	Bucket    string
+	EventType string
 	OpType    replication.Type
 	ResetID   string
 	TargetArn string
 }

+// ToMRFEntry returns the relevant info needed by MRF
+func (di DeletedObjectReplicationInfo) ToMRFEntry() MRFReplicateEntry {
+	versionID := di.DeleteMarkerVersionID
+	if versionID == "" {
+		versionID = di.VersionID
+	}
+	return MRFReplicateEntry{
+		Bucket:    di.Bucket,
+		Object:    di.ObjectName,
+		versionID: versionID,
+	}
+}
+
 // Replication specific APIName
 const (
 	ReplicateObjectAPI = "ReplicateObject"
@@ -1313,12 +1341,15 @@ const (

 	// ReplicateMRF - audit trail for replication from Most Recent Failures (MRF) queue
 	ReplicateMRF = "replicate:mrf"
-	// ReplicateIncoming - audit trail indicating replication started [could be from incoming/existing/heal activity]
+	// ReplicateIncoming - audit trail of inline replication
 	ReplicateIncoming = "replicate:incoming"
+	// ReplicateIncomingDelete - audit trail of inline replication of deletes.
+	ReplicateIncomingDelete = "replicate:incoming:delete"
+
 	// ReplicateHeal - audit trail for healing of failed/pending replications
 	ReplicateHeal = "replicate:heal"
-	// ReplicateDelete - audit trail for delete replication
-	ReplicateDelete = "replicate:delete"
+	// ReplicateHealDelete - audit trail of healing of failed/pending delete replications.
+	ReplicateHealDelete = "replicate:heal:delete"
 )

 var (
@@ -1337,13 +1368,15 @@ type ReplicationPool struct {
 	mrfReplicaCh            chan ReplicateObjectInfo
 	existingReplicaCh       chan ReplicateObjectInfo
 	existingReplicaDeleteCh chan DeletedObjectReplicationInfo
-	workerSize              int
-	mrfWorkerSize           int
-	resyncState             replicationResyncState
-	workerWg                sync.WaitGroup
-	mrfWorkerWg             sync.WaitGroup
-	once                    sync.Once
-	mu                      sync.Mutex
+	mrfSaveCh               chan MRFReplicateEntry
+
+	workerSize    int
+	mrfWorkerSize int
+	resyncState   replicationResyncState
+	workerWg      sync.WaitGroup
+	mrfWorkerWg   sync.WaitGroup
+	once          sync.Once
+	mu            sync.Mutex
 }

 // NewReplicationPool creates a pool of replication workers of specified size
@@ -1357,6 +1390,7 @@ func NewReplicationPool(ctx context.Context, o ObjectLayer, opts replicationPool
 		existingReplicaCh:       make(chan ReplicateObjectInfo, 100000),
 		existingReplicaDeleteCh: make(chan DeletedObjectReplicationInfo, 100000),
 		resyncState:             replicationResyncState{statusMap: make(map[string]BucketReplicationResyncStatus)},
+		mrfSaveCh:               make(chan MRFReplicateEntry, 100000),
 		ctx:                     ctx,
 		objLayer:                o,
 	}
@@ -1364,13 +1398,16 @@ func NewReplicationPool(ctx context.Context, o ObjectLayer, opts replicationPool
 	pool.ResizeWorkers(opts.Workers)
 	pool.ResizeFailedWorkers(opts.FailedWorkers)
 	go pool.AddExistingObjectReplicateWorker()
-	go pool.periodicResyncMetaSave(ctx, o)
+	go pool.updateResyncStatus(ctx, o)
+	go pool.processMRF()
+	go pool.persistMRF()
 	return pool
 }

 // AddMRFWorker adds a pending/failed replication worker to handle requests that could not be queued
 // to the other workers
 func (p *ReplicationPool) AddMRFWorker() {
+	defer p.mrfWorkerWg.Done()
 	for {
 		select {
 		case <-p.ctx.Done():
@@ -1379,7 +1416,7 @@ func (p *ReplicationPool) AddMRFWorker() {
 			if !ok {
 				return
 			}
-			replicateObject(p.ctx, oi, p.objLayer, ReplicateMRF)
+			replicateObject(p.ctx, oi, p.objLayer)
 		case <-p.mrfWorkerKillCh:
 			return
 		}
@@ -1397,12 +1434,12 @@ func (p *ReplicationPool) AddWorker() {
 			if !ok {
 				return
 			}
-			replicateObject(p.ctx, oi, p.objLayer, ReplicateIncoming)
+			replicateObject(p.ctx, oi, p.objLayer)
 		case doi, ok := <-p.replicaDeleteCh:
 			if !ok {
 				return
 			}
-			replicateDelete(p.ctx, doi, p.objLayer, ReplicateDelete)
+			replicateDelete(p.ctx, doi, p.objLayer)
 		case <-p.workerKillCh:
 			return
 		}
@@ -1419,12 +1456,12 @@ func (p *ReplicationPool) AddExistingObjectReplicateWorker() {
 			if !ok {
 				return
 			}
-			replicateObject(p.ctx, oi, p.objLayer, ReplicateExisting)
+			replicateObject(p.ctx, oi, p.objLayer)
 		case doi, ok := <-p.existingReplicaDeleteCh:
 			if !ok {
 				return
 			}
-			replicateDelete(p.ctx, doi, p.objLayer, ReplicateExistingDelete)
+			replicateDelete(p.ctx, doi, p.objLayer)
 		}
 	}
 }
@@ -1469,33 +1506,17 @@ func (p *ReplicationPool) suggestedWorkers(failQueue bool) int {
 	return int(float64(p.workerSize) * ReplicationWorkerMultiplier)
 }

-func (p *ReplicationPool) queueReplicaFailedTask(ri ReplicateObjectInfo) {
-	if p == nil {
-		return
-	}
-	select {
-	case <-GlobalContext.Done():
-		p.once.Do(func() {
-			close(p.replicaCh)
-			close(p.mrfReplicaCh)
-			close(p.existingReplicaCh)
-		})
-	case p.mrfReplicaCh <- ri:
-	default:
-		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Replication failed workers could not keep up with healing failures - consider increasing number of replication failed workers with `mc admin config set api replication_failed_workers=%d`", p.suggestedWorkers(true)), replicationSubsystem)
-	}
-}
-
 func (p *ReplicationPool) queueReplicaTask(ri ReplicateObjectInfo) {
 	if p == nil {
 		return
 	}
-	var ch chan ReplicateObjectInfo
+	var ch, healCh chan ReplicateObjectInfo
 	switch ri.OpType {
 	case replication.ExistingObjectReplicationType:
 		ch = p.existingReplicaCh
 	case replication.HealReplicationType:
-		fallthrough
+		ch = p.mrfReplicaCh
+		healCh = p.replicaCh
 	default:
 		ch = p.replicaCh
 	}
@@ -1506,9 +1527,10 @@ func (p *ReplicationPool) queueReplicaTask(ri ReplicateObjectInfo) {
 			close(p.mrfReplicaCh)
 			close(p.existingReplicaCh)
 		})
+	case healCh <- ri:
 	case ch <- ri:
 	default:
-		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Replication workers could not keep up with incoming traffic - consider increasing number of replication workers with `mc admin config set api replication_workers=%d`", p.suggestedWorkers(false)), replicationSubsystem)
+		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to keep up with incoming traffic - we recommend increasing number of replicate object workers with `mc admin config set api replication_workers=%d`", p.suggestedWorkers(false)), string(replicationSubsystem))
 	}
 }

@@ -1545,7 +1567,7 @@ func (p *ReplicationPool) queueReplicaDeleteTask(doi DeletedObjectReplicationInf
 		})
 	case ch <- doi:
 	default:
-		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Replication workers could not keep up with incoming traffic - consider increasing number of replication workers with `mc admin config set api replication_workers=%d`", p.suggestedWorkers(false)), replicationSubsystem)
+		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to keep up with incoming deletes - we recommend increasing number of replicate workers with `mc admin config set api replication_workers=%d`", p.suggestedWorkers(false)), string(replicationSubsystem))
 	}
 }

@@ -1649,7 +1671,7 @@ func proxyHeadToRepTarget(ctx context.Context, bucket, object string, rs *HTTPRa
 	}
 	for _, t := range proxyTargets.Targets {
 		tgt = globalBucketTargetSys.GetRemoteTargetClient(ctx, t.Arn)
-		if tgt == nil || tgt.IsOffline() {
+		if tgt == nil || globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
 			continue
 		}
 		// if proxying explicitly disabled on remote target
@@ -1722,10 +1744,11 @@ func proxyHeadToReplicationTarget(ctx context.Context, bucket, object string, rs
 }

 func scheduleReplication(ctx context.Context, objInfo ObjectInfo, o ObjectLayer, dsc ReplicateDecision, opType replication.Type) {
+	ri := ReplicateObjectInfo{ObjectInfo: objInfo, OpType: opType, Dsc: dsc, EventType: ReplicateIncoming}
 	if dsc.Synchronous() {
-		replicateObject(ctx, ReplicateObjectInfo{ObjectInfo: objInfo, OpType: opType, Dsc: dsc}, o, ReplicateIncoming)
+		replicateObject(ctx, ri, o)
 	} else {
-		globalReplicationPool.queueReplicaTask(ReplicateObjectInfo{ObjectInfo: objInfo, OpType: opType, Dsc: dsc})
+		globalReplicationPool.queueReplicaTask(ri)
 	}
 	if sz, err := objInfo.GetActualSize(); err == nil {
 		for arn := range dsc.targetsMap {
@@ -1794,6 +1817,7 @@ func (c replicationConfig) Resync(ctx context.Context, oi ObjectInfo, dsc *Repli
 	objInfo.VersionPurgeStatusInternal = ""
 	objInfo.ReplicationStatus = ""
 	objInfo.VersionPurgeStatus = ""
+	delete(objInfo.UserDefined, xhttp.AmzBucketReplicationStatus)
 	resyncdsc := mustReplicate(ctx, oi.Bucket, oi.Name, getMustReplicateOptions(objInfo, replication.ExistingObjectReplicationType, ObjectOptions{}))
 	dsc = &resyncdsc
 	return c.resync(oi, dsc, tgtStatuses)
@@ -1956,10 +1980,10 @@ func getLatestReplicationStats(bucket string, u BucketUsageInfo) (s BucketReplic
 	return calculateBucketReplicationStats(bucket, u, bucketStats)
 }

-const resyncTimeInterval = time.Minute * 10
+const resyncTimeInterval = time.Minute * 1

-// periodicResyncMetaSave saves in-memory resync meta stats to disk in periodic intervals
-func (p *ReplicationPool) periodicResyncMetaSave(ctx context.Context, objectAPI ObjectLayer) {
+// updateResyncStatus persists in-memory resync metadata stats to disk at periodic intervals
+func (p *ReplicationPool) updateResyncStatus(ctx context.Context, objectAPI ObjectLayer) {
 	resyncTimer := time.NewTimer(resyncTimeInterval)
 	defer resyncTimer.Stop()

@@ -1980,7 +2004,7 @@ func (p *ReplicationPool) periodicResyncMetaSave(ctx context.Context, objectAPI
 				if updt {
 					brs.LastUpdate = now
 					if err := saveResyncStatus(ctx, bucket, brs, objectAPI); err != nil {
-						logger.LogIf(ctx, fmt.Errorf("Could not save resync metadata to disk for %s - %w", bucket, err))
+						logger.LogIf(ctx, fmt.Errorf("Could not save resync metadata to drive for %s - %w", bucket, err))
 						continue
 					}
 				}
@@ -2041,9 +2065,9 @@ func resyncBucket(ctx context.Context, bucket, arn string, heal bool, objectAPI
 		return
 	}

-	// Walk through all object versions - note ascending order of walk needed to ensure delete marker replicated to
-	// target after object version is first created.
-	if err := objectAPI.Walk(ctx, bucket, "", objInfoCh, ObjectOptions{WalkAscending: true}); err != nil {
+	// Walk through all object versions - Walk() is always in ascending order needed to ensure
+	// delete marker replicated to target after object version is first created.
+	if err := objectAPI.Walk(ctx, bucket, "", objInfoCh, ObjectOptions{}); err != nil {
 		logger.LogIf(ctx, err)
 		return
 	}
@@ -2085,13 +2109,15 @@ func resyncBucket(ctx context.Context, bucket, arn string, heal bool, objectAPI
 					DeleteMarkerMTime:     DeleteMarkerMTime{roi.ModTime},
 					DeleteMarker:          roi.DeleteMarker,
 				},
-				Bucket: roi.Bucket,
-				OpType: replication.ExistingObjectReplicationType,
+				Bucket:    roi.Bucket,
+				OpType:    replication.ExistingObjectReplicationType,
+				EventType: ReplicateExistingDelete,
 			}
-			replicateDelete(ctx, doi, objectAPI, ReplicateDelete)
+			replicateDelete(ctx, doi, objectAPI)
 		} else {
 			roi.OpType = replication.ExistingObjectReplicationType
-			replicateObject(ctx, roi, objectAPI, ReplicateExisting)
+			roi.EventType = ReplicateExisting
+			replicateObject(ctx, roi, objectAPI)
 		}
 		_, err = tgt.StatObject(ctx, tgt.Bucket, roi.Name, miniogo.StatObjectOptions{
 			VersionID: roi.VersionID,
@@ -2205,18 +2231,21 @@ func (p *ReplicationPool) loadResync(ctx context.Context, buckets []BucketInfo,
 	for index := range buckets {
 		meta, err := loadBucketResyncMetadata(ctx, buckets[index].Name, objAPI)
 		if err != nil {
-			if errors.Is(err, errVolumeNotFound) {
-				meta = newBucketResyncStatus(buckets[index].Name)
-			} else {
+			if !errors.Is(err, errVolumeNotFound) {
 				logger.LogIf(ctx, err)
-				continue
 			}
+			continue
 		}
+		p.resyncState.Lock()
 		p.resyncState.statusMap[buckets[index].Name] = meta
+		p.resyncState.Unlock()
 	}
 	for index := range buckets {
 		bucket := buckets[index].Name
+		p.resyncState.RLock()
 		m, ok := p.resyncState.statusMap[bucket]
+		p.resyncState.RUnlock()
+
 		if ok {
 			for arn, st := range m.TargetsMap {
 				if st.ResyncStatus == ResyncFailed || st.ResyncStatus == ResyncStarted {
@@ -2283,3 +2312,365 @@ func saveResyncStatus(ctx context.Context, bucket string, brs BucketReplicationR
 	configFile := path.Join(bucketMetaPrefix, bucket, replicationDir, resyncFileName)
 	return saveConfig(ctx, objectAPI, configFile, buf)
 }
+
+// getReplicationDiff returns unreplicated objects in a channel
+func getReplicationDiff(ctx context.Context, objAPI ObjectLayer, bucket string, opts madmin.ReplDiffOpts) (diffCh chan madmin.DiffInfo, err error) {
+	objInfoCh := make(chan ObjectInfo)
+	if err := objAPI.Walk(ctx, bucket, opts.Prefix, objInfoCh, ObjectOptions{}); err != nil {
+		logger.LogIf(ctx, err)
+		return diffCh, err
+	}
+	cfg, err := getReplicationConfig(ctx, bucket)
+	if err != nil {
+		logger.LogIf(ctx, err)
+		return diffCh, err
+	}
+	tgts, err := globalBucketTargetSys.ListBucketTargets(ctx, bucket)
+	if err != nil {
+		logger.LogIf(ctx, err)
+		return diffCh, err
+	}
+	rcfg := replicationConfig{
+		Config:  cfg,
+		remotes: tgts,
+	}
+	diffCh = make(chan madmin.DiffInfo, 4000)
+	go func() {
+		defer close(diffCh)
+		for obj := range objInfoCh {
+			// Ignore object prefixes which are excluded
+			// from versioning via the MinIO bucket versioning extension.
+			if globalBucketVersioningSys.PrefixSuspended(bucket, obj.Name) {
+				continue
+			}
+			roi := getHealReplicateObjectInfo(obj, rcfg)
+			switch roi.ReplicationStatus {
+			case replication.Completed, replication.Replica:
+				if !opts.Verbose {
+					continue
+				}
+				fallthrough
+			default:
+				// ignore pre-existing objects that don't satisfy replication rule(s)
+				if roi.ReplicationStatus.Empty() && !roi.ExistingObjResync.mustResync() {
+					continue
+				}
+				tgtsMap := make(map[string]madmin.TgtDiffInfo)
+				for arn, st := range roi.TargetStatuses {
+					if opts.ARN == "" || opts.ARN == arn {
+						if !opts.Verbose && (st == replication.Completed || st == replication.Replica) {
+							continue
+						}
+						tgtsMap[arn] = madmin.TgtDiffInfo{
+							ReplicationStatus: st.String(),
+						}
+					}
+				}
+				for arn, st := range roi.TargetPurgeStatuses {
+					if opts.ARN == "" || opts.ARN == arn {
+						if !opts.Verbose && st == Complete {
+							continue
+						}
+						t, ok := tgtsMap[arn]
+						if !ok {
+							t = madmin.TgtDiffInfo{}
+						}
+						t.DeleteReplicationStatus = string(st)
+						tgtsMap[arn] = t
+					}
+				}
+				select {
+				case diffCh <- madmin.DiffInfo{
+					Object:                  obj.Name,
+					VersionID:               obj.VersionID,
+					LastModified:            obj.ModTime,
+					IsDeleteMarker:          obj.DeleteMarker,
+					ReplicationStatus:       string(roi.ReplicationStatus),
+					DeleteReplicationStatus: string(roi.VersionPurgeStatus),
+					ReplicationTimestamp:    roi.ReplicationTimestamp,
+					Targets:                 tgtsMap,
+				}:
+				case <-ctx.Done():
+					return
+				}
+			}
+		}
+	}()
+	return diffCh, nil
+}
+
+// QueueReplicationHeal is a wrapper for queueReplicationHeal
+func QueueReplicationHeal(ctx context.Context, bucket string, oi ObjectInfo) {
+	// un-versioned case
+	if oi.VersionID == "" {
+		return
+	}
+	rcfg, _, _ := globalBucketMetadataSys.GetReplicationConfig(ctx, bucket)
+	tgts, _ := globalBucketTargetSys.ListBucketTargets(ctx, bucket)
+	queueReplicationHeal(ctx, bucket, oi, replicationConfig{
+		Config:  rcfg,
+		remotes: tgts,
+	})
+}
+
+// queueReplicationHeal enqueues objects that failed replication OR eligible for resyncing through
+// an ongoing resync operation or via existing objects replication configuration setting.
+func queueReplicationHeal(ctx context.Context, bucket string, oi ObjectInfo, rcfg replicationConfig) (roi ReplicateObjectInfo) {
+	// un-versioned case
+	if oi.VersionID == "" {
+		return roi
+	}
+
+	if rcfg.Config == nil || rcfg.remotes == nil {
+		return roi
+	}
+	roi = getHealReplicateObjectInfo(oi, rcfg)
+	if !roi.Dsc.ReplicateAny() {
+		return
+	}
+	// early return if replication already done, otherwise we need to determine if this
+	// version is an existing object that needs healing.
+	if oi.ReplicationStatus == replication.Completed && oi.VersionPurgeStatus.Empty() && !roi.ExistingObjResync.mustResync() {
+		return
+	}
+
+	if roi.DeleteMarker || !roi.VersionPurgeStatus.Empty() {
+		versionID := ""
+		dmVersionID := ""
+		if roi.VersionPurgeStatus.Empty() {
+			dmVersionID = roi.VersionID
+		} else {
+			versionID = roi.VersionID
+		}
+
+		dv := DeletedObjectReplicationInfo{
+			DeletedObject: DeletedObject{
+				ObjectName:            roi.Name,
+				DeleteMarkerVersionID: dmVersionID,
+				VersionID:             versionID,
+				ReplicationState:      roi.getReplicationState(roi.Dsc.String(), versionID, true),
+				DeleteMarkerMTime:     DeleteMarkerMTime{roi.ModTime},
+				DeleteMarker:          roi.DeleteMarker,
+			},
+			Bucket:    roi.Bucket,
+			OpType:    replication.HealReplicationType,
+			EventType: ReplicateHealDelete,
+		}
+		// heal delete marker replication failure or versioned delete replication failure
+		if roi.ReplicationStatus == replication.Pending ||
+			roi.ReplicationStatus == replication.Failed ||
+			roi.VersionPurgeStatus == Failed || roi.VersionPurgeStatus == Pending {
+			globalReplicationPool.queueReplicaDeleteTask(dv)
+			return
+		}
+		// if replication status is Complete on DeleteMarker and existing object resync required
+		if roi.ExistingObjResync.mustResync() && (roi.ReplicationStatus == replication.Completed || roi.ReplicationStatus.Empty()) {
+			queueReplicateDeletesWrapper(dv, roi.ExistingObjResync)
+			return
+		}
+		return
+	}
+	if roi.ExistingObjResync.mustResync() {
+		roi.OpType = replication.ExistingObjectReplicationType
+	}
+	switch roi.ReplicationStatus {
+	case replication.Pending, replication.Failed:
+		roi.EventType = ReplicateHeal
+		globalReplicationPool.queueReplicaTask(roi)
+		return
+	}
+	if roi.ExistingObjResync.mustResync() {
+		roi.EventType = ReplicateExisting
+		globalReplicationPool.queueReplicaTask(roi)
+	}
+	return
+}
+
+const mrfTimeInterval = 5 * time.Minute
+
+func (p *ReplicationPool) persistMRF() {
+	var mu sync.Mutex
+	entries := make(map[string]MRFReplicateEntry)
+	mTimer := time.NewTimer(mrfTimeInterval)
+	defer mTimer.Stop()
+	saveMRFToDisk := func(drain bool) {
+		mu.Lock()
+		defer mu.Unlock()
+		if len(entries) == 0 {
+			return
+		}
+		cctx := p.ctx
+		if drain {
+			cctx = context.Background()
+			// drain all mrf entries and save to disk
+			for e := range p.mrfSaveCh {
+				entries[e.versionID] = e
+			}
+		}
+		if err := p.saveMRFEntries(cctx, entries); err != nil {
+			logger.LogOnceIf(p.ctx, fmt.Errorf("Unable to persist replication failures to disk:%w", err), string(replicationSubsystem))
+		}
+		entries = make(map[string]MRFReplicateEntry)
+		return
+	}
+	for {
+		select {
+		case <-mTimer.C:
+			saveMRFToDisk(false)
+			mTimer.Reset(mrfTimeInterval)
+		case <-p.ctx.Done():
+			close(p.mrfSaveCh)
+			saveMRFToDisk(true)
+			return
+		case e, ok := <-p.mrfSaveCh:
+			if !ok {
+				return
+			}
+			var cnt int
+			mu.Lock()
+			entries[e.versionID] = e
+			cnt = len(entries)
+			mu.Unlock()
+			if cnt >= cap(p.mrfSaveCh) || len(p.mrfSaveCh) >= int(0.8*float32(cap(p.mrfSaveCh))) {
+				saveMRFToDisk(true)
+			}
+		}
+	}
+}
+
+func (p *ReplicationPool) queueMRFSave(entry MRFReplicateEntry) {
+	if p == nil {
+		return
+	}
+	select {
+	case <-GlobalContext.Done():
+		return
+	case p.mrfSaveCh <- entry:
+	}
+}
+
+// save mrf entries to mrf_<uuid>.bin
+func (p *ReplicationPool) saveMRFEntries(ctx context.Context, entries map[string]MRFReplicateEntry) error {
+	if len(entries) == 0 {
+		return nil
+	}
+	v := MRFReplicateEntries{
+		Entries: entries,
+		Version: mrfMetaVersionV1,
+	}
+	data := make([]byte, 4, v.Msgsize()+4)
+
+	// Initialize the resync meta header.
+	binary.LittleEndian.PutUint16(data[0:2], resyncMetaFormat)
+	binary.LittleEndian.PutUint16(data[2:4], resyncMetaVersion)
+
+	buf, err := v.MarshalMsg(data)
+	if err != nil {
+		return err
+	}
+
+	configFile := path.Join(replicationMRFDir, mustGetUUID()+".bin")
+	err = saveConfig(ctx, p.objLayer, configFile, buf)
+	return err
+}
+
+// load mrf entries from disk
+func (p *ReplicationPool) loadMRF(fileName string) (re MRFReplicateEntries, e error) {
+	data, err := readConfig(p.ctx, p.objLayer, fileName)
+	if err != nil && err != errConfigNotFound {
+		return re, err
+	}
+	if len(data) == 0 {
+		// Seems to be empty.
+		return re, nil
+	}
+	if len(data) <= 4 {
+		return re, fmt.Errorf("replication mrf: no data")
+	}
+	// Read resync meta header
+	switch binary.LittleEndian.Uint16(data[0:2]) {
+	case mrfMetaFormat:
+	default:
+		return re, fmt.Errorf("replication mrf: unknown format: %d", binary.LittleEndian.Uint16(data[0:2]))
+	}
+	switch binary.LittleEndian.Uint16(data[2:4]) {
+	case mrfMetaVersion:
+	default:
+		return re, fmt.Errorf("replication mrf: unknown version: %d", binary.LittleEndian.Uint16(data[2:4]))
+	}
+	// OK, parse data.
+	if _, err = re.UnmarshalMsg(data[4:]); err != nil {
+		return re, err
+	}
+
+	switch re.Version {
+	case mrfMetaVersionV1:
+	default:
+		return re, fmt.Errorf("unexpected mrf meta version: %d", re.Version)
+	}
+	return re, nil
+}
+
+func (p *ReplicationPool) processMRF() {
+	if p == nil || p.objLayer == nil {
+		return
+	}
+	pTimer := time.NewTimer(mrfTimeInterval)
+	defer pTimer.Stop()
+	for {
+		select {
+		case <-pTimer.C:
+			// skip healing if all targets are offline
+			var offlineCnt int
+			tgts := globalBucketTargetSys.ListTargets(p.ctx, "", "")
+			for _, tgt := range tgts {
+				if globalBucketTargetSys.isOffline(tgt.URL()) {
+					offlineCnt++
+				}
+			}
+			if len(tgts) == offlineCnt {
+				pTimer.Reset(mrfTimeInterval)
+				continue
+			}
+			objCh := make(chan ObjectInfo)
+			cctx, cancelFn := context.WithCancel(p.ctx)
+			if err := p.objLayer.Walk(cctx, minioMetaBucket, replicationMRFDir, objCh, ObjectOptions{}); err != nil {
+				pTimer.Reset(mrfTimeInterval)
+				cancelFn()
+				logger.LogIf(p.ctx, err)
+				continue
+			}
+			for item := range objCh {
+				if err := p.queueMRFHeal(item.Name); err == nil {
+					p.objLayer.DeleteObject(p.ctx, minioMetaBucket, item.Name, ObjectOptions{})
+				}
+			}
+			pTimer.Reset(mrfTimeInterval)
+			cancelFn()
+		case <-p.ctx.Done():
+			return
+		}
+	}
+}
+
+// process sends error logs to the heal channel for an attempt to heal replication.
+func (p *ReplicationPool) queueMRFHeal(file string) error {
+	if p == nil || p.objLayer == nil {
+		return errServerNotInitialized
+	}
+
+	mrfRec, err := p.loadMRF(file)
+	if err != nil {
+		return err
+	}
+	for vID, e := range mrfRec.Entries {
+		oi, err := p.objLayer.GetObjectInfo(p.ctx, e.Bucket, e.Object, ObjectOptions{
+			VersionID: vID,
+		})
+		if err != nil {
+			continue
+		}
+		QueueReplicationHeal(p.ctx, e.Bucket, oi)
+	}
+	return nil
+}
@@ -19,6 +19,8 @@ package cmd

 import (
 	"context"
+	"fmt"
+	"net/url"
 	"sync"
 	"time"

@@ -35,7 +37,9 @@ import (
 )

 const (
-	defaultHealthCheckDuration = 30 * time.Second
+	defaultHealthCheckDuration = 5 * time.Second
+	// default interval for reload of all remote target endpoints
+	defaultHealthCheckReloadDuration = 30 * time.Minute
 )

 // BucketTargetSys represents bucket targets subsystem
@@ -43,6 +47,113 @@ type BucketTargetSys struct {
 	sync.RWMutex
 	arnRemotesMap map[string]*TargetClient
 	targetsMap    map[string][]madmin.BucketTarget
+	hMutex        sync.RWMutex
+	hc            map[string]epHealth
+	hcClient      *madmin.AnonymousClient
+}
+
+// epHealth struct represents health of a replication target endpoint.
+type epHealth struct {
+	Endpoint string
+	Scheme   string
+	Online   bool
+}
+
+// isOffline returns current liveness result of remote target. Add endpoint to
+// healthcheck map if missing and default to online status
+func (sys *BucketTargetSys) isOffline(ep *url.URL) bool {
+	sys.hMutex.RLock()
+	defer sys.hMutex.RUnlock()
+	if h, ok := sys.hc[ep.Host]; ok {
+		return !h.Online
+	}
+	go sys.initHC(ep)
+	return false
+}
+
+func (sys *BucketTargetSys) initHC(ep *url.URL) {
+	sys.hMutex.Lock()
+	sys.hc[ep.Host] = epHealth{
+		Endpoint: ep.Host,
+		Scheme:   ep.Scheme,
+		Online:   true,
+	}
+	sys.hMutex.Unlock()
+}
+
+// newHCClient initializes an anonymous client for performing health check on the remote endpoints
+func newHCClient() *madmin.AnonymousClient {
+	clnt, e := madmin.NewAnonymousClientNoEndpoint()
+	if e != nil {
+		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to initialize health check client"), string(replicationSubsystem))
+		return nil
+	}
+	clnt.SetCustomTransport(globalRemoteTargetTransport)
+	return clnt
+}
+
+// heartBeat performs liveness check on remote endpoints.
+func (sys *BucketTargetSys) heartBeat(ctx context.Context) {
+	hcTimer := time.NewTimer(defaultHealthCheckDuration)
+	defer hcTimer.Stop()
+	for {
+		select {
+		case <-hcTimer.C:
+			sys.hMutex.RLock()
+			var eps []madmin.ServerProperties
+			for _, ep := range sys.hc {
+				eps = append(eps, madmin.ServerProperties{Endpoint: ep.Endpoint, Scheme: ep.Scheme})
+			}
+			sys.hMutex.RUnlock()
+
+			if len(eps) > 0 {
+				cctx, cancel := context.WithTimeout(ctx, 30*time.Second)
+				defer cancel()
+				m := map[string]epHealth{}
+				for result := range sys.hcClient.Alive(cctx, madmin.AliveOpts{}, eps...) {
+					var online bool
+					if result.Error == nil {
+						online = result.Online
+					}
+					m[result.Endpoint.Host] = epHealth{
+						Endpoint: result.Endpoint.Host,
+						Scheme:   result.Endpoint.Scheme,
+						Online:   online,
+					}
+				}
+				sys.hMutex.Lock()
+				sys.hc = m
+				sys.hMutex.Unlock()
+			}
+			hcTimer.Reset(defaultHealthCheckDuration)
+		case <-ctx.Done():
+			return
+		}
+	}
+}
+
+// periodically rebuild the healthcheck map from list of targets to clear
+// out stale endpoints
+func (sys *BucketTargetSys) reloadHealthCheckers(ctx context.Context) {
+	m := make(map[string]epHealth)
+	tgts := sys.ListTargets(ctx, "", "")
+	for _, t := range tgts {
+		if _, ok := m[t.Endpoint]; !ok {
+			scheme := "http"
+			if t.Secure {
+				scheme = "https"
+			}
+			m[t.Endpoint] = epHealth{
+				Online:   true,
+				Endpoint: t.Endpoint,
+				Scheme:   scheme,
+			}
+		}
+	}
+	sys.hMutex.Lock()
+	// swap out the map
+	sys.hc = m
+	sys.hMutex.Unlock()
 }

 // ListTargets lists bucket targets across tenant or for individual bucket, and returns
@@ -91,9 +202,6 @@ func (sys *BucketTargetSys) Delete(bucket string) {
 		return
 	}
 	for _, t := range tgts {
-		if tgt, ok := sys.arnRemotesMap[t.Arn]; ok && tgt.healthCancelFn != nil {
-			tgt.healthCancelFn()
-		}
 		delete(sys.arnRemotesMap, t.Arn)
 	}
 	delete(sys.targetsMap, bucket)
@@ -116,7 +224,7 @@ func (sys *BucketTargetSys) SetTarget(ctx context.Context, bucket string, tgt *m
 		if minio.ToErrorResponse(err).Code == "NoSuchBucket" {
 			return BucketRemoteTargetNotFound{Bucket: tgt.TargetBucket}
 		}
-		return BucketRemoteConnectionErr{Bucket: tgt.TargetBucket, Err: err}
+		return RemoteTargetConnectionErr{Bucket: tgt.TargetBucket, Err: err}
 	}
 	if tgt.Type == madmin.ReplicationService {
 		if !globalBucketVersioningSys.Enabled(bucket) {
@@ -124,7 +232,7 @@ func (sys *BucketTargetSys) SetTarget(ctx context.Context, bucket string, tgt *m
 		}
 		vcfg, err := clnt.GetBucketVersioning(ctx, tgt.TargetBucket)
 		if err != nil {
-			return BucketRemoteConnectionErr{Bucket: tgt.TargetBucket, Err: err}
+			return RemoteTargetConnectionErr{Bucket: tgt.TargetBucket, Err: err}
 		}
 		if vcfg.Status != string(versioning.Enabled) {
 			return BucketRemoteTargetNotVersioned{Bucket: tgt.TargetBucket}
@@ -152,10 +260,7 @@ func (sys *BucketTargetSys) SetTarget(ctx context.Context, bucket string, tgt *m
 	if !found && !update {
 		newtgts = append(newtgts, *tgt)
 	}
-	// cancel health check for previous target client to avoid leak.
-	if prevClnt, ok := sys.arnRemotesMap[tgt.Arn]; ok && prevClnt.healthCancelFn != nil {
-		prevClnt.healthCancelFn()
-	}
+
 	sys.targetsMap[bucket] = newtgts
 	sys.arnRemotesMap[tgt.Arn] = clnt
 	sys.updateBandwidthLimit(bucket, tgt.BandwidthLimit)
@@ -227,9 +332,6 @@ func (sys *BucketTargetSys) RemoveTarget(ctx context.Context, bucket, arnStr str
 		return BucketRemoteTargetNotFound{Bucket: bucket}
 	}
 	sys.targetsMap[bucket] = targets
-	if tgt, ok := sys.arnRemotesMap[arnStr]; ok && tgt.healthCancelFn != nil {
-		tgt.healthCancelFn()
-	}
 	delete(sys.arnRemotesMap, arnStr)
 	sys.updateBandwidthLimit(bucket, 0)
 	return nil
@@ -258,11 +360,29 @@ func (sys *BucketTargetSys) GetRemoteBucketTargetByArn(ctx context.Context, buck
 }

