fix: list across all drives to avoid stale disks

It was observed in VMware vsphere environment during a
pod replacement, `mc admin info` might report incorrect
offline nodes for the replaced drive. This issue eventually
goes away but requires quite a lot of time for all servers
to be in sync.

This PR fixes this behavior properly.

.github/stale.yml

@@ -14,6 +14,7 @@ onlyLabels: []
 exemptLabels:
   - "security"
   - "pending discussion"
+  - "do not close"
 
 # Set to true to ignore issues in a project (defaults to false)
 exemptProjects: false

.github/workflows/go.yml

@@ -4,7 +4,6 @@ on:
   pull_request:
     branches:
     - master
-    - release
 
 jobs:
   build:
@@ -12,7 +11,7 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        go-version: [1.13.x]
+        go-version: [1.14.x, 1.15.x]
         os: [ubuntu-latest, windows-latest]
     steps:
       - uses: actions/checkout@v2
@@ -39,6 +38,9 @@ jobs:
           MINIO_CI_CD: 1
         run: |
           sudo apt-get install devscripts shellcheck
+          nancy_version=$(curl --retry 10 -Ls -o /dev/null -w "%{url_effective}" https://github.com/sonatype-nexus-community/nancy/releases/latest | sed "s/https:\/\/github.com\/sonatype-nexus-community\/nancy\/releases\/tag\///")
+          curl -L -o nancy https://github.com/sonatype-nexus-community/nancy/releases/download/${nancy_version}/nancy-linux.amd64-${nancy_version} && chmod +x nancy
+          go list -m all | ./nancy sleuth
           make
           diff -au <(gofmt -s -d cmd) <(printf "")
           diff -au <(gofmt -s -d pkg) <(printf "")

.nancy-ignore

@@ -0,0 +1,6 @@
+CVE-2018-17075
+CVE-2018-17848
+CVE-2018-17846
+CVE-2018-17847
+CVE-2018-17143
+CVE-2018-17142

CREDITS

@@ -10293,25 +10293,6 @@ Exhibit B - "Incompatible With Secondary Licenses" Notice
       the Mozilla Public License, v. 2.0.
 
 
-================================================================
-
-github.com/inconshreveable/go-update
-https://github.com/inconshreveable/go-update
-----------------------------------------------------------------
-Copyright 2015 Alan Shreve
-
-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/jcmturner/gofork
@@ -11470,8 +11451,216 @@ https://github.com/minio/md5-simd
 
 ================================================================
 
-github.com/minio/minio-go/v6
-https://github.com/minio/minio-go/v6
+github.com/minio/minio-go/v7
+https://github.com/minio/minio-go/v7
+----------------------------------------------------------------
+
+                                 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/minio/selfupdate
+https://github.com/minio/selfupdate
 ----------------------------------------------------------------
 
                                  Apache License

Dockerfile

@@ -1,4 +1,4 @@
-FROM golang:1.13-alpine as builder
+FROM golang:1.14-alpine as builder
 
 LABEL maintainer="MinIO Inc <dev@min.io>"
 

Dockerfile.arm.release

@@ -23,8 +23,7 @@ RUN \
      chmod +x /usr/bin/minio && \
      chmod +x /usr/bin/docker-entrypoint.sh && \
      chmod +x /usr/bin/verify-minio.sh && \
-     curl -s -q -O https://raw.githubusercontent.com/minio/minio/release/CREDITS && \
-     /usr/bin/verify-minio.sh
+     curl -s -q -O https://raw.githubusercontent.com/minio/minio/master/CREDITS
 
 EXPOSE 9000
 

Dockerfile.arm64.release

@@ -23,7 +23,7 @@ RUN \
      chmod +x /usr/bin/minio && \
      chmod +x /usr/bin/docker-entrypoint.sh && \
      chmod +x /usr/bin/verify-minio.sh && \
-     curl -s -q -O https://raw.githubusercontent.com/minio/minio/release/CREDITS && \
+     curl -s -q -O https://raw.githubusercontent.com/minio/minio/master/CREDITS && \
      /usr/bin/verify-minio.sh
 
 EXPOSE 9000

Dockerfile.ppc64le.release

@@ -22,7 +22,7 @@ RUN \
      chmod +x /usr/bin/minio && \
      chmod +x /usr/bin/docker-entrypoint.sh && \
      chmod +x /usr/bin/verify-minio.sh && \
-     curl -s -q -O https://raw.githubusercontent.com/minio/minio/release/CREDITS && \
+     curl -s -q -O https://raw.githubusercontent.com/minio/minio/master/CREDITS && \
      /usr/bin/verify-minio.sh
 
 EXPOSE 9000

Dockerfile.release

@@ -20,7 +20,7 @@ RUN \
      chmod +x /usr/bin/minio && \
      chmod +x /usr/bin/docker-entrypoint.sh && \
      chmod +x /usr/bin/verify-minio.sh && \
-     curl -s -q -O https://raw.githubusercontent.com/minio/minio/release/CREDITS && \
+     curl -s -q -O https://raw.githubusercontent.com/minio/minio/master/CREDITS && \
      /usr/bin/verify-minio.sh
 
 EXPOSE 9000

Dockerfile.s390x.release

@@ -22,7 +22,7 @@ RUN \
      chmod +x /usr/bin/minio && \
      chmod +x /usr/bin/docker-entrypoint.sh && \
      chmod +x /usr/bin/verify-minio.sh && \
-     curl -s -q -O https://raw.githubusercontent.com/minio/minio/release/CREDITS && \
+     curl -s -q -O https://raw.githubusercontent.com/minio/minio/master/CREDITS && \
      /usr/bin/verify-minio.sh
 
 EXPOSE 9000

Makefile

@@ -17,11 +17,12 @@ checks:
 getdeps:
 	@mkdir -p ${GOPATH}/bin
 	@which golangci-lint 1>/dev/null || (echo "Installing golangci-lint" && curl -sSfL https://raw.githubusercontent.com/golangci/golangci-lint/master/install.sh | sh -s -- -b $(GOPATH)/bin v1.27.0)
+	@which ruleguard 1>/dev/null || (echo "Installing ruleguard" && GO111MODULE=off go get github.com/quasilyte/go-ruleguard/...)
 
 crosscompile:
 	@(env bash $(PWD)/buildscripts/cross-compile.sh)
 
-verifiers: getdeps fmt lint
+verifiers: getdeps fmt lint ruleguard
 
 fmt:
 	@echo "Running $@ check"
@@ -33,6 +34,10 @@ lint:
 	@GO111MODULE=on ${GOPATH}/bin/golangci-lint cache clean
 	@GO111MODULE=on ${GOPATH}/bin/golangci-lint run --timeout=5m --config ./.golangci.yml
 
+ruleguard:
+	@echo "Running $@ check"
+	@${GOPATH}/bin/ruleguard -rules ruleguard.rules.go github.com/minio/minio/...
+
 # Builds minio, runs the verifiers then runs the tests.
 check: test
 test: verifiers build

README.md

@@ -88,7 +88,7 @@ service minio start
 ```
 
 ## Install 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 [go1.13](https://golang.org/dl/#stable)
+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 [go1.14](https://golang.org/dl/#stable)
 
 ```sh
 GO111MODULE=on go get github.com/minio/minio
@@ -177,7 +177,7 @@ mc admin update <minio alias, e.g., myminio>
 - `mc admin update` updates and restarts all servers simultaneously, applications would retry and continue their respective operations upon upgrade.
 - `mc admin update` is disabled in kubernetes/container environments, container environments provide their own mechanisms to rollout of updates.
 - In the case of federated setups `mc admin update` should be run against each cluster individually. Avoid updating `mc` to any new releases until all clusters have been successfully updated.
-- If using `kes` as KMS with MinIO, just replace the binary and restart `kes` more information about `kes` can be found [here](https://github.com/minio/kes/wiki)x
+- If using `kes` as KMS with MinIO, just replace the binary and restart `kes` more information about `kes` can be found [here](https://github.com/minio/kes/wiki)
 - If using Vault as KMS with MinIO, ensure you have followed the Vault upgrade procedure outlined here: https://www.vaultproject.io/docs/upgrading/index.html
 - If using etcd with MinIO for the federation, ensure you have followed the etcd upgrade procedure outlined here: https://github.com/etcd-io/etcd/blob/master/Documentation/upgrades/upgrading-etcd.md
 
@@ -193,4 +193,4 @@ mc admin update <minio alias, e.g., myminio>
 Please follow MinIO [Contributor's Guide](https://github.com/minio/minio/blob/master/CONTRIBUTING.md)
 
 ## License
-Use of MinIO is governed by the Apache 2.0 License found at [LICENSE](./LICENSE).
+Use of MinIO is governed by the Apache 2.0 License found at [LICENSE](https://github.com/minio/minio/blob/master/LICENSE).

README_zh_CN.md

@@ -7,29 +7,33 @@ MinIO是一个非常轻量的服务,可以很简单的和其他应用的结合
 ## Docker 容器
 ### 稳定版
 ```
-docker pull minio/minio
-docker run -p 9000:9000 minio/minio server /data
+docker run -p 9000:9000 \
+  -e "MINIO_ACCESS_KEY=AKIAIOSFODNN7EXAMPLE" \
+  -e "MINIO_SECRET_KEY=wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY" \
+  minio/minio server /data
 ```
 
 ### 尝鲜版
 ```
-docker pull minio/minio:edge
-docker run -p 9000:9000 minio/minio:edge server /data
+docker run -p 9000:9000 \
+  -e "MINIO_ACCESS_KEY=AKIAIOSFODNN7EXAMPLE" \
+  -e "MINIO_SECRET_KEY=wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY" \
+  minio/minio:edge server /data
 ```
-更多Docker部署信息请访问 [这里](https://docs.min.io/docs/minio-docker-quickstart-guide)
+
+> 提示：除非你通过`-it`(TTY交互)参数启动容器，否则Docker将不会显示默认的密钥。一般情况下，并不推荐使用容器的默认密钥，更多Docker部署信息请访问 [这里](https://docs.min.io/docs/minio-docker-quickstart-guide)
 
 ## macOS
-### Homebrew
+### Homebrew（推荐）
 使用 [Homebrew](http://brew.sh/)安装minio
 
 ```sh
 brew install minio/stable/minio
 minio server /data
 ```
-#### Note
-如果你之前使用 `brew install minio`安装过minio, 可以用 `minio/stable/minio` 官方镜像进行重装. 由于golang 1.8的bug,homebrew版本不太稳定。
 
-```
+> 提示：如果你之前使用 `brew install minio`安装过minio, 可以用 `minio/stable/minio` 官方镜像进行重装. 由于golang 1.8的bug,homebrew版本不太稳定。
+```sh
 brew uninstall minio
 brew install minio/stable/minio
 ```
@@ -49,6 +53,16 @@ chmod 755 minio
 | ---------- | --------     | ------                                                     |
 | GNU/Linux  | 64-bit Intel | https://dl.min.io/server/minio/release/linux-amd64/minio |
 ```sh
+wget https://dl.min.io/server/minio/release/linux-amd64/minio
+chmod +x minio
+./minio server /data
+```
+
+| 操作系统    | CPU架构      | 地址                                                        |
+| ---------- | --------     | ------                                                     |
+| GNU/Linux  | ppc64le      | https://dl.min.io/server/minio/release/linux-ppc64le/minio |
+```sh
+wget https://dl.min.io/server/minio/release/linux-ppc64le/minio
 chmod +x minio
 ./minio server /data
 ```
@@ -64,7 +78,7 @@ minio.exe server D:\Photos
 
 ## FreeBSD
 ### Port
-使用 [pkg](https://github.com/freebsd/pkg)进行安装。
+使用 [pkg](https://github.com/freebsd/pkg)进行安装，, MinIO官方并没有提供FreeBSD二进制文件， 它由FreeBSD上游维护，点击 [这里](https://www.freshports.org/www/minio)查看。
 
 ```sh
 pkg install minio
@@ -75,14 +89,68 @@ service minio start
 
 ## 使用源码安装
 
-采用源码安装仅供开发人员和高级用户使用,如果你还没有Golang环境， 请参考 [How to install Golang](https://golang.org/doc/install).
+采用源码安装仅供开发人员和高级用户使用,如果你还没有Golang环境， 请参考 [How to install Golang](https://golang.org/doc/install)。最低需要Golang版本为 [go1.14](https://golang.org/dl/#stable)
 
 ```sh
-go get -u github.com/minio/minio
+GO111MODULE=on go get github.com/minio/minio
+```
+
+## 为防火墙设置允许访问的端口
+
+默认情况下，MinIO 使用端口9000来侦听传入的连接。如果你的平台默认阻止了该端口，则需要启用对该端口的访问。
+
+### ufw
+
+对于启用了ufw的主机（基于Debian的发行版）, 你可以通过`ufw`命令允许指定端口上的所有流量连接. 通过如下命令允许访问端口9000
+
+```sh
+ufw allow 9000
+```
+
+如下命令允许端口9000-9010上的所有传入流量。
+
+```sh
+ufw allow 9000:9010/tcp
+```
+
+### firewall-cmd
+
+对于启用了firewall-cmd的主机（CentOS）, 你可以通过`firewall-cmd`命令允许指定端口上的所有流量连接。 通过如下命令允许访问端口9000
+
+```sh
+firewall-cmd --get-active-zones
+```
+
+这个命令获取当前正在使用的区域。 现在，就可以为以上返回的区域应用端口规则了。 假如返回的区域是 `public`, 使用如下命令
+
+```sh
+firewall-cmd --zone=public --add-port=9000/tcp --permanent
+```
+
+这里的`permanent`参数表示持久化存储规则，可用于防火墙启动、重启和重新加载。 最后，需要防火墙重新加载，让我们刚刚的修改生效。
+
+```sh
+firewall-cmd --reload
+```
+
+### iptables
+
+对于启用了iptables的主机（RHEL, CentOS, etc）, 你可以通过`iptables`命令允许指定端口上的所有流量连接。 通过如下命令允许访问端口9000
+
+```sh
+iptables -A INPUT -p tcp --dport 9000 -j ACCEPT
+service iptables restart
+```
+
+如下命令允许端口9000-9010上的所有传入流量。
+
+```sh
+iptables -A INPUT -p tcp --dport 9000:9010 -j ACCEPT
+service iptables restart
 ```
 