 // NewBucketTargetSys - creates new replication system.
-func NewBucketTargetSys() *BucketTargetSys {
-	return &BucketTargetSys{
+func NewBucketTargetSys(ctx context.Context) *BucketTargetSys {
+	sys := &BucketTargetSys{
 		arnRemotesMap: make(map[string]*TargetClient),
 		targetsMap:    make(map[string][]madmin.BucketTarget),
+		hc:            make(map[string]epHealth),
+		hcClient:      newHCClient(),
 	}
+	// reload healthcheck endpoints map periodically to remove stale endpoints from the map.
+	go func() {
+		rTimer := time.NewTimer(defaultHealthCheckReloadDuration)
+		defer rTimer.Stop()
+		for {
+			select {
+			case <-rTimer.C:
+				sys.reloadHealthCheckers(ctx)
+				rTimer.Reset(defaultHealthCheckReloadDuration)
+			case <-ctx.Done():
+				return
+			}
+		}
+	}()
+	go sys.heartBeat(ctx)
+	return sys
 }

 // UpdateAllTargets updates target to reflect metadata updates
@@ -276,9 +396,6 @@ func (sys *BucketTargetSys) UpdateAllTargets(bucket string, tgts *madmin.BucketT
 	// Remove existingtarget and arn association
 	if tgts, ok := sys.targetsMap[bucket]; ok {
 		for _, t := range tgts {
-			if tgt, ok := sys.arnRemotesMap[t.Arn]; ok && tgt.healthCancelFn != nil {
-				tgt.healthCancelFn()
-			}
 			delete(sys.arnRemotesMap, t.Arn)
 		}
 		delete(sys.targetsMap, bucket)
@@ -345,10 +462,6 @@ func (sys *BucketTargetSys) getRemoteTargetClient(tcfg *madmin.BucketTarget) (*T
 	if tcfg.HealthCheckDuration >= 1 { // require minimum health check duration of 1 sec.
 		hcDuration = tcfg.HealthCheckDuration
 	}
-	cancelFn, err := api.HealthCheck(hcDuration)
-	if err != nil {
-		return nil, err
-	}
 	tc := &TargetClient{
 		Client:              api,
 		healthCheckDuration: hcDuration,
@@ -356,9 +469,10 @@ func (sys *BucketTargetSys) getRemoteTargetClient(tcfg *madmin.BucketTarget) (*T
 		Bucket:              tcfg.TargetBucket,
 		StorageClass:        tcfg.StorageClass,
 		disableProxy:        tcfg.DisableProxy,
-		healthCancelFn:      cancelFn,
 		ARN:                 tcfg.Arn,
 		ResetID:             tcfg.ResetID,
+		Endpoint:            tcfg.Endpoint,
+		Secure:              tcfg.Secure,
 	}
 	return tc, nil
 }
@@ -432,7 +546,8 @@ type TargetClient struct {
 	replicateSync       bool
 	StorageClass        string // storage class on remote
 	disableProxy        bool
-	healthCancelFn      context.CancelFunc // cancellation function for client healthcheck
-	ARN                 string             // ARN to uniquely identify remote target
+	ARN                 string // ARN to uniquely identify remote target
 	ResetID             string
+	Endpoint            string
+	Secure              bool
 }
@@ -143,7 +143,7 @@ func (api objectAPIHandlers) GetBucketVersioningHandler(w http.ResponseWriter, r
 	}

 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -48,6 +48,7 @@ import (
 	"github.com/inconshreveable/mousetrap"
 	dns2 "github.com/miekg/dns"
 	"github.com/minio/cli"
+	consoleoauth2 "github.com/minio/console/pkg/auth/idp/oauth2"
 	consoleCerts "github.com/minio/console/pkg/certs"
 	"github.com/minio/console/restapi"
 	"github.com/minio/console/restapi/operations"
@@ -205,28 +206,6 @@ func minioConfigToConsoleFeatures() {
 	if globalLDAPConfig.Enabled {
 		os.Setenv("CONSOLE_LDAP_ENABLED", config.EnableOn)
 	}
-	// if IDP is enabled, set IDP environment variables
-	if globalOpenIDConfig.ProviderCfgs[config.Default] != nil {
-		os.Setenv("CONSOLE_IDP_URL", globalOpenIDConfig.ProviderCfgs[config.Default].URL.String())
-		os.Setenv("CONSOLE_IDP_CLIENT_ID", globalOpenIDConfig.ProviderCfgs[config.Default].ClientID)
-		os.Setenv("CONSOLE_IDP_SECRET", globalOpenIDConfig.ProviderCfgs[config.Default].ClientSecret)
-		os.Setenv("CONSOLE_IDP_HMAC_SALT", globalDeploymentID)
-		os.Setenv("CONSOLE_IDP_HMAC_PASSPHRASE", globalOpenIDConfig.ProviderCfgs[config.Default].ClientID)
-		os.Setenv("CONSOLE_IDP_SCOPES", strings.Join(globalOpenIDConfig.ProviderCfgs[config.Default].DiscoveryDoc.ScopesSupported, ","))
-		if globalOpenIDConfig.ProviderCfgs[config.Default].ClaimUserinfo {
-			os.Setenv("CONSOLE_IDP_USERINFO", config.EnableOn)
-		}
-		if globalOpenIDConfig.ProviderCfgs[config.Default].RedirectURIDynamic {
-			// Enable dynamic redirect-uri's based on incoming 'host' header,
-			// Overrides any other callback URL.
-			os.Setenv("CONSOLE_IDP_CALLBACK_DYNAMIC", config.EnableOn)
-		}
-		if globalOpenIDConfig.ProviderCfgs[config.Default].RedirectURI != "" {
-			os.Setenv("CONSOLE_IDP_CALLBACK", globalOpenIDConfig.ProviderCfgs[config.Default].RedirectURI)
-		} else {
-			os.Setenv("CONSOLE_IDP_CALLBACK", getConsoleEndpoints()[0]+"/oauth_callback")
-		}
-	}
 	os.Setenv("CONSOLE_MINIO_REGION", globalSite.Region)
 	os.Setenv("CONSOLE_CERT_PASSWD", env.Get("MINIO_CERT_PASSWD", ""))
 	if globalSubnetConfig.License != "" {
@@ -240,6 +219,29 @@ func minioConfigToConsoleFeatures() {
 	}
 }

+func buildOpenIDConsoleConfig() consoleoauth2.OpenIDPCfg {
+	m := make(map[string]consoleoauth2.ProviderConfig, len(globalOpenIDConfig.ProviderCfgs))
+	for name, cfg := range globalOpenIDConfig.ProviderCfgs {
+		callback := getConsoleEndpoints()[0] + "/oauth_callback"
+		if cfg.RedirectURI != "" {
+			callback = cfg.RedirectURI
+		}
+		m[name] = consoleoauth2.ProviderConfig{
+			URL:                     cfg.URL.String(),
+			DisplayName:             cfg.DisplayName,
+			ClientID:                cfg.ClientID,
+			ClientSecret:            cfg.ClientSecret,
+			HMACSalt:                globalDeploymentID,
+			HMACPassphrase:          cfg.ClientID,
+			Scopes:                  strings.Join(cfg.DiscoveryDoc.ScopesSupported, ","),
+			Userinfo:                cfg.ClaimUserinfo,
+			RedirectCallbackDynamic: cfg.RedirectURIDynamic,
+			RedirectCallback:        callback,
+		}
+	}
+	return m
+}
+
 func initConsoleServer() (*restapi.Server, error) {
 	// unset all console_ environment variables.
 	for _, cenv := range env.List(consolePrefix) {
@@ -273,6 +275,12 @@ func initConsoleServer() (*restapi.Server, error) {
 		api.Logger = noLog
 	}

+	// Pass in console application config. This needs to happen before the
+	// ConfigureAPI() call.
+	restapi.GlobalMinIOConfig = restapi.MinIOConfig{
+		OpenIDProviders: buildOpenIDConsoleConfig(),
+	}
+
 	server := restapi.NewServer(api)
 	// register all APIs
 	server.ConfigureAPI()
@@ -430,7 +438,6 @@ func handleCommonCmdArgs(ctx *cli.Context) {
 		if err != nil {
 			logger.FatalIf(err, "Unable to get free port for console on the host")
 		}
-		globalMinioConsolePortAuto = true
 		consoleAddr = net.JoinHostPort("", p.String())
 	}

@@ -501,8 +508,7 @@ func parsEnvEntry(envEntry string) (envKV, error) {
 			Skip: true,
 		}, nil
 	}
-	const envSeparator = "="
-	envTokens := strings.SplitN(strings.TrimSpace(strings.TrimPrefix(envEntry, "export")), envSeparator, 2)
+	envTokens := strings.SplitN(strings.TrimSpace(strings.TrimPrefix(envEntry, "export")), config.EnvSeparator, 2)
 	if len(envTokens) != 2 {
 		return envKV{}, fmt.Errorf("envEntry malformed; %s, expected to be of form 'KEY=value'", envEntry)
 	}
@@ -528,7 +534,7 @@ func parsEnvEntry(envEntry string) (envKV, error) {
 // the environment values from a file, in the form "key, value".
 // in a structured form.
 func minioEnvironFromFile(envConfigFile string) ([]envKV, error) {
-	f, err := os.Open(envConfigFile)
+	f, err := Open(envConfigFile)
 	if err != nil {
 		return nil, err
 	}
@@ -823,43 +829,56 @@ func handleKMSConfig() {
 				endpoints = append(endpoints, strings.Join(lbls, ""))
 			}
 		}
-		// Manually load the certificate and private key into memory.
-		// We need to check whether the private key is encrypted, and
-		// if so, decrypt it using the user-provided password.
-		certBytes, err := os.ReadFile(env.Get(config.EnvKESClientCert, ""))
-		if err != nil {
-			logger.Fatal(err, "Unable to load KES client certificate as specified by the shell environment")
-		}
-		keyBytes, err := os.ReadFile(env.Get(config.EnvKESClientKey, ""))
-		if err != nil {
-			logger.Fatal(err, "Unable to load KES client private key as specified by the shell environment")
-		}
-		privateKeyPEM, rest := pem.Decode(bytes.TrimSpace(keyBytes))
-		if len(rest) != 0 {
-			logger.Fatal(errors.New("private key contains additional data"), "Unable to load KES client private key as specified by the shell environment")
-		}
-		if x509.IsEncryptedPEMBlock(privateKeyPEM) {
-			keyBytes, err = x509.DecryptPEMBlock(privateKeyPEM, []byte(env.Get(config.EnvKESClientPassword, "")))
-			if err != nil {
-				logger.Fatal(err, "Unable to decrypt KES client private key as specified by the shell environment")
-			}
-			keyBytes = pem.EncodeToMemory(&pem.Block{Type: privateKeyPEM.Type, Bytes: keyBytes})
-		}
-		certificate, err := tls.X509KeyPair(certBytes, keyBytes)
-		if err != nil {
-			logger.Fatal(err, "Unable to load KES client certificate as specified by the shell environment")
-		}
 		rootCAs, err := certs.GetRootCAs(env.Get(config.EnvKESServerCA, globalCertsCADir.Get()))
 		if err != nil {
 			logger.Fatal(err, fmt.Sprintf("Unable to load X.509 root CAs for KES from %q", env.Get(config.EnvKESServerCA, globalCertsCADir.Get())))
 		}

+		loadX509KeyPair := func(certFile, keyFile string) (tls.Certificate, error) {
+			// Manually load the certificate and private key into memory.
+			// We need to check whether the private key is encrypted, and
+			// if so, decrypt it using the user-provided password.
+			certBytes, err := os.ReadFile(certFile)
+			if err != nil {
+				return tls.Certificate{}, fmt.Errorf("Unable to load KES client certificate as specified by the shell environment: %v", err)
+			}
+			keyBytes, err := os.ReadFile(keyFile)
+			if err != nil {
+				return tls.Certificate{}, fmt.Errorf("Unable to load KES client private key as specified by the shell environment: %v", err)
+			}
+			privateKeyPEM, rest := pem.Decode(bytes.TrimSpace(keyBytes))
+			if len(rest) != 0 {
+				return tls.Certificate{}, errors.New("Unable to load KES client private key as specified by the shell environment: private key contains additional data")
+			}
+			if x509.IsEncryptedPEMBlock(privateKeyPEM) {
+				keyBytes, err = x509.DecryptPEMBlock(privateKeyPEM, []byte(env.Get(config.EnvKESClientPassword, "")))
+				if err != nil {
+					return tls.Certificate{}, fmt.Errorf("Unable to decrypt KES client private key as specified by the shell environment: %v", err)
+				}
+				keyBytes = pem.EncodeToMemory(&pem.Block{Type: privateKeyPEM.Type, Bytes: keyBytes})
+			}
+			certificate, err := tls.X509KeyPair(certBytes, keyBytes)
+			if err != nil {
+				return tls.Certificate{}, fmt.Errorf("Unable to load KES client certificate as specified by the shell environment: %v", err)
+			}
+			return certificate, nil
+		}
+
+		reloadCertEvents := make(chan tls.Certificate, 1)
+		certificate, err := certs.NewCertificate(env.Get(config.EnvKESClientCert, ""), env.Get(config.EnvKESClientKey, ""), loadX509KeyPair)
+		if err != nil {
+			logger.Fatal(err, "Failed to load KES client certificate")
+		}
+		certificate.Watch(context.Background(), 15*time.Minute, syscall.SIGHUP)
+		certificate.Notify(reloadCertEvents)
+
 		defaultKeyID := env.Get(config.EnvKESKeyName, "")
 		KMS, err := kms.NewWithConfig(kms.Config{
-			Endpoints:    endpoints,
-			DefaultKeyID: defaultKeyID,
-			Certificate:  certificate,
-			RootCAs:      rootCAs,
+			Endpoints:        endpoints,
+			DefaultKeyID:     defaultKeyID,
+			Certificate:      certificate,
+			ReloadCertEvents: reloadCertEvents,
+			RootCAs:          rootCAs,
 		})
 		if err != nil {
 			logger.Fatal(err, "Unable to initialize a connection to KES as specified by the shell environment")
@@ -869,7 +888,7 @@ func handleKMSConfig() {
 		// This implicitly checks that we can communicate to KES. We don't treat
 		// a policy error as failure condition since MinIO may not have the permission
 		// to create keys - just to generate/decrypt data encryption keys.
-		if err = KMS.CreateKey(defaultKeyID); err != nil && !errors.Is(err, kes.ErrKeyExists) && !errors.Is(err, kes.ErrNotAllowed) {
+		if err = KMS.CreateKey(context.Background(), defaultKeyID); err != nil && !errors.Is(err, kes.ErrKeyExists) && !errors.Is(err, kes.ErrNotAllowed) {
 			logger.Fatal(err, "Unable to initialize a connection to KES as specified by the shell environment")
 		}
 		GlobalKMS = KMS
@@ -909,7 +928,7 @@ func getTLSConfig() (x509Certs []*x509.Certificate, manager *certs.Manager, secu
 	// Therefore, we read all filenames in the cert directory and check
 	// for each directory whether it contains a public.crt and private.key.
 	// If so, we try to add it to certificate manager.
-	root, err := os.Open(globalCertsDir.Get())
+	root, err := Open(globalCertsDir.Get())
 	if err != nil {
 		return nil, nil, false, err
 	}
@@ -928,7 +947,7 @@ func getTLSConfig() (x509Certs []*x509.Certificate, manager *certs.Manager, secu
 			continue
 		}
 		if file.Mode()&os.ModeSymlink == os.ModeSymlink {
-			file, err = os.Stat(filepath.Join(root.Name(), file.Name()))
+			file, err = Stat(filepath.Join(root.Name(), file.Name()))
 			if err != nil {
 				// not accessible ignore
 				continue
@@ -396,7 +396,7 @@ func validateSubSysConfig(s config.Config, subSys string, objAPI ObjectLayer) er
 	}

 	if config.NotifySubSystems.Contains(subSys) {
-		if err := notify.TestSubSysNotificationTargets(GlobalContext, s, NewGatewayHTTPTransport(), globalNotificationSys.ConfiguredTargetIDs(), subSys); err != nil {
+		if err := notify.TestSubSysNotificationTargets(GlobalContext, s, NewGatewayHTTPTransport(), globalEventNotifier.ConfiguredTargetIDs(), subSys); err != nil {
 			return err
 		}
 	}
@@ -471,14 +471,12 @@ func lookupConfigs(s config.Config, objAPI ObjectLayer) {
 	}

 	if etcdCfg.Enabled {
-		if globalEtcdClient == nil {
-			globalEtcdClient, err = etcd.New(etcdCfg)
-			if err != nil {
-				if globalIsGateway {
-					logger.FatalIf(err, "Unable to initialize etcd config")
-				} else {
-					logger.LogIf(ctx, fmt.Errorf("Unable to initialize etcd config: %w", err))
-				}
+		globalEtcdClient, err = etcd.New(etcdCfg)
+		if err != nil {
+			if globalIsGateway {
+				logger.FatalIf(err, "Unable to initialize etcd config")
+			} else {
+				logger.LogIf(ctx, fmt.Errorf("Unable to initialize etcd config: %w", err))
 			}
 		}

@@ -486,7 +484,7 @@ func lookupConfigs(s config.Config, objAPI ObjectLayer) {
 			if globalDNSConfig != nil {
 				// if global DNS is already configured, indicate with a warning, incase
 				// users are confused.
-				logger.LogIf(ctx, fmt.Errorf("DNS store is already configured with %s, not using etcd for DNS store", globalDNSConfig))
+				logger.LogIf(ctx, fmt.Errorf("DNS store is already configured with %s, etcd is not used for DNS store", globalDNSConfig))
 			} else {
 				globalDNSConfig, err = dns.NewCoreDNS(etcdCfg.Config,
 					dns.DomainNames(globalDomainNames),
@@ -518,16 +516,6 @@ func lookupConfigs(s config.Config, objAPI ObjectLayer) {
 		logger.LogIf(ctx, fmt.Errorf("Invalid site configuration: %w", err))
 	}

-	apiConfig, err := api.LookupConfig(s[config.APISubSys][config.Default])
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Invalid api configuration: %w", err))
-	}
-
-	// Initialize remote instance transport once.
-	getRemoteInstanceTransportOnce.Do(func() {
-		getRemoteInstanceTransport = newGatewayHTTPTransport(apiConfig.RemoteTransportDeadline)
-	})
-
 	globalCacheConfig, err = cache.LookupConfig(s[config.CacheSubSys][config.Default])
 	if err != nil {
 		if globalIsGateway {
@@ -560,56 +548,19 @@ func lookupConfigs(s config.Config, objAPI ObjectLayer) {
 		logger.LogIf(ctx, fmt.Errorf("CRITICAL: enabling %s is not recommended in a production environment", xtls.EnvIdentityTLSSkipVerify))
 	}

-	globalOpenIDConfig, err = openid.LookupConfig(s,
-		NewGatewayHTTPTransport(), xhttp.DrainBody, globalSite.Region)
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Unable to initialize OpenID: %w", err))
-	}
-
-	globalLDAPConfig, err = xldap.Lookup(s[config.IdentityLDAPSubSys][config.Default],
-		globalRootCAs)
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Unable to parse LDAP configuration: %w", err))
-	}
-
-	authNPluginCfg, err := idplugin.LookupConfig(s[config.IdentityPluginSubSys][config.Default],
-		NewGatewayHTTPTransport(), xhttp.DrainBody, globalSite.Region)
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Unable to initialize AuthNPlugin: %w", err))
-	}
-	globalAuthNPlugin = idplugin.New(authNPluginCfg)
-
-	authZPluginCfg, err := polplugin.LookupConfig(s[config.PolicyPluginSubSys][config.Default],
-		NewGatewayHTTPTransport(), xhttp.DrainBody)
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Unable to initialize AuthZPlugin: %w", err))
-	}
-	if authZPluginCfg.URL == nil {
-		opaCfg, err := opa.LookupConfig(s[config.PolicyOPASubSys][config.Default],
-			NewGatewayHTTPTransport(), xhttp.DrainBody)
-		if err != nil {
-			logger.LogIf(ctx, fmt.Errorf("Unable to initialize AuthZPlugin from legacy OPA config: %w", err))
-		} else {
-			authZPluginCfg.URL = opaCfg.URL
-			authZPluginCfg.AuthToken = opaCfg.AuthToken
-			authZPluginCfg.Transport = opaCfg.Transport
-			authZPluginCfg.CloseRespFn = opaCfg.CloseRespFn
-		}
-	}
-
-	setGlobalAuthZPlugin(polplugin.New(authZPluginCfg))
-
 	globalSubnetConfig, err = subnet.LookupConfig(s[config.SubnetSubSys][config.Default], globalProxyTransport)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("Unable to parse subnet configuration: %w", err))
 	}

-	globalConfigTargetList, err = notify.GetNotificationTargets(GlobalContext, s, NewGatewayHTTPTransport(), false)
+	transport := NewGatewayHTTPTransport()
+
+	globalConfigTargetList, err = notify.GetNotificationTargets(GlobalContext, s, transport, false)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("Unable to initialize notification target(s): %w", err))
 	}

-	globalEnvTargetList, err = notify.GetNotificationTargets(GlobalContext, newServerConfig(), NewGatewayHTTPTransport(), true)
+	globalEnvTargetList, err = notify.GetNotificationTargets(GlobalContext, newServerConfig(), transport, true)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("Unable to initialize notification target(s): %w", err))
 	}
@@ -636,6 +587,11 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 			setDriveCounts = objAPI.SetDriveCounts()
 		}
 		globalAPIConfig.init(apiConfig, setDriveCounts)
+
+		// Initialize remote instance transport once.
+		getRemoteInstanceTransportOnce.Do(func() {
+			getRemoteInstanceTransport = newGatewayHTTPTransport(apiConfig.RemoteTransportDeadline)
+		})
 	case config.CompressionSubSys:
 		cmpCfg, err := compress.LookupConfig(s[config.CompressionSubSys][config.Default])
 		if err != nil {
@@ -678,8 +634,7 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 				loggerCfg.HTTP[n] = l
 			}
 		}
-		err = logger.UpdateSystemTargets(loggerCfg)
-		if err != nil {
+		if err = logger.UpdateSystemTargets(loggerCfg); err != nil {
 			logger.LogIf(ctx, fmt.Errorf("Unable to update logger webhook config: %w", err))
 		}
 	case config.AuditWebhookSubSys:
@@ -697,8 +652,7 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 			}
 		}

-		err = logger.UpdateAuditWebhookTargets(loggerCfg)
-		if err != nil {
+		if err = logger.UpdateAuditWebhookTargets(loggerCfg); err != nil {
 			logger.LogIf(ctx, fmt.Errorf("Unable to update audit webhook targets: %w", err))
 		}
 	case config.AuditKafkaSubSys:
@@ -712,8 +666,7 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 				loggerCfg.AuditKafka[n] = l
 			}
 		}
-		err = logger.UpdateAuditKafkaTargets(loggerCfg)
-		if err != nil {
+		if err = logger.UpdateAuditKafkaTargets(loggerCfg); err != nil {
 			logger.LogIf(ctx, fmt.Errorf("Unable to update audit kafka targets: %w", err))
 		}
 	case config.StorageClassSubSys:
@@ -852,7 +805,7 @@ func newSrvConfig(objAPI ObjectLayer) error {
 	globalServerConfigMu.Unlock()

 	// Save config into file.
-	return saveServerConfig(GlobalContext, objAPI, globalServerConfig)
+	return saveServerConfig(GlobalContext, objAPI, srvCfg)
 }

 func getValidConfig(objAPI ObjectLayer) (config.Config, error) {
@@ -69,8 +69,14 @@ func migrateIAMConfigsEtcdToEncrypted(ctx context.Context, client *etcd.Client)
 		return err
 	}

+	// If backend doesn't have this file means we have already
+	// attempted then migration
+	if !encrypted {
+		return nil
+	}
+
 	if encrypted && GlobalKMS != nil {
-		stat, err := GlobalKMS.Stat()
+		stat, err := GlobalKMS.Stat(ctx)
 		if err != nil {
 			return err
 		}
@@ -109,6 +115,8 @@ func migrateIAMConfigsEtcdToEncrypted(ctx context.Context, client *etcd.Client)
 							return fmt.Errorf("Decrypting IAM config failed %w, possibly credentials are incorrect", err)
 						}
 					}
+				} else {
+					return fmt.Errorf("Decrypting IAM config failed %w, possibly credentials are incorrect", err)
 				}
 			}
 			data = pdata
@@ -131,6 +139,7 @@ func migrateIAMConfigsEtcdToEncrypted(ctx context.Context, client *etcd.Client)
 	if encrypted && GlobalKMS != nil {
 		logger.Info("Migration of encrypted IAM config data completed. All data is now encrypted on etcd.")
 	}
+
 	return deleteKeyEtcd(ctx, client, backendEncryptedFile)
 }

@@ -139,51 +148,63 @@ func migrateConfigPrefixToEncrypted(objAPI ObjectLayer, encrypted bool) error {
 		return nil
 	}
 	if encrypted && GlobalKMS != nil {
-		stat, err := GlobalKMS.Stat()
+		stat, err := GlobalKMS.Stat(context.Background())
 		if err != nil {
 			return err
 		}
 		logger.Info(fmt.Sprintf("Attempting to re-encrypt config, IAM users and policies on MinIO with %q (%s)", stat.DefaultKey, stat.Name))
 	}

-	var marker string
-	for {
-		res, err := objAPI.ListObjects(GlobalContext, minioMetaBucket, minioConfigPrefix, marker, "", maxObjectList)
+	results := make(chan ObjectInfo)
+	if err := objAPI.Walk(GlobalContext, minioMetaBucket, minioConfigPrefix, results, ObjectOptions{}); err != nil {
+		return err
+	}
+
+	for obj := range results {
+		data, err := readConfig(GlobalContext, objAPI, obj.Name)
 		if err != nil {
 			return err
 		}
-		for _, obj := range res.Objects {
-			data, err := readConfig(GlobalContext, objAPI, obj.Name)
+
+		if !utf8.Valid(data) {
+			pdata, err := madmin.DecryptData(globalActiveCred.String(), bytes.NewReader(data))
+			if err != nil {
+				if GlobalKMS != nil {
+					pdata, err = config.DecryptBytes(GlobalKMS, data, kms.Context{
+						minioMetaBucket: path.Join(minioMetaBucket, obj.Name),
+					})
+					if err != nil {
+						pdata, err = config.DecryptBytes(GlobalKMS, data, kms.Context{
+							minioMetaBucket: obj.Name,
+						})
+						if err != nil {
+							return fmt.Errorf("Decrypting IAM config failed %w, possibly credentials are incorrect", err)
+						}
+					}
+				} else {
+					return fmt.Errorf("Decrypting IAM config failed %w, possibly credentials are incorrect", err)
+				}
+			}
+			data = pdata
+		}
+
+		if GlobalKMS != nil {
+			data, err = config.EncryptBytes(GlobalKMS, data, kms.Context{
+				obj.Bucket: path.Join(obj.Bucket, obj.Name),
+			})
 			if err != nil {
 				return err
 			}
-
-			if !utf8.Valid(data) {
-				data, err = madmin.DecryptData(globalActiveCred.String(), bytes.NewReader(data))
-				if err != nil {
-					return fmt.Errorf("Decrypting config failed %w, possibly credentials are incorrect", err)
-				}
-			}
-			if GlobalKMS != nil {
-				data, err = config.EncryptBytes(GlobalKMS, data, kms.Context{
-					obj.Bucket: path.Join(obj.Bucket, obj.Name),
-				})
-				if err != nil {
-					return err
-				}
-			}
-			if err = saveConfig(GlobalContext, objAPI, obj.Name, data); err != nil {
-				return err
-			}
 		}

-		if !res.IsTruncated {
-			break
+		if err = saveConfig(GlobalContext, objAPI, obj.Name, data); err != nil {
+			return err
 		}
-		marker = res.NextMarker
 	}
+
 	if encrypted && GlobalKMS != nil {
 		logger.Info("Migration of encrypted config data completed. All config data is now encrypted.")
 	}
+
 	return deleteConfig(GlobalContext, objAPI, backendEncryptedFile)
 }
@@ -37,11 +37,7 @@ func TestServerConfigMigrateV1(t *testing.T) {
 	if err != nil {
 		t.Fatalf("Init Test config failed")
 	}
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
+	rootPath := t.TempDir()
 	globalConfigDir = &ConfigDir{path: rootPath}

 	globalObjLayerMutex.Lock()
@@ -74,13 +70,7 @@ func TestServerConfigMigrateV1(t *testing.T) {
 // Test if all migrate code returns nil when config file does not
 // exist
 func TestServerConfigMigrateInexistentConfig(t *testing.T) {
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
-
-	globalConfigDir = &ConfigDir{path: rootPath}
+	globalConfigDir = &ConfigDir{path: t.TempDir()}

 	if err := migrateV2ToV3(); err != nil {
 		t.Fatal("migrate v2 to v3 should succeed when no config file is found")
@@ -164,11 +154,7 @@ func TestServerConfigMigrateInexistentConfig(t *testing.T) {

 // Test if a config migration from v2 to v33 is successfully done
 func TestServerConfigMigrateV2toV33(t *testing.T) {
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
+	rootPath := t.TempDir()

 	globalConfigDir = &ConfigDir{path: rootPath}

@@ -234,11 +220,7 @@ func TestServerConfigMigrateV2toV33(t *testing.T) {

 // Test if all migrate code returns error with corrupted config files
 func TestServerConfigMigrateFaultyConfig(t *testing.T) {
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
+	rootPath := t.TempDir()

 	globalConfigDir = &ConfigDir{path: rootPath}
 	configPath := rootPath + SlashSeparator + minioConfigFile
@@ -331,35 +313,31 @@ func TestServerConfigMigrateFaultyConfig(t *testing.T) {

 // Test if all migrate code returns error with corrupted config files
 func TestServerConfigMigrateCorruptedConfig(t *testing.T) {
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
+	rootPath := t.TempDir()

 	globalConfigDir = &ConfigDir{path: rootPath}
 	configPath := rootPath + SlashSeparator + minioConfigFile

 	for i := 3; i <= 17; i++ {
 		// Create a corrupted config file
-		if err = ioutil.WriteFile(configPath, []byte(fmt.Sprintf("{ \"version\":\"%d\", \"credential\": { \"accessKey\": 1 } }", i)),
+		if err := ioutil.WriteFile(configPath, []byte(fmt.Sprintf("{ \"version\":\"%d\", \"credential\": { \"accessKey\": 1 } }", i)),
 			0o644); err != nil {
 			t.Fatal("Unexpected error: ", err)
 		}

 		// Test different migrate versions and be sure they are returning an error
-		if err = migrateConfig(); err == nil {
+		if err := migrateConfig(); err == nil {
 			t.Fatal("migrateConfig() should fail with a corrupted json")
 		}
 	}

 	// Create a corrupted config file for version '2'.
-	if err = ioutil.WriteFile(configPath, []byte("{ \"version\":\"2\", \"credentials\": { \"accessKeyId\": 1 } }"), 0o644); err != nil {
+	if err := ioutil.WriteFile(configPath, []byte("{ \"version\":\"2\", \"credentials\": { \"accessKeyId\": 1 } }"), 0o644); err != nil {
 		t.Fatal("Unexpected error: ", err)
 	}

 	// Test different migrate versions and be sure they are returning an error
-	if err = migrateConfig(); err == nil {
+	if err := migrateConfig(); err == nil {
 		t.Fatal("migrateConfig() should fail with a corrupted json")
 	}
 }
@@ -48,12 +48,13 @@ import (
 )

 const (
-	dataScannerSleepPerFolder     = time.Millisecond                 // Time to wait between folders.
-	dataUsageUpdateDirCycles      = 16                               // Visit all folders every n cycles.
-	dataScannerCompactLeastObject = 500                              // Compact when there is less than this many objects in a branch.
-	dataScannerCompactAtChildren  = 10000                            // Compact when there are this many children in a branch.
-	dataScannerCompactAtFolders   = dataScannerCompactAtChildren / 4 // Compact when this many subfolders in a single folder.
-	dataScannerStartDelay         = 1 * time.Minute                  // Time to wait on startup and between cycles.
+	dataScannerSleepPerFolder        = time.Millisecond                 // Time to wait between folders.
+	dataUsageUpdateDirCycles         = 16                               // Visit all folders every n cycles.
+	dataScannerCompactLeastObject    = 500                              // Compact when there is less than this many objects in a branch.
+	dataScannerCompactAtChildren     = 10000                            // Compact when there are this many children in a branch.
+	dataScannerCompactAtFolders      = dataScannerCompactAtChildren / 4 // Compact when this many subfolders in a single folder.
+	dataScannerForceCompactAtFolders = 1_000_000                        // Compact when this many subfolders in a single folder (even top level).
+	dataScannerStartDelay            = 1 * time.Minute                  // Time to wait on startup and between cycles.

 	healDeleteDangling    = true
 	healFolderIncludeProb = 32  // Include a clean folder one in n cycles.
@@ -63,7 +64,11 @@ const (
 var (
 	globalHealConfig heal.Config

-	dataScannerLeaderLockTimeout = newDynamicTimeout(30*time.Second, 10*time.Second)
+	dataScannerLeaderLockTimeout = newDynamicTimeoutWithOpts(dynamicTimeoutOpts{
+		timeout:       30 * time.Second,
+		minimum:       10 * time.Second,
+		retryInterval: time.Second,
+	})
 	// Sleeper values are updated when config is loaded.
 	scannerSleeper = newDynamicSleeper(10, 10*time.Second, true)
 	scannerCycle   = uatomic.NewDuration(dataScannerStartDelay)
@@ -566,9 +571,11 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 		}

 		// If we have many subfolders, compact ourself.
-		if !into.Compacted &&
-			f.newCache.Info.Name != folder.name &&
-			len(existingFolders)+len(newFolders) >= dataScannerCompactAtFolders {
+		shouldCompact := f.newCache.Info.Name != folder.name &&
+			len(existingFolders)+len(newFolders) >= dataScannerCompactAtFolders ||
+			len(existingFolders)+len(newFolders) >= dataScannerForceCompactAtFolders
+
+		if !into.Compacted && shouldCompact {
 			into.Compacted = true
 			newFolders = append(newFolders, existingFolders...)
 			existingFolders = nil
@@ -734,11 +741,7 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 					}
 				},
 				// Some disks have data for this.
-				partial: func(entries metaCacheEntries, nAgreed int, errs []error) {
-					if f.dataUsageScannerDebug {
-						console.Debugf(healObjectsPrefix+" got partial, %d agreed, errs: %v\n", nAgreed, errs)
-					}
-
+				partial: func(entries metaCacheEntries, errs []error) {
 					entry, ok := entries.resolve(&resolver)
 					if !ok {
 						// check if we can get one entry atleast
@@ -1226,29 +1229,16 @@ func (i *scannerItem) objectPath() string {

 // healReplication will heal a scanned item that has failed replication.
 func (i *scannerItem) healReplication(ctx context.Context, o ObjectLayer, oi ObjectInfo, sizeS *sizeSummary) {
-	roi := getHealReplicateObjectInfo(oi, i.replication)
-	if !roi.Dsc.ReplicateAny() {
+	if oi.VersionID == "" {
 		return
 	}
-
+	if i.replication.Config == nil {
+		return
+	}
+	roi := queueReplicationHeal(ctx, oi.Bucket, oi, i.replication)
 	if oi.DeleteMarker || !oi.VersionPurgeStatus.Empty() {
-		// heal delete marker replication failure or versioned delete replication failure
-		if oi.ReplicationStatus == replication.Pending ||
-			oi.ReplicationStatus == replication.Failed ||
-			oi.VersionPurgeStatus == Failed || oi.VersionPurgeStatus == Pending {
-			i.healReplicationDeletes(ctx, o, roi)
-			return
-		}
-		// if replication status is Complete on DeleteMarker and existing object resync required
-		if roi.ExistingObjResync.mustResync() && (oi.ReplicationStatus == replication.Completed || oi.ReplicationStatus.Empty()) {
-			i.healReplicationDeletes(ctx, o, roi)
-			return
-		}
 		return
 	}
-	if roi.ExistingObjResync.mustResync() {
-		roi.OpType = replication.ExistingObjectReplicationType
-	}

 	if sizeS.replTargetStats == nil && len(roi.TargetStatuses) > 0 {
 		sizeS.replTargetStats = make(map[string]replTargetSizeSummary)
@@ -1278,48 +1268,9 @@ func (i *scannerItem) healReplication(ctx context.Context, o ObjectLayer, oi Obj
 	}

 	switch oi.ReplicationStatus {
-	case replication.Pending, replication.Failed:
-		globalReplicationPool.queueReplicaTask(roi)
-		return
 	case replication.Replica:
 		sizeS.replicaSize += oi.Size
 	}
-	if roi.ExistingObjResync.mustResync() {
-		globalReplicationPool.queueReplicaTask(roi)
-	}
-}
-
-// healReplicationDeletes will heal a scanned deleted item that failed to replicate deletes.
-func (i *scannerItem) healReplicationDeletes(ctx context.Context, o ObjectLayer, roi ReplicateObjectInfo) {
-	// handle soft delete and permanent delete failures here.
-	if roi.DeleteMarker || !roi.VersionPurgeStatus.Empty() {
-		versionID := ""
-		dmVersionID := ""
-		if roi.VersionPurgeStatus.Empty() {
-			dmVersionID = roi.VersionID
-		} else {
-			versionID = roi.VersionID
-		}
-
-		doi := DeletedObjectReplicationInfo{
-			DeletedObject: DeletedObject{
-				ObjectName:            roi.Name,
-				DeleteMarkerVersionID: dmVersionID,
-				VersionID:             versionID,
-				ReplicationState:      roi.getReplicationState(roi.Dsc.String(), versionID, true),
-				DeleteMarkerMTime:     DeleteMarkerMTime{roi.ModTime},
-				DeleteMarker:          roi.DeleteMarker,
-			},
-			Bucket: roi.Bucket,
-			OpType: replication.HealReplicationType,
-		}
-		if roi.ExistingObjResync.mustResync() {
-			doi.OpType = replication.ExistingObjectReplicationType
-			queueReplicateDeletesWrapper(doi, roi.ExistingObjResync)
-			return
-		}
-		globalReplicationPool.queueReplicaDeleteTask(doi)
-	}
 }

 type dynamicSleeper struct {
@@ -1477,9 +1428,9 @@ const (
 	ILMTransition = " ilm:transition"
 )

-func auditLogLifecycle(ctx context.Context, oi ObjectInfo, trigger string) {
+func auditLogLifecycle(ctx context.Context, oi ObjectInfo, event string) {
 	var apiName string
-	switch trigger {
+	switch event {
 	case ILMExpiry:
 		apiName = "ILMExpiry"
 	case ILMFreeVersionDelete:
@@ -1488,7 +1439,7 @@ func auditLogLifecycle(ctx context.Context, oi ObjectInfo, trigger string) {
 		apiName = "ILMTransition"
 	}
 	auditLogInternal(ctx, oi.Bucket, oi.Name, AuditLogOptions{
-		Trigger:   trigger,
+		Event:     event,
 		APIName:   apiName,
 		VersionID: oi.VersionID,
 	})
@@ -227,7 +227,7 @@ func (d *dataUpdateTracker) load(ctx context.Context, drives ...string) {
 	for _, drive := range drives {

 		cacheFormatPath := pathJoin(drive, dataUpdateTrackerFilename)
-		f, err := os.Open(cacheFormatPath)
+		f, err := OpenFile(cacheFormatPath, readMode, 0o666)
 		if err != nil {
 			if osIsNotExist(err) {
 				continue
@@ -21,7 +21,6 @@ import (
 	"bytes"
 	"context"
 	"fmt"
-	"io/ioutil"
 	"math/rand"
 	"os"
 	"path"
@@ -100,15 +99,11 @@ func TestDataUpdateTracker(t *testing.T) {

 	dut.Current.bf = dut.newBloomFilter()

-	tmpDir, err := ioutil.TempDir("", "TestDataUpdateTracker")
+	tmpDir := t.TempDir()
+	err := os.MkdirAll(filepath.Dir(filepath.Join(tmpDir, dataUpdateTrackerFilename)), os.ModePerm)
 	if err != nil {
 		t.Fatal(err)
 	}
-	err = os.MkdirAll(filepath.Dir(filepath.Join(tmpDir, dataUpdateTrackerFilename)), os.ModePerm)
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(tmpDir)
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()
 	dut.start(ctx, tmpDir)
@@ -56,7 +56,8 @@ func storeDataUsageInBackend(ctx context.Context, objAPI ObjectLayer, dui <-chan
 }

 // loadPrefixUsageFromBackend returns prefix usages found in passed buckets
-//   e.g.:  /testbucket/prefix => 355601334
+//
+//	e.g.:  /testbucket/prefix => 355601334
 func loadPrefixUsageFromBackend(ctx context.Context, objAPI ObjectLayer, bucket string) (map[string]uint64, error) {
 	z, ok := objAPI.(*erasureServerPools)
 	if !ok {
@@ -35,12 +35,8 @@ type usageTestFile struct {
 }

 func TestDataUsageUpdate(t *testing.T) {
-	base, err := ioutil.TempDir("", "TestDataUsageUpdate")
-	if err != nil {
-		t.Skip(err)
-	}
+	base := t.TempDir()
 	const bucket = "bucket"
-	defer os.RemoveAll(base)
 	files := []usageTestFile{
 		{name: "rootfile", size: 10000},
 		{name: "rootfile2", size: 10000},
@@ -251,12 +247,8 @@ func TestDataUsageUpdate(t *testing.T) {
 }

 func TestDataUsageUpdatePrefix(t *testing.T) {
-	base, err := ioutil.TempDir("", "TestDataUpdateUsagePrefix")
-	if err != nil {
-		t.Skip(err)
-	}
+	base := t.TempDir()
 	scannerSleeper.Update(0, 0)
-	defer os.RemoveAll(base)
 	files := []usageTestFile{
 		{name: "bucket/rootfile", size: 10000},
 		{name: "bucket/rootfile2", size: 10000},
@@ -537,12 +529,8 @@ func generateUsageTestFiles(t *testing.T, base, bucket string, nFolders, nFiles,
 }

 func TestDataUsageCacheSerialize(t *testing.T) {
-	base, err := ioutil.TempDir("", "TestDataUsageCacheSerialize")
-	if err != nil {
-		t.Skip(err)
-	}
+	base := t.TempDir()
 	const bucket = "abucket"
-	defer os.RemoveAll(base)
 	files := []usageTestFile{
 		{name: "rootfile", size: 10000},
 		{name: "rootfile2", size: 10000},
@@ -0,0 +1,25 @@
+// Code generated by "stringer -type=decomMetric -trimprefix=decomMetric erasure-server-pool-decom.go"; DO NOT EDIT.
+
+package cmd
+
+import "strconv"
+
+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[decomMetricDecommissionBucket-0]
+	_ = x[decomMetricDecommissionObject-1]
+	_ = x[decomMetricDecommissionRemoveObject-2]
+}
+
+const _decomMetric_name = "DecommissionBucketDecommissionObjectDecommissionRemoveObject"
+
+var _decomMetric_index = [...]uint8{0, 18, 36, 60}
+
+func (i decomMetric) String() string {
+	if i >= decomMetric(len(_decomMetric_index)-1) {
+		return "decomMetric(" + strconv.FormatInt(int64(i), 10) + ")"
+	}
+	return _decomMetric_name[_decomMetric_index[i]:_decomMetric_index[i+1]]
+}
@@ -624,7 +624,7 @@ func (c *diskCache) saveMetadata(ctx context.Context, bucket, object string, met
 	if err := os.MkdirAll(cachedPath, 0o777); err != nil {
 		return err
 	}
-	f, err := os.OpenFile(metaPath, os.O_RDWR|os.O_CREATE, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|os.O_CREATE|writeMode, 0o666)
 	if err != nil {
 		return err
 	}
@@ -687,7 +687,7 @@ func (c *diskCache) updateMetadata(ctx context.Context, bucket, object, etag str
 	if err := os.MkdirAll(cachedPath, 0o777); err != nil {
 		return err
 	}
-	f, err := os.OpenFile(metaPath, os.O_RDWR, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|writeMode, 0o666)
 	if err != nil {
 		return err
 	}
@@ -797,8 +797,8 @@ func (c *diskCache) bitrotWriteToCache(cachePath, fileName string, reader io.Rea
 	return bytesWritten, base64.StdEncoding.EncodeToString(md5sumCurr), nil
 }

-func newCacheEncryptReader(content io.Reader, bucket, object string, metadata map[string]string) (r io.Reader, err error) {
-	objectEncryptionKey, err := newCacheEncryptMetadata(bucket, object, metadata)
+func newCacheEncryptReader(ctx context.Context, content io.Reader, bucket, object string, metadata map[string]string) (r io.Reader, err error) {
+	objectEncryptionKey, err := newCacheEncryptMetadata(ctx, bucket, object, metadata)
 	if err != nil {
 		return nil, err
 	}
@@ -810,12 +810,12 @@ func newCacheEncryptReader(content io.Reader, bucket, object string, metadata ma
 	return reader, nil
 }

-func newCacheEncryptMetadata(bucket, object string, metadata map[string]string) ([]byte, error) {
+func newCacheEncryptMetadata(ctx context.Context, bucket, object string, metadata map[string]string) ([]byte, error) {
 	var sealedKey crypto.SealedKey
 	if globalCacheKMS == nil {
 		return nil, errKMSNotConfigured
 	}
-	key, err := globalCacheKMS.GenerateKey("", kms.Context{bucket: pathJoin(bucket, object)})
+	key, err := globalCacheKMS.GenerateKey(ctx, "", kms.Context{bucket: pathJoin(bucket, object)})
 	if err != nil {
 		return nil, err
 	}
@@ -891,7 +891,7 @@ func (c *diskCache) put(ctx context.Context, bucket, object string, data io.Read
 	reader := data
 	actualSize := uint64(size)
 	if globalCacheKMS != nil {
-		reader, err = newCacheEncryptReader(data, bucket, object, metadata)
+		reader, err = newCacheEncryptReader(ctx, data, bucket, object, metadata)
 		if err != nil {
 			removeAll(cachePath)
 			return oi, err
@@ -948,7 +948,7 @@ func (c *diskCache) putRange(ctx context.Context, bucket, object string, data io
 	// objSize is the actual size of object (with encryption overhead if any)
 	objSize := uint64(size)
 	if globalCacheKMS != nil {
-		reader, err = newCacheEncryptReader(data, bucket, object, metadata)
+		reader, err = newCacheEncryptReader(ctx, data, bucket, object, metadata)
 		if err != nil {
 			return err
 		}
@@ -1275,12 +1275,12 @@ func (c *diskCache) NewMultipartUpload(ctx context.Context, bucket, object, uID

 	cachePath := getMultipartCacheSHADir(c.dir, bucket, object)
 	uploadIDDir := path.Join(cachePath, uploadID)
-	if err := os.MkdirAll(uploadIDDir, 0o777); err != nil {
+	if err := mkdirAll(uploadIDDir, 0o777); err != nil {
 		return uploadID, err
 	}
 	metaPath := pathJoin(uploadIDDir, cacheMetaJSONFile)

-	f, err := os.OpenFile(metaPath, os.O_RDWR|os.O_CREATE, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|os.O_CREATE|writeMode, 0o666)
 	if err != nil {
 		return uploadID, err
 	}
@@ -1301,7 +1301,7 @@ func (c *diskCache) NewMultipartUpload(ctx context.Context, bucket, object, uID
 	m.Stat.ModTime = UTCNow()
 	if globalCacheKMS != nil {
 		m.Meta[ReservedMetadataPrefix+"Encrypted-Multipart"] = ""
-		if _, err := newCacheEncryptMetadata(bucket, object, m.Meta); err != nil {
+		if _, err := newCacheEncryptMetadata(ctx, bucket, object, m.Meta); err != nil {
 			return uploadID, err
 		}
 	}
@@ -1386,7 +1386,7 @@ func (c *diskCache) SavePartMetadata(ctx context.Context, bucket, object, upload
 	defer uploadLock.Unlock(ulkctx.Cancel)

 	metaPath := pathJoin(uploadDir, cacheMetaJSONFile)
-	f, err := os.OpenFile(metaPath, os.O_RDWR, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|writeMode, 0o666)
 	if err != nil {
 		return err
 	}
@@ -1465,7 +1465,7 @@ func newCachePartEncryptReader(ctx context.Context, bucket, object string, partI
 func (c *diskCache) uploadIDExists(bucket, object, uploadID string) (err error) {
 	mpartCachePath := getMultipartCacheSHADir(c.dir, bucket, object)
 	uploadIDDir := path.Join(mpartCachePath, uploadID)
-	if _, err := os.Stat(uploadIDDir); err != nil {
+	if _, err := Stat(uploadIDDir); err != nil {
 		return err
 	}
 	return nil
@@ -1564,7 +1564,7 @@ func (c *diskCache) CompleteMultipartUpload(ctx context.Context, bucket, object,
 	uploadMeta.Hits++
 	metaPath := pathJoin(uploadIDDir, cacheMetaJSONFile)

-	f, err := os.OpenFile(metaPath, os.O_RDWR|os.O_CREATE, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|os.O_CREATE|writeMode, 0o666)
 	if err != nil {
 		return oi, err
 	}
@@ -1634,7 +1634,7 @@ func (c *diskCache) cleanupStaleUploads(ctx context.Context) {
 			readDirFn(pathJoin(c.dir, minioMetaBucket, cacheMultipartDir), func(shaDir string, typ os.FileMode) error {
 				return readDirFn(pathJoin(c.dir, minioMetaBucket, cacheMultipartDir, shaDir), func(uploadIDDir string, typ os.FileMode) error {
 					uploadIDPath := pathJoin(c.dir, minioMetaBucket, cacheMultipartDir, shaDir, uploadIDDir)
-					fi, err := os.Stat(uploadIDPath)
+					fi, err := Stat(uploadIDPath)
 					if err != nil {
 						return nil
 					}
@@ -1649,8 +1649,8 @@ func (c *diskCache) cleanupStaleUploads(ctx context.Context) {
 			readDirFn(pathJoin(c.dir, minioMetaBucket, cacheWritebackDir), func(shaDir string, typ os.FileMode) error {
 				wbdir := pathJoin(c.dir, minioMetaBucket, cacheWritebackDir, shaDir)
 				cachedir := pathJoin(c.dir, shaDir)
-				if _, err := os.Stat(cachedir); os.IsNotExist(err) {
-					fi, err := os.Stat(wbdir)
+				if _, err := Stat(cachedir); os.IsNotExist(err) {
+					fi, err := Stat(wbdir)
 					if err != nil {
 						return nil
 					}
@@ -607,12 +607,12 @@ func newCache(config cache.Config) ([]*diskCache, bool, error) {
 				warningMsg = fmt.Sprintf("Invalid cache dir %s err : %s", dir, err.Error())
 			}
 			if rootDsk {
-				warningMsg = fmt.Sprintf("cache dir cannot be part of root disk: %s", dir)
+				warningMsg = fmt.Sprintf("cache dir cannot be part of root drive: %s", dir)
 			}
 		}

 		if err := checkAtimeSupport(dir); err != nil {
-			return nil, false, fmt.Errorf("Atime support required for disk caching, atime check failed with %w", err)
+			return nil, false, fmt.Errorf("Atime support required for drive caching, atime check failed with %w", err)
 		}

 		cache, err := newDiskCache(ctx, dir, config)
@@ -622,7 +622,7 @@ func newCache(config cache.Config) ([]*diskCache, bool, error) {
 		caches = append(caches, cache)
 	}
 	if warningMsg != "" {
-		logger.Info(color.Yellow(fmt.Sprintf("WARNING: Usage of root disk for disk caching is deprecated: %s", warningMsg)))
+		logger.Info(color.Yellow(fmt.Sprintf("WARNING: Usage of root drive for drive caching is deprecated: %s", warningMsg)))
 	}
 	return caches, migrating, nil
 }
@@ -45,10 +45,10 @@ type DummyDataGen struct {
 //
 // Given the function:
 //
-// f := func(r io.Reader) string {
-//           b, _ := ioutil.ReadAll(r)
-//           return string(b)
-// }
+//	f := func(r io.Reader) string {
+//	          b, _ := ioutil.ReadAll(r)
+//	          return string(b)
+//	}
 //
 // for example, the following is true:
 //
@@ -52,7 +52,7 @@ func (api objectAPIHandlers) GetBucketWebsiteHandler(w http.ResponseWriter, r *h
 	}

 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -84,7 +84,7 @@ func (api objectAPIHandlers) GetBucketAccelerateHandler(w http.ResponseWriter, r
 	}

 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -117,7 +117,7 @@ func (api objectAPIHandlers) GetBucketRequestPaymentHandler(w http.ResponseWrite
 	}

 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -151,7 +151,7 @@ func (api objectAPIHandlers) GetBucketLoggingHandler(w http.ResponseWriter, r *h
 	}

 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -189,7 +189,7 @@ func (api objectAPIHandlers) GetBucketCorsHandler(w http.ResponseWriter, r *http
 	}

 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -34,11 +34,24 @@ const (

 // timeouts that are dynamically adapted based on actual usage results
 type dynamicTimeout struct {
-	timeout int64
-	minimum int64
-	entries int64
-	log     [dynamicTimeoutLogSize]time.Duration
-	mutex   sync.Mutex
+	timeout       int64
+	minimum       int64
+	entries       int64
+	log           [dynamicTimeoutLogSize]time.Duration
+	mutex         sync.Mutex
+	retryInterval time.Duration
+}
+
+type dynamicTimeoutOpts struct {
+	timeout       time.Duration
+	minimum       time.Duration
+	retryInterval time.Duration
+}
+
+func newDynamicTimeoutWithOpts(opts dynamicTimeoutOpts) *dynamicTimeout {
+	dt := newDynamicTimeout(opts.timeout, opts.minimum)
+	dt.retryInterval = opts.retryInterval
+	return dt
 }

 // newDynamicTimeout returns a new dynamic timeout initialized with timeout value
@@ -57,6 +70,10 @@ func (dt *dynamicTimeout) Timeout() time.Duration {
 	return time.Duration(atomic.LoadInt64(&dt.timeout))
 }

+func (dt *dynamicTimeout) RetryInterval() time.Duration {
+	return dt.retryInterval
+}
+
 // LogSuccess logs the duration of a successful action that
 // did not hit the timeout
 func (dt *dynamicTimeout) LogSuccess(duration time.Duration) {
@@ -82,7 +82,7 @@ func (o *MultipartInfo) KMSKeyID() string { return kmsKeyIDFromMetadata(o.UserDe
 // metadata, if any. It returns an empty ID if no key ID is
 // present.
 func kmsKeyIDFromMetadata(metadata map[string]string) string {
-	const ARNPrefix = "arn:aws:kms:"
+	const ARNPrefix = crypto.ARNPrefix
 	if len(metadata) == 0 {
 		return ""
 	}
@@ -255,7 +255,7 @@ func ParseSSECustomerHeader(header http.Header) (key []byte, err error) {
 }

 // This function rotates old to new key.
-func rotateKey(oldKey []byte, newKeyID string, newKey []byte, bucket, object string, metadata map[string]string, ctx kms.Context) error {
+func rotateKey(ctx context.Context, oldKey []byte, newKeyID string, newKey []byte, bucket, object string, metadata map[string]string, cryptoCtx kms.Context) error {
 	kind, _ := crypto.IsEncrypted(metadata)
 	switch kind {
 	case crypto.S3:
@@ -275,7 +275,7 @@ func rotateKey(oldKey []byte, newKeyID string, newKey []byte, bucket, object str
 			return err
 		}

-		newKey, err := GlobalKMS.GenerateKey("", kms.Context{bucket: path.Join(bucket, object)})
+		newKey, err := GlobalKMS.GenerateKey(ctx, "", kms.Context{bucket: path.Join(bucket, object)})
 		if err != nil {
 			return err
 		}
@@ -291,8 +291,8 @@ func rotateKey(oldKey []byte, newKeyID string, newKey []byte, bucket, object str
 			return err
 		}

-		if len(ctx) == 0 {
-			_, _, _, ctx, err = crypto.S3KMS.ParseMetadata(metadata)
+		if len(cryptoCtx) == 0 {
+			_, _, _, cryptoCtx, err = crypto.S3KMS.ParseMetadata(metadata)
 			if err != nil {
 				return err
 			}
@@ -305,19 +305,19 @@ func rotateKey(oldKey []byte, newKeyID string, newKey []byte, bucket, object str
 		// of the client provided context and add the bucket
 		// key, if not present.
 		kmsCtx := kms.Context{}
-		for k, v := range ctx {
+		for k, v := range cryptoCtx {
 			kmsCtx[k] = v
 		}
 		if _, ok := kmsCtx[bucket]; !ok {
 			kmsCtx[bucket] = path.Join(bucket, object)
 		}
-		newKey, err := GlobalKMS.GenerateKey(newKeyID, kmsCtx)
+		newKey, err := GlobalKMS.GenerateKey(ctx, newKeyID, kmsCtx)
 		if err != nil {
 			return err
 		}

 		sealedKey := objectKey.Seal(newKey.Plaintext, crypto.GenerateIV(rand.Reader), crypto.S3KMS.String(), bucket, object)
-		crypto.S3KMS.CreateMetadata(metadata, newKey.KeyID, newKey.Ciphertext, sealedKey, ctx)
+		crypto.S3KMS.CreateMetadata(metadata, newKey.KeyID, newKey.Ciphertext, sealedKey, cryptoCtx)
 		return nil
 	case crypto.SSEC:
 		sealedKey, err := crypto.SSEC.ParseMetadata(metadata)
@@ -344,14 +344,14 @@ func rotateKey(oldKey []byte, newKeyID string, newKey []byte, bucket, object str
 	}
 }

-func newEncryptMetadata(kind crypto.Type, keyID string, key []byte, bucket, object string, metadata map[string]string, ctx kms.Context) (crypto.ObjectKey, error) {
+func newEncryptMetadata(ctx context.Context, kind crypto.Type, keyID string, key []byte, bucket, object string, metadata map[string]string, cryptoCtx kms.Context) (crypto.ObjectKey, error) {
 	var sealedKey crypto.SealedKey
 	switch kind {
 	case crypto.S3:
 		if GlobalKMS == nil {
 			return crypto.ObjectKey{}, errKMSNotConfigured
 		}
-		key, err := GlobalKMS.GenerateKey("", kms.Context{bucket: path.Join(bucket, object)})
+		key, err := GlobalKMS.GenerateKey(ctx, "", kms.Context{bucket: path.Join(bucket, object)})
 		if err != nil {
 			return crypto.ObjectKey{}, err
 		}
@@ -372,13 +372,13 @@ func newEncryptMetadata(kind crypto.Type, keyID string, key []byte, bucket, obje
 		// of the client provided context and add the bucket
 		// key, if not present.
 		kmsCtx := kms.Context{}
-		for k, v := range ctx {
+		for k, v := range cryptoCtx {
 			kmsCtx[k] = v
 		}
 		if _, ok := kmsCtx[bucket]; !ok {
 			kmsCtx[bucket] = path.Join(bucket, object)
 		}
-		key, err := GlobalKMS.GenerateKey(keyID, kmsCtx)
+		key, err := GlobalKMS.GenerateKey(ctx, keyID, kmsCtx)
 		if err != nil {
 			if errors.Is(err, kes.ErrKeyNotFound) {
 				return crypto.ObjectKey{}, errKMSKeyNotFound
@@ -388,7 +388,7 @@ func newEncryptMetadata(kind crypto.Type, keyID string, key []byte, bucket, obje

 		objectKey := crypto.GenerateKey(key.Plaintext, rand.Reader)
 		sealedKey = objectKey.Seal(key.Plaintext, crypto.GenerateIV(rand.Reader), crypto.S3KMS.String(), bucket, object)
-		crypto.S3KMS.CreateMetadata(metadata, key.KeyID, key.Ciphertext, sealedKey, ctx)
+		crypto.S3KMS.CreateMetadata(metadata, key.KeyID, key.Ciphertext, sealedKey, cryptoCtx)
 		return objectKey, nil
 	case crypto.SSEC:
 		objectKey := crypto.GenerateKey(key, rand.Reader)
@@ -400,8 +400,8 @@ func newEncryptMetadata(kind crypto.Type, keyID string, key []byte, bucket, obje
 	}
 }

-func newEncryptReader(content io.Reader, kind crypto.Type, keyID string, key []byte, bucket, object string, metadata map[string]string, ctx kms.Context) (io.Reader, crypto.ObjectKey, error) {
-	objectEncryptionKey, err := newEncryptMetadata(kind, keyID, key, bucket, object, metadata, ctx)
+func newEncryptReader(ctx context.Context, content io.Reader, kind crypto.Type, keyID string, key []byte, bucket, object string, metadata map[string]string, cryptoCtx kms.Context) (io.Reader, crypto.ObjectKey, error) {
+	objectEncryptionKey, err := newEncryptMetadata(ctx, kind, keyID, key, bucket, object, metadata, cryptoCtx)
 	if err != nil {
 		return nil, crypto.ObjectKey{}, err
 	}
@@ -418,9 +418,9 @@ func newEncryptReader(content io.Reader, kind crypto.Type, keyID string, key []b
 // SSE-S3
 func setEncryptionMetadata(r *http.Request, bucket, object string, metadata map[string]string) (err error) {
 	var (
-		key   []byte
-		keyID string
-		ctx   kms.Context
+		key    []byte
+		keyID  string
+		kmsCtx kms.Context
 	)
 	kind, _ := crypto.IsRequested(r.Header)
 	switch kind {
@@ -430,12 +430,12 @@ func setEncryptionMetadata(r *http.Request, bucket, object string, metadata map[
 			return err
 		}
 	case crypto.S3KMS:
-		keyID, ctx, err = crypto.S3KMS.ParseHTTP(r.Header)
+		keyID, kmsCtx, err = crypto.S3KMS.ParseHTTP(r.Header)
 		if err != nil {
 			return err
 		}
 	}
-	_, err = newEncryptMetadata(kind, keyID, key, bucket, object, metadata, ctx)
+	_, err = newEncryptMetadata(r.Context(), kind, keyID, key, bucket, object, metadata, kmsCtx)
 	return
 }

@@ -468,7 +468,7 @@ func EncryptRequest(content io.Reader, r *http.Request, bucket, object string, m
 			return nil, crypto.ObjectKey{}, err
 		}
 	}
-	return newEncryptReader(content, kind, keyID, key, bucket, object, metadata, ctx)
+	return newEncryptReader(r.Context(), content, kind, keyID, key, bucket, object, metadata, ctx)
 }

 func decryptObjectInfo(key []byte, bucket, object string, metadata map[string]string) ([]byte, error) {
@@ -156,7 +156,7 @@ func getSetIndexes(args []string, totalSizes []uint64, customSetDriveCount uint6

 	setCounts := possibleSetCounts(commonSize)
 	if len(setCounts) == 0 {
-		msg := fmt.Sprintf("Incorrect number of endpoints provided %s, number of disks %d is not divisible by any supported erasure set sizes %d", args, commonSize, setSizes)
+		msg := fmt.Sprintf("Incorrect number of endpoints provided %s, number of drives %d is not divisible by any supported erasure set sizes %d", args, commonSize, setSizes)
 		return nil, config.ErrInvalidNumberOfErasureEndpoints(nil).Msg(msg)
 	}

@@ -183,7 +183,7 @@ func getSetIndexes(args []string, totalSizes []uint64, customSetDriveCount uint6
 		setCounts = possibleSetCountsWithSymmetry(setCounts, argPatterns)

 		if len(setCounts) == 0 {
-			msg := fmt.Sprintf("No symmetric distribution detected with input endpoints provided %s, disks %d cannot be spread symmetrically by any supported erasure set sizes %d", args, commonSize, setSizes)
+			msg := fmt.Sprintf("No symmetric distribution detected with input endpoints provided %s, drives %d cannot be spread symmetrically by any supported erasure set sizes %d", args, commonSize, setSizes)
 			return nil, config.ErrInvalidNumberOfErasureEndpoints(nil).Msg(msg)
 		}

@@ -193,7 +193,7 @@ func getSetIndexes(args []string, totalSizes []uint64, customSetDriveCount uint6

 	// Check whether setSize is with the supported range.
 	if !isValidSetSize(setSize) {
-		msg := fmt.Sprintf("Incorrect number of endpoints provided %s, number of disks %d is not divisible by any supported erasure set sizes %d", args, commonSize, setSizes)
+		msg := fmt.Sprintf("Incorrect number of endpoints provided %s, number of drives %d is not divisible by any supported erasure set sizes %d", args, commonSize, setSizes)
 		return nil, config.ErrInvalidNumberOfErasureEndpoints(nil).Msg(msg)
 	}

@@ -292,6 +292,19 @@ func (l EndpointServerPools) LocalDisksPaths() []string {
 	return disks
 }

+// NLocalDisksPathsPerPool returns the disk paths of the local disks per pool
+func (l EndpointServerPools) NLocalDisksPathsPerPool() []int {
+	localDisksCount := make([]int, len(l))
+	for i, ep := range l {
+		for _, endpoint := range ep.Endpoints {
+			if endpoint.IsLocal {
+				localDisksCount[i]++
+			}
+		}
+	}
+	return localDisksCount
+}
+
 // FirstLocal returns true if the first endpoint is local.
 func (l EndpointServerPools) FirstLocal() bool {
 	return l[0].Endpoints[0].IsLocal
@@ -302,7 +315,7 @@ func (l EndpointServerPools) HTTPS() bool {
 	return l[0].Endpoints.HTTPS()
 }

-// NEndpoints - returns all nodes count
+// NEndpoints - returns number of endpoints
 func (l EndpointServerPools) NEndpoints() (count int) {
 	for _, ep := range l {
 		count += len(ep.Endpoints)
@@ -689,6 +702,17 @@ func CreateEndpoints(serverAddr string, foundLocal bool, args ...[]string) (Endp
 		}
 	}

+	// Add missing port in all endpoints.
+	for i := range endpoints {
+		_, port, err := net.SplitHostPort(endpoints[i].Host)
+		if err != nil {
+			endpoints[i].Host = net.JoinHostPort(endpoints[i].Host, serverAddrPort)
+		} else if endpoints[i].IsLocal && serverAddrPort != port {
+			// If endpoint is local, but port is different than serverAddrPort, then make it as remote.
+			endpoints[i].IsLocal = false
+		}
+	}
+
 	// All endpoints are pointing to local host
 	if len(endpoints) == localEndpointCount {
 		// If all endpoints have same port number, Just treat it as local erasure setup
@@ -705,17 +729,6 @@ func CreateEndpoints(serverAddr string, foundLocal bool, args ...[]string) (Endp
 		// This means it is DistErasure setup.
 	}

-	// Add missing port in all endpoints.
-	for i := range endpoints {
-		_, port, err := net.SplitHostPort(endpoints[i].Host)
-		if err != nil {
-			endpoints[i].Host = net.JoinHostPort(endpoints[i].Host, serverAddrPort)
-		} else if endpoints[i].IsLocal && serverAddrPort != port {
-			// If endpoint is local, but port is different than serverAddrPort, then make it as remote.
-			endpoints[i].IsLocal = false
-		}
-	}
-
 	uniqueArgs := set.NewStringSet()
 	for _, endpoint := range endpoints {
 		uniqueArgs.Add(endpoint.Host)
@@ -252,10 +252,10 @@ func TestCreateEndpoints(t *testing.T) {
 		},
 		// DistErasure Setup with URLEndpointType
 		{":9000", [][]string{{"http://localhost/d1", "http://localhost/d2", "http://localhost/d3", "http://localhost/d4"}}, ":9000", Endpoints{
-			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost", Path: "/d1"}, IsLocal: true},
-			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost", Path: "/d2"}, IsLocal: true},
-			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost", Path: "/d3"}, IsLocal: true},
-			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost", Path: "/d4"}, IsLocal: true},
+			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost:9000", Path: "/d1"}, IsLocal: true},
+			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost:9000", Path: "/d2"}, IsLocal: true},
+			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost:9000", Path: "/d3"}, IsLocal: true},
+			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost:9000", Path: "/d4"}, IsLocal: true},
 		}, ErasureSetupType, nil},
 		// DistErasure Setup with URLEndpointType having mixed naming to local host.
 		{"127.0.0.1:10000", [][]string{{"http://localhost/d1", "http://localhost/d2", "http://127.0.0.1/d3", "http://127.0.0.1/d4"}}, "", Endpoints{}, -1, fmt.Errorf("all local endpoints should not have different hostnames/ips")},
@@ -35,7 +35,7 @@ var bucketMetadataOpIgnoredErrs = append(bucketOpIgnoredErrs, errVolumeNotFound)
 // Bucket operations

 // MakeBucket - make a bucket.
-func (er erasureObjects) MakeBucketWithLocation(ctx context.Context, bucket string, opts BucketOptions) error {
+func (er erasureObjects) MakeBucketWithLocation(ctx context.Context, bucket string, opts MakeBucketOptions) error {
 	defer NSUpdated(bucket, slashSeparator)

 	// Verify if bucket is valid.
@@ -94,7 +94,7 @@ func undoDeleteBucket(storageDisks []StorageAPI, bucket string) {
 }