 ## 使用MinIO浏览器进行验证
-安装后使用浏览器访问[http://127.0.0.1:9000](http://127.0.0.1:9000)，如果可以访问，则表示minio已经安装成功。
+MinIO Server带有一个嵌入的Web对象浏览器，安装后使用浏览器访问[http://127.0.0.1:9000](http://127.0.0.1:9000)，如果可以访问，则表示minio已经安装成功。
 
 ![Screenshot](https://github.com/minio/minio/blob/master/docs/screenshots/minio-browser.png?raw=true)
 
@@ -94,6 +162,25 @@ go get -u github.com/minio/minio
 
 上述描述对所有网关后端同样有效。
 
+## 升级 MinIO
+MinIO 服务端支持滚动升级, 也就是说你可以一次更新分布式集群中的一个MinIO实例。 这样可以在不停机的情况下进行升级。可以通过将二进制文件替换为最新版本并以滚动方式重新启动所有服务器来手动完成升级。但是, 我们建议所有用户从客户端使用 [`mc admin update`](https://docs.min.io/docs/minio-admin-complete-guide.html#update) 命令升级。 这将同时更新集群中的所有节点并重新启动它们, 如下命令所示:
+
+```
+mc admin update <minio alias, e.g., myminio>
+```
+
+> 注意: 有些发行版可能不允许滚动升级，这通常在发行说明中提到，所以建议在升级之前阅读发行说明。在这种情况下，建议使用`mc admin update`升级机制来一次升级所有服务器。
+
+### MinIO升级时要记住的重要事项
+
+- `mc admin update` 命令仅当运行MinIO的用户对二进制文件所在的父目录具有写权限时才工作, 比如当前二进制文件位于`/usr/local/bin/minio`, 你需要具备`/usr/local/bin`目录的写权限.
+- `mc admin update` 命令同时更新并重新启动所有服务器，应用程序将在升级后重试并继续各自的操作。
+- `mc admin update` 命令在 kubernetes/container 环境下是不能用的, 容器环境提供了它自己的更新机制来更新。
+- 对于联盟部署模式，应分别针对每个群集运行`mc admin update`。 在成功更新所有群集之前，不要将`mc`更新为任何新版本。
+- 如果将`kes`用作MinIO的KMS，只需替换二进制文件并重新启动`kes`，可以在 [这里](https://github.com/minio/kes/wiki) 找到有关`kes`的更多信息。
+- 如果将Vault作为MinIO的KMS，请确保已遵循如下Vault升级过程的概述：https://www.vaultproject.io/docs/upgrading/index.html
+- 如果将MindIO与etcd配合使用, 请确保已遵循如下etcd升级过程的概述: https://github.com/etcd-io/etcd/blob/master/Documentation/upgrades/upgrading-etcd.md
+
 ## 了解更多
 - [MinIO纠删码入门](https://docs.min.io/docs/minio-erasure-code-quickstart-guide)
 - [`mc`快速入门](https://docs.min.io/docs/minio-client-quickstart-guide)
@@ -104,3 +191,6 @@ go get -u github.com/minio/minio
 
 ## 如何参与到MinIO项目
 请参考 [贡献者指南](https://github.com/minio/minio/blob/master/CONTRIBUTING.md)。欢迎各位中国程序员加到MinIO项目中。
+
+## 授权许可
+MinIO的使用受 Apache 2.0 License 约束，你可以在 [LICENSE](./LICENSE) 查看许可。

buildscripts/verify-build.sh

@@ -45,88 +45,63 @@ FUNCTIONAL_TESTS="$WORK_DIR/functional-tests.sh"
 function start_minio_fs()
 {
     "${MINIO[@]}" server "${WORK_DIR}/fs-disk" >"$WORK_DIR/fs-minio.log" 2>&1 &
-    minio_pid=$!
     sleep 10
-
-    echo "$minio_pid"
 }
 
 function start_minio_erasure()
 {
     "${MINIO[@]}" server "${WORK_DIR}/erasure-disk1" "${WORK_DIR}/erasure-disk2" "${WORK_DIR}/erasure-disk3" "${WORK_DIR}/erasure-disk4" >"$WORK_DIR/erasure-minio.log" 2>&1 &
-    minio_pid=$!
     sleep 15
-
-    echo "$minio_pid"
 }
 
 function start_minio_erasure_sets()
 {
     "${MINIO[@]}" server "${WORK_DIR}/erasure-disk-sets{1...32}" >"$WORK_DIR/erasure-minio-sets.log" 2>&1 &
-    minio_pid=$!
     sleep 15
-
-    echo "$minio_pid"
 }
 
 function start_minio_zone_erasure_sets()
 {
-    declare -a minio_pids
     export MINIO_ACCESS_KEY=$ACCESS_KEY
     export MINIO_SECRET_KEY=$SECRET_KEY
 
     "${MINIO[@]}" server --address=:9000 "http://127.0.0.1:9000${WORK_DIR}/zone-disk-sets{1...4}" "http://127.0.0.1:9001${WORK_DIR}/zone-disk-sets{5...8}" >"$WORK_DIR/zone-minio-9000.log" 2>&1 &
-    minio_pids[0]=$!
-
     "${MINIO[@]}" server --address=:9001 "http://127.0.0.1:9000${WORK_DIR}/zone-disk-sets{1...4}" "http://127.0.0.1:9001${WORK_DIR}/zone-disk-sets{5...8}" >"$WORK_DIR/zone-minio-9001.log" 2>&1 &
-    minio_pids[1]=$!
 
     sleep 40
-    echo "${minio_pids[@]}"
 }
 
 function start_minio_zone_erasure_sets_ipv6()
 {
-    declare -a minio_pids
     export MINIO_ACCESS_KEY=$ACCESS_KEY
     export MINIO_SECRET_KEY=$SECRET_KEY
 
     "${MINIO[@]}" server --address="[::1]:9000" "http://[::1]:9000${WORK_DIR}/zone-disk-sets{1...4}" "http://[::1]:9001${WORK_DIR}/zone-disk-sets{5...8}" >"$WORK_DIR/zone-minio-9000.log" 2>&1 &
-    minio_pids[0]=$!
-
     "${MINIO[@]}" server --address="[::1]:9001" "http://[::1]:9000${WORK_DIR}/zone-disk-sets{1...4}" "http://[::1]:9001${WORK_DIR}/zone-disk-sets{5...8}" >"$WORK_DIR/zone-minio-9001.log" 2>&1 &
-    minio_pids[1]=$!
 
     sleep 40
-    echo "${minio_pids[@]}"
 }
 
 function start_minio_dist_erasure()
 {
-    declare -a minio_pids
     export MINIO_ACCESS_KEY=$ACCESS_KEY
     export MINIO_SECRET_KEY=$SECRET_KEY
     "${MINIO[@]}" server --address=:9000 "http://127.0.0.1:9000${WORK_DIR}/dist-disk1" "http://127.0.0.1:9001${WORK_DIR}/dist-disk2" "http://127.0.0.1:9002${WORK_DIR}/dist-disk3" "http://127.0.0.1:9003${WORK_DIR}/dist-disk4" >"$WORK_DIR/dist-minio-9000.log" 2>&1 &
-    minio_pids[0]=$!
     "${MINIO[@]}" server --address=:9001 "http://127.0.0.1:9000${WORK_DIR}/dist-disk1" "http://127.0.0.1:9001${WORK_DIR}/dist-disk2" "http://127.0.0.1:9002${WORK_DIR}/dist-disk3" "http://127.0.0.1:9003${WORK_DIR}/dist-disk4" >"$WORK_DIR/dist-minio-9001.log" 2>&1 &
-    minio_pids[1]=$!
     "${MINIO[@]}" server --address=:9002 "http://127.0.0.1:9000${WORK_DIR}/dist-disk1" "http://127.0.0.1:9001${WORK_DIR}/dist-disk2" "http://127.0.0.1:9002${WORK_DIR}/dist-disk3" "http://127.0.0.1:9003${WORK_DIR}/dist-disk4" >"$WORK_DIR/dist-minio-9002.log" 2>&1 &
-    minio_pids[2]=$!
     "${MINIO[@]}" server --address=:9003 "http://127.0.0.1:9000${WORK_DIR}/dist-disk1" "http://127.0.0.1:9001${WORK_DIR}/dist-disk2" "http://127.0.0.1:9002${WORK_DIR}/dist-disk3" "http://127.0.0.1:9003${WORK_DIR}/dist-disk4" >"$WORK_DIR/dist-minio-9003.log" 2>&1 &
-    minio_pids[3]=$!
 
     sleep 40
-    echo "${minio_pids[@]}"
 }
 
 function run_test_fs()
 {
-    minio_pid="$(start_minio_fs)"
+    start_minio_fs
 
     (cd "$WORK_DIR" && "$FUNCTIONAL_TESTS")
     rv=$?
 
-    kill "$minio_pid"
+    pkill minio
     sleep 3
 
     if [ "$rv" -ne 0 ]; then
@@ -138,12 +113,12 @@ function run_test_fs()
 }
 
 function run_test_erasure_sets() {
-    minio_pid="$(start_minio_erasure_sets)"
+    start_minio_erasure_sets
 
     (cd "$WORK_DIR" && "$FUNCTIONAL_TESTS")
     rv=$?
 
-    kill "$minio_pid"
+    pkill minio
     sleep 3
 
     if [ "$rv" -ne 0 ]; then
@@ -156,14 +131,12 @@ function run_test_erasure_sets() {
 
 function run_test_zone_erasure_sets()
 {
-    minio_pids=( $(start_minio_zone_erasure_sets) )
+    start_minio_zone_erasure_sets
 
     (cd "$WORK_DIR" && "$FUNCTIONAL_TESTS")
     rv=$?
 
-    for pid in "${minio_pids[@]}"; do
-        kill "$pid"
-    done
+    pkill minio
     sleep 3
 
     if [ "$rv" -ne 0 ]; then
@@ -182,16 +155,14 @@ function run_test_zone_erasure_sets()
 
 function run_test_zone_erasure_sets_ipv6()
 {
-    minio_pids=( $(start_minio_zone_erasure_sets_ipv6) )
+    start_minio_zone_erasure_sets_ipv6
 
     export SERVER_ENDPOINT="[::1]:9000"
 
     (cd "$WORK_DIR" && "$FUNCTIONAL_TESTS")
     rv=$?
 
-    for pid in "${minio_pids[@]}"; do
-        kill "$pid"
-    done
+    pkill minio
     sleep 3
 
     if [ "$rv" -ne 0 ]; then
@@ -210,12 +181,12 @@ function run_test_zone_erasure_sets_ipv6()
 
 function run_test_erasure()
 {
-    minio_pid="$(start_minio_erasure)"
+    start_minio_erasure
 
     (cd "$WORK_DIR" && "$FUNCTIONAL_TESTS")
     rv=$?
 
-    kill "$minio_pid"
+    pkill minio
     sleep 3
 
     if [ "$rv" -ne 0 ]; then
@@ -228,14 +199,12 @@ function run_test_erasure()
 
 function run_test_dist_erasure()
 {
-    minio_pids=( $(start_minio_dist_erasure) )
+    start_minio_dist_erasure
 
     (cd "$WORK_DIR" && "$FUNCTIONAL_TESTS")
     rv=$?
 
-    for pid in "${minio_pids[@]}"; do
-        kill "$pid"
-    done
+    pkill minio
     sleep 3
 
     if [ "$rv" -ne 0 ]; then

cmd/admin-bucket-handlers.go

@@ -20,20 +20,20 @@ import (
 	"encoding/json"
 	"io"
 	"io/ioutil"
+	"net"
 	"net/http"
 
 	"github.com/gorilla/mux"
 	"github.com/minio/minio/cmd/config"
 	"github.com/minio/minio/cmd/logger"
-	"github.com/minio/minio/pkg/auth"
 	"github.com/minio/minio/pkg/env"
 	iampolicy "github.com/minio/minio/pkg/iam/policy"
 	"github.com/minio/minio/pkg/madmin"
 )
 
 const (
-	bucketQuotaConfigFile        = "quota.json"
-	bucketReplicationTargetsFile = "replication-targets.json"
+	bucketQuotaConfigFile = "quota.json"
+	bucketTargetsFile     = "bucket-targets.json"
 )
 
 // PutBucketQuotaConfigHandler - PUT Bucket quota configuration.
@@ -121,16 +121,16 @@ func (a adminAPIHandlers) GetBucketQuotaConfigHandler(w http.ResponseWriter, r *
 	writeSuccessResponseJSON(w, configData)
 }
 
-// SetBucketReplicationTargetHandler - sets a replication target for bucket
-func (a adminAPIHandlers) SetBucketReplicationTargetHandler(w http.ResponseWriter, r *http.Request) {
-	ctx := newContext(r, w, "SetBucketReplicationTarget")
+// SetRemoteTargetHandler - sets a remote target for bucket
+func (a adminAPIHandlers) SetRemoteTargetHandler(w http.ResponseWriter, r *http.Request) {
+	ctx := newContext(r, w, "SetBucketTarget")
 
-	defer logger.AuditLog(w, r, "SetBucketReplicationTarget", mustGetClaimsFromToken(r))
+	defer logger.AuditLog(w, r, "SetBucketTarget", mustGetClaimsFromToken(r))
 	vars := mux.Vars(r)
 	bucket := vars["bucket"]
 