 // getBucketInfo - returns the BucketInfo from one of the load balanced disks.
-func (er erasureObjects) getBucketInfo(ctx context.Context, bucketName string) (bucketInfo BucketInfo, err error) {
+func (er erasureObjects) getBucketInfo(ctx context.Context, bucketName string, opts BucketOptions) (bucketInfo BucketInfo, err error) {
 	storageDisks := er.getDisks()

 	g := errgroup.WithNErrs(len(storageDisks))
@@ -108,9 +108,17 @@ func (er erasureObjects) getBucketInfo(ctx context.Context, bucketName string) (
 			}
 			volInfo, err := storageDisks[index].StatVol(ctx, bucketName)
 			if err != nil {
+				if opts.Deleted {
+					dvi, derr := storageDisks[index].StatVol(ctx, pathJoin(minioMetaBucket, bucketMetaPrefix, deletedBucketsPrefix, bucketName))
+					if derr != nil {
+						return err
+					}
+					bucketsInfo[index] = BucketInfo{Name: bucketName, Deleted: dvi.Created}
+					return nil
+				}
 				return err
 			}
-			bucketsInfo[index] = BucketInfo(volInfo)
+			bucketsInfo[index] = BucketInfo{Name: volInfo.Name, Created: volInfo.Created}
 			return nil
 		}, index)
 	}
@@ -131,8 +139,8 @@ func (er erasureObjects) getBucketInfo(ctx context.Context, bucketName string) (
 }

 // GetBucketInfo - returns BucketInfo for a bucket.
-func (er erasureObjects) GetBucketInfo(ctx context.Context, bucket string) (bi BucketInfo, e error) {
-	bucketInfo, err := er.getBucketInfo(ctx, bucket)
+func (er erasureObjects) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (bi BucketInfo, e error) {
+	bucketInfo, err := er.getBucketInfo(ctx, bucket, opts)
 	if err != nil {
 		return bi, toObjectErr(err, bucket)
 	}
@@ -191,12 +199,58 @@ func (er erasureObjects) DeleteBucket(ctx context.Context, bucket string, opts D
 		if purgedDangling {
 			err = nil
 		}
-
+		if opts.SRDeleteOp == MarkDelete {
+			er.markDelete(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, deletedBucketsPrefix, bucket))
+		}
 	}

 	return toObjectErr(err, bucket)
 }

+// markDelete creates a vol entry in .minio.sys/buckets/.deleted until site replication
+// syncs the delete to peers
+func (er erasureObjects) markDelete(ctx context.Context, bucket, prefix string) error {
+	storageDisks := er.getDisks()
+	g := errgroup.WithNErrs(len(storageDisks))
+	// Make a volume entry on all underlying storage disks.
+	for index := range storageDisks {
+		index := index
+		if storageDisks[index] == nil {
+			continue
+		}
+		g.Go(func() error {
+			if err := storageDisks[index].MakeVol(ctx, pathJoin(bucket, prefix)); err != nil {
+				if errors.Is(err, errVolumeExists) {
+					return nil
+				}
+				return err
+			}
+			return nil
+		}, index)
+	}
+	err := reduceWriteQuorumErrs(ctx, g.Wait(), bucketOpIgnoredErrs, er.defaultWQuorum())
+	return toObjectErr(err, bucket)
+}
+
+// purgeDelete deletes vol entry in .minio.sys/buckets/.deleted after site replication
+// syncs the delete to peers OR on a new MakeBucket call.
+func (er erasureObjects) purgeDelete(ctx context.Context, bucket, prefix string) error {
+	storageDisks := er.getDisks()
+	g := errgroup.WithNErrs(len(storageDisks))
+	// Make a volume entry on all underlying storage disks.
+	for index := range storageDisks {
+		index := index
+		g.Go(func() error {
+			if storageDisks[index] != nil {
+				return storageDisks[index].DeleteVol(ctx, pathJoin(bucket, prefix), true)
+			}
+			return errDiskNotFound
+		}, index)
+	}
+	err := reduceWriteQuorumErrs(ctx, g.Wait(), bucketOpIgnoredErrs, er.defaultWQuorum())
+	return toObjectErr(err, bucket)
+}
+
 // IsNotificationSupported returns whether bucket notification is applicable for this layer.
 func (er erasureObjects) IsNotificationSupported() bool {
 	return true
@@ -19,7 +19,11 @@ package cmd

 import (
 	"context"
+	"fmt"
 	"sync"
+
+	"github.com/minio/minio/internal/logger"
+	"github.com/minio/minio/internal/sync/errgroup"
 )

 func (er erasureObjects) getOnlineDisks() (newDisks []StorageAPI) {
@@ -119,3 +123,84 @@ func (er erasureObjects) getLoadBalancedDisks(optimized bool) []StorageAPI {
 	// Return disks which have maximum disk usage common.
 	return newDisks[max]
 }
+
+// readMultipleFiles Reads raw data from all specified files from all disks.
+func readMultipleFiles(ctx context.Context, disks []StorageAPI, req ReadMultipleReq, readQuorum int) ([]ReadMultipleResp, error) {
+	resps := make([]chan ReadMultipleResp, len(disks))
+	for i := range resps {
+		resps[i] = make(chan ReadMultipleResp, len(req.Files))
+	}
+	g := errgroup.WithNErrs(len(disks))
+	// Read files in parallel across disks.
+	for index := range disks {
+		index := index
+		g.Go(func() (err error) {
+			if disks[index] == nil {
+				return errDiskNotFound
+			}
+			return disks[index].ReadMultiple(ctx, req, resps[index])
+		}, index)
+	}
+
+	dataArray := make([]ReadMultipleResp, 0, len(req.Files))
+	// Merge results. They should come in order from each.
+	for _, wantFile := range req.Files {
+		quorum := 0
+		toAdd := ReadMultipleResp{
+			Bucket: req.Bucket,
+			Prefix: req.Prefix,
+			File:   wantFile,
+		}
+		for i := range resps {
+			if disks[i] == nil {
+				continue
+			}
+			select {
+			case <-ctx.Done():
+			case gotFile, ok := <-resps[i]:
+				if !ok {
+					continue
+				}
+				if gotFile.Error != "" || !gotFile.Exists {
+					continue
+				}
+				if gotFile.File != wantFile || gotFile.Bucket != req.Bucket || gotFile.Prefix != req.Prefix {
+					continue
+				}
+				quorum++
+				if toAdd.Modtime.After(gotFile.Modtime) || len(gotFile.Data) < len(toAdd.Data) {
+					// Pick latest, or largest to avoid possible truncated entries.
+					continue
+				}
+				toAdd = gotFile
+			}
+		}
+		if quorum < readQuorum {
+			toAdd.Exists = false
+			toAdd.Error = errErasureReadQuorum.Error()
+			toAdd.Data = nil
+		}
+		dataArray = append(dataArray, toAdd)
+	}
+
+	errs := g.Wait()
+	for index, err := range errs {
+		if err == nil {
+			continue
+		}
+		if !IsErr(err, []error{
+			errFileNotFound,
+			errVolumeNotFound,
+			errFileVersionNotFound,
+			errDiskNotFound,
+			errUnformattedDisk,
+		}...) {
+			logger.LogOnceIf(ctx, fmt.Errorf("Drive %s, path (%s/%s) returned an error (%w)",
+				disks[index], req.Bucket, req.Prefix, err),
+				disks[index].String())
+		}
+	}
+
+	// Return all the metadata.
+	return dataArray, nil
+}
@@ -85,19 +85,17 @@ var erasureDecodeTests = []struct {

 func TestErasureDecode(t *testing.T) {
 	for i, test := range erasureDecodeTests {
-		setup, err := newErasureTestSetup(test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
+		setup, err := newErasureTestSetup(t, test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
 		if err != nil {
 			t.Fatalf("Test %d: failed to create test setup: %v", i, err)
 		}
 		erasure, err := NewErasure(context.Background(), test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create ErasureStorage: %v", i, err)
 		}
 		disks := setup.disks
 		data := make([]byte, test.data)
 		if _, err = io.ReadFull(crand.Reader, data); err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to generate random test data: %v", i, err)
 		}

@@ -113,11 +111,9 @@ func TestErasureDecode(t *testing.T) {
 		n, err := erasure.Encode(context.Background(), bytes.NewReader(data), writers, buffer, erasure.dataBlocks+1)
 		closeBitrotWriters(writers)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create erasure test file: %v", i, err)
 		}
 		if n != test.data {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create erasure test file", i)
 		}
 		for i, w := range writers {
@@ -195,7 +191,6 @@ func TestErasureDecode(t *testing.T) {
 				}
 			}
 		}
-		setup.Remove()
 	}
 }

@@ -210,12 +205,11 @@ func TestErasureDecodeRandomOffsetLength(t *testing.T) {
 	dataBlocks := 7
 	parityBlocks := 7
 	blockSize := int64(1 * humanize.MiByte)
-	setup, err := newErasureTestSetup(dataBlocks, parityBlocks, blockSize)
+	setup, err := newErasureTestSetup(t, dataBlocks, parityBlocks, blockSize)
 	if err != nil {
 		t.Error(err)
 		return
 	}
-	defer setup.Remove()
 	disks := setup.disks
 	erasure, err := NewErasure(context.Background(), dataBlocks, parityBlocks, blockSize)
 	if err != nil {
@@ -288,11 +282,10 @@ func TestErasureDecodeRandomOffsetLength(t *testing.T) {
 // Benchmarks

 func benchmarkErasureDecode(data, parity, dataDown, parityDown int, size int64, b *testing.B) {
-	setup, err := newErasureTestSetup(data, parity, blockSizeV2)
+	setup, err := newErasureTestSetup(b, data, parity, blockSizeV2)
 	if err != nil {
 		b.Fatalf("failed to create test setup: %v", err)
 	}
-	defer setup.Remove()
 	disks := setup.disks
 	erasure, err := NewErasure(context.Background(), data, parity, blockSizeV2)
 	if err != nil {
@@ -87,21 +87,19 @@ var erasureEncodeTests = []struct {

 func TestErasureEncode(t *testing.T) {
 	for i, test := range erasureEncodeTests {
-		setup, err := newErasureTestSetup(test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
+		setup, err := newErasureTestSetup(t, test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
 		if err != nil {
 			t.Fatalf("Test %d: failed to create test setup: %v", i, err)
 		}
 		disks := setup.disks
 		erasure, err := NewErasure(context.Background(), test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create ErasureStorage: %v", i, err)
 		}
 		buffer := make([]byte, test.blocksize, 2*test.blocksize)

 		data := make([]byte, test.data)
 		if _, err = io.ReadFull(rand.Reader, data); err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to generate random test data: %v", i, err)
 		}
 		writers := make([]io.Writer, len(disks))
@@ -160,18 +158,16 @@ func TestErasureEncode(t *testing.T) {
 				}
 			}
 		}
-		setup.Remove()
 	}
 }

 // Benchmarks

 func benchmarkErasureEncode(data, parity, dataDown, parityDown int, size int64, b *testing.B) {
-	setup, err := newErasureTestSetup(data, parity, blockSizeV2)
+	setup, err := newErasureTestSetup(b, data, parity, blockSizeV2)
 	if err != nil {
 		b.Fatalf("failed to create test setup: %v", err)
 	}
-	defer setup.Remove()
 	erasure, err := NewErasure(context.Background(), data, parity, blockSizeV2)
 	if err != nil {
 		b.Fatalf("failed to create ErasureStorage: %v", err)
@@ -196,7 +192,10 @@ func benchmarkErasureEncode(data, parity, dataDown, parityDown int, size int64,
 			if disk == OfflineDisk {
 				continue
 			}
-			disk.Delete(context.Background(), "testbucket", "object", false)
+			disk.Delete(context.Background(), "testbucket", "object", DeleteOptions{
+				Recursive: false,
+				Force:     false,
+			})
 			writers[i] = newBitrotWriter(disk, "testbucket", "object", erasure.ShardFileSize(size), DefaultBitrotAlgorithm, erasure.ShardSize())
 		}
 		_, err := erasure.Encode(context.Background(), bytes.NewReader(content), writers, buffer, erasure.dataBlocks+1)
@@ -20,10 +20,10 @@ package cmd
 import "errors"

 // errErasureReadQuorum - did not meet read quorum.
-var errErasureReadQuorum = errors.New("Read failed. Insufficient number of disks online")
+var errErasureReadQuorum = errors.New("Read failed. Insufficient number of drives online")

 // errErasureWriteQuorum - did not meet write quorum.
-var errErasureWriteQuorum = errors.New("Write failed. Insufficient number of disks online")
+var errErasureWriteQuorum = errors.New("Write failed. Insufficient number of drives online")

 // errNoHealRequired - returned when healing is attempted on a previously healed disks.
 var errNoHealRequired = errors.New("No healing is required")
@@ -66,23 +66,21 @@ func TestErasureHeal(t *testing.T) {
 	for i, test := range erasureHealTests {
 		if test.offDisks < test.badStaleDisks {
 			// test case sanity check
-			t.Fatalf("Test %d: Bad test case - number of stale disks cannot be less than number of badstale disks", i)
+			t.Fatalf("Test %d: Bad test case - number of stale drives cannot be less than number of badstale drives", i)
 		}

 		// create some test data
-		setup, err := newErasureTestSetup(test.dataBlocks, test.disks-test.dataBlocks, test.blocksize)
+		setup, err := newErasureTestSetup(t, test.dataBlocks, test.disks-test.dataBlocks, test.blocksize)
 		if err != nil {
 			t.Fatalf("Test %d: failed to setup Erasure environment: %v", i, err)
 		}
 		disks := setup.disks
 		erasure, err := NewErasure(context.Background(), test.dataBlocks, test.disks-test.dataBlocks, test.blocksize)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create ErasureStorage: %v", i, err)
 		}
 		data := make([]byte, test.size)
 		if _, err = io.ReadFull(rand.Reader, data); err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create random test data: %v", i, err)
 		}
 		buffer := make([]byte, test.blocksize, 2*test.blocksize)
@@ -93,7 +91,6 @@ func TestErasureHeal(t *testing.T) {
 		_, err = erasure.Encode(context.Background(), bytes.NewReader(data), writers, buffer, erasure.dataBlocks+1)
 		closeBitrotWriters(writers)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create random test data: %v", i, err)
 		}

@@ -156,6 +153,5 @@ func TestErasureHeal(t *testing.T) {
 				}
 			}
 		}
-		setup.Remove()
 	}
 }
@@ -180,9 +180,15 @@ func listOnlineDisks(disks []StorageAPI, partsMetadata []FileInfo, errs []error)
 }

 // Returns the latest updated FileInfo files and error in case of failure.
-func getLatestFileInfo(ctx context.Context, partsMetadata []FileInfo, errs []error) (FileInfo, error) {
+func getLatestFileInfo(ctx context.Context, partsMetadata []FileInfo, defaultParityCount int, errs []error) (FileInfo, error) {
 	// There should be atleast half correct entries, if not return failure
-	reducedErr := reduceReadQuorumErrs(ctx, errs, objectOpIgnoredErrs, len(partsMetadata)/2)
+	expectedRQuorum := len(partsMetadata) / 2
+	if defaultParityCount == 0 {
+		// if parity count is '0', we expected all entries to be present.
+		expectedRQuorum = len(partsMetadata)
+	}
+
+	reducedErr := reduceReadQuorumErrs(ctx, errs, objectOpIgnoredErrs, expectedRQuorum)
 	if reducedErr != nil {
 		return FileInfo{}, reducedErr
 	}
@@ -225,8 +231,8 @@ func getLatestFileInfo(ctx context.Context, partsMetadata []FileInfo, errs []err
 //
 // - disks which have all parts specified in the latest xl.meta.
 //
-// - slice of errors about the state of data files on disk - can have
-//   a not-found error or a hash-mismatch error.
+//   - slice of errors about the state of data files on disk - can have
+//     a not-found error or a hash-mismatch error.
 func disksWithAllParts(ctx context.Context, onlineDisks []StorageAPI, partsMetadata []FileInfo,
 	errs []error, latestMeta FileInfo, bucket, object string,
 	scanMode madmin.HealScanMode) ([]StorageAPI, []error, time.Time,
@@ -176,7 +176,7 @@ func TestListOnlineDisks(t *testing.T) {
 	}

 	bucket := "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket %v", err)
 	}
@@ -194,7 +194,7 @@ func TestListOnlineDisks(t *testing.T) {
 			}

 			partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
-			fi, err := getLatestFileInfo(ctx, partsMetadata, errs)
+			fi, err := getLatestFileInfo(ctx, partsMetadata, z.serverPools[0].sets[0].defaultParityCount, errs)
 			if err != nil {
 				t.Fatalf("Failed to getLatestFileInfo %v", err)
 			}
@@ -218,7 +218,10 @@ func TestListOnlineDisks(t *testing.T) {
 					// and check if that disk
 					// appears in outDatedDisks.
 					tamperedIndex = index
-					dErr := erasureDisks[index].Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), false)
+					dErr := erasureDisks[index].Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), DeleteOptions{
+						Recursive: false,
+						Force:     false,
+					})
 					if dErr != nil {
 						t.Fatalf("Failed to delete %s - %v", filepath.Join(object, "part.1"), dErr)
 					}
@@ -257,7 +260,7 @@ func TestListOnlineDisks(t *testing.T) {

 			if test._tamperBackend != noTamper {
 				if tamperedIndex != -1 && availableDisks[tamperedIndex] != nil {
-					t.Fatalf("disk (%v) with part.1 missing is not a disk with available data",
+					t.Fatalf("Drive (%v) with part.1 missing is not a drive with available data",
 						erasureDisks[tamperedIndex])
 				}
 			}
@@ -348,7 +351,7 @@ func TestListOnlineDisksSmallObjects(t *testing.T) {
 	}

 	bucket := "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket %v", err)
 	}
@@ -367,7 +370,7 @@ func TestListOnlineDisksSmallObjects(t *testing.T) {
 			}

 			partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", true)
-			_, err = getLatestFileInfo(ctx, partsMetadata, errs)
+			_, err = getLatestFileInfo(ctx, partsMetadata, z.serverPools[0].sets[0].defaultParityCount, errs)
 			if err != nil {
 				t.Fatalf("Failed to getLatestFileInfo %v", err)
 			}
@@ -395,7 +398,10 @@ func TestListOnlineDisksSmallObjects(t *testing.T) {
 					// and check if that disk
 					// appears in outDatedDisks.
 					tamperedIndex = index
-					dErr := erasureDisks[index].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), false)
+					dErr := erasureDisks[index].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
+						Recursive: false,
+						Force:     false,
+					})
 					if dErr != nil {
 						t.Fatalf("Failed to delete %s - %v", pathJoin(object, xlStorageFormatFile), dErr)
 					}
@@ -423,7 +429,7 @@ func TestListOnlineDisksSmallObjects(t *testing.T) {

 			}
 			partsMetadata, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", true)
-			fi, err := getLatestFileInfo(ctx, partsMetadata, errs)
+			fi, err := getLatestFileInfo(ctx, partsMetadata, z.serverPools[0].sets[0].defaultParityCount, errs)
 			if !errors.Is(err, errErasureReadQuorum) {
 				t.Fatalf("Failed to getLatestFileInfo, expected %v, got %v", errErasureReadQuorum, err)
 			}
@@ -440,7 +446,7 @@ func TestListOnlineDisksSmallObjects(t *testing.T) {

 			if test._tamperBackend != noTamper {
 				if tamperedIndex != -1 && availableDisks[tamperedIndex] != nil {
-					t.Fatalf("disk (%v) with part.1 missing is not a disk with available data",
+					t.Fatalf("Drive (%v) with part.1 missing is not a drive with available data",
 						erasureDisks[tamperedIndex])
 				}
 			}
@@ -466,7 +472,7 @@ func TestDisksWithAllParts(t *testing.T) {
 	z := obj.(*erasureServerPools)
 	s := z.serverPools[0].sets[0]
 	erasureDisks := s.getDisks()
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket %v", err)
 	}
@@ -489,7 +495,7 @@ func TestDisksWithAllParts(t *testing.T) {
 		t.Fatalf("Failed to read xl meta data %v", err)
 	}

-	fi, err := getLatestFileInfo(ctx, partsMetadata, errs)
+	fi, err := getLatestFileInfo(ctx, partsMetadata, s.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to get quorum consistent fileInfo %v", err)
 	}
@@ -500,7 +506,7 @@ func TestDisksWithAllParts(t *testing.T) {
 		errs, fi, bucket, object, madmin.HealDeepScan)

 	if len(filteredDisks) != len(erasureDisks) {
-		t.Errorf("Unexpected number of disks: %d", len(filteredDisks))
+		t.Errorf("Unexpected number of drives: %d", len(filteredDisks))
 	}

 	for diskIndex, disk := range filteredDisks {
@@ -509,7 +515,7 @@ func TestDisksWithAllParts(t *testing.T) {
 		}

 		if disk == nil {
-			t.Errorf("Disk erroneously filtered, diskIndex: %d", diskIndex)
+			t.Errorf("Drive erroneously filtered, driveIndex: %d", diskIndex)
 		}
 	}

@@ -522,14 +528,14 @@ func TestDisksWithAllParts(t *testing.T) {
 		errs, fi, bucket, object, madmin.HealDeepScan)

 	if len(filteredDisks) != len(erasureDisks) {
-		t.Errorf("Unexpected number of disks: %d", len(filteredDisks))
+		t.Errorf("Unexpected number of drives: %d", len(filteredDisks))
 	}
 	for diskIndex, disk := range filteredDisks {
 		if diskIndex == 0 && disk != nil {
-			t.Errorf("Disk not filtered as expected, disk: %d", diskIndex)
+			t.Errorf("Drive not filtered as expected, drive: %d", diskIndex)
 		}
 		if diskIndex != 0 && disk == nil {
-			t.Errorf("Disk erroneously filtered, diskIndex: %d", diskIndex)
+			t.Errorf("Drive erroneously filtered, driveIndex: %d", diskIndex)
 		}
 	}
 	partsMetadata[0] = partsMetadataBackup // Revert before going to the next test
@@ -543,14 +549,14 @@ func TestDisksWithAllParts(t *testing.T) {
 		errs, fi, bucket, object, madmin.HealDeepScan)

 	if len(filteredDisks) != len(erasureDisks) {
-		t.Errorf("Unexpected number of disks: %d", len(filteredDisks))
+		t.Errorf("Unexpected number of drives: %d", len(filteredDisks))
 	}
 	for diskIndex, disk := range filteredDisks {
 		if diskIndex == 1 && disk != nil {
-			t.Errorf("Disk not filtered as expected, disk: %d", diskIndex)
+			t.Errorf("Drive not filtered as expected, drive: %d", diskIndex)
 		}
 		if diskIndex != 1 && disk == nil {
-			t.Errorf("Disk erroneously filtered, diskIndex: %d", diskIndex)
+			t.Errorf("Drive erroneously filtered, driveIndex: %d", diskIndex)
 		}
 	}
 	partsMetadata[1] = partsMetadataBackup // Revert before going to the next test
@@ -580,23 +586,23 @@ func TestDisksWithAllParts(t *testing.T) {
 		errs, fi, bucket, object, madmin.HealDeepScan)

 	if len(filteredDisks) != len(erasureDisks) {
-		t.Errorf("Unexpected number of disks: %d", len(filteredDisks))
+		t.Errorf("Unexpected number of drives: %d", len(filteredDisks))
 	}

 	for diskIndex, disk := range filteredDisks {
 		if _, ok := diskFailures[diskIndex]; ok {
 			if disk != nil {
-				t.Errorf("Disk not filtered as expected, disk: %d", diskIndex)
+				t.Errorf("Drive not filtered as expected, drive: %d", diskIndex)
 			}
 			if errs[diskIndex] == nil {
-				t.Errorf("Expected error not received, diskIndex: %d", diskIndex)
+				t.Errorf("Expected error not received, driveIndex: %d", diskIndex)
 			}
 		} else {
 			if disk == nil {
-				t.Errorf("Disk erroneously filtered, diskIndex: %d", diskIndex)
+				t.Errorf("Drive erroneously filtered, driveIndex: %d", diskIndex)
 			}
 			if errs[diskIndex] != nil {
-				t.Errorf("Unexpected error, %s, diskIndex: %d", errs[diskIndex], diskIndex)
+				t.Errorf("Unexpected error, %s, driveIndex: %d", errs[diskIndex], diskIndex)
 			}

 		}
@@ -243,7 +243,7 @@ func listAllBuckets(ctx context.Context, storageDisks []StorageAPI, healBuckets

 // Only heal on disks where we are sure that healing is needed. We can expand
 // this list as and when we figure out more errors can be added to this list safely.
-func shouldHealObjectOnDisk(erErr, dataErr error, meta FileInfo, latestMeta FileInfo) bool {
+func shouldHealObjectOnDisk(erErr, dataErr error, meta FileInfo, latestMeta FileInfo, doinline bool) bool {
 	switch {
 	case errors.Is(erErr, errFileNotFound) || errors.Is(erErr, errFileVersionNotFound):
 		return true
@@ -256,6 +256,10 @@ func shouldHealObjectOnDisk(erErr, dataErr error, meta FileInfo, latestMeta File
 			// always check first.
 			return true
 		}
+		if doinline {
+			// convert small files to 'inline'
+			return true
+		}
 		if !meta.Deleted && !meta.IsRemote() {
 			// If xl.meta was read fine but there may be problem with the part.N files.
 			if IsErr(dataErr, []error{
@@ -275,10 +279,6 @@ func shouldHealObjectOnDisk(erErr, dataErr error, meta FileInfo, latestMeta File

 // Heals an object by re-writing corrupt/missing erasure blocks.
 func (er erasureObjects) healObject(ctx context.Context, bucket string, object string, versionID string, opts madmin.HealOpts) (result madmin.HealResultItem, err error) {
-	if !opts.DryRun {
-		defer NSUpdated(bucket, object)
-	}
-
 	dryRun := opts.DryRun
 	scanMode := opts.ScanMode

@@ -346,6 +346,26 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 	availableDisks, dataErrs, diskMTime := disksWithAllParts(ctx, onlineDisks, partsMetadata,
 		errs, latestMeta, bucket, object, scanMode)

+	var erasure Erasure
+	var recreate bool
+	if !latestMeta.Deleted && !latestMeta.IsRemote() {
+		// Initialize erasure coding
+		erasure, err = NewErasure(ctx, latestMeta.Erasure.DataBlocks,
+			latestMeta.Erasure.ParityBlocks, latestMeta.Erasure.BlockSize)
+		if err != nil {
+			return result, err
+		}
+
+		// Is only 'true' if the opts.Recreate is true and
+		// the object shardSize < smallFileThreshold do not
+		// set this to 'true' arbitrarily and must be only
+		// 'true' with caller ask.
+		recreate = (opts.Recreate &&
+			!latestMeta.InlineData() &&
+			len(latestMeta.Parts) == 1 &&
+			erasure.ShardFileSize(latestMeta.Parts[0].ActualSize) < smallFileThreshold)
+	}
+
 	// Loop to find number of disks with valid data, per-drive
 	// data state and a list of outdated disks on which data needs
 	// to be healed.
@@ -372,7 +392,7 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 			driveState = madmin.DriveStateCorrupt
 		}

-		if shouldHealObjectOnDisk(errs[i], dataErrs[i], partsMetadata[i], latestMeta) {
+		if shouldHealObjectOnDisk(errs[i], dataErrs[i], partsMetadata[i], latestMeta, recreate) {
 			outDatedDisks[i] = storageDisks[i]
 			disksToHealCount++
 			result.Before.Drives = append(result.Before.Drives, madmin.HealDriveInfo{
@@ -426,9 +446,9 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 		return result, nil
 	}

-	if !latestMeta.XLV1 && !latestMeta.Deleted && disksToHealCount > latestMeta.Erasure.ParityBlocks {
+	if !latestMeta.XLV1 && !latestMeta.Deleted && !recreate && disksToHealCount > latestMeta.Erasure.ParityBlocks {
 		// When disk to heal count is greater than parity blocks we should simply error out.
-		err := fmt.Errorf("more disks are expected to heal than parity, returned errors: %v (dataErrs %v) -> %s/%s(%s)", errs, dataErrs, bucket, object, versionID)
+		err := fmt.Errorf("more drives are expected to heal than parity, returned errors: %v (dataErrs %v) -> %s/%s(%s)", errs, dataErrs, bucket, object, versionID)
 		logger.LogIf(ctx, err)
 		return er.defaultHealResult(latestMeta, storageDisks, storageEndpoints, errs,
 			bucket, object, versionID), err
@@ -482,25 +502,18 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 	}

 	var inlineBuffers []*bytes.Buffer
-	if latestMeta.InlineData() {
-		inlineBuffers = make([]*bytes.Buffer, len(outDatedDisks))
-	}
-
 	if !latestMeta.Deleted && !latestMeta.IsRemote() {
-		// Heal each part. erasureHealFile() will write the healed
-		// part to .minio/tmp/uuid/ which needs to be renamed later to
-		// the final location.
-		erasure, err := NewErasure(ctx, latestMeta.Erasure.DataBlocks,
-			latestMeta.Erasure.ParityBlocks, latestMeta.Erasure.BlockSize)
-		if err != nil {
-			return result, err
+		if latestMeta.InlineData() || recreate {
+			inlineBuffers = make([]*bytes.Buffer, len(outDatedDisks))
 		}

 		erasureInfo := latestMeta.Erasure
 		for partIndex := 0; partIndex < len(latestMeta.Parts); partIndex++ {
 			partSize := latestMeta.Parts[partIndex].Size
 			partActualSize := latestMeta.Parts[partIndex].ActualSize
+			partModTime := latestMeta.Parts[partIndex].ModTime
 			partNumber := latestMeta.Parts[partIndex].Number
+			partIdx := latestMeta.Parts[partIndex].Index
 			tillOffset := erasure.ShardFileOffset(0, partSize, partSize)
 			readers := make([]io.ReaderAt, len(latestDisks))
 			checksumAlgo := erasureInfo.GetChecksumInfo(partNumber).Algorithm
@@ -527,6 +540,10 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 						tillOffset, DefaultBitrotAlgorithm, erasure.ShardSize())
 				}
 			}
+
+			// Heal each part. erasure.Heal() will write the healed
+			// part to .minio/tmp/uuid/ which needs to be renamed
+			// later to the final location.
 			err = erasure.Heal(ctx, writers, readers, partSize)
 			closeBitrotReaders(readers)
 			closeBitrotWriters(writers)
@@ -550,7 +567,7 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 				}

 				partsMetadata[i].DataDir = dstDataDir
-				partsMetadata[i].AddObjectPart(partNumber, "", partSize, partActualSize)
+				partsMetadata[i].AddObjectPart(partNumber, "", partSize, partActualSize, partModTime, partIdx)
 				partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{
 					PartNumber: partNumber,
 					Algorithm:  checksumAlgo,
@@ -558,6 +575,7 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 				})
 				if len(inlineBuffers) > 0 && inlineBuffers[i] != nil {
 					partsMetadata[i].Data = inlineBuffers[i].Bytes()
+					partsMetadata[i].SetInlineData()
 				} else {
 					partsMetadata[i].Data = nil
 				}
@@ -565,7 +583,7 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s

 			// If all disks are having errors, we give up.
 			if disksToHealCount == 0 {
-				return result, fmt.Errorf("all disks had write errors, unable to heal %s/%s", bucket, object)
+				return result, fmt.Errorf("all drives had write errors, unable to heal %s/%s", bucket, object)
 			}

 		}
@@ -589,8 +607,9 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 			return result, err
 		}

-		// Remove any remaining parts from outdated disks from before transition.
-		if partsMetadata[i].IsRemote() {
+		// - Remove any parts from healed disks after its been inlined.
+		// - Remove any remaining parts from outdated disks from before transition.
+		if recreate || partsMetadata[i].IsRemote() {
 			rmDataDir := partsMetadata[i].DataDir
 			disk.DeleteVol(ctx, pathJoin(bucket, encodeDirObject(object), rmDataDir), true)
 		}
@@ -655,7 +674,10 @@ func (er erasureObjects) healObjectDir(ctx context.Context, bucket, object strin
 				wg.Add(1)
 				go func(index int, disk StorageAPI) {
 					defer wg.Done()
-					_ = disk.Delete(ctx, bucket, object, false)
+					_ = disk.Delete(ctx, bucket, object, DeleteOptions{
+						Recursive: false,
+						Force:     false,
+					})
 				}(index, disk)
 			}
 			wg.Wait()
@@ -236,7 +236,7 @@ func TestHealing(t *testing.T) {
 	er := z.serverPools[0].sets[0]

 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -359,7 +359,7 @@ func TestHealingDanglingObject(t *testing.T) {
 	object := getRandomObjectName()
 	data := bytes.Repeat([]byte("a"), 128*1024)

-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -554,7 +554,7 @@ func TestHealCorrectQuorum(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions

-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -580,7 +580,7 @@ func TestHealCorrectQuorum(t *testing.T) {

 	_, err = objLayer.CompleteMultipartUpload(ctx, bucket, object, uploadID, uploadedParts, ObjectOptions{})
 	if err != nil {
-		t.Fatalf("Failed to complete multipart upload - %v", err)
+		t.Fatalf("Failed to complete multipart upload - got: %v", err)
 	}

 	cfgFile := pathJoin(bucketMetaPrefix, bucket, ".test.bin")
@@ -601,7 +601,7 @@ func TestHealCorrectQuorum(t *testing.T) {
 		erasureDisks := er.getDisks()

 		fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
-		nfi, err := getLatestFileInfo(ctx, fileInfos, errs)
+		nfi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 		if errors.Is(err, errFileNotFound) {
 			continue
 		}
@@ -610,7 +610,10 @@ func TestHealCorrectQuorum(t *testing.T) {
 		}

 		for i := 0; i < nfi.Erasure.ParityBlocks; i++ {
-			erasureDisks[i].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), false)
+			erasureDisks[i].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
+				Recursive: false,
+				Force:     false,
+			})
 		}

 		// Try healing now, it should heal the content properly.
@@ -625,7 +628,7 @@ func TestHealCorrectQuorum(t *testing.T) {
 		}

 		fileInfos, errs = readAllFileInfo(ctx, erasureDisks, minioMetaBucket, cfgFile, "", false)
-		nfi, err = getLatestFileInfo(ctx, fileInfos, errs)
+		nfi, err = getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 		if errors.Is(err, errFileNotFound) {
 			continue
 		}
@@ -634,7 +637,10 @@ func TestHealCorrectQuorum(t *testing.T) {
 		}

 		for i := 0; i < nfi.Erasure.ParityBlocks; i++ {
-			erasureDisks[i].Delete(context.Background(), minioMetaBucket, pathJoin(cfgFile, xlStorageFormatFile), false)
+			erasureDisks[i].Delete(context.Background(), minioMetaBucket, pathJoin(cfgFile, xlStorageFormatFile), DeleteOptions{
+				Recursive: false,
+				Force:     false,
+			})
 		}

 		// Try healing now, it should heal the content properly.
@@ -679,7 +685,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions

-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -714,7 +720,10 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	er := set.sets[0]
 	erasureDisks := er.getDisks()
 	firstDisk := erasureDisks[0]
-	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), false)
+	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
+		Recursive: false,
+		Force:     false,
+	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
 	}
@@ -725,7 +734,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	}

 	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
-	fi, err := getLatestFileInfo(ctx, fileInfos, errs)
+	fi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
 	}
@@ -734,7 +743,10 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 		t.Errorf("Expected xl.meta file to be present but stat failed - %v", err)
 	}

-	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), false)
+	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), DeleteOptions{
+		Recursive: false,
+		Force:     false,
+	})
 	if err != nil {
 		t.Errorf("Failure during deleting part.1 - %v", err)
 	}
@@ -750,7 +762,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	}

 	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
-	nfi, err := getLatestFileInfo(ctx, fileInfos, errs)
+	nfi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
 	}
@@ -761,7 +773,10 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 		t.Fatalf("FileInfo not equal after healing: %v != %v", fi, nfi)
 	}

-	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), false)
+	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), DeleteOptions{
+		Recursive: false,
+		Force:     false,
+	})
 	if err != nil {
 		t.Errorf("Failure during deleting part.1 - %v", err)
 	}
@@ -778,7 +793,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	}

 	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
-	nfi, err = getLatestFileInfo(ctx, fileInfos, errs)
+	nfi, err = getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
 	}
@@ -792,7 +807,10 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	// Test 4: checks if HealObject returns an error when xl.meta is not found
 	// in more than read quorum number of disks, to create a corrupted situation.
 	for i := 0; i <= nfi.Erasure.DataBlocks; i++ {
-		erasureDisks[i].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), false)
+		erasureDisks[i].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
+			Recursive: false,
+			Force:     false,
+		})
 	}

 	// Try healing now, expect to receive errFileNotFound.
@@ -843,7 +861,7 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions

-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -879,12 +897,15 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {

 	// Test 1: Remove the object backend files from the first disk.
 	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
-	fi, err := getLatestFileInfo(ctx, fileInfos, errs)
+	fi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
 	}

-	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), false)
+	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
+		Recursive: false,
+		Force:     false,
+	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
 	}
@@ -899,7 +920,7 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {
 	}

 	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
-	nfi1, err := getLatestFileInfo(ctx, fileInfos, errs)
+	nfi1, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
 	}
@@ -922,7 +943,7 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {
 	}

 	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
-	nfi2, err := getLatestFileInfo(ctx, fileInfos, errs)
+	nfi2, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
 	}
@@ -936,7 +957,10 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {
 	// Test 3: checks if HealObject returns an error when xl.meta is not found
 	// in more than read quorum number of disks, to create a corrupted situation.
 	for i := 0; i <= nfi2.Erasure.DataBlocks; i++ {
-		erasureDisks[i].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), false)
+		erasureDisks[i].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
+			Recursive: false,
+			Force:     false,
+		})
 	}

 	// Try healing now, expect to receive errFileNotFound.
@@ -980,7 +1004,7 @@ func TestHealObjectCorruptedParts(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions

-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -1017,7 +1041,7 @@ func TestHealObjectCorruptedParts(t *testing.T) {
 	secondDisk := erasureDisks[1]

 	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
-	fi, err := getLatestFileInfo(ctx, fileInfos, errs)
+	fi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
 	}
@@ -1033,7 +1057,10 @@ func TestHealObjectCorruptedParts(t *testing.T) {
 	}

 	// Test 1, remove part.1
-	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), false)
+	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), DeleteOptions{
+		Recursive: false,
+		Force:     false,
+	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
 	}
@@ -1078,7 +1105,10 @@ func TestHealObjectCorruptedParts(t *testing.T) {
 		t.Fatalf("Failed to write a file - %v", err)
 	}

-	err = secondDisk.Delete(context.Background(), bucket, object, true)
+	err = secondDisk.Delete(context.Background(), bucket, object, DeleteOptions{
+		Recursive: true,
+		Force:     false,
+	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
 	}
@@ -1131,7 +1161,7 @@ func TestHealObjectErasure(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions

-	err = obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -1166,7 +1196,10 @@ func TestHealObjectErasure(t *testing.T) {
 	}

 	// Delete the whole object folder
-	err = firstDisk.Delete(context.Background(), bucket, object, true)
+	err = firstDisk.Delete(context.Background(), bucket, object, DeleteOptions{
+		Recursive: true,
+		Force:     false,
+	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
 	}
@@ -1225,7 +1258,7 @@ func TestHealEmptyDirectoryErasure(t *testing.T) {
 	object := "empty-dir/"
 	var opts ObjectOptions

-	err = obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -1326,7 +1359,7 @@ func TestHealLastDataShard(t *testing.T) {
 			}
 			var opts ObjectOptions

-			err = obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+			err = obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 			if err != nil {
 				t.Fatalf("Failed to make a bucket - %v", err)
 			}
@@ -275,7 +275,7 @@ func shuffleDisks(disks []StorageAPI, distribution []int) (shuffledDisks []Stora
 // the corresponding error in errs slice is not nil
 func evalDisks(disks []StorageAPI, errs []error) []StorageAPI {
 	if len(errs) != len(disks) {
-		logger.LogIf(GlobalContext, errors.New("unexpected disks/errors slice length"))
+		logger.LogIf(GlobalContext, errors.New("unexpected drives/errors slice length"))
 		return nil
 	}
 	newDisks := make([]StorageAPI, len(disks))
@@ -237,12 +237,14 @@ func objectPartIndex(parts []ObjectPartInfo, partNumber int) int {
 }

 // AddObjectPart - add a new object part in order.
-func (fi *FileInfo) AddObjectPart(partNumber int, partETag string, partSize int64, actualSize int64) {
+func (fi *FileInfo) AddObjectPart(partNumber int, partETag string, partSize, actualSize int64, modTime time.Time, idx []byte) {
 	partInfo := ObjectPartInfo{
 		Number:     partNumber,
 		ETag:       partETag,
 		Size:       partSize,
 		ActualSize: actualSize,
+		ModTime:    modTime,
+		Index:      idx,
 	}

 	// Update part info if it already exists.
@@ -398,17 +400,17 @@ func writeUniqueFileInfo(ctx context.Context, disks []StorageAPI, bucket, prefix
 // readQuorum is the min required disks to read data.
 // writeQuorum is the min required disks to write data.
 func objectQuorumFromMeta(ctx context.Context, partsMetaData []FileInfo, errs []error, defaultParityCount int) (objectReadQuorum, objectWriteQuorum int, err error) {
-	if defaultParityCount == 0 {
-		return 1, 1, nil
-	}
-
 	// get the latest updated Metadata and a count of all the latest updated FileInfo(s)
-	latestFileInfo, err := getLatestFileInfo(ctx, partsMetaData, errs)
+	latestFileInfo, err := getLatestFileInfo(ctx, partsMetaData, defaultParityCount, errs)
 	if err != nil {
 		return 0, 0, err
 	}

 	if latestFileInfo.Deleted {
+		// special case when parity is '0'
+		if defaultParityCount == 0 {
+			return len(partsMetaData), len(partsMetaData), nil
+		}
 		// For delete markers do not use 'defaultParityCount' as it is not expected to be the case.
 		// Use maximum allowed read quorum instead, writeQuorum+1 is returned for compatibility sake
 		// but there are no callers that shall be using this.
@@ -23,7 +23,7 @@ import (
 	"testing"
 	"time"

-	humanize "github.com/dustin/go-humanize"
+	"github.com/dustin/go-humanize"
 )

 const ActualSize = 1000
@@ -58,7 +58,7 @@ func TestAddObjectPart(t *testing.T) {
 	for _, testCase := range testCases {
 		if testCase.expectedIndex > -1 {
 			partNumString := strconv.Itoa(testCase.partNum)
-			fi.AddObjectPart(testCase.partNum, "etag."+partNumString, int64(testCase.partNum+humanize.MiByte), ActualSize)
+			fi.AddObjectPart(testCase.partNum, "etag."+partNumString, int64(testCase.partNum+humanize.MiByte), ActualSize, UTCNow(), nil)
 		}

 		if index := objectPartIndex(fi.Parts, testCase.partNum); index != testCase.expectedIndex {
@@ -91,7 +91,7 @@ func TestObjectPartIndex(t *testing.T) {
 	// Add some parts for testing.
 	for _, testCase := range testCases {
 		partNumString := strconv.Itoa(testCase.partNum)
-		fi.AddObjectPart(testCase.partNum, "etag."+partNumString, int64(testCase.partNum+humanize.MiByte), ActualSize)
+		fi.AddObjectPart(testCase.partNum, "etag."+partNumString, int64(testCase.partNum+humanize.MiByte), ActualSize, UTCNow(), nil)
 	}

 	// Add failure test case.
@@ -121,7 +121,7 @@ func TestObjectToPartOffset(t *testing.T) {
 	// Total size of all parts is 5,242,899 bytes.
 	for _, partNum := range []int{1, 2, 4, 5, 7} {
 		partNumString := strconv.Itoa(partNum)
-		fi.AddObjectPart(partNum, "etag."+partNumString, int64(partNum+humanize.MiByte), ActualSize)
+		fi.AddObjectPart(partNum, "etag."+partNumString, int64(partNum+humanize.MiByte), ActualSize, UTCNow(), nil)
 	}

 	testCases := []struct {
@@ -160,7 +160,7 @@ func TestObjectToPartOffset(t *testing.T) {
 func TestFindFileInfoInQuorum(t *testing.T) {
 	getNFInfo := func(n int, quorum int, t int64, dataDir string) []FileInfo {
 		fi := newFileInfo("test", 8, 8)
-		fi.AddObjectPart(1, "etag", 100, 100)
+		fi.AddObjectPart(1, "etag", 100, 100, UTCNow(), nil)
 		fi.ModTime = time.Unix(t, 0)
 		fi.DataDir = dataDir
 		fis := make([]FileInfo, n)
@@ -19,6 +19,7 @@ package cmd

 import (
 	"context"
+	"errors"
 	"fmt"
 	"io"
 	"os"
@@ -48,7 +49,7 @@ func (er erasureObjects) getMultipartSHADir(bucket, object string) string {
 // checkUploadIDExists - verify if a given uploadID exists and is valid.
 func (er erasureObjects) checkUploadIDExists(ctx context.Context, bucket, object, uploadID string) (err error) {
 	defer func() {
-		if err == errFileNotFound {
+		if errors.Is(err, errFileNotFound) || errors.Is(err, errVolumeNotFound) {
 			err = errUploadIDNotFound
 		}
 	}()
@@ -75,6 +76,30 @@ func (er erasureObjects) checkUploadIDExists(ctx context.Context, bucket, object
 	return err
 }

+// Removes part.meta given by partName belonging to a mulitpart upload from minioMetaBucket
+func (er erasureObjects) removePartMeta(bucket, object, uploadID, dataDir string, partNumber int) {
+	uploadIDPath := er.getUploadIDDir(bucket, object, uploadID)
+	curpartPath := pathJoin(uploadIDPath, dataDir, fmt.Sprintf("part.%d", partNumber))
+	storageDisks := er.getDisks()
+
+	g := errgroup.WithNErrs(len(storageDisks))
+	for index, disk := range storageDisks {
+		if disk == nil {
+			continue
+		}
+		index := index
+		g.Go(func() error {
+			_ = storageDisks[index].Delete(context.TODO(), minioMetaMultipartBucket, curpartPath+".meta", DeleteOptions{
+				Recursive: false,
+				Force:     false,
+			})
+
+			return nil
+		}, index)
+	}
+	g.Wait()
+}
+
 // Removes part given by partName belonging to a mulitpart upload from minioMetaBucket
 func (er erasureObjects) removeObjectPart(bucket, object, uploadID, dataDir string, partNumber int) {
 	uploadIDPath := er.getUploadIDDir(bucket, object, uploadID)
@@ -91,7 +116,15 @@ func (er erasureObjects) removeObjectPart(bucket, object, uploadID, dataDir stri
 			// Ignoring failure to remove parts that weren't present in CompleteMultipartUpload
 			// requests. xl.meta is the authoritative source of truth on which parts constitute
 			// the object. The presence of parts that don't belong in the object doesn't affect correctness.
-			_ = storageDisks[index].Delete(context.TODO(), minioMetaMultipartBucket, curpartPath, false)
+			_ = storageDisks[index].Delete(context.TODO(), minioMetaMultipartBucket, curpartPath, DeleteOptions{
+				Recursive: false,
+				Force:     false,
+			})
+			_ = storageDisks[index].Delete(context.TODO(), minioMetaMultipartBucket, curpartPath+".meta", DeleteOptions{
+				Recursive: false,
+				Force:     false,
+			})
+
 			return nil
 		}, index)
 	}
@@ -138,7 +171,10 @@ func (er erasureObjects) deleteAll(ctx context.Context, bucket, prefix string) {
 		wg.Add(1)
 		go func(disk StorageAPI) {
 			defer wg.Done()
-			disk.Delete(ctx, bucket, prefix, true)
+			disk.Delete(ctx, bucket, prefix, DeleteOptions{
+				Recursive: true,
+				Force:     false,
+			})
 		}(disk)
 	}
 	wg.Wait()
@@ -202,10 +238,10 @@ func (er erasureObjects) ListMultipartUploads(ctx context.Context, bucket, objec
 	for _, disk = range er.getLoadBalancedDisks(true) {
 		uploadIDs, err = disk.ListDir(ctx, minioMetaMultipartBucket, er.getMultipartSHADir(bucket, object), -1)
 		if err != nil {
-			if err == errDiskNotFound {
+			if errors.Is(err, errDiskNotFound) {
 				continue
 			}
-			if err == errFileNotFound {
+			if errors.Is(err, errFileNotFound) || errors.Is(err, errVolumeNotFound) {
 				return result, nil
 			}
 			logger.LogIf(ctx, err)
@@ -230,7 +266,11 @@ func (er erasureObjects) ListMultipartUploads(ctx context.Context, bucket, objec
 		}
 		fi, err := disk.ReadVersion(ctx, minioMetaMultipartBucket, pathJoin(er.getUploadIDDir(bucket, object, uploadID)), "", false)
 		if err != nil {
-			return result, toObjectErr(err, bucket, object)
+			if !IsErrIgnored(err, errFileNotFound, errDiskNotFound) {
+				logger.LogIf(ctx, err)
+			}
+			// Ignore this invalid upload-id since we are listing here
+			continue
 		}
 		populatedUploadIds.Add(uploadID)
 		uploads = append(uploads, MultipartInfo{
@@ -442,6 +482,47 @@ func renamePart(ctx context.Context, disks []StorageAPI, srcBucket, srcEntry, ds
 	return evalDisks(disks, errs), err
 }

+// writeAllDisks - writes 'b' to all provided disks.
+// If write cannot reach quorum, the files will be deleted from all disks.
+func writeAllDisks(ctx context.Context, disks []StorageAPI, dstBucket, dstEntry string, b []byte, writeQuorum int) ([]StorageAPI, error) {
+	g := errgroup.WithNErrs(len(disks))
+
+	// Write file to all underlying storage disks.
+	for index := range disks {
+		index := index
+		g.Go(func() error {
+			if disks[index] == nil {
+				return errDiskNotFound
+			}
+			return disks[index].WriteAll(ctx, dstBucket, dstEntry, b)
+		}, index)
+	}
+
+	// Wait for all renames to finish.
+	errs := g.Wait()
+
+	// We can safely allow RenameFile errors up to len(er.getDisks()) - writeQuorum
+	// otherwise return failure. Cleanup successful renames.
+	err := reduceWriteQuorumErrs(ctx, errs, objectOpIgnoredErrs, writeQuorum)
+	if err == errErasureWriteQuorum {
+		// Remove all written
+		g := errgroup.WithNErrs(len(disks))
+		for index := range disks {
+			if disks[index] == nil || errs[index] != nil {
+				continue
+			}
+			index := index
+			g.Go(func() error {
+				return disks[index].Delete(ctx, dstBucket, dstEntry, DeleteOptions{Force: true})
+			}, index)
+		}
+		// Ignore these errors.
+		g.WaitErr()
+	}
+
+	return evalDisks(disks, errs), err
+}
+
 // PutObjectPart - reads incoming stream and internally erasure codes
 // them. This call is similar to single put operation but it is part
 // of the multipart transaction.
@@ -468,11 +549,7 @@ func (er erasureObjects) PutObjectPart(ctx context.Context, bucket, object, uplo
 		return PartInfo{}, err
 	}
 	rctx := rlkctx.Context()
-	defer func() {
-		if uploadIDRLock != nil {
-			uploadIDRLock.RUnlock(rlkctx.Cancel)
-		}
-	}()
+	defer uploadIDRLock.RUnlock(rlkctx.Cancel)

 	data := r.Reader
 	// Validate input data size and it can never be less than zero.
@@ -496,10 +573,6 @@ func (er erasureObjects) PutObjectPart(ctx context.Context, bucket, object, uplo
 	partsMetadata, errs = readAllFileInfo(rctx, storageDisks, minioMetaMultipartBucket,
 		uploadIDPath, "", false)

-	// Unlock upload id locks before, so others can get it.
-	uploadIDRLock.RUnlock(rlkctx.Cancel)
-	uploadIDRLock = nil
-
 	// get Quorum for this object
 	_, writeQuorum, err := objectQuorumFromMeta(pctx, partsMetadata, errs, er.defaultParityCount)
 	if err != nil {
@@ -606,79 +679,50 @@ func (er erasureObjects) PutObjectPart(ctx context.Context, bucket, object, uplo
 		}
 	}

-	// Acquire write lock to update metadata.
-	uploadIDWLock := er.NewNSLock(bucket, pathJoin(object, uploadID))
-	wlkctx, err := uploadIDWLock.GetLock(pctx, globalOperationTimeout)
-	if err != nil {
-		return PartInfo{}, err
-	}
-	wctx := wlkctx.Context()
-	defer uploadIDWLock.Unlock(wlkctx.Cancel)
-
-	// Validates if upload ID exists.
-	if err = er.checkUploadIDExists(wctx, bucket, object, uploadID); err != nil {
-		return pi, toObjectErr(err, bucket, object, uploadID)
-	}
-
 	// Rename temporary part file to its final location.
 	partPath := pathJoin(uploadIDPath, fi.DataDir, partSuffix)
-	onlineDisks, err = renamePart(wctx, onlineDisks, minioMetaTmpBucket, tmpPartPath, minioMetaMultipartBucket, partPath, writeQuorum)
+	onlineDisks, err = renamePart(ctx, onlineDisks, minioMetaTmpBucket, tmpPartPath, minioMetaMultipartBucket, partPath, writeQuorum)
 	if err != nil {
 		return pi, toObjectErr(err, minioMetaMultipartBucket, partPath)
 	}

-	// Read metadata again because it might be updated with parallel upload of another part.
-	partsMetadata, errs = readAllFileInfo(wctx, onlineDisks, minioMetaMultipartBucket, uploadIDPath, "", false)
-	reducedErr = reduceWriteQuorumErrs(wctx, errs, objectOpIgnoredErrs, writeQuorum)
-	if reducedErr == errErasureWriteQuorum {
-		return pi, toObjectErr(reducedErr, bucket, object)
-	}
-
-	// Get current highest version based on re-read partsMetadata.
-	onlineDisks, modTime = listOnlineDisks(onlineDisks, partsMetadata, errs)
-
-	// Pick one from the first valid metadata.
-	fi, err = pickValidFileInfo(wctx, partsMetadata, modTime, writeQuorum)
-	if err != nil {
-		return pi, err
-	}
-
-	// Once part is successfully committed, proceed with updating erasure metadata.
-	fi.ModTime = UTCNow()
-
 	md5hex := r.MD5CurrentHexString()
-
-	// Add the current part.
-	fi.AddObjectPart(partID, md5hex, n, data.ActualSize())
-
-	for i, disk := range onlineDisks {
-		if disk == OfflineDisk {
-			continue
-		}
-		partsMetadata[i].Size = fi.Size
-		partsMetadata[i].ModTime = fi.ModTime
-		partsMetadata[i].Parts = fi.Parts
-		partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{
-			PartNumber: partID,
-			Algorithm:  DefaultBitrotAlgorithm,
-			Hash:       bitrotWriterSum(writers[i]),
-		})
+	if opts.PreserveETag != "" {
+		md5hex = opts.PreserveETag
 	}

-	// Writes update `xl.meta` format for each disk.
-	if _, err = writeUniqueFileInfo(wctx, onlineDisks, minioMetaMultipartBucket, uploadIDPath, partsMetadata, writeQuorum); err != nil {
-		return pi, toObjectErr(err, minioMetaMultipartBucket, uploadIDPath)
+	var index []byte
+	if opts.IndexCB != nil {
+		index = opts.IndexCB()
 	}

-	online = countOnlineDisks(onlineDisks)
+	part := ObjectPartInfo{
+		Number:     partID,
+		ETag:       md5hex,
+		Size:       n,
+		ActualSize: data.ActualSize(),
+		ModTime:    UTCNow(),
+		Index:      index,
+	}
+
+	partMsg, err := part.MarshalMsg(nil)
+	if err != nil {
+		return pi, toObjectErr(err, minioMetaMultipartBucket, partPath)
+	}
+
+	// Write part metadata to all disks.
+	onlineDisks, err = writeAllDisks(ctx, onlineDisks, minioMetaMultipartBucket, partPath+".meta", partMsg, writeQuorum)
+	if err != nil {
+		return pi, toObjectErr(err, minioMetaMultipartBucket, partPath)
+	}

 	// Return success.
 	return PartInfo{
-		PartNumber:   partID,
-		ETag:         md5hex,
-		LastModified: fi.ModTime,
-		Size:         n,
-		ActualSize:   data.ActualSize(),
+		PartNumber:   part.Number,
+		ETag:         part.ETag,
+		LastModified: part.ModTime,
+		Size:         part.Size,
+		ActualSize:   part.ActualSize,
 	}, nil
 }

@@ -686,6 +730,7 @@ func (er erasureObjects) PutObjectPart(ctx context.Context, bucket, object, uplo
 // by callers to verify object states
 // - encrypted
 // - compressed
+// Does not contain currently uploaded parts by design.
 func (er erasureObjects) GetMultipartInfo(ctx context.Context, bucket, object, uploadID string, opts ObjectOptions) (MultipartInfo, error) {
 	auditObjectErasureSet(ctx, object, &er)

@@ -777,7 +822,7 @@ func (er erasureObjects) ListObjectParts(ctx context.Context, bucket, object, up
 		return result, toObjectErr(reducedErr, minioMetaMultipartBucket, uploadIDPath)
 	}

-	_, modTime := listOnlineDisks(storageDisks, partsMetadata, errs)
+	onlineDisks, modTime := listOnlineDisks(storageDisks, partsMetadata, errs)

 	// Pick one from the first valid metadata.
 	fi, err := pickValidFileInfo(ctx, partsMetadata, modTime, writeQuorum)
@@ -785,6 +830,41 @@ func (er erasureObjects) ListObjectParts(ctx context.Context, bucket, object, up
 		return result, err
 	}

+	if maxParts == 0 {
+		return result, nil
+	}
+
+	if partNumberMarker < 0 {
+		partNumberMarker = 0
+	}
+
+	// Limit output to maxPartsList.
+	if maxParts > maxPartsList-partNumberMarker {
+		maxParts = maxPartsList - partNumberMarker
+	}
+
+	// Read Part info for all parts
+	partPath := pathJoin(uploadIDPath, fi.DataDir) + "/"
+	req := ReadMultipleReq{
+		Bucket:     minioMetaMultipartBucket,
+		Prefix:     partPath,
+		MaxSize:    1 << 20, // Each part should realistically not be > 1MiB.
+		MaxResults: maxParts + 1,
+	}
+
+	start := partNumberMarker + 1
+	end := start + maxParts
+
+	// Parts are 1 based, so index 0 is part one, etc.
+	for i := start; i <= end; i++ {
+		req.Files = append(req.Files, fmt.Sprintf("part.%d.meta", i))
+	}
+
+	partInfoFiles, err := readMultipleFiles(ctx, onlineDisks, req, writeQuorum)
+	if err != nil {
+		return result, err
+	}
+
 	// Populate the result stub.
 	result.Bucket = bucket
 	result.Object = object
@@ -794,34 +874,49 @@ func (er erasureObjects) ListObjectParts(ctx context.Context, bucket, object, up
 	result.UserDefined = cloneMSS(fi.Metadata)

 	// For empty number of parts or maxParts as zero, return right here.
-	if len(fi.Parts) == 0 || maxParts == 0 {
+	if len(partInfoFiles) == 0 || maxParts == 0 {
 		return result, nil
 	}

-	// Limit output to maxPartsList.
-	if maxParts > maxPartsList {
-		maxParts = maxPartsList
+	var partI ObjectPartInfo
+	for i, part := range partInfoFiles {
+		partN := i + partNumberMarker + 1
+		if part.Error != "" || !part.Exists {
+			continue
+		}
+
+		_, err := partI.UnmarshalMsg(part.Data)
+		if err != nil {
+			// Maybe crash or similar.
+			logger.LogIf(ctx, err)
+			continue
+		}
+
+		if partN != partI.Number {
+			logger.LogIf(ctx, fmt.Errorf("part.%d.meta has incorrect corresponding part number: expected %d, got %d", i+1, i+1, partI.Number))
+			continue
+		}
+
+		// Add the current part.
+		fi.AddObjectPart(partI.Number, partI.ETag, partI.Size, partI.ActualSize, partI.ModTime, partI.Index)
 	}

 	// Only parts with higher part numbers will be listed.
-	partIdx := objectPartIndex(fi.Parts, partNumberMarker)
 	parts := fi.Parts
-	if partIdx != -1 {
-		parts = fi.Parts[partIdx+1:]
-	}
-	count := maxParts
+	result.Parts = make([]PartInfo, 0, len(parts))
 	for _, part := range parts {
 		result.Parts = append(result.Parts, PartInfo{
 			PartNumber:   part.Number,
 			ETag:         part.ETag,
-			LastModified: fi.ModTime,
+			LastModified: part.ModTime,
+			ActualSize:   part.ActualSize,
 			Size:         part.Size,
 		})
-		count--
-		if count == 0 {
+		if len(result.Parts) >= maxParts {
 			break
 		}
 	}
+
 	// If listed entries are more than maxParts, we set IsTruncated as true.
 	if len(parts) > len(result.Parts) {
 		result.IsTruncated = true
@@ -842,17 +937,17 @@ func (er erasureObjects) ListObjectParts(ctx context.Context, bucket, object, up
 func (er erasureObjects) CompleteMultipartUpload(ctx context.Context, bucket string, object string, uploadID string, parts []CompletePart, opts ObjectOptions) (oi ObjectInfo, err error) {
 	auditObjectErasureSet(ctx, object, &er)

-	// Hold read-locks to verify uploaded parts, also disallows
-	// parallel part uploads as well.
+	// Hold write locks to verify uploaded parts, also disallows any
+	// parallel PutObjectPart() requests.
 	uploadIDLock := er.NewNSLock(bucket, pathJoin(object, uploadID))
-	rlkctx, err := uploadIDLock.GetRLock(ctx, globalOperationTimeout)
+	wlkctx, err := uploadIDLock.GetLock(ctx, globalOperationTimeout)
 	if err != nil {
 		return oi, err
 	}
-	rctx := rlkctx.Context()
-	defer uploadIDLock.RUnlock(rlkctx.Cancel)
+	wctx := wlkctx.Context()
+	defer uploadIDLock.Unlock(wlkctx.Cancel)

-	if err = er.checkUploadIDExists(rctx, bucket, object, uploadID); err != nil {
+	if err = er.checkUploadIDExists(wctx, bucket, object, uploadID); err != nil {
 		return oi, toObjectErr(err, bucket, object, uploadID)
 	}

@@ -861,15 +956,15 @@ func (er erasureObjects) CompleteMultipartUpload(ctx context.Context, bucket str
 	storageDisks := er.getDisks()

 	// Read metadata associated with the object from all disks.
-	partsMetadata, errs := readAllFileInfo(rctx, storageDisks, minioMetaMultipartBucket, uploadIDPath, "", false)
+	partsMetadata, errs := readAllFileInfo(wctx, storageDisks, minioMetaMultipartBucket, uploadIDPath, "", false)

 	// get Quorum for this object
-	_, writeQuorum, err := objectQuorumFromMeta(rctx, partsMetadata, errs, er.defaultParityCount)
+	_, writeQuorum, err := objectQuorumFromMeta(wctx, partsMetadata, errs, er.defaultParityCount)
 	if err != nil {
 		return oi, toObjectErr(err, bucket, object)
 	}

-	reducedErr := reduceWriteQuorumErrs(rctx, errs, objectOpIgnoredErrs, writeQuorum)
+	reducedErr := reduceWriteQuorumErrs(wctx, errs, objectOpIgnoredErrs, writeQuorum)
 	if reducedErr == errErasureWriteQuorum {
 		return oi, toObjectErr(reducedErr, bucket, object)
 	}
@@ -877,11 +972,63 @@ func (er erasureObjects) CompleteMultipartUpload(ctx context.Context, bucket str
 	onlineDisks, modTime := listOnlineDisks(storageDisks, partsMetadata, errs)

 	// Pick one from the first valid metadata.
-	fi, err := pickValidFileInfo(rctx, partsMetadata, modTime, writeQuorum)
+	fi, err := pickValidFileInfo(wctx, partsMetadata, modTime, writeQuorum)
 	if err != nil {
 		return oi, err
 	}

+	// Read Part info for all parts
+	partPath := pathJoin(uploadIDPath, fi.DataDir) + "/"
+	req := ReadMultipleReq{
+		Bucket:     minioMetaMultipartBucket,
+		Prefix:     partPath,
+		MaxSize:    1 << 20, // Each part should realistically not be > 1MiB.
+		Files:      make([]string, 0, len(parts)),
+		AbortOn404: true,
+	}
+	for _, part := range parts {
+		req.Files = append(req.Files, fmt.Sprintf("part.%d.meta", part.PartNumber))
+	}
+	partInfoFiles, err := readMultipleFiles(ctx, onlineDisks, req, writeQuorum)
+	if err != nil {
+		return oi, err
+	}
+	if len(partInfoFiles) != len(parts) {
+		// Should only happen through internal error
+		err := fmt.Errorf("unexpected part result count: %d, want %d", len(partInfoFiles), len(parts))
+		logger.LogIf(ctx, err)
+		return oi, toObjectErr(err, bucket, object)
+	}
+
+	var partI ObjectPartInfo
+	for i, part := range partInfoFiles {
+		partID := parts[i].PartNumber
+		if part.Error != "" || !part.Exists {
+			return oi, InvalidPart{
+				PartNumber: partID,
+			}
+		}
+
+		_, err := partI.UnmarshalMsg(part.Data)
+		if err != nil {
+			// Maybe crash or similar.
+			logger.LogIf(ctx, err)
+			return oi, InvalidPart{
+				PartNumber: partID,
+			}
+		}
+
+		if partID != partI.Number {
+			logger.LogIf(ctx, fmt.Errorf("part.%d.meta has incorrect corresponding part number: expected %d, got %d", partID, partID, partI.Number))
+			return oi, InvalidPart{
+				PartNumber: partID,
+			}
+		}
+
+		// Add the current part.
+		fi.AddObjectPart(partI.Number, partI.ETag, partI.Size, partI.ActualSize, partI.ModTime, partI.Index)
+	}
+
 	// Calculate full object size.
 	var objectSize int64

@@ -921,7 +1068,7 @@ func (er erasureObjects) CompleteMultipartUpload(ctx context.Context, bucket str
 			return oi, invp
 		}

-		// All parts except the last part has to be atleast 5MB.
+		// All parts except the last part has to be at least 5MB.
 		if (i < len(parts)-1) && !isMinAllowedPartSize(currentFI.Parts[partIdx].ActualSize) {
 			return oi, PartTooSmall{
 				PartNumber: part.PartNumber,
@@ -941,6 +1088,8 @@ func (er erasureObjects) CompleteMultipartUpload(ctx context.Context, bucket str
 			Number:     part.PartNumber,
 			Size:       currentFI.Parts[partIdx].Size,
 			ActualSize: currentFI.Parts[partIdx].ActualSize,
+			ModTime:    currentFI.Parts[partIdx].ModTime,
+			Index:      currentFI.Parts[partIdx].Index,
 		}
 	}

@@ -999,12 +1148,19 @@ func (er erasureObjects) CompleteMultipartUpload(ctx context.Context, bucket str
 		}
 	}

+	// Remove part.meta which is not needed anymore.
+	for _, part := range fi.Parts {
+		er.removePartMeta(bucket, object, uploadID, fi.DataDir, part.Number)
+	}
+
 	// Rename the multipart object to final location.
 	if onlineDisks, err = renameData(ctx, onlineDisks, minioMetaMultipartBucket, uploadIDPath,
 		partsMetadata, bucket, object, writeQuorum); err != nil {
 		return oi, toObjectErr(err, bucket, object)
 	}

+	defer NSUpdated(bucket, object)
+
 	// Check if there is any offline disk and add it to the MRF list
 	for _, disk := range onlineDisks {
 		if disk != nil && disk.IsOnline() {
@@ -448,15 +448,26 @@ func (er erasureObjects) deleteIfDangling(ctx context.Context, bucket, object st
 		}
 		defer NSUpdated(bucket, object)

-		if opts.VersionID != "" {
-			er.deleteObjectVersion(ctx, bucket, object, 1, FileInfo{
-				VersionID: opts.VersionID,
-			}, false)
-		} else {
-			er.deleteObjectVersion(ctx, bucket, object, 1, FileInfo{
-				VersionID: m.VersionID,
-			}, false)
+		fi := FileInfo{
+			VersionID: m.VersionID,
 		}
+		if opts.VersionID != "" {
+			fi.VersionID = opts.VersionID
+		}
+
+		disks := er.getDisks()
+		g := errgroup.WithNErrs(len(disks))
+		for index := range disks {
+			index := index
+			g.Go(func() error {
+				if disks[index] == nil {
+					return errDiskNotFound
+				}
+				return disks[index].DeleteVersion(ctx, bucket, object, fi, false)
+			}, index)
+		}
+
+		g.Wait()
 	}
 	return m, err
 }
@@ -534,6 +545,11 @@ func readAllXL(ctx context.Context, disks []StorageAPI, bucket, object string, r
 		return metaFileInfos, errs
 	}

+	versionID := lfi.VersionID
+	if versionID == "" {
+		versionID = nullVersionID
+	}
+
 	for index := range metadataArray {
 		if metadataArray[index] == nil {
 			continue
@@ -541,21 +557,15 @@ func readAllXL(ctx context.Context, disks []StorageAPI, bucket, object string, r

 		// make sure to preserve this for diskmtime based healing bugfix.
 		diskMTime := metaFileInfos[index].DiskMTime
-		metaFileInfos[index], errs[index] = metadataArray[index].ToFileInfo(bucket, object, "")
+		metaFileInfos[index], errs[index] = metadataArray[index].ToFileInfo(bucket, object, versionID)
 		if errs[index] != nil {
 			continue
 		}

-		if metaFileInfos[index].IsValid() && metaFileInfos[index].ModTime.Equal(lfi.ModTime) {
-			versionID := metaFileInfos[index].VersionID
-			if versionID == "" {
-				versionID = nullVersionID
-			}
-			if readData {
-				metaFileInfos[index].Data = metadataArray[index].data.find(versionID)
-			}
-			metaFileInfos[index].DiskMTime = diskMTime
+		if readData {
+			metaFileInfos[index].Data = metadataArray[index].data.find(versionID)
 		}
+		metaFileInfos[index].DiskMTime = diskMTime
 	}