 	// Get current object layer instance.
-	objectAPI, _ := validateAdminUsersReq(ctx, w, r, iampolicy.SetBucketReplicationTargetAction)
+	objectAPI, _ := validateAdminUsersReq(ctx, w, r, iampolicy.SetBucketTargetAction)
 	if objectAPI == nil {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
 		return
@@ -139,11 +139,6 @@ func (a adminAPIHandlers) SetBucketReplicationTargetHandler(w http.ResponseWrite
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
 		return
 	}
-	// Turn off replication if disk crawl is unavailable.
-	if env.Get(envDataUsageCrawlConf, config.EnableOn) == config.EnableOff {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrBucketReplicationDisabledError), r.URL)
-		return
-	}
 
 	// Check if bucket exists.
 	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
@@ -151,17 +146,11 @@ func (a adminAPIHandlers) SetBucketReplicationTargetHandler(w http.ResponseWrite
 		return
 	}
 
-	if versioned := globalBucketVersioningSys.Enabled(bucket); !versioned {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrReplicationBucketNeedsVersioningError), r.URL)
-		return
-	}
-
 	cred, _, _, s3Err := validateAdminSignature(ctx, r, "")
 	if s3Err != ErrNone {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(s3Err), r.URL)
 		return
 	}
-
 	password := cred.SecretKey
 
 	reqBytes, err := madmin.DecryptData(password, io.LimitReader(r.Body, r.ContentLength))
@@ -169,84 +158,134 @@ func (a adminAPIHandlers) SetBucketReplicationTargetHandler(w http.ResponseWrite
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err), r.URL)
 		return
 	}
-	var target madmin.BucketReplicationTarget
+	var target madmin.BucketTarget
 	if err = json.Unmarshal(reqBytes, &target); err != nil {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err), r.URL)
 		return
 	}
-	target.Arn = globalBucketReplicationSys.getReplicationARN(target.URL())
-	tgtBytes, err := json.Marshal(&target)
+	host, port, err := net.SplitHostPort(target.Endpoint)
 	if err != nil {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err), r.URL)
 		return
 	}
-	if err = globalBucketMetadataSys.Update(bucket, bucketReplicationTargetsFile, tgtBytes); err != nil {
+	sameTarget, _ := isLocalHost(host, port, globalMinioPort)
+	if sameTarget && bucket == target.TargetBucket {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrBucketRemoteIdenticalToSource), r.URL)
+		return
+	}
+	target.SourceBucket = bucket
+	target.Arn = globalBucketTargetSys.getRemoteARN(bucket, &target)
+	if target.Arn == "" {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err), r.URL)
+		return
+	}
+	if err = globalBucketTargetSys.SetTarget(ctx, bucket, &target); err != nil {
+		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
+		return
+	}
+	targets, err := globalBucketTargetSys.ListBucketTargets(ctx, bucket)
+	if err != nil {
+		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
+		return
+
+	}
+	tgtBytes, err := json.Marshal(&targets)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err), r.URL)
+		return
+	}
+
+	if err = globalBucketMetadataSys.Update(bucket, bucketTargetsFile, tgtBytes); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
 		return
 	}
-	if err = globalBucketReplicationSys.SetTarget(ctx, bucket, &target); err != nil {
+
+	data, err := json.Marshal(target.Arn)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+	// Write success response.
+	writeSuccessResponseJSON(w, data)
+}
+
+// ListRemoteTargetsHandler - lists remote target(s) for a bucket or gets a target
+// for a particular ARN type
+func (a adminAPIHandlers) ListRemoteTargetsHandler(w http.ResponseWriter, r *http.Request) {
+	ctx := newContext(r, w, "ListBucketTargets")
+
+	defer logger.AuditLog(w, r, "ListBucketTargets", mustGetClaimsFromToken(r))
+	vars := mux.Vars(r)
+	bucket := vars["bucket"]
+	arnType := vars["type"]
+
+	// Get current object layer instance.
+	objectAPI, _ := validateAdminUsersReq(ctx, w, r, iampolicy.GetBucketTargetAction)
+	if objectAPI == nil {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
+		return
+	}
+	if bucket != "" {
+		if _, err := globalBucketMetadataSys.GetBucketTargetsConfig(bucket); err != nil {
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+	}
+	targets := globalBucketTargetSys.ListTargets(ctx, bucket, arnType)
+	data, err := json.Marshal(targets)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+	// Write success response.
+	writeSuccessResponseJSON(w, data)
+}
+
+// RemoveRemoteTargetHandler - removes a remote target for bucket with specified ARN
+func (a adminAPIHandlers) RemoveRemoteTargetHandler(w http.ResponseWriter, r *http.Request) {
+	ctx := newContext(r, w, "RemoveBucketTarget")
+
+	defer logger.AuditLog(w, r, "RemoveBucketTarget", mustGetClaimsFromToken(r))
+	vars := mux.Vars(r)
+	bucket := vars["bucket"]
+	arn := vars["arn"]
+
+	// Get current object layer instance.
+	objectAPI, _ := validateAdminUsersReq(ctx, w, r, iampolicy.SetBucketTargetAction)
+	if objectAPI == nil {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
+		return
+	}
+	if !globalIsErasure {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
+		return
+	}
+
+	// Check if bucket exists.
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
 
-	// Write success response.
-	writeSuccessResponseHeadersOnly(w)
-}
-
-// GetBucketReplicationTargetsHandler - gets bucket replication targets for a particular bucket
-func (a adminAPIHandlers) GetBucketReplicationTargetsHandler(w http.ResponseWriter, r *http.Request) {
-	ctx := newContext(r, w, "GetBucketReplicationTarget")
-
-	defer logger.AuditLog(w, r, "GetBucketReplicationTarget", mustGetClaimsFromToken(r))
-	vars := mux.Vars(r)
-	bucket := vars["bucket"]
-	// Get current object layer instance.
-	objectAPI, _ := validateAdminUsersReq(ctx, w, r, iampolicy.GetBucketReplicationTargetAction)
-	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
+	if err := globalBucketTargetSys.RemoveTarget(ctx, bucket, arn); err != nil {
+		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
-
-	target, err := globalBucketMetadataSys.GetReplicationTargetConfig(bucket)
+	targets, err := globalBucketTargetSys.ListBucketTargets(ctx, bucket)
 	if err != nil {
-		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
 		return
 	}
-
-	// remove secretKey from creds
-	var tgt madmin.BucketReplicationTarget
-	if !target.Empty() {
-		var creds auth.Credentials
-		creds.AccessKey = target.Credentials.AccessKey
-		tgt = madmin.BucketReplicationTarget{Endpoint: target.Endpoint, TargetBucket: target.TargetBucket, Credentials: &creds, Arn: target.Arn}
-	}
-	data, err := json.Marshal(tgt)
+	tgtBytes, err := json.Marshal(&targets)
 	if err != nil {
-		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err), r.URL)
+		return
+	}
+	if err = globalBucketMetadataSys.Update(bucket, bucketTargetsFile, tgtBytes); err != nil {
+		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
 		return
 	}
-	// Write success response.
-	writeSuccessResponseJSON(w, data)
-}
 
-// GetBucketReplicationARNHandler - gets replication ARN for a particular remote
-func (a adminAPIHandlers) GetBucketReplicationARNHandler(w http.ResponseWriter, r *http.Request) {
-	ctx := newContext(r, w, "GetBucketReplicationARN")
-
-	defer logger.AuditLog(w, r, "GetBucketReplicationARN", mustGetClaimsFromToken(r))
-	vars := mux.Vars(r)
-	rURL := vars["url"]
-	// Get current object layer instance.
-	objectAPI, _ := validateAdminUsersReq(ctx, w, r, iampolicy.GetBucketReplicationTargetAction)
-	if objectAPI == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
-		return
-	}
-	data, err := json.Marshal(globalBucketReplicationSys.getARN(rURL))
-	if err != nil {
-		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
-		return
-	}
 	// Write success response.
-	writeSuccessResponseJSON(w, data)
+	writeSuccessNoContent(w)
 }

cmd/admin-handlers-config-kv.go

@@ -136,7 +136,7 @@ func (a adminAPIHandlers) SetConfigKVHandler(w http.ResponseWriter, r *http.Requ
 		return
 	}
 
-	if err = validateConfig(cfg); err != nil {
+	if err = validateConfig(cfg, objectAPI.SetDriveCount()); err != nil {
 		writeCustomErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAdminConfigBadJSON), err.Error(), r.URL)
 		return
 	}
@@ -271,7 +271,7 @@ func (a adminAPIHandlers) RestoreConfigHistoryKVHandler(w http.ResponseWriter, r
 		return
 	}
 
-	if err = validateConfig(cfg); err != nil {
+	if err = validateConfig(cfg, objectAPI.SetDriveCount()); err != nil {
 		writeCustomErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAdminConfigBadJSON), err.Error(), r.URL)
 		return
 	}
@@ -383,7 +383,7 @@ func (a adminAPIHandlers) SetConfigHandler(w http.ResponseWriter, r *http.Reques
 		return
 	}
 
-	if err = validateConfig(cfg); err != nil {
+	if err = validateConfig(cfg, objectAPI.SetDriveCount()); err != nil {
 		writeCustomErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAdminConfigBadJSON), err.Error(), r.URL)
 		return
 	}

cmd/admin-handlers.go

@@ -695,7 +695,7 @@ func (a adminAPIHandlers) HealHandler(w http.ResponseWriter, r *http.Request) {
 					// Start writing response to client
 					started = true
 					setCommonHeaders(w)
-					w.Header().Set(xhttp.ContentType, "text/event-stream")
+					setEventStreamHeaders(w)
 					// Set 200 OK status
 					w.WriteHeader(200)
 				}
@@ -799,6 +799,59 @@ func (a adminAPIHandlers) HealHandler(w http.ResponseWriter, r *http.Request) {
 	keepConnLive(w, r, respCh)
 }
 
+func getAggregatedBackgroundHealState(ctx context.Context) (madmin.BgHealState, error) {
+	var bgHealStates []madmin.BgHealState
+
+	localHealState, ok := getLocalBackgroundHealStatus()
+	if !ok {
+		return madmin.BgHealState{}, errServerNotInitialized
+	}
+
+	// Get local heal status first
+	bgHealStates = append(bgHealStates, localHealState)
+
+	if globalIsDistErasure {
+		// Get heal status from other peers
+		peersHealStates, nerrs := globalNotificationSys.BackgroundHealStatus()
+		var errCount int
+		for _, nerr := range nerrs {
+			if nerr.Err != nil {
+				logger.LogIf(ctx, nerr.Err)
+				errCount++
+			}
+		}
+		if errCount == len(nerrs) {
+			return madmin.BgHealState{}, fmt.Errorf("all remote servers failed to report heal status, cluster is unhealthy")
+		}
+		bgHealStates = append(bgHealStates, peersHealStates...)
+	}
+
+	// Aggregate healing result
+	var aggregatedHealStateResult = madmin.BgHealState{
+		ScannedItemsCount: bgHealStates[0].ScannedItemsCount,
+		LastHealActivity:  bgHealStates[0].LastHealActivity,
+		NextHealRound:     bgHealStates[0].NextHealRound,
+		HealDisks:         bgHealStates[0].HealDisks,
+	}
+
+	bgHealStates = bgHealStates[1:]
+
+	for _, state := range bgHealStates {
+		aggregatedHealStateResult.ScannedItemsCount += state.ScannedItemsCount
+		aggregatedHealStateResult.HealDisks = append(aggregatedHealStateResult.HealDisks, state.HealDisks...)
+		if !state.LastHealActivity.IsZero() && aggregatedHealStateResult.LastHealActivity.Before(state.LastHealActivity) {
+			aggregatedHealStateResult.LastHealActivity = state.LastHealActivity
+			// The node which has the last heal activity means its
+			// is the node that is orchestrating self healing operations,
+			// which also means it is the same node which decides when
+			// the next self healing operation will be done.
+			aggregatedHealStateResult.NextHealRound = state.NextHealRound
+		}
+	}
+
+	return aggregatedHealStateResult, nil
+}
+
 func (a adminAPIHandlers) BackgroundHealStatusHandler(w http.ResponseWriter, r *http.Request) {
 	ctx := newContext(r, w, "HealBackgroundStatus")
 
@@ -815,39 +868,13 @@ func (a adminAPIHandlers) BackgroundHealStatusHandler(w http.ResponseWriter, r *
 		return
 	}
 
-	var bgHealStates []madmin.BgHealState
-
-	// Get local heal status first
-	bgHealStates = append(bgHealStates, getLocalBackgroundHealStatus())
-
-	if globalIsDistErasure {
-		// Get heal status from other peers
-		peersHealStates := globalNotificationSys.BackgroundHealStatus()
-		bgHealStates = append(bgHealStates, peersHealStates...)
+	aggregateHealStateResult, err := getAggregatedBackgroundHealState(r.Context())
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
 	}
 
-	// Aggregate healing result
-	var aggregatedHealStateResult = madmin.BgHealState{
-		ScannedItemsCount: bgHealStates[0].ScannedItemsCount,
-		LastHealActivity:  bgHealStates[0].LastHealActivity,
-		NextHealRound:     bgHealStates[0].NextHealRound,
-	}
-
-	bgHealStates = bgHealStates[1:]
-
-	for _, state := range bgHealStates {
-		aggregatedHealStateResult.ScannedItemsCount += state.ScannedItemsCount
-		if !state.LastHealActivity.IsZero() && aggregatedHealStateResult.LastHealActivity.Before(state.LastHealActivity) {
-			aggregatedHealStateResult.LastHealActivity = state.LastHealActivity
-			// The node which has the last heal activity means its
-			// is the node that is orchestrating self healing operations,
-			// which also means it is the same node which decides when
-			// the next self healing operation will be done.
-			aggregatedHealStateResult.NextHealRound = state.NextHealRound
-		}
-	}
-
-	if err := json.NewEncoder(w).Encode(aggregatedHealStateResult); err != nil {
+	if err := json.NewEncoder(w).Encode(aggregateHealStateResult); err != nil {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
 	}
@@ -995,7 +1022,7 @@ func (a adminAPIHandlers) TraceHandler(w http.ResponseWriter, r *http.Request) {
 		return
 	}
 