 	// Return all the metadata.
@@ -687,8 +697,6 @@ func (er erasureObjects) getObjectInfoAndQuorum(ctx context.Context, bucket, obj

 // Similar to rename but renames data from srcEntry to dstEntry at dataDir
 func renameData(ctx context.Context, disks []StorageAPI, srcBucket, srcEntry string, metadata []FileInfo, dstBucket, dstEntry string, writeQuorum int) ([]StorageAPI, error) {
-	defer NSUpdated(dstBucket, dstEntry)
-
 	g := errgroup.WithNErrs(len(disks))

 	fvID := mustGetUUID()
@@ -815,6 +823,12 @@ func (er erasureObjects) putMetacacheObject(ctx context.Context, key string, r *
 	if n < data.Size() {
 		return ObjectInfo{}, IncompleteBody{Bucket: minioMetaBucket, Object: key}
 	}
+	var index []byte
+	if opts.IndexCB != nil {
+		index = opts.IndexCB()
+	}
+
+	modTime := UTCNow()

 	for i, w := range writers {
 		if w == nil {
@@ -823,7 +837,7 @@ func (er erasureObjects) putMetacacheObject(ctx context.Context, key string, r *
 			continue
 		}
 		partsMetadata[i].Data = inlineBuffers[i].Bytes()
-		partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize())
+		partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize(), modTime, index)
 		partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{
 			PartNumber: 1,
 			Algorithm:  DefaultBitrotAlgorithm,
@@ -831,8 +845,6 @@ func (er erasureObjects) putMetacacheObject(ctx context.Context, key string, r *
 		})
 	}

-	modTime := UTCNow()
-
 	// Fill all the necessary metadata.
 	// Update `xl.meta` content on each disks.
 	for index := range partsMetadata {
@@ -972,14 +984,7 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 	switch size := data.Size(); {
 	case size == 0:
 		buffer = make([]byte, 1) // Allocate atleast a byte to reach EOF
-	case size == -1:
-		if size := data.ActualSize(); size > 0 && size < fi.Erasure.BlockSize {
-			buffer = make([]byte, data.ActualSize()+256, data.ActualSize()*2+512)
-		} else {
-			buffer = er.bp.Get()
-			defer er.bp.Put(buffer)
-		}
-	case size >= fi.Erasure.BlockSize:
+	case size >= fi.Erasure.BlockSize || size == -1:
 		buffer = er.bp.Get()
 		defer er.bp.Put(buffer)
 	case size < fi.Erasure.BlockSize:
@@ -1071,6 +1076,10 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 		return ObjectInfo{}, IncompleteBody{Bucket: bucket, Object: object}
 	}

+	var compIndex []byte
+	if opts.IndexCB != nil {
+		compIndex = opts.IndexCB()
+	}
 	if !opts.NoLock {
 		lk := er.NewNSLock(bucket, object)
 		lkctx, err := lk.GetLock(ctx, globalOperationTimeout)
@@ -1081,6 +1090,11 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 		defer lk.Unlock(lkctx.Cancel)
 	}

+	modTime := opts.MTime
+	if opts.MTime.IsZero() {
+		modTime = UTCNow()
+	}
+
 	for i, w := range writers {
 		if w == nil {
 			onlineDisks[i] = nil
@@ -1091,15 +1105,19 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 		} else {
 			partsMetadata[i].Data = nil
 		}
-		partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize())
+		partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize(), modTime, compIndex)
 		partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{
 			PartNumber: 1,
 			Algorithm:  DefaultBitrotAlgorithm,
 			Hash:       bitrotWriterSum(w),
 		})
 	}
+
 	if userDefined["etag"] == "" {
 		userDefined["etag"] = r.MD5CurrentHexString()
+		if opts.PreserveETag != "" {
+			userDefined["etag"] = opts.PreserveETag
+		}
 	}

 	// Guess content-type from the extension if possible.
@@ -1107,11 +1125,6 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 		userDefined["content-type"] = mimedb.TypeByExtension(path.Ext(object))
 	}

-	modTime := opts.MTime
-	if opts.MTime.IsZero() {
-		modTime = UTCNow()
-	}
-
 	// Fill all the necessary metadata.
 	// Update `xl.meta` content on each disks.
 	for index := range partsMetadata {
@@ -1136,6 +1149,8 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 		return ObjectInfo{}, toObjectErr(err, bucket, object)
 	}

+	defer NSUpdated(bucket, object)
+
 	for i := 0; i < len(onlineDisks); i++ {
 		if onlineDisks[i] != nil && onlineDisks[i].IsOnline() {
 			// Object info is the same in all disks, so we can pick
@@ -1168,8 +1183,16 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 	return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
 }

-func (er erasureObjects) deleteObjectVersion(ctx context.Context, bucket, object string, writeQuorum int, fi FileInfo, forceDelMarker bool) error {
+func (er erasureObjects) deleteObjectVersion(ctx context.Context, bucket, object string, fi FileInfo, forceDelMarker bool) error {
 	disks := er.getDisks()
+	// Assume (N/2 + 1) quorum for Delete()
+	// this is a theoretical assumption such that
+	// for delete's we do not need to honor storage
+	// class for objects that have reduced quorum
+	// due to storage class - this only needs to be honored
+	// for Read() requests alone that we already do.
+	writeQuorum := len(disks)/2 + 1
+
 	g := errgroup.WithNErrs(len(disks))
 	for index := range disks {
 		index := index
@@ -1367,8 +1390,14 @@ func (er erasureObjects) deletePrefix(ctx context.Context, bucket, prefix string
 			// Deletes
 			// - The prefix and its children
 			// - The prefix__XLDIR__
-			defer disks[index].Delete(ctx, bucket, dirPrefix, true)
-			return disks[index].Delete(ctx, bucket, prefix, true)
+			defer disks[index].Delete(ctx, bucket, dirPrefix, DeleteOptions{
+				Recursive: true,
+				Force:     true,
+			})
+			return disks[index].Delete(ctx, bucket, prefix, DeleteOptions{
+				Recursive: true,
+				Force:     true,
+			})
 		}, index)
 	}
 	for _, err := range g.Wait() {
@@ -1413,69 +1442,49 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		}
 	}

-	// Acquire a write lock before deleting the object.
-	lk := er.NewNSLock(bucket, object)
-	lkctx, err := lk.GetLock(ctx, globalDeleteOperationTimeout)
-	if err != nil {
-		return ObjectInfo{}, err
-	}
-	ctx = lkctx.Context()
-	defer lk.Unlock(lkctx.Cancel)
-
-	versionFound := true
 	objInfo = ObjectInfo{VersionID: opts.VersionID} // version id needed in Delete API response.
-	goi, writeQuorum, gerr := er.getObjectInfoAndQuorum(ctx, bucket, object, opts)
-	if gerr != nil && goi.Name == "" {
-		switch gerr.(type) {
-		case InsufficientReadQuorum:
-			return objInfo, InsufficientWriteQuorum{}
-		}
-		// For delete marker replication, versionID being replicated will not exist on disk
-		if opts.DeleteMarker {
-			versionFound = false
-		} else {
-			return objInfo, gerr
-		}
-	}
-
-	if opts.Expiration.Expire {
-		action := evalActionFromLifecycle(ctx, *lc, rcfg, goi, false)
-		var isErr bool
-		switch action {
-		case lifecycle.NoneAction:
-			isErr = true
-		case lifecycle.TransitionAction, lifecycle.TransitionVersionAction:
-			isErr = true
-		}
-		if isErr {
-			if goi.VersionID != "" {
-				return goi, VersionNotFound{
-					Bucket:    bucket,
-					Object:    object,
-					VersionID: goi.VersionID,
-				}
-			}
-			return goi, ObjectNotFound{
-				Bucket: bucket,
-				Object: object,
-			}
-		}
-	}
-
 	defer NSUpdated(bucket, object)

 	storageDisks := er.getDisks()

-	var markDelete bool
-	// Determine whether to mark object deleted for replication
-	if goi.VersionID != "" {
-		markDelete = true
+	if opts.Expiration.Expire {
+		goi, _, err := er.getObjectInfoAndQuorum(ctx, bucket, object, opts)
+		if err == nil {
+			action := evalActionFromLifecycle(ctx, *lc, rcfg, goi, false)
+			var isErr bool
+			switch action {
+			case lifecycle.NoneAction:
+				isErr = true
+			case lifecycle.TransitionAction, lifecycle.TransitionVersionAction:
+				isErr = true
+			}
+			if isErr {
+				if goi.VersionID != "" {
+					return goi, VersionNotFound{
+						Bucket:    bucket,
+						Object:    object,
+						VersionID: goi.VersionID,
+					}
+				}
+				return goi, ObjectNotFound{
+					Bucket: bucket,
+					Object: object,
+				}
+			}
+		}
 	}

-	// Default deleteMarker to true if object is under versioning
-	deleteMarker := opts.Versioned
+	versionFound := !(opts.DeleteMarker && opts.VersionID != "")

-	if opts.VersionID != "" {
+	// Determine whether to mark object deleted for replication
+	markDelete := !opts.DeleteMarker && opts.VersionID != ""
+
+	// Default deleteMarker to true if object is under versioning
+	// versioning suspended means we add `null` version as
+	// delete marker, if its not decided already.
+	deleteMarker := (opts.Versioned || opts.VersionSuspended) && opts.VersionID == "" || (opts.DeleteMarker && opts.VersionID != "")
+
+	if markDelete {
 		// case where replica version needs to be deleted on target cluster
 		if versionFound && opts.DeleteMarkerReplicationStatus() == replication.Replica {
 			markDelete = false
@@ -1486,22 +1495,6 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		if opts.VersionPurgeStatus() == Complete {
 			markDelete = false
 		}
-
-		// Version is found but we do not wish to create more delete markers
-		// now, since VersionPurgeStatus() is already set, we can let the
-		// lower layers decide this. This fixes a regression that was introduced
-		// in PR #14555 where !VersionPurgeStatus.Empty() is automatically
-		// considered as Delete marker true to avoid listing such objects by
-		// regular ListObjects() calls. However for delete replication this
-		// ends up being a problem because "upon" a successful delete this
-		// ends up creating a new delete marker that is spurious and unnecessary.
-		if versionFound {
-			if !goi.VersionPurgeStatus.Empty() {
-				deleteMarker = false
-			} else if !goi.DeleteMarker { // implies a versioned delete of object
-				deleteMarker = false
-			}
-		}
 	}

 	modTime := opts.MTime
@@ -1509,38 +1502,32 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		modTime = UTCNow()
 	}
 	fvID := mustGetUUID()
-	if markDelete {
-		if opts.Versioned || opts.VersionSuspended {
-			if !deleteMarker {
-				// versioning suspended means we add `null` version as
-				// delete marker, if its not decided already.
-				deleteMarker = opts.VersionSuspended && opts.VersionID == ""
-			}
-			fi := FileInfo{
-				Name:             object,
-				Deleted:          deleteMarker,
-				MarkDeleted:      markDelete,
-				ModTime:          modTime,
-				ReplicationState: opts.DeleteReplication,
-				TransitionStatus: opts.Transition.Status,
-				ExpireRestored:   opts.Transition.ExpireRestored,
-			}
-			fi.SetTierFreeVersionID(fvID)
-			if opts.Versioned {
-				fi.VersionID = mustGetUUID()
-				if opts.VersionID != "" {
-					fi.VersionID = opts.VersionID
-				}
-			}
-			// versioning suspended means we add `null` version as
-			// delete marker. Add delete marker, since we don't have
-			// any version specified explicitly. Or if a particular
-			// version id needs to be replicated.
-			if err = er.deleteObjectVersion(ctx, bucket, object, writeQuorum, fi, opts.DeleteMarker); err != nil {
-				return objInfo, toObjectErr(err, bucket, object)
-			}
-			return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
+
+	if markDelete && (opts.Versioned || opts.VersionSuspended) {
+		fi := FileInfo{
+			Name:             object,
+			Deleted:          true,
+			MarkDeleted:      markDelete,
+			ModTime:          modTime,
+			ReplicationState: opts.DeleteReplication,
+			TransitionStatus: opts.Transition.Status,
+			ExpireRestored:   opts.Transition.ExpireRestored,
 		}
+		fi.SetTierFreeVersionID(fvID)
+		if opts.Versioned {
+			fi.VersionID = mustGetUUID()
+			if opts.VersionID != "" {
+				fi.VersionID = opts.VersionID
+			}
+		}
+		// versioning suspended means we add `null` version as
+		// delete marker. Add delete marker, since we don't have
+		// any version specified explicitly. Or if a particular
+		// version id needs to be replicated.
+		if err = er.deleteObjectVersion(ctx, bucket, object, fi, opts.DeleteMarker); err != nil {
+			return objInfo, toObjectErr(err, bucket, object)
+		}
+		return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
 	}

 	// Delete the object version on all disks.
@@ -1555,7 +1542,7 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		ExpireRestored:   opts.Transition.ExpireRestored,
 	}
 	dfi.SetTierFreeVersionID(fvID)
-	if err = er.deleteObjectVersion(ctx, bucket, object, writeQuorum, dfi, opts.DeleteMarker); err != nil {
+	if err = er.deleteObjectVersion(ctx, bucket, object, dfi, opts.DeleteMarker); err != nil {
 		return objInfo, toObjectErr(err, bucket, object)
 	}

@@ -1567,13 +1554,7 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		break
 	}

-	return ObjectInfo{
-		Bucket:                     bucket,
-		Name:                       object,
-		VersionID:                  opts.VersionID,
-		VersionPurgeStatusInternal: opts.DeleteReplication.VersionPurgeStatusInternal,
-		ReplicationStatusInternal:  opts.DeleteReplication.ReplicationStatusInternal,
-	}, nil
+	return dfi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
 }

 // Send the successful but partial upload/delete, however ignore
@@ -1843,14 +1824,8 @@ func (er erasureObjects) TransitionObject(ctx context.Context, bucket, object st
 	eventName := event.ObjectTransitionComplete

 	storageDisks := er.getDisks()
-	// we now know the number of blocks this object needs for data and parity.
-	// writeQuorum is dataBlocks + 1
-	writeQuorum := fi.Erasure.DataBlocks
-	if fi.Erasure.DataBlocks == fi.Erasure.ParityBlocks {
-		writeQuorum++
-	}

-	if err = er.deleteObjectVersion(ctx, bucket, object, writeQuorum, fi, false); err != nil {
+	if err = er.deleteObjectVersion(ctx, bucket, object, fi, false); err != nil {
 		eventName = event.ObjectTransitionFailed
 	}

@@ -53,7 +53,7 @@ func TestRepeatPutObjectPart(t *testing.T) {
 	defer objLayer.Shutdown(context.Background())
 	defer removeRoots(disks)

-	err = objLayer.MakeBucketWithLocation(ctx, "bucket1", BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, "bucket1", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -101,7 +101,7 @@ func TestErasureDeleteObjectBasic(t *testing.T) {
 	}
 	defer xl.Shutdown(context.Background())

-	err = xl.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = xl.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -155,7 +155,7 @@ func TestDeleteObjectsVersioned(t *testing.T) {
 		{bucketName, "dir/obj1"},
 	}

-	err = obj.MakeBucketWithLocation(ctx, bucketName, BucketOptions{
+	err = obj.MakeBucketWithLocation(ctx, bucketName, MakeBucketOptions{
 		VersioningEnabled: true,
 	})
 	if err != nil {
@@ -244,7 +244,7 @@ func TestErasureDeleteObjectsErasureSet(t *testing.T) {
 		{bucketName, "obj_4"},
 	}

-	err := erasureSets.MakeBucketWithLocation(ctx, bucketName, BucketOptions{})
+	err := erasureSets.MakeBucketWithLocation(ctx, bucketName, MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -305,7 +305,7 @@ func TestErasureDeleteObjectDiskNotFound(t *testing.T) {
 	xl := z.serverPools[0].sets[0]

 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -374,7 +374,7 @@ func TestErasureDeleteObjectDiskNotFoundErasure4(t *testing.T) {
 	xl := z.serverPools[0].sets[0]

 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -434,7 +434,7 @@ func TestErasureDeleteObjectDiskNotFoundErr(t *testing.T) {
 	xl := z.serverPools[0].sets[0]

 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -505,7 +505,7 @@ func TestGetObjectNoQuorum(t *testing.T) {
 	xl := z.serverPools[0].sets[0]

 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -528,7 +528,10 @@ func TestGetObjectNoQuorum(t *testing.T) {
 		files, _ := disk.ListDir(ctx, bucket, object, -1)
 		for _, file := range files {
 			if file != "xl.meta" {
-				disk.Delete(ctx, bucket, pathJoin(object, file), true)
+				disk.Delete(ctx, bucket, pathJoin(object, file), DeleteOptions{
+					Recursive: true,
+					Force:     false,
+				})
 			}
 		}
 	}
@@ -611,7 +614,7 @@ func TestHeadObjectNoQuorum(t *testing.T) {
 	xl := z.serverPools[0].sets[0]

 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -629,7 +632,10 @@ func TestHeadObjectNoQuorum(t *testing.T) {
 		files, _ := disk.ListDir(ctx, bucket, object, -1)
 		for _, file := range files {
 			if file != "xl.meta" {
-				disk.Delete(ctx, bucket, pathJoin(object, file), true)
+				disk.Delete(ctx, bucket, pathJoin(object, file), DeleteOptions{
+					Recursive: true,
+					Force:     false,
+				})
 			}
 		}
 	}
@@ -686,7 +692,7 @@ func TestPutObjectNoQuorum(t *testing.T) {
 	xl := z.serverPools[0].sets[0]

 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -749,7 +755,7 @@ func TestPutObjectNoQuorumSmall(t *testing.T) {
 	xl := z.serverPools[0].sets[0]

 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -818,7 +824,7 @@ func TestPutObjectSmallInlineData(t *testing.T) {
 	object := "object"

 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -886,7 +892,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 	ctx, cancel := context.WithCancel(GlobalContext)
 	defer cancel()

-	err := obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err := obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket %v", err)
 	}
@@ -1067,10 +1073,7 @@ func TestGetObjectInlineNotInline(t *testing.T) {
 	// Create a backend with 4 disks named disk{1...4}, this name convention
 	// because we will unzip some object data from a sample archive.
 	const numDisks = 4
-	path, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
+	path := t.TempDir()

 	var fsDirs []string
 	for i := 1; i <= numDisks; i++ {
@@ -1088,7 +1091,7 @@ func TestGetObjectInlineNotInline(t *testing.T) {
 	defer removeRoots(fsDirs)

 	// Create a testbucket
-	err = objLayer.MakeBucketWithLocation(ctx, "testbucket", BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, "testbucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -1154,7 +1157,7 @@ func TestGetObjectWithOutdatedDisks(t *testing.T) {

 	for i, testCase := range testCases {
 		// Step 1: create a bucket
-		err = z.MakeBucketWithLocation(ctx, testCase.bucket, BucketOptions{VersioningEnabled: testCase.versioned})
+		err = z.MakeBucketWithLocation(ctx, testCase.bucket, MakeBucketOptions{VersioningEnabled: testCase.versioned})
 		if err != nil {
 			t.Fatalf("Test %d: Failed to create a bucket: %v", i+1, err)
 		}
@@ -22,6 +22,7 @@ import (
 	"encoding/binary"
 	"errors"
 	"fmt"
+	"math/rand"
 	"net/http"
 	"sort"
 	"strconv"
@@ -30,6 +31,8 @@ import (
 	"time"

 	"github.com/dustin/go-humanize"
+	"github.com/minio/madmin-go"
+	"github.com/minio/minio/internal/bucket/lifecycle"
 	"github.com/minio/minio/internal/hash"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/console"
@@ -454,7 +457,7 @@ func (p *poolMeta) updateAfter(ctx context.Context, idx int, pools []*erasureSet
 	now := UTCNow()
 	if now.Sub(p.Pools[idx].LastUpdate) >= duration {
 		if serverDebugLog {
-			console.Debugf("decommission: persisting poolMeta on disk: threshold:%s, poolMeta:%#v\n", now.Sub(p.Pools[idx].LastUpdate), p.Pools[idx])
+			console.Debugf("decommission: persisting poolMeta on drive: threshold:%s, poolMeta:%#v\n", now.Sub(p.Pools[idx].LastUpdate), p.Pools[idx])
 		}
 		p.Pools[idx].LastUpdate = now
 		if err := p.save(ctx, pools); err != nil {
@@ -523,14 +526,26 @@ func (z *erasureServerPools) Init(ctx context.Context) error {
 			}
 			if globalEndpoints[idx].Endpoints[0].IsLocal {
 				go func(pool PoolStatus) {
-					switch err := z.Decommission(ctx, pool.ID); err {
-					case nil:
-						// we already started decommission
-					case errDecommissionAlreadyRunning:
-						// A previous decommission running found restart it.
-						z.doDecommissionInRoutine(ctx, idx)
-					default:
-						logger.LogIf(ctx, fmt.Errorf("Unable to resume decommission of pool %v: %w", pool, err))
+					r := rand.New(rand.NewSource(time.Now().UnixNano()))
+					for {
+						if err := z.Decommission(ctx, pool.ID); err != nil {
+							switch err {
+							// we already started decommission
+							case errDecommissionAlreadyRunning:
+								// A previous decommission running found restart it.
+								z.doDecommissionInRoutine(ctx, idx)
+								return
+							default:
+								if configRetriableErrors(err) {
+									logger.LogIf(ctx, fmt.Errorf("Unable to resume decommission of pool %v: %w: retrying..", pool, err))
+									time.Sleep(time.Second + time.Duration(r.Float64()*float64(5*time.Second)))
+									continue
+								}
+								logger.LogIf(ctx, fmt.Errorf("Unable to resume decommission of pool %v: %w", pool, err))
+								return
+							}
+						}
+						break
 					}
 				}(pool)
 			}
@@ -566,6 +581,11 @@ func (z *erasureServerPools) decommissionObject(ctx context.Context, bucket stri
 		auditLogDecom(ctx, "DecomCopyData", objInfo.Bucket, objInfo.Name, objInfo.VersionID, err)
 	}()

+	actualSize, err := objInfo.GetActualSize()
+	if err != nil {
+		return err
+	}
+
 	if objInfo.isMultipart() {
 		uploadID, err := z.NewMultipartUpload(ctx, bucket, objInfo.Name, ObjectOptions{
 			VersionID:   objInfo.VersionID,
@@ -573,21 +593,26 @@ func (z *erasureServerPools) decommissionObject(ctx context.Context, bucket stri
 			UserDefined: objInfo.UserDefined,
 		})
 		if err != nil {
-			return err
+			return fmt.Errorf("decommissionObject: NewMultipartUpload() %w", err)
 		}
 		defer z.AbortMultipartUpload(ctx, bucket, objInfo.Name, uploadID, ObjectOptions{})
 		parts := make([]CompletePart, len(objInfo.Parts))
 		for i, part := range objInfo.Parts {
-			hr, err := hash.NewReader(gr, part.Size, "", "", part.Size)
+			hr, err := hash.NewReader(gr, part.Size, "", "", part.ActualSize)
 			if err != nil {
-				return err
+				return fmt.Errorf("decommissionObject: hash.NewReader() %w", err)
 			}
 			pi, err := z.PutObjectPart(ctx, bucket, objInfo.Name, uploadID,
 				part.Number,
 				NewPutObjReader(hr),
-				ObjectOptions{})
+				ObjectOptions{
+					PreserveETag: part.ETag, // Preserve original ETag to ensure same metadata.
+					IndexCB: func() []byte {
+						return part.Index // Preserve part Index to ensure decompression works.
+					},
+				})
 			if err != nil {
-				return err
+				return fmt.Errorf("decommissionObject: PutObjectPart() %w", err)
 			}
 			parts[i] = CompletePart{
 				ETag:       pi.ETag,
@@ -597,25 +622,37 @@ func (z *erasureServerPools) decommissionObject(ctx context.Context, bucket stri
 		_, err = z.CompleteMultipartUpload(ctx, bucket, objInfo.Name, uploadID, parts, ObjectOptions{
 			MTime: objInfo.ModTime,
 		})
+		if err != nil {
+			err = fmt.Errorf("decommissionObject: CompleteMultipartUpload() %w", err)
+		}
 		return err
 	}
-	hr, err := hash.NewReader(gr, objInfo.Size, "", "", objInfo.Size)
+
+	hr, err := hash.NewReader(gr, objInfo.Size, "", "", actualSize)
 	if err != nil {
-		return err
+		return fmt.Errorf("decommissionObject: hash.NewReader() %w", err)
 	}
 	_, err = z.PutObject(ctx,
 		bucket,
 		objInfo.Name,
 		NewPutObjReader(hr),
 		ObjectOptions{
-			VersionID:   objInfo.VersionID,
-			MTime:       objInfo.ModTime,
-			UserDefined: objInfo.UserDefined,
+			VersionID:    objInfo.VersionID,
+			MTime:        objInfo.ModTime,
+			UserDefined:  objInfo.UserDefined,
+			PreserveETag: objInfo.ETag, // Preserve original ETag to ensure same metadata.
+			IndexCB: func() []byte {
+				return objInfo.Parts[0].Index // Preserve part Index to ensure decompression works.
+			},
 		})
+	if err != nil {
+		err = fmt.Errorf("decommissionObject: PutObject() %w", err)
+	}
 	return err
 }

 // versionsSorter sorts FileInfo slices by version.
+//
 //msgp:ignore versionsSorter
 type versionsSorter []FileInfo

@@ -641,12 +678,39 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 		set := set
 		disks := set.getOnlineDisks()
 		if len(disks) == 0 {
-			logger.LogIf(GlobalContext, fmt.Errorf("no online disks found for set with endpoints %s",
+			logger.LogIf(GlobalContext, fmt.Errorf("no online drives found for set with endpoints %s",
 				set.getEndpoints()))
 			continue
 		}

 		vc, _ := globalBucketVersioningSys.Get(bi.Name)
+
+		// Check if the current bucket has a configured lifecycle policy
+		lc, _ := globalLifecycleSys.Get(bi.Name)
+
+		// Check if bucket is object locked.
+		lr, _ := globalBucketObjectLockSys.Get(bi.Name)
+
+		filterLifecycle := func(bucket, object string, fi FileInfo) bool {
+			if lc == nil {
+				return false
+			}
+			versioned := vc != nil && vc.Versioned(object)
+			objInfo := fi.ToObjectInfo(bucket, object, versioned)
+			action := evalActionFromLifecycle(ctx, *lc, lr, objInfo, false)
+			switch action {
+			case lifecycle.DeleteVersionAction, lifecycle.DeleteAction:
+				globalExpiryState.enqueueByDays(objInfo, false, action == lifecycle.DeleteVersionAction)
+				// Skip this entry.
+				return true
+			case lifecycle.DeleteRestoredAction, lifecycle.DeleteRestoredVersionAction:
+				globalExpiryState.enqueueByDays(objInfo, true, action == lifecycle.DeleteRestoredVersionAction)
+				// Skip this entry.
+				return true
+			}
+			return false
+		}
+
 		decommissionEntry := func(entry metaCacheEntry) {
 			defer func() {
 				<-parallelWorkers
@@ -673,6 +737,13 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 					logger.LogIf(ctx, fmt.Errorf("found %s/%s transitioned object, transitioned object won't be decommissioned", bi.Name, version.Name))
 					continue
 				}
+
+				// Apply lifecycle rules on the objects that are expired.
+				if filterLifecycle(bi.Name, version.Name, version) {
+					logger.LogIf(ctx, fmt.Errorf("found %s/%s (%s) expired object based on ILM rules, skipping and scheduled for deletion", bi.Name, version.Name, version.VersionID))
+					continue
+				}
+
 				// We will skip decommissioning delete markers
 				// with single version, its as good as there
 				// is no data associated with the object.
@@ -680,16 +751,18 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 					logger.LogIf(ctx, fmt.Errorf("found %s/%s delete marked object with no other versions, skipping since there is no content left", bi.Name, version.Name))
 					continue
 				}
+
 				if version.Deleted {
 					_, err := z.DeleteObject(ctx,
 						bi.Name,
 						version.Name,
 						ObjectOptions{
-							Versioned:         vc.PrefixEnabled(version.Name),
-							VersionID:         version.VersionID,
-							MTime:             version.ModTime,
-							DeleteReplication: version.ReplicationState,
-							DeleteMarker:      true, // make sure we create a delete marker
+							Versioned:          vc.PrefixEnabled(version.Name),
+							VersionID:          version.VersionID,
+							MTime:              version.ModTime,
+							DeleteReplication:  version.ReplicationState,
+							DeleteMarker:       true, // make sure we create a delete marker
+							SkipDecommissioned: true, // make sure we skip the decommissioned pool
 						})
 					var failure bool
 					if err != nil {
@@ -716,7 +789,8 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 						http.Header{},
 						noLock, // all mutations are blocked reads are safe without locks.
 						ObjectOptions{
-							VersionID: version.VersionID,
+							VersionID:    version.VersionID,
+							NoDecryption: true,
 						})
 					if isErrObjectNotFound(err) {
 						// object deleted by the application, nothing to do here we move on.
@@ -727,11 +801,14 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 						logger.LogIf(ctx, err)
 						continue
 					}
+					stopFn := globalDecommissionMetrics.log(decomMetricDecommissionObject, idx, bi.Name, version.Name, version.VersionID)
 					if err = z.decommissionObject(ctx, bi.Name, gr); err != nil {
+						stopFn(err)
 						failure = true
 						logger.LogIf(ctx, err)
 						continue
 					}
+					stopFn(nil)
 					failure = false
 					break
 				}
@@ -746,6 +823,7 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool

 			// if all versions were decommissioned, then we can delete the object versions.
 			if decommissionedCount == len(fivs.Versions) {
+				stopFn := globalDecommissionMetrics.log(decomMetricDecommissionRemoveObject, idx, bi.Name, entry.name)
 				_, err := set.DeleteObject(ctx,
 					bi.Name,
 					encodeDirObject(entry.name),
@@ -753,6 +831,7 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 						DeletePrefix: true, // use prefix delete to delete all versions at once.
 					},
 				)
+				stopFn(err)
 				auditLogDecom(ctx, "DecomDeleteObject", bi.Name, entry.name, "", err)
 				if err != nil {
 					logger.LogIf(ctx, err)
@@ -791,7 +870,7 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 					wg.Add(1)
 					go decommissionEntry(entry)
 				},
-				partial: func(entries metaCacheEntries, nAgreed int, errs []error) {
+				partial: func(entries metaCacheEntries, _ []error) {
 					entry, ok := entries.resolve(&resolver)
 					if ok {
 						parallelWorkers <- struct{}{}
@@ -808,6 +887,47 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 	return nil
 }

+//msgp:ignore decomMetrics
+type decomMetrics struct{}
+
+var globalDecommissionMetrics decomMetrics
+
+//msgp:ignore decomMetric
+//go:generate stringer -type=decomMetric -trimprefix=decomMetric $GOFILE
+type decomMetric uint8
+
+const (
+	decomMetricDecommissionBucket decomMetric = iota
+	decomMetricDecommissionObject
+	decomMetricDecommissionRemoveObject
+)
+
+func decomTrace(d decomMetric, poolIdx int, startTime time.Time, duration time.Duration, path string, err error) madmin.TraceInfo {
+	var errStr string
+	if err != nil {
+		errStr = err.Error()
+	}
+	return madmin.TraceInfo{
+		TraceType: madmin.TraceDecommission,
+		Time:      startTime,
+		NodeName:  globalLocalNodeName,
+		FuncName:  fmt.Sprintf("decommission.%s (pool-id=%d)", d.String(), poolIdx),
+		Duration:  duration,
+		Path:      path,
+		Error:     errStr,
+	}
+}
+
+func (m *decomMetrics) log(d decomMetric, poolIdx int, paths ...string) func(err error) {
+	startTime := time.Now()
+	return func(err error) {
+		duration := time.Since(startTime)
+		if globalTrace.NumSubscribers(madmin.TraceDecommission) > 0 {
+			globalTrace.Publish(decomTrace(d, poolIdx, startTime, duration, strings.Join(paths, " "), err))
+		}
+	}
+}
+
 func (z *erasureServerPools) decommissionInBackground(ctx context.Context, idx int) error {
 	pool := z.serverPools[idx]
 	for _, bucket := range z.poolMeta.PendingBuckets(idx) {
@@ -825,9 +945,13 @@ func (z *erasureServerPools) decommissionInBackground(ctx context.Context, idx i
 		if serverDebugLog {
 			console.Debugln("decommission: currently on bucket", bucket.Name)
 		}
+		stopFn := globalDecommissionMetrics.log(decomMetricDecommissionBucket, idx, bucket.Name)
 		if err := z.decommissionPool(ctx, idx, pool, bucket); err != nil {
+			stopFn(err)
 			return err
 		}
+		stopFn(nil)
+
 		z.poolMetaMutex.Lock()
 		z.poolMeta.BucketDone(idx, bucket)
 		z.poolMeta.save(ctx, z.serverPools)
@@ -912,15 +1036,9 @@ func (z *erasureServerPools) getDecommissionPoolSpaceInfo(idx int) (pi poolSpace
 	if idx+1 > len(z.serverPools) {
 		return pi, errInvalidArgument
 	}
+
 	info, _ := z.serverPools[idx].StorageInfo(context.Background())
 	info.Backend = z.BackendInfo()
-	for _, disk := range info.Disks {
-		if disk.Healing {
-			return pi, decomError{
-				Err: fmt.Sprintf("%s drive is healing, decommission will not be started", disk.Endpoint),
-			}
-		}
-	}

 	usableTotal := int64(GetTotalUsableCapacity(info.Disks, info))
 	usableFree := int64(GetTotalUsableCapacityFree(info.Disks, info))
@@ -984,7 +1102,9 @@ func (z *erasureServerPools) DecommissionCancel(ctx context.Context, idx int) (e
 	defer z.poolMetaMutex.Unlock()

 	if z.poolMeta.DecommissionCancel(idx) {
-		defer z.decommissionCancelers[idx]() // cancel any active thread.
+		if fn := z.decommissionCancelers[idx]; fn != nil {
+			defer fn() // cancel any active thread.
+		}
 		if err = z.poolMeta.save(ctx, z.serverPools); err != nil {
 			return err
 		}
@@ -1006,7 +1126,9 @@ func (z *erasureServerPools) DecommissionFailed(ctx context.Context, idx int) (e
 	defer z.poolMetaMutex.Unlock()

 	if z.poolMeta.DecommissionFailed(idx) {
-		defer z.decommissionCancelers[idx]() // cancel any active thread.
+		if fn := z.decommissionCancelers[idx]; fn != nil {
+			defer fn() // cancel any active thread.
+		}
 		if err = z.poolMeta.save(ctx, z.serverPools); err != nil {
 			return err
 		}
@@ -1028,7 +1150,9 @@ func (z *erasureServerPools) CompleteDecommission(ctx context.Context, idx int)
 	defer z.poolMetaMutex.Unlock()

 	if z.poolMeta.DecommissionComplete(idx) {
-		defer z.decommissionCancelers[idx]() // cancel any active thread.
+		if fn := z.decommissionCancelers[idx]; fn != nil {
+			defer fn() // cancel any active thread.
+		}
 		if err = z.poolMeta.save(ctx, z.serverPools); err != nil {
 			return err
 		}
@@ -1046,11 +1170,18 @@ func (z *erasureServerPools) StartDecommission(ctx context.Context, idx int) (er
 		return errInvalidArgument
 	}

-	buckets, err := z.ListBuckets(ctx)
+	buckets, err := z.ListBuckets(ctx, BucketOptions{})
 	if err != nil {
 		return err
 	}

+	// Make sure to heal the buckets to ensure the new
+	// pool has the new buckets, this is to avoid
+	// failures later.
+	for _, bucket := range buckets {
+		z.HealBucket(ctx, bucket.Name, madmin.HealOpts{})
+	}
+
 	decomBuckets := make([]decomBucketInfo, len(buckets))
 	for i := range buckets {
 		decomBuckets[i] = decomBucketInfo{
@@ -1077,7 +1208,7 @@ func (z *erasureServerPools) StartDecommission(ctx context.Context, idx int) (er
 		pathJoin(minioMetaBucket, bucketMetaPrefix),
 	} {
 		var bucketExists BucketExists
-		if err = z.MakeBucketWithLocation(ctx, metaBucket, BucketOptions{}); err != nil {
+		if err = z.MakeBucketWithLocation(ctx, metaBucket, MakeBucketOptions{}); err != nil {
 			if !errors.As(err, &bucketExists) {
 				return err
 			}
@@ -1132,7 +1263,7 @@ func auditLogDecom(ctx context.Context, apiName, bucket, object, versionID strin
 		errStr = err.Error()
 	}
 	auditLogInternal(ctx, bucket, object, AuditLogOptions{
-		Trigger:   "decommissioning",
+		Event:     "decommission",
 		APIName:   apiName,
 		VersionID: versionID,
 		Error:     errStr,
@@ -380,19 +380,15 @@ func (z *erasureServerPools) getServerPoolsAvailableSpace(ctx context.Context, b
 	return serverPools
 }

-// poolObjInfo represents the state of an object per pool
-type poolObjInfo struct {
-	PoolIndex int
-	ObjInfo   ObjectInfo
-	Err       error
+// PoolObjInfo represents the state of current object version per pool
+type PoolObjInfo struct {
+	Index   int
+	ObjInfo ObjectInfo
+	Err     error
 }

-func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, bucket, object string, opts ObjectOptions) (idx int, err error) {
-	if z.SinglePool() {
-		return 0, nil
-	}
-
-	poolObjInfos := make([]poolObjInfo, len(z.serverPools))
+func (z *erasureServerPools) getPoolInfoExistingWithOpts(ctx context.Context, bucket, object string, opts ObjectOptions) (PoolObjInfo, error) {
+	poolObjInfos := make([]PoolObjInfo, len(z.serverPools))
 	poolOpts := make([]ObjectOptions, len(z.serverPools))
 	for i := range z.serverPools {
 		poolOpts[i] = opts
@@ -404,8 +400,8 @@ func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, buc
 		go func(i int, pool *erasureSets, opts ObjectOptions) {
 			defer wg.Done()
 			// remember the pool index, we may sort the slice original index might be lost.
-			pinfo := poolObjInfo{
-				PoolIndex: i,
+			pinfo := PoolObjInfo{
+				Index: i,
 			}
 			// do not remove this check as it can lead to inconsistencies
 			// for all callers of bucket replication.
@@ -427,20 +423,21 @@ func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, buc
 	})

 	for _, pinfo := range poolObjInfos {
-		// skip all objects from suspended pools for mutating calls.
-		if z.IsSuspended(pinfo.PoolIndex) && opts.Mutate {
+		// skip all objects from suspended pools if asked by the
+		// caller.
+		if z.IsSuspended(pinfo.Index) && opts.SkipDecommissioned {
 			continue
 		}

 		if pinfo.Err != nil && !isErrObjectNotFound(pinfo.Err) {
-			return -1, pinfo.Err
+			return pinfo, pinfo.Err
 		}

 		if isErrObjectNotFound(pinfo.Err) {
 			// No object exists or its a delete marker,
 			// check objInfo to confirm.
 			if pinfo.ObjInfo.DeleteMarker && pinfo.ObjInfo.Name != "" {
-				return pinfo.PoolIndex, nil
+				return pinfo, nil
 			}

 			// objInfo is not valid, truly the object doesn't
@@ -448,10 +445,18 @@ func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, buc
 			continue
 		}

-		return pinfo.PoolIndex, nil
+		return pinfo, nil
 	}

-	return -1, toObjectErr(errFileNotFound, bucket, object)
+	return PoolObjInfo{}, toObjectErr(errFileNotFound, bucket, object)
+}
+
+func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, bucket, object string, opts ObjectOptions) (idx int, err error) {
+	pinfo, err := z.getPoolInfoExistingWithOpts(ctx, bucket, object, opts)
+	if err != nil {
+		return -1, err
+	}
+	return pinfo.Index, nil
 }

 // getPoolIdxExistingNoLock returns the (first) found object pool index containing an object.
@@ -461,8 +466,8 @@ func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, buc
 // The check is skipped if there is only one pool, and 0, nil is always returned in that case.
 func (z *erasureServerPools) getPoolIdxExistingNoLock(ctx context.Context, bucket, object string) (idx int, err error) {
 	return z.getPoolIdxExistingWithOpts(ctx, bucket, object, ObjectOptions{
-		NoLock: true,
-		Mutate: true,
+		NoLock:             true,
+		SkipDecommissioned: true,
 	})
 }

@@ -486,7 +491,7 @@ func (z *erasureServerPools) getPoolIdxNoLock(ctx context.Context, bucket, objec
 // if none are found falls back to most available space pool, this function is
 // designed to be only used by PutObject, CopyObject (newObject creation) and NewMultipartUpload.
 func (z *erasureServerPools) getPoolIdx(ctx context.Context, bucket, object string, size int64) (idx int, err error) {
-	idx, err = z.getPoolIdxExistingWithOpts(ctx, bucket, object, ObjectOptions{Mutate: true})
+	idx, err = z.getPoolIdxExistingWithOpts(ctx, bucket, object, ObjectOptions{SkipDecommissioned: true})
 	if err != nil && !isErrObjectNotFound(err) {
 		return idx, err
 	}
@@ -532,9 +537,11 @@ func (z *erasureServerPools) BackendInfo() (b madmin.BackendInfo) {
 	rrSCParity := globalStorageClass.GetParityForSC(storageclass.RRS)

 	// Data blocks can vary per pool, but parity is same.
-	for _, setDriveCount := range z.SetDriveCounts() {
+	for i, setDriveCount := range z.SetDriveCounts() {
 		b.StandardSCData = append(b.StandardSCData, setDriveCount-scParity)
 		b.RRSCData = append(b.RRSCData, setDriveCount-rrSCParity)
+		b.DrivesPerSet = append(b.DrivesPerSet, setDriveCount)
+		b.TotalSets = append(b.TotalSets, z.serverPools[i].setCount)
 	}

 	b.StandardSCParity = scParity
@@ -612,7 +619,7 @@ func (z *erasureServerPools) NSScanner(ctx context.Context, bf *bloomFilter, upd
 	var results []dataUsageCache
 	var firstErr error

-	allBuckets, err := z.ListBuckets(ctx)
+	allBuckets, err := z.ListBuckets(ctx, BucketOptions{})
 	if err != nil {
 		return err
 	}
@@ -717,7 +724,7 @@ func (z *erasureServerPools) NSScanner(ctx context.Context, bf *bloomFilter, upd
 // MakeBucketWithLocation - creates a new bucket across all serverPools simultaneously
 // even if one of the sets fail to create buckets, we proceed all the successful
 // operations.
-func (z *erasureServerPools) MakeBucketWithLocation(ctx context.Context, bucket string, opts BucketOptions) error {
+func (z *erasureServerPools) MakeBucketWithLocation(ctx context.Context, bucket string, opts MakeBucketOptions) error {
 	g := errgroup.WithNErrs(len(z.serverPools))

 	// Lock the bucket name before creating.
@@ -757,6 +764,7 @@ func (z *erasureServerPools) MakeBucketWithLocation(ctx context.Context, bucket

 	// If it doesn't exist we get a new, so ignore errors
 	meta := newBucketMetadata(bucket)
+	meta.SetCreatedAt(opts.CreatedAt)
 	if opts.LockEnabled {
 		meta.VersioningConfigXML = enabledBucketVersioningConfig
 		meta.ObjectLockConfigXML = enabledBucketObjectLockConfig
@@ -976,13 +984,8 @@ func (z *erasureServerPools) PutObject(ctx context.Context, bucket string, objec
 }

 func (z *erasureServerPools) deletePrefix(ctx context.Context, bucket string, prefix string) error {
-	for idx, pool := range z.serverPools {
-		if z.IsSuspended(idx) {
-			logger.LogIf(ctx, fmt.Errorf("pool %d is suspended, all writes are suspended", idx+1))
-			continue
-		}
-		_, err := pool.DeleteObject(ctx, bucket, prefix, ObjectOptions{DeletePrefix: true})
-		if err != nil {
+	for _, pool := range z.serverPools {
+		if _, err := pool.DeleteObject(ctx, bucket, prefix, ObjectOptions{DeletePrefix: true}); err != nil {
 			return err
 		}
 	}
@@ -1000,16 +1003,36 @@ func (z *erasureServerPools) DeleteObject(ctx context.Context, bucket string, ob
 	}

 	object = encodeDirObject(object)
-	if z.SinglePool() {
-		return z.serverPools[0].DeleteObject(ctx, bucket, object, opts)
-	}

-	idx, err := z.getPoolIdxExistingWithOpts(ctx, bucket, object, opts)
+	// Acquire a write lock before deleting the object.
+	lk := z.NewNSLock(bucket, object)
+	lkctx, err := lk.GetLock(ctx, globalDeleteOperationTimeout)
 	if err != nil {
+		return ObjectInfo{}, err
+	}
+	ctx = lkctx.Context()
+	defer lk.Unlock(lkctx.Cancel)
+
+	gopts := opts
+	gopts.NoLock = true
+	pinfo, err := z.getPoolInfoExistingWithOpts(ctx, bucket, object, gopts)
+	if err != nil {
+		switch err.(type) {
+		case InsufficientReadQuorum:
+			return objInfo, InsufficientWriteQuorum{}
+		}
 		return objInfo, err
 	}

-	return z.serverPools[idx].DeleteObject(ctx, bucket, object, opts)
+	// Delete marker already present we are not going to create new delete markers.
+	if pinfo.ObjInfo.DeleteMarker && opts.VersionID == "" {
+		pinfo.ObjInfo.Name = decodeDirObject(object)
+		return pinfo.ObjInfo, nil
+	}
+
+	objInfo, err = z.serverPools[pinfo.Index].DeleteObject(ctx, bucket, object, opts)
+	objInfo.Name = decodeDirObject(object)
+	return objInfo, err
 }

 func (z *erasureServerPools) DeleteObjects(ctx context.Context, bucket string, objects []ObjectToDelete, opts ObjectOptions) ([]DeletedObject, []error) {
@@ -1035,64 +1058,84 @@ func (z *erasureServerPools) DeleteObjects(ctx context.Context, bucket string, o
 	ctx = lkctx.Context()
 	defer multiDeleteLock.Unlock(lkctx.Cancel)

-	if z.SinglePool() {
-		deleteObjects, dErrs := z.serverPools[0].DeleteObjects(ctx, bucket, objects, opts)
-		for i := range deleteObjects {
-			deleteObjects[i].ObjectName = decodeDirObject(deleteObjects[i].ObjectName)
-		}
-		return deleteObjects, dErrs
-	}
-
 	// Fetch location of up to 10 objects concurrently.
 	poolObjIdxMap := map[int][]ObjectToDelete{}
 	origIndexMap := map[int][]int{}

+	// Always perform 1/10th of the number of objects per delete
+	concurrent := len(objects) / 10
+	if concurrent <= 10 {
+		// if we cannot get 1/10th then choose the number of
+		// objects as concurrent.
+		concurrent = len(objects)
+	}
+
 	var mu sync.Mutex
-	eg := errgroup.WithNErrs(len(objects)).WithConcurrency(10)
+	eg := errgroup.WithNErrs(len(objects)).WithConcurrency(concurrent)
 	for j, obj := range objects {
 		j := j
 		obj := obj
 		eg.Go(func() error {
-			idx, err := z.getPoolIdxExistingWithOpts(ctx, bucket, obj.ObjectName, ObjectOptions{
+			pinfo, err := z.getPoolInfoExistingWithOpts(ctx, bucket, obj.ObjectName, ObjectOptions{
 				NoLock: true,
 			})
 			if err != nil {
 				derrs[j] = err
+				dobjects[j] = DeletedObject{
+					ObjectName: obj.ObjectName,
+				}
 				return nil
 			}
+
+			// Delete marker already present we are not going to create new delete markers.
+			if pinfo.ObjInfo.DeleteMarker && obj.VersionID == "" {
+				dobjects[j] = DeletedObject{
+					DeleteMarker:          pinfo.ObjInfo.DeleteMarker,
+					DeleteMarkerVersionID: pinfo.ObjInfo.VersionID,
+					DeleteMarkerMTime:     DeleteMarkerMTime{pinfo.ObjInfo.ModTime},
+					ObjectName:            pinfo.ObjInfo.Name,
+				}
+				return nil
+			}
+
+			idx := pinfo.Index
+
 			mu.Lock()
+			defer mu.Unlock()
+
 			poolObjIdxMap[idx] = append(poolObjIdxMap[idx], obj)
 			origIndexMap[idx] = append(origIndexMap[idx], j)
-			mu.Unlock()
 			return nil
 		}, j)
 	}

 	eg.Wait() // wait to check all the pools.

-	// Delete concurrently in all server pools.
-	var wg sync.WaitGroup
-	wg.Add(len(z.serverPools))
-	for idx, pool := range z.serverPools {
-		go func(idx int, pool *erasureSets) {
-			defer wg.Done()
-			objs := poolObjIdxMap[idx]
-			if len(objs) > 0 {
-				orgIndexes := origIndexMap[idx]
-				deletedObjects, errs := pool.DeleteObjects(ctx, bucket, objs, opts)
-				mu.Lock()
-				for i, derr := range errs {
-					if derr != nil {
-						derrs[orgIndexes[i]] = derr
+	if len(poolObjIdxMap) > 0 {
+		// Delete concurrently in all server pools.
+		var wg sync.WaitGroup
+		wg.Add(len(z.serverPools))
+		for idx, pool := range z.serverPools {
+			go func(idx int, pool *erasureSets) {
+				defer wg.Done()
+				objs := poolObjIdxMap[idx]
+				if len(objs) > 0 {
+					orgIndexes := origIndexMap[idx]
+					deletedObjects, errs := pool.DeleteObjects(ctx, bucket, objs, opts)
+					mu.Lock()
+					for i, derr := range errs {
+						if derr != nil {
+							derrs[orgIndexes[i]] = derr
+						}
+						deletedObjects[i].ObjectName = decodeDirObject(deletedObjects[i].ObjectName)
+						dobjects[orgIndexes[i]] = deletedObjects[i]
 					}
-					deletedObjects[i].ObjectName = decodeDirObject(deletedObjects[i].ObjectName)
-					dobjects[orgIndexes[i]] = deletedObjects[i]
+					mu.Unlock()
 				}
-				mu.Unlock()
-			}
-		}(idx, pool)
+			}(idx, pool)
+		}
+		wg.Wait()
 	}
-	wg.Wait()

 	return dobjects, derrs
 }
@@ -1179,7 +1222,6 @@ func (z *erasureServerPools) ListObjectVersions(ctx context.Context, bucket, pre
 	if marker == "" && versionMarker != "" {
 		return loi, NotImplemented{}
 	}
-
 	opts := listPathOptions{
 		Bucket:      bucket,
 		Prefix:      prefix,
@@ -1190,6 +1232,8 @@ func (z *erasureServerPools) ListObjectVersions(ctx context.Context, bucket, pre
 		AskDisks:    globalAPIConfig.getListQuorum(),
 		Versioned:   true,
 	}
+	// set bucket metadata in opts
+	opts.setBucketMeta(ctx)

 	merged, err := z.listPath(ctx, &opts)
 	if err != nil && err != io.EOF {
@@ -1236,12 +1280,16 @@ func maxKeysPlusOne(maxKeys int, addOne bool) int {

 func (z *erasureServerPools) ListObjects(ctx context.Context, bucket, prefix, marker, delimiter string, maxKeys int) (ListObjectsInfo, error) {
 	var loi ListObjectsInfo
-
-	// Automatically remove the object/version is an expiry lifecycle rule can be applied
-	lc, _ := globalLifecycleSys.Get(bucket)
-
-	// Check if bucket is object locked.
-	rcfg, _ := globalBucketObjectLockSys.Get(bucket)
+	opts := listPathOptions{
+		Bucket:      bucket,
+		Prefix:      prefix,
+		Separator:   delimiter,
+		Limit:       maxKeysPlusOne(maxKeys, marker != ""),
+		Marker:      marker,
+		InclDeleted: false,
+		AskDisks:    globalAPIConfig.getListQuorum(),
+	}
+	opts.setBucketMeta(ctx)

 	if len(prefix) > 0 && maxKeys == 1 && delimiter == "" && marker == "" {
 		// Optimization for certain applications like
@@ -1253,8 +1301,8 @@ func (z *erasureServerPools) ListObjects(ctx context.Context, bucket, prefix, ma
 		// to avoid the need for ListObjects().
 		objInfo, err := z.GetObjectInfo(ctx, bucket, prefix, ObjectOptions{NoLock: true})
 		if err == nil {
-			if lc != nil {
-				action := evalActionFromLifecycle(ctx, *lc, rcfg, objInfo, false)
+			if opts.Lifecycle != nil {
+				action := evalActionFromLifecycle(ctx, *opts.Lifecycle, opts.Retention, objInfo, false)
 				switch action {
 				case lifecycle.DeleteVersionAction, lifecycle.DeleteAction:
 					fallthrough
@@ -1267,18 +1315,6 @@ func (z *erasureServerPools) ListObjects(ctx context.Context, bucket, prefix, ma
 		}
 	}

-	opts := listPathOptions{
-		Bucket:      bucket,
-		Prefix:      prefix,
-		Separator:   delimiter,
-		Limit:       maxKeysPlusOne(maxKeys, marker != ""),
-		Marker:      marker,
-		InclDeleted: false,
-		AskDisks:    globalAPIConfig.getListQuorum(),
-		Lifecycle:   lc,
-		Retention:   rcfg,
-	}
-
 	merged, err := z.listPath(ctx, &opts)
 	if err != nil && err != io.EOF {
 		if !isErrBucketNotFound(err) {
@@ -1325,7 +1361,10 @@ func (z *erasureServerPools) ListMultipartUploads(ctx context.Context, bucket, p
 	poolResult.KeyMarker = keyMarker
 	poolResult.Prefix = prefix
 	poolResult.Delimiter = delimiter
-	for _, pool := range z.serverPools {
+	for idx, pool := range z.serverPools {
+		if z.IsSuspended(idx) {
+			continue
+		}
 		result, err := pool.ListMultipartUploads(ctx, bucket, prefix, keyMarker, uploadIDMarker,
 			delimiter, maxUploads)
 		if err != nil {
@@ -1350,6 +1389,9 @@ func (z *erasureServerPools) NewMultipartUpload(ctx context.Context, bucket, obj
 	}

 	for idx, pool := range z.serverPools {
+		if z.IsSuspended(idx) {
+			continue
+		}
 		result, err := pool.ListMultipartUploads(ctx, bucket, object, "", "", "", maxUploadsList)
 		if err != nil {
 			return "", err
@@ -1392,7 +1434,10 @@ func (z *erasureServerPools) PutObjectPart(ctx context.Context, bucket, object,
 		return z.serverPools[0].PutObjectPart(ctx, bucket, object, uploadID, partID, data, opts)
 	}

-	for _, pool := range z.serverPools {
+	for idx, pool := range z.serverPools {
+		if z.IsSuspended(idx) {
+			continue
+		}
 		_, err := pool.GetMultipartInfo(ctx, bucket, object, uploadID, opts)
 		if err == nil {
 			return pool.PutObjectPart(ctx, bucket, object, uploadID, partID, data, opts)
@@ -1421,7 +1466,10 @@ func (z *erasureServerPools) GetMultipartInfo(ctx context.Context, bucket, objec
 	if z.SinglePool() {
 		return z.serverPools[0].GetMultipartInfo(ctx, bucket, object, uploadID, opts)
 	}
-	for _, pool := range z.serverPools {
+	for idx, pool := range z.serverPools {
+		if z.IsSuspended(idx) {
+			continue
+		}
 		mi, err := pool.GetMultipartInfo(ctx, bucket, object, uploadID, opts)
 		if err == nil {
 			return mi, nil
@@ -1450,7 +1498,10 @@ func (z *erasureServerPools) ListObjectParts(ctx context.Context, bucket, object
 	if z.SinglePool() {
 		return z.serverPools[0].ListObjectParts(ctx, bucket, object, uploadID, partNumberMarker, maxParts, opts)
 	}
-	for _, pool := range z.serverPools {
+	for idx, pool := range z.serverPools {
+		if z.IsSuspended(idx) {
+			continue
+		}
 		_, err := pool.GetMultipartInfo(ctx, bucket, object, uploadID, opts)
 		if err == nil {
 			return pool.ListObjectParts(ctx, bucket, object, uploadID, partNumberMarker, maxParts, opts)
@@ -1478,7 +1529,10 @@ func (z *erasureServerPools) AbortMultipartUpload(ctx context.Context, bucket, o
 		return z.serverPools[0].AbortMultipartUpload(ctx, bucket, object, uploadID, opts)
 	}

-	for _, pool := range z.serverPools {
+	for idx, pool := range z.serverPools {
+		if z.IsSuspended(idx) {
+			continue
+		}
 		_, err := pool.GetMultipartInfo(ctx, bucket, object, uploadID, opts)
 		if err == nil {
 			return pool.AbortMultipartUpload(ctx, bucket, object, uploadID, opts)
@@ -1507,7 +1561,10 @@ func (z *erasureServerPools) CompleteMultipartUpload(ctx context.Context, bucket
 		return z.serverPools[0].CompleteMultipartUpload(ctx, bucket, object, uploadID, uploadedParts, opts)
 	}

-	for _, pool := range z.serverPools {
+	for idx, pool := range z.serverPools {
+		if z.IsSuspended(idx) {
+			continue
+		}
 		_, err := pool.GetMultipartInfo(ctx, bucket, object, uploadID, opts)
 		if err == nil {
 			return pool.CompleteMultipartUpload(ctx, bucket, object, uploadID, uploadedParts, opts)
@@ -1522,9 +1579,9 @@ func (z *erasureServerPools) CompleteMultipartUpload(ctx context.Context, bucket
 }

 // GetBucketInfo - returns bucket info from one of the erasure coded serverPools.
-func (z *erasureServerPools) GetBucketInfo(ctx context.Context, bucket string) (bucketInfo BucketInfo, err error) {
+func (z *erasureServerPools) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (bucketInfo BucketInfo, err error) {
 	if z.SinglePool() {
-		bucketInfo, err = z.serverPools[0].GetBucketInfo(ctx, bucket)
+		bucketInfo, err = z.serverPools[0].GetBucketInfo(ctx, bucket, opts)
 		if err != nil {
 			return bucketInfo, err
 		}
@@ -1535,7 +1592,7 @@ func (z *erasureServerPools) GetBucketInfo(ctx context.Context, bucket string) (
 		return bucketInfo, nil
 	}
 	for _, pool := range z.serverPools {
-		bucketInfo, err = pool.GetBucketInfo(ctx, bucket)
+		bucketInfo, err = pool.GetBucketInfo(ctx, bucket, opts)
 		if err != nil {
 			if isErrBucketNotFound(err) {
 				continue
@@ -1609,7 +1666,11 @@ func (z *erasureServerPools) DeleteBucket(ctx context.Context, bucket string, op

 	// Purge the entire bucket metadata entirely.
 	z.renameAll(context.Background(), minioMetaBucket, pathJoin(bucketMetaPrefix, bucket))
-
+	// If site replication is configured, hold on to deleted bucket state until sites sync
+	switch opts.SRDeleteOp {
+	case MarkDelete:
+		z.markDelete(context.Background(), minioMetaBucket, pathJoin(bucketMetaPrefix, deletedBucketsPrefix, bucket))
+	}
 	// Success.
 	return nil
 }
@@ -1627,6 +1688,27 @@ func (z *erasureServerPools) renameAll(ctx context.Context, bucket, prefix strin
 	}
 }

+// markDelete will create a directory of deleted bucket in .minio.sys/buckets/.deleted across all disks
+// in situations where the deleted bucket needs to be held on to until all sites are in sync for
+// site replication
+func (z *erasureServerPools) markDelete(ctx context.Context, bucket, prefix string) {
+	for _, servers := range z.serverPools {
+		for _, set := range servers.sets {
+			set.markDelete(ctx, bucket, prefix)
+		}
+	}
+}
+
+// purgeDelete deletes vol entry in .minio.sys/buckets/.deleted after site replication
+// syncs the delete to peers.
+func (z *erasureServerPools) purgeDelete(ctx context.Context, bucket, prefix string) {
+	for _, servers := range z.serverPools {
+		for _, set := range servers.sets {
+			set.purgeDelete(ctx, bucket, prefix)
+		}
+	}
+}
+
 // This function is used to undo a successful DeleteBucket operation.
 func undoDeleteBucketServerPools(ctx context.Context, bucket string, serverPools []*erasureSets, errs []error) {
 	g := errgroup.WithNErrs(len(serverPools))
@@ -1636,7 +1718,7 @@ func undoDeleteBucketServerPools(ctx context.Context, bucket string, serverPools
 		index := index
 		g.Go(func() error {
 			if errs[index] == nil {
-				return serverPools[index].MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+				return serverPools[index].MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 			}
 			return nil
 		}, index)
@@ -1648,12 +1730,15 @@ func undoDeleteBucketServerPools(ctx context.Context, bucket string, serverPools
 // List all buckets from one of the serverPools, we are not doing merge
 // sort here just for simplification. As per design it is assumed
 // that all buckets are present on all serverPools.
-func (z *erasureServerPools) ListBuckets(ctx context.Context) (buckets []BucketInfo, err error) {
+func (z *erasureServerPools) ListBuckets(ctx context.Context, opts BucketOptions) (buckets []BucketInfo, err error) {
 	if z.SinglePool() {
-		buckets, err = z.serverPools[0].ListBuckets(ctx)
+		buckets, err = z.serverPools[0].ListBuckets(ctx, opts)
 	} else {
-		for _, pool := range z.serverPools {
-			buckets, err = pool.ListBuckets(ctx)
+		for idx, pool := range z.serverPools {
+			if z.IsSuspended(idx) {
+				continue
+			}
+			buckets, err = pool.ListBuckets(ctx, opts)
 			if err != nil {
 				logger.LogIf(ctx, err)
 				continue
@@ -1665,9 +1750,9 @@ func (z *erasureServerPools) ListBuckets(ctx context.Context) (buckets []BucketI
 		return nil, err
 	}
 	for i := range buckets {
-		meta, err := globalBucketMetadataSys.Get(buckets[i].Name)
+		createdAt, err := globalBucketMetadataSys.CreatedAt(buckets[i].Name)
 		if err == nil {
-			buckets[i].Created = meta.Created
+			buckets[i].Created = createdAt
 		}
 	}
 	return buckets, nil
@@ -1721,7 +1806,9 @@ func (z *erasureServerPools) HealBucket(ctx context.Context, bucket string, opts
 	}

 	// Attempt heal on the bucket metadata, ignore any failures
-	defer z.HealObject(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, bucket, bucketMetadataFile), "", opts)
+	hopts := opts
+	hopts.Recreate = false
+	defer z.HealObject(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, bucket, bucketMetadataFile), "", hopts)

 	for _, pool := range z.serverPools {
 		result, err := pool.HealBucket(ctx, bucket, opts)
@@ -1784,18 +1871,17 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 							cancel()
 							return
 						}
-						if opts.WalkAscending {
-							for i := len(fivs.Versions) - 1; i >= 0; i-- {
-								version := fivs.Versions[i]
-								versioned := vcfg != nil && vcfg.Versioned(version.Name)

-								results <- version.ToObjectInfo(bucket, version.Name, versioned)
-							}
-							return
-						}
+						versionsSorter(fivs.Versions).reverse()
+
 						for _, version := range fivs.Versions {
 							versioned := vcfg != nil && vcfg.Versioned(version.Name)
-							results <- version.ToObjectInfo(bucket, version.Name, versioned)
+
+							select {
+							case <-ctx.Done():
+								return
+							case results <- version.ToObjectInfo(bucket, version.Name, versioned):
+							}
 						}
 					}

@@ -1822,7 +1908,7 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 						minDisks:       1,
 						reportNotFound: false,
 						agreed:         loadEntry,
-						partial: func(entries metaCacheEntries, nAgreed int, errs []error) {
+						partial: func(entries metaCacheEntries, _ []error) {
 							entry, ok := entries.resolve(&resolver)
 							if !ok {
 								// check if we can get one entry atleast
@@ -1837,6 +1923,7 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re

 					if err := listPathRaw(ctx, lopts); err != nil {
 						logger.LogIf(ctx, fmt.Errorf("listPathRaw returned %w: opts(%#v)", err, lopts))
+						cancel()
 						return
 					}
 				}()
@@ -1857,7 +1944,7 @@ func listAndHeal(ctx context.Context, bucket, prefix string, set *erasureObjects

 	disks, _ := set.getOnlineDisksWithHealing()
 	if len(disks) == 0 {
-		return errors.New("listAndHeal: No non-healing disks found")
+		return errors.New("listAndHeal: No non-healing drives found")
 	}

 	// How to resolve partial results.
@@ -1889,7 +1976,7 @@ func listAndHeal(ctx context.Context, bucket, prefix string, set *erasureObjects
 				cancel()
 			}
 		},
-		partial: func(entries metaCacheEntries, nAgreed int, errs []error) {
+		partial: func(entries metaCacheEntries, _ []error) {
 			entry, ok := entries.resolve(&resolver)
 			if !ok {
 				// check if we can get one entry atleast
@@ -2048,7 +2135,7 @@ func (z *erasureServerPools) getPoolAndSet(id string) (poolIdx, setIdx, diskIdx
 			}
 		}
 	}
-	return -1, -1, -1, fmt.Errorf("DiskID(%s) %w", id, errDiskNotFound)
+	return -1, -1, -1, fmt.Errorf("DriveID(%s) %w", id, errDiskNotFound)
 }

 // HealthOptions takes input options to return sepcific information
@@ -2256,7 +2343,6 @@ func (z *erasureServerPools) GetObjectTags(ctx context.Context, bucket, object s
 		return z.serverPools[0].GetObjectTags(ctx, bucket, object, opts)
 	}

-	opts.Mutate = false
 	idx, err := z.getPoolIdxExistingWithOpts(ctx, bucket, object, opts)
 	if err != nil {
 		return nil, err
@@ -2272,7 +2358,8 @@ func (z *erasureServerPools) TransitionObject(ctx context.Context, bucket, objec
 		return z.serverPools[0].TransitionObject(ctx, bucket, object, opts)
 	}

-	opts.Mutate = true // Avoid transitioning an object from a pool being decommissioned.
+	// Avoid transitioning an object from a pool being decommissioned.
+	opts.SkipDecommissioned = true
 	idx, err := z.getPoolIdxExistingWithOpts(ctx, bucket, object, opts)
 	if err != nil {
 		return err
@@ -2288,7 +2375,8 @@ func (z *erasureServerPools) RestoreTransitionedObject(ctx context.Context, buck
 		return z.serverPools[0].RestoreTransitionedObject(ctx, bucket, object, opts)
 	}

-	opts.Mutate = true // Avoid restoring object from a pool being decommissioned.
+	// Avoid restoring object from a pool being decommissioned.
+	opts.SkipDecommissioned = true
 	idx, err := z.getPoolIdxExistingWithOpts(ctx, bucket, object, opts)
 	if err != nil {
 		return err
@@ -20,7 +20,6 @@ package cmd
 import (
 	"context"
 	"encoding/binary"
-	"encoding/json"
 	"errors"
 	"fmt"
 	"hash/crc32"
@@ -28,6 +27,7 @@ import (
 	"net/http"
 	"reflect"
 	"sort"
+	"strings"
 	"sync"
 	"time"

@@ -129,10 +129,10 @@ func connectEndpoint(endpoint Endpoint) (StorageAPI, *formatErasureV3, error) {
 		if errors.Is(err, errUnformattedDisk) {
 			info, derr := disk.DiskInfo(context.TODO())
 			if derr != nil && info.RootDisk {
-				return nil, nil, fmt.Errorf("Disk: %s is a root disk", disk)
+				return nil, nil, fmt.Errorf("Drive: %s is a root drive", disk)
 			}
 		}
-		return nil, nil, fmt.Errorf("Disk: %s returned %w", disk, err) // make sure to '%w' to wrap the error
+		return nil, nil, fmt.Errorf("Drive: %s returned %w", disk, err) // make sure to '%w' to wrap the error
 	}

 	return disk, format, nil
@@ -147,7 +147,7 @@ func findDiskIndexByDiskID(refFormat *formatErasureV3, diskID string) (int, int,
 		return -1, -1, errDiskNotFound
 	}
 	if diskID == offlineDiskUUID {
-		return -1, -1, fmt.Errorf("diskID: %s is offline", diskID)
+		return -1, -1, fmt.Errorf("DriveID: %s is offline", diskID)
 	}
 	for i := 0; i < len(refFormat.Erasure.Sets); i++ {
 		for j := 0; j < len(refFormat.Erasure.Sets[0]); j++ {
@@ -157,7 +157,7 @@ func findDiskIndexByDiskID(refFormat *formatErasureV3, diskID string) (int, int,
 		}
 	}

-	return -1, -1, fmt.Errorf("diskID: %s not found", diskID)
+	return -1, -1, fmt.Errorf("DriveID: %s not found", diskID)
 }

 // findDiskIndex - returns the i,j'th position of the input `format` against the reference
@@ -170,7 +170,7 @@ func findDiskIndex(refFormat, format *formatErasureV3) (int, int, error) {
 	}

 	if format.Erasure.This == offlineDiskUUID {
-		return -1, -1, fmt.Errorf("diskID: %s is offline", format.Erasure.This)
+		return -1, -1, fmt.Errorf("DriveID: %s is offline", format.Erasure.This)
 	}

 	for i := 0; i < len(refFormat.Erasure.Sets); i++ {
@@ -181,7 +181,7 @@ func findDiskIndex(refFormat, format *formatErasureV3) (int, int, error) {
 		}
 	}