-	w.Header().Set(xhttp.ContentType, "text/event-stream")
+	setEventStreamHeaders(w)
 
 	// Trace Publisher and peer-trace-client uses nonblocking send and hence does not wait for slow receivers.
 	// Use buffered channel to take care of burst sends or slow w.Write()
@@ -1063,8 +1090,9 @@ func (a adminAPIHandlers) ConsoleLogHandler(w http.ResponseWriter, r *http.Reque
 	// Avoid reusing tcp connection if read timeout is hit
 	// This is needed to make r.Context().Done() work as
 	// expected in case of read timeout
-	w.Header().Add("Connection", "close")
-	w.Header().Set(xhttp.ContentType, "text/event-stream")
+	w.Header().Set("Connection", "close")
+
+	setEventStreamHeaders(w)
 
 	logCh := make(chan interface{}, 4000)
 
@@ -1213,7 +1241,7 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 		return
 	}
 
-	vars := mux.Vars(r)
+	query := r.URL.Query()
 	obdInfo := madmin.OBDInfo{}
 	obdInfoCh := make(chan madmin.OBDInfo)
 
@@ -1223,7 +1251,9 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 	}
 
 	setCommonHeaders(w)
-	w.Header().Set(xhttp.ContentType, "text/event-stream")
+
+	setEventStreamHeaders(w)
+
 	w.WriteHeader(http.StatusOK)
 
 	errResp := func(err error) {
@@ -1245,10 +1275,9 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 	}
 
 	deadlinedCtx, cancel := context.WithTimeout(ctx, deadline)
-
 	defer cancel()
 
-	nsLock := objectAPI.NewNSLock(deadlinedCtx, minioMetaBucket, "obd-in-progress")
+	nsLock := objectAPI.NewNSLock(ctx, minioMetaBucket, "obd-in-progress")
 	if err := nsLock.GetLock(newDynamicTimeout(deadline, deadline)); err != nil { // returns a locked lock
 		errResp(err)
 		return
@@ -1258,7 +1287,13 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 	go func() {
 		defer close(obdInfoCh)
 
-		if cpu, ok := vars["syscpu"]; ok && cpu == "true" {
+		if log := query.Get("log"); log == "true" {
+			obdInfo.Logging.ServersLog = append(obdInfo.Logging.ServersLog, getLocalLogOBD(deadlinedCtx, r))
+			obdInfo.Logging.ServersLog = append(obdInfo.Logging.ServersLog, globalNotificationSys.LogOBDInfo(deadlinedCtx)...)
+			partialWrite(obdInfo)
+		}
+
+		if cpu := query.Get("syscpu"); cpu == "true" {
 			cpuInfo := getLocalCPUOBDInfo(deadlinedCtx, r)
 
 			obdInfo.Sys.CPUInfo = append(obdInfo.Sys.CPUInfo, cpuInfo)
@@ -1266,7 +1301,7 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 			partialWrite(obdInfo)
 		}
 
-		if diskHw, ok := vars["sysdiskhw"]; ok && diskHw == "true" {
+		if diskHw := query.Get("sysdiskhw"); diskHw == "true" {
 			diskHwInfo := getLocalDiskHwOBD(deadlinedCtx, r)
 
 			obdInfo.Sys.DiskHwInfo = append(obdInfo.Sys.DiskHwInfo, diskHwInfo)
@@ -1274,7 +1309,7 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 			partialWrite(obdInfo)
 		}
 
-		if osInfo, ok := vars["sysosinfo"]; ok && osInfo == "true" {
+		if osInfo := query.Get("sysosinfo"); osInfo == "true" {
 			osInfo := getLocalOsInfoOBD(deadlinedCtx, r)
 
 			obdInfo.Sys.OsInfo = append(obdInfo.Sys.OsInfo, osInfo)
@@ -1282,7 +1317,7 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 			partialWrite(obdInfo)
 		}
 
-		if mem, ok := vars["sysmem"]; ok && mem == "true" {
+		if mem := query.Get("sysmem"); mem == "true" {
 			memInfo := getLocalMemOBD(deadlinedCtx, r)
 
 			obdInfo.Sys.MemInfo = append(obdInfo.Sys.MemInfo, memInfo)
@@ -1290,7 +1325,7 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 			partialWrite(obdInfo)
 		}
 
-		if proc, ok := vars["sysprocess"]; ok && proc == "true" {
+		if proc := query.Get("sysprocess"); proc == "true" {
 			procInfo := getLocalProcOBD(deadlinedCtx, r)
 
 			obdInfo.Sys.ProcInfo = append(obdInfo.Sys.ProcInfo, procInfo)
@@ -1298,14 +1333,14 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 			partialWrite(obdInfo)
 		}
 
-		if config, ok := vars["minioconfig"]; ok && config == "true" {
+		if config := query.Get("minioconfig"); config == "true" {
 			cfg, err := readServerConfig(ctx, objectAPI)
 			logger.LogIf(ctx, err)
 			obdInfo.Minio.Config = cfg
 			partialWrite(obdInfo)
 		}
 
-		if drive, ok := vars["perfdrive"]; ok && drive == "true" {
+		if drive := query.Get("perfdrive"); drive == "true" {
 			// Get drive obd details from local server's drive(s)
 			driveOBDSerial := getLocalDrivesOBD(deadlinedCtx, false, globalEndpoints, r)
 			driveOBDParallel := getLocalDrivesOBD(deadlinedCtx, true, globalEndpoints, r)
@@ -1336,7 +1371,7 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 			partialWrite(obdInfo)
 		}
 
-		if net, ok := vars["perfnet"]; ok && net == "true" && globalIsDistErasure {
+		if net := query.Get("perfnet"); net == "true" && globalIsDistErasure {
 			obdInfo.Perf.Net = append(obdInfo.Perf.Net, globalNotificationSys.NetOBDInfo(deadlinedCtx))
 			partialWrite(obdInfo)
 
@@ -1350,6 +1385,7 @@ func (a adminAPIHandlers) OBDInfoHandler(w http.ResponseWriter, r *http.Request)
 			obdInfo.Perf.NetParallel = globalNotificationSys.NetOBDParallelInfo(deadlinedCtx)
 			partialWrite(obdInfo)
 		}
+
 	}()
 
 	ticker := time.NewTicker(30 * time.Second)
@@ -1465,17 +1501,34 @@ func (a adminAPIHandlers) ServerInfoHandler(w http.ResponseWriter, r *http.Reque
 		Notifications: notifyTarget,
 	}
 
+	// Collect any disk healing.
+	healing, _ := getAggregatedBackgroundHealState(ctx)
+	healDisks := make(map[string]struct{}, len(healing.HealDisks))
+	for _, disk := range healing.HealDisks {
+		healDisks[disk] = struct{}{}
+	}
+
 	// find all disks which belong to each respective endpoints
 	for i := range servers {
 		for _, disk := range storageInfo.Disks {
 			if strings.Contains(disk.Endpoint, servers[i].Endpoint) {
+				if _, ok := healDisks[disk.Endpoint]; ok {
+					disk.Healing = true
+				}
 				servers[i].Disks = append(servers[i].Disks, disk)
 			}
 		}
 	}
+
 	// add all the disks local to this server.
 	for _, disk := range storageInfo.Disks {
+		if disk.DrivePath == "" && disk.Endpoint == "" {
+			continue
+		}
 		if disk.Endpoint == disk.DrivePath {
+			if _, ok := healDisks[disk.Endpoint]; ok {
+				disk.Healing = true
+			}
 			servers[len(servers)-1].Disks = append(servers[len(servers)-1].Disks, disk)
 		}
 	}
@@ -1511,7 +1564,7 @@ func fetchLambdaInfo(cfg config.Config) []map[string][]madmin.TargetIDStatus {
 	// Fetch the configured targets
 	tr := NewGatewayHTTPTransport()
 	defer tr.CloseIdleConnections()
-	targetList, err := notify.FetchRegisteredTargets(cfg, GlobalContext.Done(), tr, true, false)
+	targetList, err := notify.FetchRegisteredTargets(GlobalContext, cfg, tr, true, false)
 	if err != nil && err != notify.ErrTargetsOffline {
 		logger.LogIf(GlobalContext, err)
 		return nil
@@ -1555,12 +1608,12 @@ func fetchVaultStatus(cfg config.Config) madmin.Vault {
 	keyID := GlobalKMS.DefaultKeyID()
 	kmsInfo := GlobalKMS.Info()
 
-	if kmsInfo.Endpoint == "" {
+	if len(kmsInfo.Endpoints) == 0 {
 		vault.Status = "KMS configured using master key"
 		return vault
 	}
 
-	if err := checkConnection(kmsInfo.Endpoint, 15*time.Second); err != nil {
+	if err := checkConnection(kmsInfo.Endpoints[0], 15*time.Second); err != nil {
 		vault.Status = "offline"
 	} else {
 		vault.Status = "online"
@@ -1628,23 +1681,32 @@ func fetchLoggerInfo(cfg config.Config) ([]madmin.Logger, []madmin.Audit) {
 
 // checkConnection - ping an endpoint , return err in case of no connection
 func checkConnection(endpointStr string, timeout time.Duration) error {
-	tr := newCustomHTTPTransport(&tls.Config{RootCAs: globalRootCAs}, timeout)()
-	defer tr.CloseIdleConnections()
-
 	ctx, cancel := context.WithTimeout(GlobalContext, timeout)
 	defer cancel()
 
-	req, err := http.NewRequest(http.MethodHead, endpointStr, nil)
+	client := &http.Client{Transport: &http.Transport{
+		Proxy:                 http.ProxyFromEnvironment,
+		DialContext:           xhttp.NewCustomDialContext(timeout),
+		ResponseHeaderTimeout: 5 * time.Second,
+		TLSHandshakeTimeout:   5 * time.Second,
+		ExpectContinueTimeout: 5 * time.Second,
+		TLSClientConfig:       &tls.Config{RootCAs: globalRootCAs},
+		// Go net/http automatically unzip if content-type is
+		// gzip disable this feature, as we are always interested
+		// in raw stream.
+		DisableCompression: true,
+	}}
+	defer client.CloseIdleConnections()
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodHead, endpointStr, nil)
 	if err != nil {
 		return err
 	}
 
-	client := &http.Client{Transport: tr}
-	resp, err := client.Do(req.WithContext(ctx))
+	resp, err := client.Do(req)
 	if err != nil {
 		return err
 	}
 	defer xhttp.DrainBody(resp.Body)
-	resp.Body.Close()
 	return nil
 }

cmd/admin-heal-ops.go

@@ -88,13 +88,15 @@ type allHealState struct {
 	sync.Mutex
 
 	// map of heal path to heal sequence
-	healSeqMap map[string]*healSequence
+	healSeqMap     map[string]*healSequence
+	healLocalDisks map[Endpoint]struct{}
 }
 
 // newHealState - initialize global heal state management
 func newHealState() *allHealState {
 	healState := &allHealState{
-		healSeqMap: make(map[string]*healSequence),
+		healSeqMap:     make(map[string]*healSequence),
+		healLocalDisks: map[Endpoint]struct{}{},
 	}
 
 	go healState.periodicHealSeqsClean(GlobalContext)
@@ -102,6 +104,42 @@ func newHealState() *allHealState {
 	return healState
 }
 
+func (ahs *allHealState) healDriveCount() int {
+	ahs.Lock()
+	defer ahs.Unlock()
+
+	return len(ahs.healLocalDisks)
+}
+
+func (ahs *allHealState) getHealLocalDisks() Endpoints {
+	ahs.Lock()
+	defer ahs.Unlock()
+
+	var healLocalDisks Endpoints
+	for ep := range ahs.healLocalDisks {
+		healLocalDisks = append(healLocalDisks, ep)
+	}
+	return healLocalDisks
+}
+
+func (ahs *allHealState) popHealLocalDisks(healLocalDisks ...Endpoint) {
+	ahs.Lock()
+	defer ahs.Unlock()
+
+	for _, ep := range healLocalDisks {
+		delete(ahs.healLocalDisks, ep)
+	}
+}
+
+func (ahs *allHealState) pushHealLocalDisks(healLocalDisks ...Endpoint) {
+	ahs.Lock()
+	defer ahs.Unlock()
+
+	for _, ep := range healLocalDisks {
+		ahs.healLocalDisks[ep] = struct{}{}
+	}
+}
+
 func (ahs *allHealState) periodicHealSeqsClean(ctx context.Context) {
 	// Launch clean-up routine to remove this heal sequence (after
 	// it ends) from the global state after timeout has elapsed.
@@ -485,9 +523,12 @@ func (h *healSequence) isQuitting() bool {
 // check if the heal sequence has ended
 func (h *healSequence) hasEnded() bool {
 	h.mutex.RLock()
-	ended := len(h.currentStatus.Items) == 0 || h.currentStatus.Summary == healStoppedStatus || h.currentStatus.Summary == healFinishedStatus
-	h.mutex.RUnlock()
-	return ended
+	defer h.mutex.RUnlock()
+	// background heal never ends
+	if h.clientToken == bgHealingUUID {
+		return false
+	}
+	return len(h.currentStatus.Items) == 0 || h.currentStatus.Summary == healStoppedStatus || h.currentStatus.Summary == healFinishedStatus
 }
 