-	return -1, -1, fmt.Errorf("diskID: %s not found", format.Erasure.This)
+	return -1, -1, fmt.Errorf("DriveID: %s not found", format.Erasure.This)
 }

 // connectDisks - attempt to connect all the endpoints, loads format
@@ -239,7 +239,7 @@ func (s *erasureSets) connectDisks() {
 			s.erasureDisksMu.Lock()
 			if currentDisk := s.erasureDisks[setIndex][diskIndex]; currentDisk != nil {
 				if !reflect.DeepEqual(currentDisk.Endpoint(), disk.Endpoint()) {
-					err = fmt.Errorf("Detected unexpected disk ordering refusing to use the disk: expecting %s, found %s, refusing to use the disk",
+					err = fmt.Errorf("Detected unexpected drive ordering refusing to use the drive: expecting %s, found %s, refusing to use the drive",
 						currentDisk.Endpoint(), disk.Endpoint())
 					printEndpointError(endpoint, err, false)
 					disk.Close()
@@ -300,7 +300,7 @@ func (s *erasureSets) monitorAndConnectEndpoints(ctx context.Context, monitorInt
 			return
 		case <-monitor.C:
 			if serverDebugLog {
-				console.Debugln("running disk monitoring")
+				console.Debugln("running drive monitoring")
 			}

 			s.connectDisks()
@@ -446,7 +446,7 @@ func newErasureSets(ctx context.Context, endpoints PoolEndpoints, storageDisks [
 						return
 					}
 					if m != i || n != j {
-						logger.LogIf(ctx, fmt.Errorf("Detected unexpected disk ordering refusing to use the disk - poolID: %s, found disk mounted at (set=%s, disk=%s) expected mount at (set=%s, disk=%s): %s(%s)", humanize.Ordinal(poolIdx+1), humanize.Ordinal(m+1), humanize.Ordinal(n+1), humanize.Ordinal(i+1), humanize.Ordinal(j+1), disk, diskID))
+						logger.LogIf(ctx, fmt.Errorf("Detected unexpected drive ordering refusing to use the drive - poolID: %s, found drive mounted at (set=%s, drive=%s) expected mount at (set=%s, drive=%s): %s(%s)", humanize.Ordinal(poolIdx+1), humanize.Ordinal(m+1), humanize.Ordinal(n+1), humanize.Ordinal(i+1), humanize.Ordinal(j+1), disk, diskID))
 						s.erasureDisks[i][j] = &unrecognizedDisk{storage: disk}
 						return
 					}
@@ -549,29 +549,13 @@ func (s *erasureSets) cleanupStaleUploads(ctx context.Context) {
 	}
 }

-const objectErasureMapKey = "objectErasureMap"
-
 type auditObjectOp struct {
+	Name  string   `json:"name"`
 	Pool  int      `json:"poolId"`
 	Set   int      `json:"setId"`
 	Disks []string `json:"disks"`
 }

-type auditObjectErasureMap struct {
-	sync.Map
-}
-
-// Define how to marshal auditObjectErasureMap so it can be
-// printed in the audit webhook notification request.
-func (a *auditObjectErasureMap) MarshalJSON() ([]byte, error) {
-	mapCopy := make(map[string]auditObjectOp)
-	a.Range(func(k, v interface{}) bool {
-		mapCopy[k.(string)] = v.(auditObjectOp)
-		return true
-	})
-	return json.Marshal(mapCopy)
-}
-
 // Add erasure set information to the current context
 func auditObjectErasureSet(ctx context.Context, object string, set *erasureObjects) {
 	if len(logger.AuditTargets()) == 0 {
@@ -579,33 +563,20 @@ func auditObjectErasureSet(ctx context.Context, object string, set *erasureObjec
 	}

 	object = decodeDirObject(object)
-
-	var disksEndpoints []string
-	for _, endpoint := range set.getEndpoints() {
+	endpoints := set.getEndpoints()
+	disksEndpoints := make([]string, 0, len(endpoints))
+	for _, endpoint := range endpoints {
 		disksEndpoints = append(disksEndpoints, endpoint.String())
 	}

 	op := auditObjectOp{
+		Name:  object,
 		Pool:  set.poolIndex + 1,
 		Set:   set.setIndex + 1,
 		Disks: disksEndpoints,
 	}

-	var objectErasureSetTag *auditObjectErasureMap
-	reqInfo := logger.GetReqInfo(ctx)
-	for _, kv := range reqInfo.GetTags() {
-		if kv.Key == objectErasureMapKey {
-			objectErasureSetTag = kv.Val.(*auditObjectErasureMap)
-			break
-		}
-	}
-
-	if objectErasureSetTag == nil {
-		objectErasureSetTag = &auditObjectErasureMap{}
-	}
-
-	objectErasureSetTag.Store(object, op)
-	reqInfo.SetTags(objectErasureMapKey, objectErasureSetTag)
+	logger.GetReqInfo(ctx).AppendTags("objectLocation", op)
 }

 // NewNSLock - initialize a new namespace RWLocker instance.
@@ -719,7 +690,7 @@ func (s *erasureSets) Shutdown(ctx context.Context) error {

 // MakeBucketLocation - creates a new bucket across all sets simultaneously,
 // then return the first encountered error
-func (s *erasureSets) MakeBucketWithLocation(ctx context.Context, bucket string, opts BucketOptions) error {
+func (s *erasureSets) MakeBucketWithLocation(ctx context.Context, bucket string, opts MakeBucketOptions) error {
 	g := errgroup.WithNErrs(len(s.sets))

 	// Create buckets in parallel across all sets.
@@ -790,8 +761,8 @@ func (s *erasureSets) getHashedSet(input string) (set *erasureObjects) {
 }

 // GetBucketInfo - returns bucket info from one of the erasure coded set.
-func (s *erasureSets) GetBucketInfo(ctx context.Context, bucket string) (bucketInfo BucketInfo, err error) {
-	return s.getHashedSet("").GetBucketInfo(ctx, bucket)
+func (s *erasureSets) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (bucketInfo BucketInfo, err error) {
+	return s.getHashedSet("").GetBucketInfo(ctx, bucket, opts)
 }

 // IsNotificationSupported returns whether bucket notification is applicable for this layer.
@@ -855,7 +826,7 @@ func undoDeleteBucketSets(ctx context.Context, bucket string, sets []*erasureObj
 		index := index
 		g.Go(func() error {
 			if errs[index] == nil {
-				return sets[index].MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+				return sets[index].MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 			}
 			return nil
 		}, index)
@@ -867,7 +838,7 @@ func undoDeleteBucketSets(ctx context.Context, bucket string, sets []*erasureObj
 // List all buckets from one of the set, we are not doing merge
 // sort here just for simplification. As per design it is assumed
 // that all buckets are present on all sets.
-func (s *erasureSets) ListBuckets(ctx context.Context) (buckets []BucketInfo, err error) {
+func (s *erasureSets) ListBuckets(ctx context.Context, opts BucketOptions) (buckets []BucketInfo, err error) {
 	var listBuckets []BucketInfo
 	healBuckets := map[string]VolInfo{}
 	for _, set := range s.sets {
@@ -877,8 +848,34 @@ func (s *erasureSets) ListBuckets(ctx context.Context) (buckets []BucketInfo, er
 		}
 	}

+	// include deleted buckets in listBuckets output
+	deletedBuckets := map[string]VolInfo{}
+
+	if opts.Deleted {
+		for _, set := range s.sets {
+			// lists all unique buckets across drives.
+			if err := listDeletedBuckets(ctx, set.getDisks(), deletedBuckets, s.defaultParityCount); err != nil {
+				return nil, err
+			}
+		}
+	}
 	for _, v := range healBuckets {
-		listBuckets = append(listBuckets, BucketInfo(v))
+		bi := BucketInfo{
+			Name:    v.Name,
+			Created: v.Created,
+		}
+		if vi, ok := deletedBuckets[v.Name]; ok {
+			bi.Deleted = vi.Created
+		}
+		listBuckets = append(listBuckets, bi)
+	}
+	for _, v := range deletedBuckets {
+		if _, ok := healBuckets[v.Name]; !ok {
+			listBuckets = append(listBuckets, BucketInfo{
+				Name:    v.Name,
+				Deleted: v.Created,
+			})
+		}
 	}

 	sort.Slice(listBuckets, func(i, j int) bool {
@@ -888,6 +885,45 @@ func (s *erasureSets) ListBuckets(ctx context.Context) (buckets []BucketInfo, er
 	return listBuckets, nil
 }

+// listDeletedBuckets lists deleted buckets from all disks.
+func listDeletedBuckets(ctx context.Context, storageDisks []StorageAPI, delBuckets map[string]VolInfo, readQuorum int) error {
+	g := errgroup.WithNErrs(len(storageDisks))
+	var mu sync.Mutex
+	for index := range storageDisks {
+		index := index
+		g.Go(func() error {
+			if storageDisks[index] == nil {
+				// we ignore disk not found errors
+				return nil
+			}
+			volsInfo, err := storageDisks[index].ListDir(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, deletedBucketsPrefix), -1)
+			if err != nil {
+				if err == errFileNotFound {
+					return nil
+				}
+				return err
+			}
+			for _, volName := range volsInfo {
+				mu.Lock()
+				if _, ok := delBuckets[volName]; !ok {
+					vi, err := storageDisks[index].StatVol(ctx, pathJoin(minioMetaBucket, bucketMetaPrefix, deletedBucketsPrefix, volName))
+					if err != nil {
+						return err
+					}
+					bkt := strings.TrimPrefix(vi.Name, pathJoin(minioMetaBucket, bucketMetaPrefix, deletedBucketsPrefix))
+					bkt = strings.TrimPrefix(bkt, slashSeparator)
+					bkt = strings.TrimSuffix(bkt, slashSeparator)
+					vi.Name = bkt
+					delBuckets[bkt] = vi
+				}
+				mu.Unlock()
+			}
+			return nil
+		}, index)
+	}
+	return reduceReadQuorumErrs(ctx, g.Wait(), bucketMetadataOpIgnoredErrs, readQuorum)
+}
+
 // --- Object Operations ---

 // GetObjectNInfo - returns object info and locked object ReadCloser
@@ -1195,12 +1231,16 @@ func markRootDisksAsDown(storageDisks []StorageAPI, errs []error) {
 		// Do nothing
 		return
 	}
-	infos, _ := getHealDiskInfos(storageDisks, errs)
+	infos, ierrs := getHealDiskInfos(storageDisks, errs)
 	for i := range storageDisks {
+		if ierrs[i] != nil && ierrs[i] != errUnformattedDisk {
+			storageDisks[i] = nil
+			continue
+		}
 		if storageDisks[i] != nil && infos[i].RootDisk {
 			// We should not heal on root disk. i.e in a situation where the minio-administrator has unmounted a
 			// defective drive we should not heal a path on the root disk.
-			logger.LogIf(GlobalContext, fmt.Errorf("Disk `%s` is part of root disk, will not be used", storageDisks[i]))
+			logger.LogIf(GlobalContext, fmt.Errorf("Drive `%s` is part of root drive, will not be used", storageDisks[i]))
 			storageDisks[i] = nil
 		}
 	}
@@ -1274,7 +1314,7 @@ func (s *erasureSets) HealFormat(ctx context.Context, dryRun bool) (res madmin.H
 				continue
 			}
 			if err := saveFormatErasure(storageDisks[index], format, true); err != nil {
-				logger.LogIf(ctx, fmt.Errorf("Disk %s failed to write updated 'format.json': %v", storageDisks[index], err))
+				logger.LogIf(ctx, fmt.Errorf("Drive %s failed to write updated 'format.json': %v", storageDisks[index], err))
 				tmpNewFormats[index] = nil // this disk failed to write new format
 			}
 		}
@@ -187,7 +187,7 @@ func TestNewErasureSets(t *testing.T) {
 	// Initializes all erasure disks
 	storageDisks, format, err := waitForFormatErasure(true, endpoints, 1, 1, 16, "", "")
 	if err != nil {
-		t.Fatalf("Unable to format disks for erasure, %s", err)
+		t.Fatalf("Unable to format drives for erasure, %s", err)
 	}

 	ep := PoolEndpoints{Endpoints: endpoints}
@@ -108,16 +108,40 @@ func newErasureSingle(ctx context.Context, storageDisk StorageAPI, format *forma
 // List all buckets from one of the set, we are not doing merge
 // sort here just for simplification. As per design it is assumed
 // that all buckets are present on all sets.
-func (es *erasureSingle) ListBuckets(ctx context.Context) (buckets []BucketInfo, err error) {
+func (es *erasureSingle) ListBuckets(ctx context.Context, opts BucketOptions) (buckets []BucketInfo, err error) {
 	var listBuckets []BucketInfo
 	healBuckets := map[string]VolInfo{}
 	// lists all unique buckets across drives.
 	if err := listAllBuckets(ctx, []StorageAPI{es.disk}, healBuckets, 0); err != nil {
 		return nil, err
 	}
+	// include deleted buckets in listBuckets output
+	deletedBuckets := map[string]VolInfo{}

+	if opts.Deleted {
+		// lists all deleted buckets across drives.
+		if err := listDeletedBuckets(ctx, []StorageAPI{es.disk}, deletedBuckets, 0); err != nil {
+			return nil, err
+		}
+	}
 	for _, v := range healBuckets {
-		listBuckets = append(listBuckets, BucketInfo(v))
+		bi := BucketInfo{
+			Name:    v.Name,
+			Created: v.Created,
+		}
+		if vi, ok := deletedBuckets[v.Name]; ok {
+			bi.Deleted = vi.Created
+		}
+		listBuckets = append(listBuckets, bi)
+	}
+
+	for _, v := range deletedBuckets {
+		if _, ok := healBuckets[v.Name]; !ok {
+			listBuckets = append(listBuckets, BucketInfo{
+				Name:    v.Name,
+				Deleted: v.Created,
+			})
+		}
 	}

 	sort.Slice(listBuckets, func(i, j int) bool {
@@ -256,7 +280,7 @@ type renameAllStorager interface {

 // Bucket operations
 // MakeBucket - make a bucket.
-func (es *erasureSingle) MakeBucketWithLocation(ctx context.Context, bucket string, opts BucketOptions) error {
+func (es *erasureSingle) MakeBucketWithLocation(ctx context.Context, bucket string, opts MakeBucketOptions) error {
 	defer NSUpdated(bucket, slashSeparator)

 	// Lock the bucket name before creating.
@@ -289,6 +313,7 @@ func (es *erasureSingle) MakeBucketWithLocation(ctx context.Context, bucket stri

 	// If it doesn't exist we get a new, so ignore errors
 	meta := newBucketMetadata(bucket)
+	meta.SetCreatedAt(opts.CreatedAt)
 	if opts.LockEnabled {
 		meta.VersioningConfigXML = enabledBucketVersioningConfig
 		meta.ObjectLockConfigXML = enabledBucketObjectLockConfig
@@ -308,12 +333,17 @@ func (es *erasureSingle) MakeBucketWithLocation(ctx context.Context, bucket stri
 }

 // GetBucketInfo - returns BucketInfo for a bucket.
-func (es *erasureSingle) GetBucketInfo(ctx context.Context, bucket string) (bi BucketInfo, e error) {
+func (es *erasureSingle) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (bi BucketInfo, e error) {
 	volInfo, err := es.disk.StatVol(ctx, bucket)
 	if err != nil {
+		if opts.Deleted {
+			if dvi, derr := es.disk.StatVol(ctx, pathJoin(minioMetaBucket, bucketMetaPrefix, deletedBucketsPrefix, bucket)); derr == nil {
+				return BucketInfo{Name: bucket, Deleted: dvi.Created}, nil
+			}
+		}
 		return bi, toObjectErr(err, bucket)
 	}
-	return BucketInfo(volInfo), nil
+	return BucketInfo{Name: volInfo.Name, Created: volInfo.Created}, nil
 }

 // DeleteBucket - deletes a bucket.
@@ -322,6 +352,32 @@ func (es *erasureSingle) DeleteBucket(ctx context.Context, bucket string, opts D
 	defer NSUpdated(bucket, slashSeparator)

 	err := es.disk.DeleteVol(ctx, bucket, opts.Force)
+	// Purge the entire bucket metadata entirely.
+	deleteBucketMetadata(ctx, es, bucket)
+	globalBucketMetadataSys.Remove(bucket)
+
+	if err == nil || errors.Is(err, errVolumeNotFound) {
+		if opts.SRDeleteOp == MarkDelete {
+			es.markDelete(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, deletedBucketsPrefix, bucket))
+		}
+	}
+	return toObjectErr(err, bucket)
+}
+
+// markDelete creates a vol entry in .minio.sys/buckets/.deleted until site replication
+// syncs the delete to peers
+func (es *erasureSingle) markDelete(ctx context.Context, bucket, prefix string) error {
+	err := es.disk.MakeVol(ctx, pathJoin(bucket, prefix))
+	if err != nil && errors.Is(err, errVolumeExists) {
+		return nil
+	}
+	return toObjectErr(err, bucket)
+}
+
+// purgeDelete deletes vol entry in .minio.sys/buckets/.deleted after site replication
+// syncs the delete to peers OR on a new MakeBucket call.
+func (es *erasureSingle) purgeDelete(ctx context.Context, bucket, prefix string) error {
+	err := es.disk.DeleteVol(ctx, pathJoin(bucket, prefix), true)
 	return toObjectErr(err, bucket)
 }

@@ -849,6 +905,13 @@ func (es *erasureSingle) putMetacacheObject(ctx context.Context, key string, r *
 		return ObjectInfo{}, IncompleteBody{Bucket: minioMetaBucket, Object: key}
 	}

+	var index []byte
+	if opts.IndexCB != nil {
+		index = opts.IndexCB()
+	}
+
+	modTime := UTCNow()
+
 	for i, w := range writers {
 		if w == nil {
 			// Make sure to avoid writing to disks which we couldn't complete in erasure.Encode()
@@ -856,7 +919,7 @@ func (es *erasureSingle) putMetacacheObject(ctx context.Context, key string, r *
 			continue
 		}
 		partsMetadata[i].Data = inlineBuffers[i].Bytes()
-		partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize())
+		partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize(), modTime, index)
 		partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{
 			PartNumber: 1,
 			Algorithm:  DefaultBitrotAlgorithm,
@@ -864,8 +927,6 @@ func (es *erasureSingle) putMetacacheObject(ctx context.Context, key string, r *
 		})
 	}

-	modTime := UTCNow()
-
 	// Fill all the necessary metadata.
 	// Update `xl.meta` content on each disks.
 	for index := range partsMetadata {
@@ -1082,6 +1143,16 @@ func (es *erasureSingle) putObject(ctx context.Context, bucket string, object st
 		defer lk.Unlock(lkctx.Cancel)
 	}

+	var index []byte
+	if opts.IndexCB != nil {
+		index = opts.IndexCB()
+	}
+
+	modTime := opts.MTime
+	if opts.MTime.IsZero() {
+		modTime = UTCNow()
+	}
+
 	for i, w := range writers {
 		if w == nil {
 			onlineDisks[i] = nil
@@ -1092,7 +1163,7 @@ func (es *erasureSingle) putObject(ctx context.Context, bucket string, object st
 		} else {
 			partsMetadata[i].Data = nil
 		}
-		partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize())
+		partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize(), modTime, index)
 		partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{
 			PartNumber: 1,
 			Algorithm:  DefaultBitrotAlgorithm,
@@ -1108,11 +1179,6 @@ func (es *erasureSingle) putObject(ctx context.Context, bucket string, object st
 		opts.UserDefined["content-type"] = mimedb.TypeByExtension(path.Ext(object))
 	}

-	modTime := opts.MTime
-	if opts.MTime.IsZero() {
-		modTime = UTCNow()
-	}
-
 	// Fill all the necessary metadata.
 	// Update `xl.meta` content on each disks.
 	for index := range partsMetadata {
@@ -1137,6 +1203,8 @@ func (es *erasureSingle) putObject(ctx context.Context, bucket string, object st
 		return ObjectInfo{}, toObjectErr(err, bucket, object)
 	}

+	defer NSUpdated(bucket, object)
+
 	for i := 0; i < len(onlineDisks); i++ {
 		if onlineDisks[i] != nil && onlineDisks[i].IsOnline() {
 			// Object info is the same in all disks, so we can pick
@@ -1155,7 +1223,7 @@ func (es *erasureSingle) putObject(ctx context.Context, bucket string, object st
 	return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
 }

-func (es *erasureSingle) deleteObjectVersion(ctx context.Context, bucket, object string, writeQuorum int, fi FileInfo, forceDelMarker bool) error {
+func (es *erasureSingle) deleteObjectVersion(ctx context.Context, bucket, object string, fi FileInfo, forceDelMarker bool) error {
 	return es.disk.DeleteVersion(ctx, bucket, object, fi, forceDelMarker)
 }

@@ -1242,12 +1310,12 @@ func (es *erasureSingle) DeleteObjects(ctx context.Context, bucket string, objec
 				DeleteMarker:          vr.Deleted,
 				DeleteMarkerVersionID: vr.VersionID,
 				DeleteMarkerMTime:     DeleteMarkerMTime{vr.ModTime},
-				ObjectName:            vr.Name,
+				ObjectName:            decodeDirObject(vr.Name),
 				ReplicationState:      vr.ReplicationState,
 			}
 		} else {
 			dobjects[i] = DeletedObject{
-				ObjectName:       vr.Name,
+				ObjectName:       decodeDirObject(vr.Name),
 				VersionID:        vr.VersionID,
 				ReplicationState: vr.ReplicationState,
 			}
@@ -1321,8 +1389,14 @@ func (es *erasureSingle) DeleteObjects(ctx context.Context, bucket string, objec

 func (es *erasureSingle) deletePrefix(ctx context.Context, bucket, prefix string) error {
 	dirPrefix := encodeDirObject(prefix)
-	defer es.disk.Delete(ctx, bucket, dirPrefix, true)
-	return es.disk.Delete(ctx, bucket, prefix, true)
+	defer es.disk.Delete(ctx, bucket, dirPrefix, DeleteOptions{
+		Recursive: true,
+		Force:     true,
+	})
+	return es.disk.Delete(ctx, bucket, prefix, DeleteOptions{
+		Recursive: true,
+		Force:     true,
+	})
 }

 // DeleteObject - deletes an object, this call doesn't necessary reply
@@ -1373,7 +1447,7 @@ func (es *erasureSingle) DeleteObject(ctx context.Context, bucket, object string

 	versionFound := true
 	objInfo = ObjectInfo{VersionID: opts.VersionID} // version id needed in Delete API response.
-	goi, writeQuorum, gerr := es.getObjectInfoAndQuorum(ctx, bucket, object, opts)
+	goi, _, gerr := es.getObjectInfoAndQuorum(ctx, bucket, object, opts)
 	if gerr != nil && goi.Name == "" {
 		switch gerr.(type) {
 		case InsufficientReadQuorum:
@@ -1387,6 +1461,11 @@ func (es *erasureSingle) DeleteObject(ctx context.Context, bucket, object string
 		}
 	}

+	// Do not create new delete markers.
+	if goi.DeleteMarker && opts.VersionID == "" {
+		return goi, nil
+	}
+
 	if opts.Expiration.Expire {
 		action := evalActionFromLifecycle(ctx, *lc, rcfg, goi, false)
 		var isErr bool
@@ -1483,7 +1562,7 @@ func (es *erasureSingle) DeleteObject(ctx context.Context, bucket, object string
 			// delete marker. Add delete marker, since we don't have
 			// any version specified explicitly. Or if a particular
 			// version id needs to be replicated.
-			if err = es.deleteObjectVersion(ctx, bucket, object, writeQuorum, fi, opts.DeleteMarker); err != nil {
+			if err = es.deleteObjectVersion(ctx, bucket, object, fi, opts.DeleteMarker); err != nil {
 				return objInfo, toObjectErr(err, bucket, object)
 			}
 			return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
@@ -1502,7 +1581,7 @@ func (es *erasureSingle) DeleteObject(ctx context.Context, bucket, object string
 		ExpireRestored:   opts.Transition.ExpireRestored,
 	}
 	dfi.SetTierFreeVersionID(fvID)
-	if err = es.deleteObjectVersion(ctx, bucket, object, writeQuorum, dfi, opts.DeleteMarker); err != nil {
+	if err = es.deleteObjectVersion(ctx, bucket, object, dfi, opts.DeleteMarker); err != nil {
 		return objInfo, toObjectErr(err, bucket, object)
 	}

@@ -1739,14 +1818,7 @@ func (es *erasureSingle) TransitionObject(ctx context.Context, bucket, object st
 	fi.TransitionVersionID = string(rv)
 	eventName := event.ObjectTransitionComplete

-	// we now know the number of blocks this object needs for data and parity.
-	// writeQuorum is dataBlocks + 1
-	writeQuorum := fi.Erasure.DataBlocks
-	if fi.Erasure.DataBlocks == fi.Erasure.ParityBlocks {
-		writeQuorum++
-	}
-
-	if err = es.deleteObjectVersion(ctx, bucket, object, writeQuorum, fi, false); err != nil {
+	if err = es.deleteObjectVersion(ctx, bucket, object, fi, false); err != nil {
 		eventName = event.ObjectTransitionFailed
 	}

@@ -1916,7 +1988,10 @@ func (es *erasureSingle) removeObjectPart(bucket, object, uploadID, dataDir stri
 			// Ignoring failure to remove parts that weren't present in CompleteMultipartUpload
 			// requests. xl.meta is the authoritative source of truth on which parts constitute
 			// the object. The presence of parts that don't belong in the object doesn't affect correctness.
-			_ = storageDisks[index].Delete(context.TODO(), minioMetaMultipartBucket, curpartPath, false)
+			_ = storageDisks[index].Delete(context.TODO(), minioMetaMultipartBucket, curpartPath, DeleteOptions{
+				Recursive: false,
+				Force:     false,
+			})
 			return nil
 		}, index)
 	}
@@ -1954,7 +2029,10 @@ func (es *erasureSingle) cleanupStaleUploadsOnDisk(ctx context.Context, disk Sto
 		}
 		wait := es.deletedCleanupSleeper.Timer(ctx)
 		if now.Sub(vi.Created) > expiry {
-			disk.Delete(ctx, minioMetaTmpBucket, tmpDir, true)
+			disk.Delete(ctx, minioMetaTmpBucket, tmpDir, DeleteOptions{
+				Recursive: true,
+				Force:     false,
+			})
 		}
 		wait()
 		return nil
@@ -2357,8 +2435,13 @@ func (es *erasureSingle) PutObjectPart(ctx context.Context, bucket, object, uplo

 	md5hex := r.MD5CurrentHexString()

+	var index []byte
+	if opts.IndexCB != nil {
+		index = opts.IndexCB()
+	}
+
 	// Add the current part.
-	fi.AddObjectPart(partID, md5hex, n, data.ActualSize())
+	fi.AddObjectPart(partID, md5hex, n, data.ActualSize(), fi.ModTime, index)

 	for i, disk := range onlineDisks {
 		if disk == OfflineDisk {
@@ -2656,6 +2739,7 @@ func (es *erasureSingle) CompleteMultipartUpload(ctx context.Context, bucket str
 			Number:     part.PartNumber,
 			Size:       currentFI.Parts[partIdx].Size,
 			ActualSize: currentFI.Parts[partIdx].ActualSize,
+			Index:      currentFI.Parts[partIdx].Index,
 		}
 	}

@@ -2720,6 +2804,8 @@ func (es *erasureSingle) CompleteMultipartUpload(ctx context.Context, bucket str
 		return oi, toObjectErr(err, bucket, object)
 	}

+	defer NSUpdated(bucket, object)
+
 	for i := 0; i < len(onlineDisks); i++ {
 		if onlineDisks[i] != nil && onlineDisks[i].IsOnline() {
 			// Object info is the same in all disks, so we can pick
@@ -2771,11 +2857,16 @@ func (es *erasureSingle) AbortMultipartUpload(ctx context.Context, bucket, objec
 func (es *erasureSingle) ListObjects(ctx context.Context, bucket, prefix, marker, delimiter string, maxKeys int) (ListObjectsInfo, error) {
 	var loi ListObjectsInfo

-	// Automatically remove the object/version is an expiry lifecycle rule can be applied
-	lc, _ := globalLifecycleSys.Get(bucket)
-
-	// Check if bucket is object locked.
-	rcfg, _ := globalBucketObjectLockSys.Get(bucket)
+	opts := listPathOptions{
+		Bucket:      bucket,
+		Prefix:      prefix,
+		Separator:   delimiter,
+		Limit:       maxKeysPlusOne(maxKeys, marker != ""),
+		Marker:      marker,
+		InclDeleted: false,
+		AskDisks:    globalAPIConfig.getListQuorum(),
+	}
+	opts.setBucketMeta(ctx)

 	if len(prefix) > 0 && maxKeys == 1 && delimiter == "" && marker == "" {
 		// Optimization for certain applications like
@@ -2787,8 +2878,8 @@ func (es *erasureSingle) ListObjects(ctx context.Context, bucket, prefix, marker
 		// to avoid the need for ListObjects().
 		objInfo, err := es.GetObjectInfo(ctx, bucket, prefix, ObjectOptions{NoLock: true})
 		if err == nil {
-			if lc != nil {
-				action := evalActionFromLifecycle(ctx, *lc, rcfg, objInfo, false)
+			if opts.Lifecycle != nil {
+				action := evalActionFromLifecycle(ctx, *opts.Lifecycle, opts.Retention, objInfo, false)
 				switch action {
 				case lifecycle.DeleteVersionAction, lifecycle.DeleteAction:
 					fallthrough
@@ -2801,18 +2892,6 @@ func (es *erasureSingle) ListObjects(ctx context.Context, bucket, prefix, marker
 		}
 	}

-	opts := listPathOptions{
-		Bucket:      bucket,
-		Prefix:      prefix,
-		Separator:   delimiter,
-		Limit:       maxKeysPlusOne(maxKeys, marker != ""),
-		Marker:      marker,
-		InclDeleted: false,
-		AskDisks:    globalAPIConfig.getListQuorum(),
-		Lifecycle:   lc,
-		Retention:   rcfg,
-	}
-
 	merged, err := es.listPath(ctx, &opts)
 	if err != nil && err != io.EOF {
 		if !isErrBucketNotFound(err) {
@@ -2871,7 +2950,6 @@ func (es *erasureSingle) ListObjectVersions(ctx context.Context, bucket, prefix,
 	if marker == "" && versionMarker != "" {
 		return loi, NotImplemented{}
 	}
-
 	opts := listPathOptions{
 		Bucket:      bucket,
 		Prefix:      prefix,
@@ -2882,6 +2960,7 @@ func (es *erasureSingle) ListObjectVersions(ctx context.Context, bucket, prefix,
 		AskDisks:    "strict",
 		Versioned:   true,
 	}
+	opts.setBucketMeta(ctx)

 	merged, err := es.listPath(ctx, &opts)
 	if err != nil && err != io.EOF {
@@ -2928,13 +3007,13 @@ func (es *erasureSingle) Walk(ctx context.Context, bucket, prefix string, result
 		return err
 	}

+	vcfg, _ := globalBucketVersioningSys.Get(bucket)
+
 	ctx, cancel := context.WithCancel(ctx)
 	go func() {
 		defer cancel()
 		defer close(results)

-		versioned := opts.Versioned || opts.VersionSuspended
-
 		var wg sync.WaitGroup
 		wg.Add(1)
 		go func() {
@@ -2949,15 +3028,17 @@ func (es *erasureSingle) Walk(ctx context.Context, bucket, prefix string, result
 					cancel()
 					return
 				}
-				if opts.WalkAscending {
-					for i := len(fivs.Versions) - 1; i >= 0; i-- {
-						version := fivs.Versions[i]
-						results <- version.ToObjectInfo(bucket, version.Name, versioned)
-					}
-					return
-				}
+
+				versionsSorter(fivs.Versions).reverse()
+
 				for _, version := range fivs.Versions {
-					results <- version.ToObjectInfo(bucket, version.Name, versioned)
+					versioned := vcfg != nil && vcfg.Versioned(version.Name)
+
+					select {
+					case <-ctx.Done():
+						return
+					case results <- version.ToObjectInfo(bucket, version.Name, versioned):
+					}
 				}
 			}

@@ -2984,7 +3065,7 @@ func (es *erasureSingle) Walk(ctx context.Context, bucket, prefix string, result
 				minDisks:       1,
 				reportNotFound: false,
 				agreed:         loadEntry,
-				partial: func(entries metaCacheEntries, nAgreed int, errs []error) {
+				partial: func(entries metaCacheEntries, _ []error) {
 					entry, ok := entries.resolve(&resolver)
 					if !ok {
 						// check if we can get one entry atleast
@@ -2999,6 +3080,7 @@ func (es *erasureSingle) Walk(ctx context.Context, bucket, prefix string, result

 			if err := listPathRaw(ctx, lopts); err != nil {
 				logger.LogIf(ctx, fmt.Errorf("listPathRaw returned %w: opts(%#v)", err, lopts))
+				cancel()
 				return
 			}
 		}()
@@ -3008,6 +3090,33 @@ func (es *erasureSingle) Walk(ctx context.Context, bucket, prefix string, result
 	return nil
 }

+// Health - returns current status of the object layer health, for single drive
+// its as simple as returning healthy as long as drive is accessible.
+func (es *erasureSingle) Health(ctx context.Context, opts HealthOptions) HealthResult {
+	_, err := es.disk.DiskInfo(ctx)
+	if err != nil {
+		return HealthResult{}
+	}
+	if opts.Maintenance {
+		// Single drive cannot be put under maintenance.
+		return HealthResult{
+			Healthy:     false,
+			WriteQuorum: 1,
+		}
+	}
+	return HealthResult{
+		Healthy:     true,
+		WriteQuorum: 1,
+	}
+}
+
+// ReadHealth - returns current status of the object layer health for reads,
+// for single drive its as simple as returning healthy as long as drive is accessible.
+func (es *erasureSingle) ReadHealth(ctx context.Context) bool {
+	res := es.Health(ctx, HealthOptions{})
+	return res.Healthy
+}
+
 // nsScanner will start scanning buckets and send updated totals as they are traversed.
 // Updates are sent on a regular basis and the caller *must* consume them.
 func (es *erasureSingle) nsScanner(ctx context.Context, buckets []BucketInfo, bf *bloomFilter, wantCycle uint32, updates chan<- dataUsageCache, healScanMode madmin.HealScanMode) error {
@@ -3203,7 +3312,7 @@ func (es *erasureSingle) NSScanner(ctx context.Context, bf *bloomFilter, updates
 	results := make([]dataUsageCache, 1)
 	var firstErr error

-	allBuckets, err := es.ListBuckets(ctx)
+	allBuckets, err := es.ListBuckets(ctx, BucketOptions{})
 	if err != nil {
 		return err
 	}
--- a/Show More
+++ b/Show More