 // stops the heal sequence - safe to call multiple times.
@@ -627,6 +668,12 @@ func (h *healSequence) queueHealTask(source healSource, healType madmin.HealItem
 	if source.opts != nil {
 		task.opts = *source.opts
 	}
+
+	h.mutex.Lock()
+	h.scannedItemsMap[healType]++
+	h.lastHealActivity = UTCNow()
+	h.mutex.Unlock()
+
 	globalBackgroundHealRoutine.queueHealTask(task)
 
 	select {
@@ -634,9 +681,11 @@ func (h *healSequence) queueHealTask(source healSource, healType madmin.HealItem
 		if !h.reportProgress {
 			// Object might have been deleted, by the time heal
 			// was attempted, we should ignore this object and
-			// return success.
+			// return the error and not calculate this object
+			// as part of the metrics.
 			if isErrObjectNotFound(res.err) || isErrVersionNotFound(res.err) {
-				return nil
+				// Return the error so that caller can handle it.
+				return res.err
 			}
 
 			h.mutex.Lock()
@@ -700,14 +749,13 @@ func (h *healSequence) healItemsFromSourceCh() error {
 			if err := h.queueHealTask(source, itemType); err != nil {
 				switch err.(type) {
 				case ObjectExistsAsDirectory:
+				case ObjectNotFound:
+				case VersionNotFound:
 				default:
 					logger.LogIf(h.ctx, fmt.Errorf("Heal attempt failed for %s: %w",
 						pathJoin(source.bucket, source.object), err))
 				}
 			}
-
-			h.scannedItemsMap[itemType]++
-			h.lastHealActivity = UTCNow()
 		case <-h.ctx.Done():
 			return nil
 		}
@@ -773,17 +821,17 @@ func (h *healSequence) healMinioSysMeta(metaPrefix string) func() error {
 				return errHealStopSignalled
 			}
 
-			herr := h.queueHealTask(healSource{
+			err := h.queueHealTask(healSource{
 				bucket:    bucket,
 				object:    object,
 				versionID: versionID,
 			}, madmin.HealItemBucketMetadata)
 			// Object might have been deleted, by the time heal
 			// was attempted we ignore this object an move on.
-			if isErrObjectNotFound(herr) || isErrVersionNotFound(herr) {
+			if isErrObjectNotFound(err) || isErrVersionNotFound(err) {
 				return nil
 			}
-			return herr
+			return err
 		})
 	}
 }
@@ -844,7 +892,9 @@ func (h *healSequence) healBucket(bucket string, bucketsOnly bool) error {
 	}
 
 	if err := h.queueHealTask(healSource{bucket: bucket}, madmin.HealItemBucket); err != nil {
-		return err
+		if !isErrObjectNotFound(err) && !isErrVersionNotFound(err) {
+			return err
+		}
 	}
 
 	if bucketsOnly {
@@ -855,9 +905,12 @@ func (h *healSequence) healBucket(bucket string, bucketsOnly bool) error {
 		if h.object != "" {
 			// Check if an object named as the objPrefix exists,
 			// and if so heal it.
-			_, err := objectAPI.GetObjectInfo(h.ctx, bucket, h.object, ObjectOptions{})
+			oi, err := objectAPI.GetObjectInfo(h.ctx, bucket, h.object, ObjectOptions{})
 			if err == nil {
-				if err = h.healObject(bucket, h.object, ""); err != nil {
+				if err = h.healObject(bucket, h.object, oi.VersionID); err != nil {
+					if isErrObjectNotFound(err) || isErrVersionNotFound(err) {
+						return nil
+					}
 					return err
 				}
 			}
@@ -867,7 +920,11 @@ func (h *healSequence) healBucket(bucket string, bucketsOnly bool) error {
 	}
 
 	if err := objectAPI.HealObjects(h.ctx, bucket, h.object, h.settings, h.healObject); err != nil {
-		return errFnHealFromAPIErr(h.ctx, err)
+		// Object might have been deleted, by the time heal
+		// was attempted we ignore this object an move on.
+		if !isErrObjectNotFound(err) && !isErrVersionNotFound(err) {
+			return errFnHealFromAPIErr(h.ctx, err)
+		}
 	}
 	return nil
 }
@@ -884,9 +941,10 @@ func (h *healSequence) healObject(bucket, object, versionID string) error {
 		return errHealStopSignalled
 	}
 
-	return h.queueHealTask(healSource{
+	err := h.queueHealTask(healSource{
 		bucket:    bucket,
 		object:    object,
 		versionID: versionID,
 	}, madmin.HealItemObject)
+	return err
 }

cmd/admin-router.go

@@ -182,15 +182,15 @@ func registerAdminRouter(router *mux.Router, enableConfigOps, enableIAMOps bool)
 					httpTraceHdrs(adminAPI.PutBucketQuotaConfigHandler)).Queries("bucket", "{bucket:.*}")
 			}
 			// Bucket replication operations
-			// GetBucketReplicationTargetHandler
-			adminRouter.Methods(http.MethodGet).Path(adminVersion+"/get-bucket-replication-target").HandlerFunc(
-				httpTraceHdrs(adminAPI.GetBucketReplicationTargetsHandler)).Queries("bucket", "{bucket:.*}")
-			// GetBucketReplicationARN	Handler
-			adminRouter.Methods(http.MethodGet).Path(adminVersion+"/get-bucket-replication-arn").HandlerFunc(
-				httpTraceHdrs(adminAPI.GetBucketReplicationARNHandler)).Queries("url", "{url:.*}")
-			// SetBucketReplicationTargetHandler
-			adminRouter.Methods(http.MethodPut).Path(adminVersion+"/set-bucket-replication-target").HandlerFunc(
-				httpTraceHdrs(adminAPI.SetBucketReplicationTargetHandler)).Queries("bucket", "{bucket:.*}")
+			// GetBucketTargetHandler
+			adminRouter.Methods(http.MethodGet).Path(adminVersion+"/list-remote-targets").HandlerFunc(
+				httpTraceHdrs(adminAPI.ListRemoteTargetsHandler)).Queries("bucket", "{bucket:.*}", "type", "{type:.*}")
+			// SetRemoteTargetHandler
+			adminRouter.Methods(http.MethodPut).Path(adminVersion+"/set-remote-target").HandlerFunc(
+				httpTraceHdrs(adminAPI.SetRemoteTargetHandler)).Queries("bucket", "{bucket:.*}")
+			// SetRemoteTargetHandler
+			adminRouter.Methods(http.MethodDelete).Path(adminVersion+"/remove-remote-target").HandlerFunc(
+				httpTraceHdrs(adminAPI.RemoveRemoteTargetHandler)).Queries("bucket", "{bucket:.*}", "arn", "{arn:.*}")
 		}
 
 		// -- Top APIs --
@@ -212,18 +212,8 @@ func registerAdminRouter(router *mux.Router, enableConfigOps, enableIAMOps bool)
 
 		if !globalIsGateway {
 			// -- OBD API --
-			adminRouter.Methods(http.MethodGet).Path(adminVersion+"/obdinfo").
-				HandlerFunc(httpTraceHdrs(adminAPI.OBDInfoHandler)).
-				Queries("perfdrive", "{perfdrive:true|false}",
-					"perfnet", "{perfnet:true|false}",
-					"minioinfo", "{minioinfo:true|false}",
-					"minioconfig", "{minioconfig:true|false}",
-					"syscpu", "{syscpu:true|false}",
-					"sysdiskhw", "{sysdiskhw:true|false}",
-					"sysosinfo", "{sysosinfo:true|false}",
-					"sysmem", "{sysmem:true|false}",
-					"sysprocess", "{sysprocess:true|false}",
-				)
+			adminRouter.Methods(http.MethodGet).Path(adminVersion + "/obdinfo").
+				HandlerFunc(httpTraceHdrs(adminAPI.OBDInfoHandler))
 		}
 	}
 

cmd/api-errors.go

@@ -29,7 +29,7 @@ import (
 
 	minio "github.com/minio/minio-go/v7"
 	"github.com/minio/minio-go/v7/pkg/tags"
-	"github.com/minio/minio/cmd/config/etcd/dns"
+	"github.com/minio/minio/cmd/config/dns"
 	"github.com/minio/minio/cmd/crypto"
 	"github.com/minio/minio/cmd/logger"
 	"github.com/minio/minio/pkg/auth"
@@ -107,8 +107,16 @@ const (
 	ErrNoSuchWebsiteConfiguration
 	ErrReplicationConfigurationNotFoundError
 	ErrReplicationDestinationNotFoundError
+	ErrReplicationDestinationMissingLock
 	ErrReplicationTargetNotFoundError
+	ErrReplicationRemoteConnectionError
+	ErrBucketRemoteIdenticalToSource
+	ErrBucketRemoteAlreadyExists
+	ErrBucketRemoteArnTypeInvalid
+	ErrBucketRemoteArnInvalid
+	ErrBucketRemoteRemoveDisallowed
 	ErrReplicationTargetNotVersionedError
+	ErrReplicationSourceNotVersionedError
 	ErrReplicationNeedsVersioningError
 	ErrReplicationBucketNeedsVersioningError
 	ErrBucketReplicationDisabledError
@@ -656,20 +664,9 @@ var errorCodes = errorCodeMap{
 		Description:    "X-Amz-Date must be in the ISO8601 Long Format \"yyyyMMdd'T'HHmmss'Z'\"",
 		HTTPStatusCode: http.StatusBadRequest,
 	},
-	// FIXME: Should contain the invalid param set as seen in https://github.com/minio/minio/issues/2385.
-	// right Description:    "Error parsing the X-Amz-Credential parameter; incorrect date format \"%s\". This date in the credential must be in the format \"yyyyMMdd\".",
-	// Need changes to make sure variable messages can be constructed.
 	ErrMalformedCredentialDate: {
 		Code:           "AuthorizationQueryParametersError",
-		Description:    "Error parsing the X-Amz-Credential parameter; incorrect date format \"%s\". This date in the credential must be in the format \"yyyyMMdd\".",
-		HTTPStatusCode: http.StatusBadRequest,
-	},
-	// FIXME: Should contain the invalid param set as seen in https://github.com/minio/minio/issues/2385.
-	// right Description:    "Error parsing the X-Amz-Credential parameter; the region 'us-east-' is wrong; expecting 'us-east-1'".
-	// Need changes to make sure variable messages can be constructed.
-	ErrMalformedCredentialRegion: {
-		Code:           "AuthorizationQueryParametersError",
-		Description:    "Error parsing the X-Amz-Credential parameter; the region is wrong;",
+		Description:    "Error parsing the X-Amz-Credential parameter; incorrect date format. This date in the credential must be in the format \"yyyyMMdd\".",
 		HTTPStatusCode: http.StatusBadRequest,
 	},
 	ErrInvalidRegion: {
@@ -677,9 +674,6 @@ var errorCodes = errorCodeMap{
 		Description:    "Region does not match.",
 		HTTPStatusCode: http.StatusBadRequest,
 	},
-	// FIXME: Should contain the invalid param set as seen in https://github.com/minio/minio/issues/2385.
-	// right Description:   "Error parsing the X-Amz-Credential parameter; incorrect service \"s4\". This endpoint belongs to \"s3\".".
-	// Need changes to make sure variable messages can be constructed.
 	ErrInvalidServiceS3: {
 		Code:           "AuthorizationParametersError",
 		Description:    "Error parsing the Credential/X-Amz-Credential parameter; incorrect service. This endpoint belongs to \"s3\".",
@@ -690,9 +684,6 @@ var errorCodes = errorCodeMap{
 		Description:    "Error parsing the Credential parameter; incorrect service. This endpoint belongs to \"sts\".",
 		HTTPStatusCode: http.StatusBadRequest,
 	},
-	// FIXME: Should contain the invalid param set as seen in https://github.com/minio/minio/issues/2385.
-	// Description:   "Error parsing the X-Amz-Credential parameter; incorrect terminal "aws4_reque". This endpoint uses "aws4_request".
-	// Need changes to make sure variable messages can be constructed.
 	ErrInvalidRequestVersion: {
 		Code:           "AuthorizationQueryParametersError",
 		Description:    "Error parsing the X-Amz-Credential parameter; incorrect terminal. This endpoint uses \"aws4_request\".",
@@ -763,8 +754,6 @@ var errorCodes = errorCodeMap{
 		Description:    "Your key is too long",
 		HTTPStatusCode: http.StatusBadRequest,
 	},
-
-	// FIXME: Actual XML error response also contains the header which missed in list of signed header parameters.
 	ErrUnsignedHeaders: {
 		Code:           "AccessDenied",
 		Description:    "There were headers present in the request which were not signed",
@@ -825,15 +814,55 @@ var errorCodes = errorCodeMap{
 		Description:    "The replication destination bucket does not exist",
 		HTTPStatusCode: http.StatusNotFound,
 	},
+	ErrReplicationDestinationMissingLock: {
+		Code:           "ReplicationDestinationMissingLockError",
+		Description:    "The replication destination bucket does not have object locking enabled",
+		HTTPStatusCode: http.StatusBadRequest,
+	},
 	ErrReplicationTargetNotFoundError: {
-		Code:           "ReplicationTargetNotFoundError",
+		Code:           "XminioAdminReplicationTargetNotFoundError",
 		Description:    "The replication target does not exist",
 		HTTPStatusCode: http.StatusNotFound,
 	},
+	ErrReplicationRemoteConnectionError: {
+		Code:           "XminioAdminReplicationRemoteConnectionError",
+		Description:    "Remote service endpoint or target bucket not available",
+		HTTPStatusCode: http.StatusNotFound,
+	},
+	ErrBucketRemoteIdenticalToSource: {
+		Code:           "XminioAdminRemoteIdenticalToSource",
+		Description:    "The remote target cannot be identical to source",
+		HTTPStatusCode: http.StatusBadRequest,
+	},
+	ErrBucketRemoteAlreadyExists: {
+		Code:           "XminioAdminBucketRemoteAlreadyExists",
+		Description:    "The remote target already exists",
+		HTTPStatusCode: http.StatusBadRequest,
+	},
+	ErrBucketRemoteRemoveDisallowed: {
+		Code:           "XMinioAdminRemoteRemoveDisallowed",
+		Description:    "Replication configuration exists with this ARN.",
+		HTTPStatusCode: http.StatusBadRequest,
+	},
+	ErrBucketRemoteArnTypeInvalid: {
+		Code:           "XMinioAdminRemoteARNTypeInvalid",
+		Description:    "The bucket remote ARN type is not valid",
+		HTTPStatusCode: http.StatusBadRequest,
+	},
+	ErrBucketRemoteArnInvalid: {
+		Code:           "XMinioAdminRemoteArnInvalid",
+		Description:    "The bucket remote ARN does not have correct format",
+		HTTPStatusCode: http.StatusBadRequest,
+	},
 	ErrReplicationTargetNotVersionedError: {
 		Code:           "ReplicationTargetNotVersionedError",
 		Description:    "The replication target does not have versioning enabled",
-		HTTPStatusCode: http.StatusNotFound,
+		HTTPStatusCode: http.StatusBadRequest,
+	},
+	ErrReplicationSourceNotVersionedError: {
+		Code:           "ReplicationSourceNotVersionedError",
+		Description:    "The replication source does not have versioning enabled",
+		HTTPStatusCode: http.StatusBadRequest,
 	},
 	ErrReplicationNeedsVersioningError: {
 		Code:           "InvalidRequest",
@@ -942,7 +971,7 @@ var errorCodes = errorCodeMap{
 		HTTPStatusCode: http.StatusBadRequest,
 	},
 	ErrMetadataTooLarge: {
-		Code:           "InvalidArgument",
+		Code:           "MetadataTooLarge",
 		Description:    "Your metadata headers exceed the maximum allowed metadata size.",
 		HTTPStatusCode: http.StatusBadRequest,
 	},
@@ -1879,10 +1908,24 @@ func toAPIErrorCode(ctx context.Context, err error) (apiErr APIErrorCode) {
 		apiErr = ErrReplicationConfigurationNotFoundError
 	case BucketReplicationDestinationNotFound:
 		apiErr = ErrReplicationDestinationNotFoundError
-	case BucketReplicationTargetNotFound:
+	case BucketReplicationDestinationMissingLock:
+		apiErr = ErrReplicationDestinationMissingLock
+	case BucketRemoteTargetNotFound:
 		apiErr = ErrReplicationTargetNotFoundError
+	case BucketRemoteConnectionErr:
+		apiErr = ErrReplicationRemoteConnectionError
+	case BucketRemoteAlreadyExists:
+		apiErr = ErrBucketRemoteAlreadyExists
+	case BucketRemoteArnTypeInvalid:
+		apiErr = ErrBucketRemoteArnTypeInvalid
+	case BucketRemoteArnInvalid:
+		apiErr = ErrBucketRemoteArnInvalid
+	case BucketRemoteRemoveDisallowed:
+		apiErr = ErrBucketRemoteRemoveDisallowed
 	case BucketReplicationTargetNotVersioned:
 		apiErr = ErrReplicationTargetNotVersionedError
+	case BucketReplicationSourceNotVersioned:
+		apiErr = ErrReplicationSourceNotVersionedError
 	case BucketQuotaExceeded:
 		apiErr = ErrAdminBucketQuotaExceeded
 	case *event.ErrInvalidEventName:

cmd/api-headers.go

@@ -38,6 +38,13 @@ func mustGetRequestID(t time.Time) string {
 	return fmt.Sprintf("%X", t.UnixNano())
 }
 
+// setEventStreamHeaders to allow proxies to avoid buffering proxy responses
+func setEventStreamHeaders(w http.ResponseWriter) {
+	w.Header().Set(xhttp.ContentType, "text/event-stream")
+	w.Header().Set(xhttp.CacheControl, "no-cache") // nginx to turn off buffering
+	w.Header().Set("X-Accel-Buffering", "no")      // nginx to turn off buffering
+}
+
 // Write http common headers
 func setCommonHeaders(w http.ResponseWriter) {
 	w.Header().Set(xhttp.ServerInfo, "MinIO/"+ReleaseTag)
@@ -121,6 +128,11 @@ func setObjectHeaders(w http.ResponseWriter, objInfo ObjectInfo, rs *HTTPRangeSp
 			// values to client.
 			continue
 		}
+
+		// https://github.com/google/security-research/security/advisories/GHSA-76wf-9vgp-pj7w
+		if strings.EqualFold(k, xhttp.AmzMetaUnencryptedContentLength) || strings.EqualFold(k, xhttp.AmzMetaUnencryptedContentMD5) {
+			continue
+		}
 		var isSet bool
 		for _, userMetadataPrefix := range userMetadataKeyPrefixes {
 			if !strings.HasPrefix(k, userMetadataPrefix) {

cmd/api-response.go

@@ -20,6 +20,7 @@ import (
 	"context"
 	"encoding/base64"
 	"encoding/xml"
+	"fmt"
 	"net/http"
 	"net/url"
 	"path"
@@ -35,7 +36,7 @@ import (
 const (
 	// RFC3339 a subset of the ISO8601 timestamp format. e.g 2014-04-29T18:30:38Z
 	iso8601TimeFormat = "2006-01-02T15:04:05.000Z" // Reply date format with nanosecond precision.
-	maxObjectList     = 10000                      // Limit number of objects in a listObjectsResponse/listObjectsVersionsResponse.
+	maxObjectList     = 1000                       // Limit number of objects in a listObjectsResponse/listObjectsVersionsResponse.
 	maxDeleteList     = 10000                      // Limit number of objects deleted in a delete call.
 	maxUploadsList    = 10000                      // Limit number of uploads in a listUploadsResponse.
 	maxPartsList      = 10000                      // Limit number of parts in a listPartsResponse.
@@ -407,7 +408,7 @@ func getObjectLocation(r *http.Request, domains []string, bucket, object string)
 // generates ListBucketsResponse from array of BucketInfo which can be
 // serialized to match XML and JSON API spec output.
 func generateListBucketsResponse(buckets []BucketInfo) ListBucketsResponse {
-	var listbuckets []Bucket
+	listbuckets := make([]Bucket, 0, len(buckets))
 	var data = ListBucketsResponse{}
 	var owner = Owner{}
 
@@ -427,8 +428,7 @@ 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 {
-	var versions []ObjectVersion
-	var prefixes []CommonPrefix
+	versions := make([]ObjectVersion, 0, len(resp.Objects))
 	var owner = Owner{}
 	var data = ListVersionsResponse{}
 
@@ -472,6 +472,7 @@ func generateListVersionsResponse(bucket, prefix, marker, versionIDMarker, delim
 	data.VersionIDMarker = versionIDMarker
 	data.IsTruncated = resp.IsTruncated
 
+	prefixes := make([]CommonPrefix, 0, len(resp.Prefixes))
 	for _, prefix := range resp.Prefixes {
 		var prefixItem = CommonPrefix{}
 		prefixItem.Prefix = s3EncodeName(prefix, encodingType)
@@ -483,8 +484,7 @@ func generateListVersionsResponse(bucket, prefix, marker, versionIDMarker, delim
 
 // generates an ListObjectsV1 response for the said bucket with other enumerated options.
 func generateListObjectsV1Response(bucket, prefix, marker, delimiter, encodingType string, maxKeys int, resp ListObjectsInfo) ListObjectsResponse {
-	var contents []Object
-	var prefixes []CommonPrefix
+	contents := make([]Object, 0, len(resp.Objects))
 	var owner = Owner{}
 	var data = ListObjectsResponse{}
 
@@ -516,9 +516,10 @@ func generateListObjectsV1Response(bucket, prefix, marker, delimiter, encodingTy
 	data.Marker = s3EncodeName(marker, encodingType)
 	data.Delimiter = s3EncodeName(delimiter, encodingType)
 	data.MaxKeys = maxKeys
-
 	data.NextMarker = s3EncodeName(resp.NextMarker, encodingType)
 	data.IsTruncated = resp.IsTruncated
+
+	prefixes := make([]CommonPrefix, 0, len(resp.Prefixes))
 	for _, prefix := range resp.Prefixes {
 		var prefixItem = CommonPrefix{}
 		prefixItem.Prefix = s3EncodeName(prefix, encodingType)
@@ -530,8 +531,7 @@ func generateListObjectsV1Response(bucket, prefix, marker, delimiter, encodingTy
 
 // generates an ListObjectsV2 response for the said bucket with other enumerated options.
 func generateListObjectsV2Response(bucket, prefix, token, nextToken, startAfter, delimiter, encodingType string, fetchOwner, isTruncated bool, maxKeys int, objects []ObjectInfo, prefixes []string, metadata bool) ListObjectsV2Response {
-	var contents []Object
-	var commonPrefixes []CommonPrefix
+	contents := make([]Object, 0, len(objects))
 	var owner = Owner{}
 	var data = ListObjectsV2Response{}
 
@@ -564,6 +564,10 @@ func generateListObjectsV2Response(bucket, prefix, token, nextToken, startAfter,
 					// values to client.
 					continue
 				}
+				// https://github.com/google/security-research/security/advisories/GHSA-76wf-9vgp-pj7w
+				if strings.EqualFold(k, xhttp.AmzMetaUnencryptedContentLength) || strings.EqualFold(k, xhttp.AmzMetaUnencryptedContentMD5) {
+					continue
+				}
 				content.UserMetadata[k] = v
 			}
 		}
@@ -580,6 +584,8 @@ func generateListObjectsV2Response(bucket, prefix, token, nextToken, startAfter,
 	data.ContinuationToken = base64.StdEncoding.EncodeToString([]byte(token))
 	data.NextContinuationToken = base64.StdEncoding.EncodeToString([]byte(nextToken))
 	data.IsTruncated = isTruncated
+
+	commonPrefixes := make([]CommonPrefix, 0, len(prefixes))
 	for _, prefix := range prefixes {
 		var prefixItem = CommonPrefix{}
 		prefixItem.Prefix = s3EncodeName(prefix, encodingType)
@@ -759,6 +765,10 @@ func writeErrorResponse(ctx context.Context, w http.ResponseWriter, err APIError
 		// Set retry-after header to indicate user-agents to retry request after 120secs.
 		// https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Retry-After
 		w.Header().Set(xhttp.RetryAfter, "120")
+	case "InvalidRegion":
+		err.Description = fmt.Sprintf("Region does not match; expecting '%s'.", globalServerRegion)
+	case "AuthorizationHeaderMalformed":
+		err.Description = fmt.Sprintf("The authorization header is malformed; the region is wrong; expecting '%s'.", globalServerRegion)
 	case "AccessDenied":
 		// The request is from browser and also if browser
 		// is enabled we need to redirect.
@@ -817,38 +827,3 @@ func writeCustomErrorResponseJSON(ctx context.Context, w http.ResponseWriter, er
 	encodedErrorResponse := encodeResponseJSON(errorResponse)
 	writeResponse(w, err.HTTPStatusCode, encodedErrorResponse, mimeJSON)
 }
-
-// writeCustomErrorResponseXML - similar to writeErrorResponse,
-// but accepts the error message directly (this allows messages to be
-// dynamically generated.)
-func writeCustomErrorResponseXML(ctx context.Context, w http.ResponseWriter, err APIError, errBody string, reqURL *url.URL, browser bool) {
-
-	switch err.Code {
-	case "SlowDown", "XMinioServerNotInitialized", "XMinioReadQuorum", "XMinioWriteQuorum":
-		// Set retry-after header to indicate user-agents to retry request after 120secs.
-		// https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Retry-After
-		w.Header().Set(xhttp.RetryAfter, "120")
-	case "AccessDenied":
-		// The request is from browser and also if browser
-		// is enabled we need to redirect.
-		if browser && globalBrowserEnabled {
-			w.Header().Set(xhttp.Location, minioReservedBucketPath+reqURL.Path)
-			w.WriteHeader(http.StatusTemporaryRedirect)
-			return
-		}
-	}
-
-	reqInfo := logger.GetReqInfo(ctx)
-	errorResponse := APIErrorResponse{
-		Code:       err.Code,
-		Message:    errBody,
-		Resource:   reqURL.Path,
-		BucketName: reqInfo.BucketName,
-		Key:        reqInfo.ObjectName,
-		RequestID:  w.Header().Get(xhttp.AmzRequestID),
-		HostID:     globalDeploymentID,
-	}
-
-	encodedErrorResponse := encodeResponse(errorResponse)
-	writeResponse(w, err.HTTPStatusCode, encodedErrorResponse, mimeXML)
-}

cmd/api-router.go

@@ -17,6 +17,7 @@
 package cmd
 
 import (
+	"net"
 	"net/http"
 
 	"github.com/gorilla/mux"
@@ -60,31 +61,50 @@ func newCachedObjectLayerFn() CacheObjectLayer {
 type objectAPIHandlers struct {
 	ObjectAPI func() ObjectLayer
 	CacheAPI  func() CacheObjectLayer
-	// Returns true of handlers should interpret encryption.
-	EncryptionEnabled func() bool
-	// Returns true if handlers allow SSE-KMS encryption headers.
-	AllowSSEKMS func() bool
+}
+
+// getHost tries its best to return the request host.
+// According to section 14.23 of RFC 2616 the Host header
+// can include the port number if the default value of 80 is not used.
+func getHost(r *http.Request) string {
+	if r.URL.IsAbs() {
+		return r.URL.Host
+	}
+	return r.Host
 }
 
 // registerAPIRouter - registers S3 compatible APIs.
-func registerAPIRouter(router *mux.Router, encryptionEnabled, allowSSEKMS bool) {
+func registerAPIRouter(router *mux.Router) {
 	// Initialize API.
 	api := objectAPIHandlers{
 		ObjectAPI: newObjectLayerFn,
 		CacheAPI:  newCachedObjectLayerFn,
-		EncryptionEnabled: func() bool {
-			return encryptionEnabled
-		},
-		AllowSSEKMS: func() bool {
-			return allowSSEKMS
-		},
 	}
 
 	// API Router
 	apiRouter := router.PathPrefix(SlashSeparator).Subrouter()
+
 	var routers []*mux.Router
 	for _, domainName := range globalDomainNames {
-		routers = append(routers, apiRouter.Host("{bucket:.+}."+domainName).Subrouter())
+		if IsKubernetes() {
+			routers = append(routers, apiRouter.MatcherFunc(func(r *http.Request, match *mux.RouteMatch) bool {
+				host, _, _ := net.SplitHostPort(getHost(r))
+				// Make sure to skip matching minio.<domain>` this is
+				// specifically meant for operator/k8s deployment
+				// The reason we need to skip this is for a special
+				// usecase where we need to make sure that
+				// minio.<namespace>.svc.<cluster_domain> is ignored
+				// by the bucketDNS style to ensure that path style
+				// is available and honored at this domain.
+				//
+				// All other `<bucket>.<namespace>.svc.<cluster_domain>`
+				// makes sure that buckets are routed through this matcher
+				// to match for `<bucket>`
+				return host != minioReservedBucket+"."+domainName
+			}).Host("{bucket:.+}."+domainName).Subrouter())
+		} else {
+			routers = append(routers, apiRouter.Host("{bucket:.+}."+domainName).Subrouter())
+		}
 	}
 	routers = append(routers, apiRouter.PathPrefix("/{bucket}").Subrouter())
 
@@ -94,7 +114,10 @@ func registerAPIRouter(router *mux.Router, encryptionEnabled, allowSSEKMS bool)
 		bucket.Methods(http.MethodHead).Path("/{object:.+}").HandlerFunc(
 			maxClients(collectAPIStats("headobject", httpTraceAll(api.HeadObjectHandler))))
 		// CopyObjectPart
-		bucket.Methods(http.MethodPut).Path("/{object:.+}").HeadersRegexp(xhttp.AmzCopySource, ".*?(\\/|%2F).*?").HandlerFunc(maxClients(collectAPIStats("copyobjectpart", httpTraceAll(api.CopyObjectPartHandler)))).Queries("partNumber", "{partNumber:[0-9]+}", "uploadId", "{uploadId:.*}")
+		bucket.Methods(http.MethodPut).Path("/{object:.+}").
+			HeadersRegexp(xhttp.AmzCopySource, ".*?(\\/|%2F).*?").
+			HandlerFunc(maxClients(collectAPIStats("copyobjectpart", httpTraceAll(api.CopyObjectPartHandler)))).
+			Queries("partNumber", "{partNumber:[0-9]+}", "uploadId", "{uploadId:.*}")
 		// PutObjectPart
 		bucket.Methods(http.MethodPut).Path("/{object:.+}").HandlerFunc(
 			maxClients(collectAPIStats("putobjectpart", httpTraceHdrs(api.PutObjectPartHandler)))).Queries("partNumber", "{partNumber:[0-9]+}", "uploadId", "{uploadId:.*}")
@@ -138,7 +161,8 @@ func registerAPIRouter(router *mux.Router, encryptionEnabled, allowSSEKMS bool)
 		bucket.Methods(http.MethodGet).Path("/{object:.+}").HandlerFunc(
 			maxClients(collectAPIStats("getobject", httpTraceHdrs(api.GetObjectHandler))))
 		// CopyObject
-		bucket.Methods(http.MethodPut).Path("/{object:.+}").HeadersRegexp(xhttp.AmzCopySource, ".*?(\\/|%2F).*?").HandlerFunc(maxClients(collectAPIStats("copyobject", httpTraceAll(api.CopyObjectHandler))))
+		bucket.Methods(http.MethodPut).Path("/{object:.+}").HeadersRegexp(xhttp.AmzCopySource, ".*?(\\/|%2F).*?").
+			HandlerFunc(maxClients(collectAPIStats("copyobject", httpTraceAll(api.CopyObjectHandler))))
 		// PutObjectRetention
 		bucket.Methods(http.MethodPut).Path("/{object:.+}").HandlerFunc(
 			maxClients(collectAPIStats("putobjectretention", httpTraceAll(api.PutObjectRetentionHandler)))).Queries("retention", "")

cmd/auth-handler.go

@@ -346,8 +346,26 @@ func checkRequestAuthTypeToAccessKey(ctx context.Context, r *http.Request, actio
 			// Request is allowed return the appropriate access key.
 			return cred.AccessKey, owner, ErrNone
 		}
+
+		if action == policy.ListBucketVersionsAction {
+			// In AWS S3 s3:ListBucket permission is same as s3:ListBucketVersions permission
+			// verify as a fallback.
+			if globalPolicySys.IsAllowed(policy.Args{
+				AccountName:     cred.AccessKey,
+				Action:          policy.ListBucketAction,
+				BucketName:      bucketName,
+				ConditionValues: getConditionValues(r, locationConstraint, "", nil),
+				IsOwner:         false,
+				ObjectName:      objectName,
+			}) {
+				// Request is allowed return the appropriate access key.
+				return cred.AccessKey, owner, ErrNone
+			}
+		}
+
 		return cred.AccessKey, owner, ErrAccessDenied
 	}
+
 	if globalIAMSys.IsAllowed(iampolicy.Args{
 		AccountName:     cred.AccessKey,
 		Action:          iampolicy.Action(action),
@@ -360,6 +378,22 @@ func checkRequestAuthTypeToAccessKey(ctx context.Context, r *http.Request, actio
 		// Request is allowed return the appropriate access key.
 		return cred.AccessKey, owner, ErrNone
 	}
+	if action == policy.ListBucketVersionsAction {
+		// In AWS S3 s3:ListBucket permission is same as s3:ListBucketVersions permission
+		// verify as a fallback.
+		if globalIAMSys.IsAllowed(iampolicy.Args{
+			AccountName:     cred.AccessKey,
+			Action:          iampolicy.Action(policy.ListBucketAction),
+			BucketName:      bucketName,
+			ConditionValues: getConditionValues(r, "", cred.AccessKey, claims),
+			ObjectName:      objectName,
+			IsOwner:         owner,
+			Claims:          claims,
+		}) {
+			// Request is allowed return the appropriate access key.
+			return cred.AccessKey, owner, ErrNone
+		}
+	}
 
 	return cred.AccessKey, owner, ErrAccessDenied
 }

cmd/background-heal-ops.go

@@ -21,7 +21,6 @@ import (
 	"path"
 	"time"
 
-	"github.com/minio/minio/cmd/logger"
 	"github.com/minio/minio/pkg/madmin"
 )
 
@@ -55,19 +54,20 @@ func (h *healRoutine) queueHealTask(task healTask) {
 	h.tasks <- task
 }
 
-func waitForLowHTTPReq(tolerance int32) {
+func waitForLowHTTPReq(tolerance int32, maxWait time.Duration) {
+	const wait = 10 * time.Millisecond
+	waitCount := maxWait / wait
+
 	// Bucket notification and http trace are not costly, it is okay to ignore them
 	// while counting the number of concurrent connections
 	tolerance += int32(globalHTTPListen.NumSubscribers() + globalHTTPTrace.NumSubscribers())
 
 	if httpServer := newHTTPServerFn(); httpServer != nil {
-		// Wait at max 10 minute for an inprogress request before proceeding to heal
-		waitCount := 600
 		// Any requests in progress, delay the heal.
 		for (httpServer.GetRequestCount() >= tolerance) &&
 			waitCount > 0 {
 			waitCount--
-			time.Sleep(1 * time.Second)
+			time.Sleep(wait)
 		}
 	}
 }
@@ -82,7 +82,7 @@ func (h *healRoutine) run(ctx context.Context, objAPI ObjectLayer) {
 			}
 
 			// Wait and proceed if there are active requests
-			waitForLowHTTPReq(int32(globalEndpoints.NEndpoints()))
+			waitForLowHTTPReq(int32(globalEndpoints.NEndpoints()), time.Second)
 
 			var res madmin.HealResultItem
 			var err error
@@ -100,6 +100,7 @@ func (h *healRoutine) run(ctx context.Context, objAPI ObjectLayer) {
 				ObjectPathUpdated(path.Join(task.bucket, task.object))
 			}
 			task.responseCh <- healResult{result: res, err: err}
+
 		case <-h.doneCh:
 			return
 		case <-ctx.Done():
@@ -116,24 +117,6 @@ func newHealRoutine() *healRoutine {
 
 }
 
-func initBackgroundHealing(ctx context.Context, objAPI ObjectLayer) {
-	// Run the background healer
-	globalBackgroundHealRoutine = newHealRoutine()
-	go globalBackgroundHealRoutine.run(ctx, objAPI)
-
-	nh := newBgHealSequence()
-	// Heal any disk format and metadata early, if possible.
-	if err := nh.healDiskMeta(); err != nil {
-		if newObjectLayerFn() != nil {
-			// log only in situations, when object layer
-			// has fully initialized.
-			logger.LogIf(nh.ctx, err)
-		}
-	}
-
-	globalBackgroundHealState.LaunchNewHealSequence(nh)
-}
-
 // healDiskFormat - heals format.json, return value indicates if a
 // failure error occurred.
 func healDiskFormat(ctx context.Context, objAPI ObjectLayer, opts madmin.HealOpts) (madmin.HealResultItem, error) {
@@ -145,24 +128,5 @@ func healDiskFormat(ctx context.Context, objAPI ObjectLayer, opts madmin.HealOpt
 		return madmin.HealResultItem{}, err
 	}
 
-	// Healing succeeded notify the peers to reload format and re-initialize disks.
-	// We will not notify peers if healing is not required.
-	if err == nil {
-		// Notify servers in background and retry if needed.
-		go func() {
-		retry:
-			for _, nerr := range globalNotificationSys.ReloadFormat(opts.DryRun) {
-				if nerr.Err != nil {
-					if nerr.Err.Error() == errServerNotInitialized.Error() {
-						time.Sleep(time.Second)
-						goto retry
-					}
-					logger.GetReqInfo(ctx).SetTags("peerAddress", nerr.Host.String())
-					logger.LogIf(ctx, nerr.Err)
-				}
-			}
-		}()
-	}
-
 	return res, nil
 }

cmd/background-newdisks-heal-ops.go

@@ -18,27 +18,24 @@ package cmd
 
 import (
 	"context"
+	"errors"
+	"fmt"
 	"time"
 
 	"github.com/dustin/go-humanize"
 	"github.com/minio/minio/cmd/logger"
 )
 
-const defaultMonitorNewDiskInterval = time.Minute * 3
+const defaultMonitorNewDiskInterval = time.Second * 10
 
-func initLocalDisksAutoHeal(ctx context.Context, objAPI ObjectLayer) {
-	go monitorLocalDisksAndHeal(ctx, objAPI)
-}
-
-// monitorLocalDisksAndHeal - ensures that detected new disks are healed
-//  1. Only the concerned erasure set will be listed and healed
-//  2. Only the node hosting the disk is responsible to perform the heal
-func monitorLocalDisksAndHeal(ctx context.Context, objAPI ObjectLayer) {
+func initAutoHeal(ctx context.Context, objAPI ObjectLayer) {
 	z, ok := objAPI.(*erasureZones)
 	if !ok {
 		return
 	}
 
+	initBackgroundHealing(ctx, objAPI) // start quick background healing
+
 	var bgSeq *healSequence
 	var found bool
 
@@ -50,82 +47,119 @@ func monitorLocalDisksAndHeal(ctx context.Context, objAPI ObjectLayer) {
 		time.Sleep(time.Second)
 	}
 
+	for _, ep := range getLocalDisksToHeal() {
+		globalBackgroundHealState.pushHealLocalDisks(ep)
+	}
+
+	if drivesToHeal := globalBackgroundHealState.healDriveCount(); drivesToHeal > 0 {
+		logger.Info(fmt.Sprintf("Found drives to heal %d, waiting until %s to heal the content...",
+			drivesToHeal, defaultMonitorNewDiskInterval))
+
+		// Heal any disk format and metadata early, if possible.
+		if err := bgSeq.healDiskMeta(); err != nil {
+			if newObjectLayerFn() != nil {
+				// log only in situations, when object layer
+				// has fully initialized.
+				logger.LogIf(bgSeq.ctx, err)
+			}
+		}
+	}
+
+	go monitorLocalDisksAndHeal(ctx, z, bgSeq)
+}
+
+func getLocalDisksToHeal() (disksToHeal Endpoints) {
+	for _, ep := range globalEndpoints {
+		for _, endpoint := range ep.Endpoints {
+			if !endpoint.IsLocal {
+				continue
+			}
+			// Try to connect to the current endpoint
+			// and reformat if the current disk is not formatted
+			_, _, err := connectEndpoint(endpoint)
+			if errors.Is(err, errUnformattedDisk) {
+				disksToHeal = append(disksToHeal, endpoint)
+			}
+		}
+	}
+	return disksToHeal
+
+}
+
+func initBackgroundHealing(ctx context.Context, objAPI ObjectLayer) {
+	// Run the background healer
+	globalBackgroundHealRoutine = newHealRoutine()
+	go globalBackgroundHealRoutine.run(ctx, objAPI)
+
+	globalBackgroundHealState.LaunchNewHealSequence(newBgHealSequence())
+}
+
+// monitorLocalDisksAndHeal - ensures that detected new disks are healed
+//  1. Only the concerned erasure set will be listed and healed
+//  2. Only the node hosting the disk is responsible to perform the heal
+func monitorLocalDisksAndHeal(ctx context.Context, z *erasureZones, bgSeq *healSequence) {
 	// Perform automatic disk healing when a disk is replaced locally.
 	for {
 		select {
 		case <-ctx.Done():
 			return
 		case <-time.After(defaultMonitorNewDiskInterval):
-			// Attempt a heal as the server starts-up first.
-			localDisksInZoneHeal := make([]Endpoints, len(z.zones))
-			var healNewDisks bool
-			for i, ep := range globalEndpoints {
-				localDisksToHeal := Endpoints{}
-				for _, endpoint := range ep.Endpoints {
-					if !endpoint.IsLocal {
-						continue
-					}
-					// Try to connect to the current endpoint
-					// and reformat if the current disk is not formatted
-					_, _, err := connectEndpoint(endpoint)
-					if err == errUnformattedDisk {
-						localDisksToHeal = append(localDisksToHeal, endpoint)
-					}
+			waitForLowHTTPReq(int32(globalEndpoints.NEndpoints()), time.Second)
+
+			var erasureSetInZoneEndpointToHeal []map[int]Endpoints
+			healDisks := globalBackgroundHealState.getHealLocalDisks()
+			if len(healDisks) > 0 {
+				// Reformat disks
+				bgSeq.sourceCh <- healSource{bucket: SlashSeparator}
+
+				// Ensure that reformatting disks is finished
+				bgSeq.sourceCh <- healSource{bucket: nopHeal}
+
+				logger.Info(fmt.Sprintf("Found drives to heal %d, proceeding to heal content...",
+					len(healDisks)))
+
+				erasureSetInZoneEndpointToHeal = make([]map[int]Endpoints, len(z.zones))
+				for i := range z.zones {
+					erasureSetInZoneEndpointToHeal[i] = map[int]Endpoints{}
 				}
-				if len(localDisksToHeal) == 0 {
+			}
+
+			// heal only if new disks found.
+			for _, endpoint := range healDisks {
+				// Load the new format of this passed endpoint
+				_, format, err := connectEndpoint(endpoint)
+				if err != nil {
+					printEndpointError(endpoint, err, true)
 					continue
 				}
-				localDisksInZoneHeal[i] = localDisksToHeal
-				healNewDisks = true
-			}
 
-			// Reformat disks only if needed.
-			if !healNewDisks {
-				continue
-			}
-
-			logger.Info("New unformatted drives detected attempting to heal...")
-			for i, disks := range localDisksInZoneHeal {
-				for _, disk := range disks {
-					logger.Info("Healing disk '%s' on %s zone", disk, humanize.Ordinal(i+1))
+				zoneIdx := globalEndpoints.GetLocalZoneIdx(endpoint)
+				if zoneIdx < 0 {
+					continue
 				}
-			}
 
-			// Reformat disks
-			bgSeq.sourceCh <- healSource{bucket: SlashSeparator}
-
-			// Ensure that reformatting disks is finished
-			bgSeq.sourceCh <- healSource{bucket: nopHeal}
-
-			var erasureSetInZoneToHeal = make([][]int, len(localDisksInZoneHeal))
-			// Compute the list of erasure set to heal
-			for i, localDisksToHeal := range localDisksInZoneHeal {
-				var erasureSetToHeal []int
-				for _, endpoint := range localDisksToHeal {
-					// Load the new format of this passed endpoint
-					_, format, err := connectEndpoint(endpoint)
-					if err != nil {
-						printEndpointError(endpoint, err, true)
-						continue
-					}
-					// Calculate the set index where the current endpoint belongs
-					setIndex, _, err := findDiskIndex(z.zones[i].format, format)
-					if err != nil {
-						printEndpointError(endpoint, err, false)
-						continue
-					}
-
-					erasureSetToHeal = append(erasureSetToHeal, setIndex)
+				// Calculate the set index where the current endpoint belongs
+				setIndex, _, err := findDiskIndex(z.zones[zoneIdx].format, format)
+				if err != nil {
+					printEndpointError(endpoint, err, false)
+					continue
 				}
-				erasureSetInZoneToHeal[i] = erasureSetToHeal
+
+				erasureSetInZoneEndpointToHeal[zoneIdx][setIndex] = append(erasureSetInZoneEndpointToHeal[zoneIdx][setIndex], endpoint)
 			}
 
-			// Heal all erasure sets that need
-			for i, erasureSetToHeal := range erasureSetInZoneToHeal {
-				for _, setIndex := range erasureSetToHeal {
-					err := healErasureSet(ctx, setIndex, z.zones[i].sets[setIndex], z.zones[i].drivesPerSet)
-					if err != nil {
-						logger.LogIf(ctx, err)
+			for i, setMap := range erasureSetInZoneEndpointToHeal {
+				for setIndex, endpoints := range setMap {
+					for _, ep := range endpoints {
+						logger.Info("Healing disk '%s' on %s zone", ep, humanize.Ordinal(i+1))
+
+						if err := healErasureSet(ctx, setIndex, z.zones[i].sets[setIndex], z.zones[i].setDriveCount); err != nil {
+							logger.LogIf(ctx, err)
+							continue
+						}
+
+						// Only upon success pop the healed disk.
+						globalBackgroundHealState.popHealLocalDisks(ep)
 					}
 				}
 			}

cmd/bitrot-streaming.go

@@ -18,6 +18,7 @@ package cmd
 
 import (
 	"bytes"
+	"context"
 	"encoding/hex"
 	"fmt"
 	"hash"
@@ -80,7 +81,7 @@ func newStreamingBitrotWriter(disk StorageAPI, volume, filePath string, length i
 			bitrotSumsTotalSize := ceilFrac(length, shardSize) * int64(h.Size()) // Size used for storing bitrot checksums.
 			totalFileSize = bitrotSumsTotalSize + length
 		}
-		err := disk.CreateFile(volume, filePath, totalFileSize, r)
+		err := disk.CreateFile(context.TODO(), volume, filePath, totalFileSize, r)
 		r.CloseWithError(err)
 		close(bw.canClose)
 	}()
@@ -118,7 +119,7 @@ func (b *streamingBitrotReader) ReadAt(buf []byte, offset int64) (int, error) {
 		// For the first ReadAt() call we need to open the stream for reading.
 		b.currOffset = offset
 		streamOffset := (offset/b.shardSize)*int64(b.h.Size()) + offset
-		b.rc, err = b.disk.ReadFileStream(b.volume, b.filePath, streamOffset, b.tillOffset-streamOffset)
+		b.rc, err = b.disk.ReadFileStream(context.TODO(), b.volume, b.filePath, streamOffset, b.tillOffset-streamOffset)
 		if err != nil {
 			return 0, err
 		}
@@ -139,8 +140,8 @@ func (b *streamingBitrotReader) ReadAt(buf []byte, offset int64) (int, error) {
 	b.h.Write(buf)
 
 	if !bytes.Equal(b.h.Sum(nil), b.hashBytes) {
-		err := &errHashMismatch{fmt.Sprintf("hashes do not match expected %s, got %s",
-			hex.EncodeToString(b.hashBytes), hex.EncodeToString(b.h.Sum(nil)))}
+		err := &errHashMismatch{fmt.Sprintf("Disk: %s - content hash does not match - expected %s, got %s",
+			b.disk, hex.EncodeToString(b.hashBytes), hex.EncodeToString(b.h.Sum(nil)))}
 		logger.LogIf(GlobalContext, err)
 		return 0, err
 	}

cmd/bitrot-whole.go

@@ -17,6 +17,8 @@
 package cmd
 
 import (
+	"context"
+	"fmt"
 	"hash"
 	"io"
 
@@ -33,14 +35,14 @@ type wholeBitrotWriter struct {
 }
 
 func (b *wholeBitrotWriter) Write(p []byte) (int, error) {
-	err := b.disk.AppendFile(b.volume, b.filePath, p)
+	err := b.disk.AppendFile(context.TODO(), b.volume, b.filePath, p)
 	if err != nil {
-		logger.LogIf(GlobalContext, err)
+		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s returned %w", b.disk, err))
 		return 0, err
 	}
 	_, err = b.Hash.Write(p)
 	if err != nil {
-		logger.LogIf(GlobalContext, err)
+		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s returned %w", b.disk, err))
 		return 0, err
 	}
 	return len(p), nil
@@ -68,10 +70,8 @@ type wholeBitrotReader struct {
 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(b.volume, b.filePath, offset, b.buf, b.verifier); err != nil {
-			ctx := GlobalContext
-			logger.GetReqInfo(ctx).AppendTags("disk", b.disk.String())
-			logger.LogIf(ctx, err)
+		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 returned %w", b.disk, err))
 			return 0, err
 		}
 	}

cmd/bitrot_test.go

@@ -17,6 +17,7 @@
 package cmd
 
 import (
+	"context"
 	"io"
 	"io/ioutil"
 	"log"
@@ -34,12 +35,12 @@ func testBitrotReaderWriterAlgo(t *testing.T, bitrotAlgo BitrotAlgorithm) {
 	volume := "testvol"
 	filePath := "testfile"
 
-	disk, err := newXLStorage(tmpDir, "")
+	disk, err := newLocalXLStorage(tmpDir)
 	if err != nil {
 		t.Fatal(err)
 	}
 
-	disk.MakeVol(volume)
+	disk.MakeVol(context.Background(), volume)
 
 	writer := newBitrotWriter(disk, volume, filePath, 35, bitrotAlgo, 10)
 

cmd/bootstrap-peer-server.go

@@ -131,7 +131,7 @@ func (client *bootstrapRESTClient) callWithContext(ctx context.Context, method s
 		values = make(url.Values)
 	}
 
-	respBody, err = client.restClient.CallWithContext(ctx, method, values, body, length)
+	respBody, err = client.restClient.Call(ctx, method, values, body, length)
 	if err == nil {
 		return respBody, nil
 	}
@@ -178,17 +178,22 @@ func verifyServerSystemConfig(ctx context.Context, endpointZones EndpointZones)
 			}
 			onlineServers++
 		}
-		// Sleep for a while - so that we don't go into
-		// 100% CPU when half the endpoints are offline.
-		time.Sleep(500 * time.Millisecond)
-		retries++
-		// after 5 retries start logging that servers are not reachable yet
-		if retries >= 5 {
-			logger.Info(fmt.Sprintf("Waiting for atleast %d servers to be online for bootstrap check", len(clnts)/2))
-			logger.Info(fmt.Sprintf("Following servers are currently offline or unreachable %s", offlineEndpoints))
-			retries = 0 // reset to log again after 5 retries.
+		select {
+		case <-ctx.Done():
+			return ctx.Err()
+		default:
+			// Sleep for a while - so that we don't go into
+			// 100% CPU when half the endpoints are offline.
+			time.Sleep(100 * time.Millisecond)
+			retries++
+			// after 5 retries start logging that servers are not reachable yet
+			if retries >= 5 {
+				logger.Info(fmt.Sprintf("Waiting for atleast %d servers to be online for bootstrap check", len(clnts)/2))
+				logger.Info(fmt.Sprintf("Following servers are currently offline or unreachable %s", offlineEndpoints))
+				retries = 0 // reset to log again after 5 retries.
+			}
+			offlineEndpoints = nil
 		}
-		offlineEndpoints = nil
 	}
 	return nil
 }
@@ -228,13 +233,13 @@ func newBootstrapRESTClient(endpoint Endpoint) *bootstrapRESTClient {
 		}
 	}
 
-	trFn := newCustomHTTPTransport(tlsConfig, rest.DefaultRESTTimeout)
+	trFn := newInternodeHTTPTransport(tlsConfig, rest.DefaultRESTTimeout)
 	restClient := rest.NewClient(serverURL, trFn, newAuthToken)
 	restClient.HealthCheckFn = func() bool {
 		ctx, cancel := context.WithTimeout(GlobalContext, restClient.HealthCheckTimeout)
 		// Instantiate a new rest client for healthcheck
 		// to avoid recursive healthCheckFn()
-		respBody, err := rest.NewClient(serverURL, trFn, newAuthToken).CallWithContext(ctx, bootstrapRESTMethodHealth, nil, nil, -1)
+		respBody, err := rest.NewClient(serverURL, trFn, newAuthToken).Call(ctx, bootstrapRESTMethodHealth, nil, nil, -1)
 		xhttp.DrainBody(respBody)
 		cancel()
 		var ne *rest.NetworkError

cmd/bucket-handlers.go

@@ -33,7 +33,7 @@ import (
 	"github.com/minio/minio-go/v7/pkg/set"
 	"github.com/minio/minio-go/v7/pkg/tags"
 	"github.com/minio/minio/cmd/config"
-	"github.com/minio/minio/cmd/config/etcd/dns"
+	"github.com/minio/minio/cmd/config/dns"
 	"github.com/minio/minio/cmd/crypto"
 	xhttp "github.com/minio/minio/cmd/http"
 	"github.com/minio/minio/cmd/logger"
@@ -72,7 +72,7 @@ func initFederatorBackend(buckets []BucketInfo, objLayer ObjectLayer) {
 
 	// Get buckets in the DNS
 	dnsBuckets, err := globalDNSConfig.List()
-	if err != nil && err != dns.ErrNoEntriesFound {
+	if err != nil && err != dns.ErrNoEntriesFound && err != dns.ErrNotImplemented {
 		logger.LogIf(GlobalContext, err)
 		return
 	}
@@ -80,33 +80,35 @@ func initFederatorBackend(buckets []BucketInfo, objLayer ObjectLayer) {
 	bucketsSet := set.NewStringSet()
 	bucketsToBeUpdated := set.NewStringSet()
 	bucketsInConflict := set.NewStringSet()
-	for _, bucket := range buckets {
-		bucketsSet.Add(bucket.Name)
-		r, ok := dnsBuckets[bucket.Name]
-		if !ok {
-			bucketsToBeUpdated.Add(bucket.Name)
-			continue
-		}
-		if !globalDomainIPs.Intersection(set.CreateStringSet(getHostsSlice(r)...)).IsEmpty() {
-			if globalDomainIPs.Difference(set.CreateStringSet(getHostsSlice(r)...)).IsEmpty() {
-				// No difference in terms of domainIPs and nothing
-				// has changed so we don't change anything on the etcd.
+	if dnsBuckets != nil {
+		for _, bucket := range buckets {
+			bucketsSet.Add(bucket.Name)
+			r, ok := dnsBuckets[bucket.Name]
+			if !ok {
+				bucketsToBeUpdated.Add(bucket.Name)
 				continue
 			}
-			// if domain IPs intersect then it won't be an empty set.
-			// such an intersection means that bucket exists on etcd.
-			// but if we do see a difference with local domain IPs with
-			// hostSlice from etcd then we should update with newer
-			// domainIPs, we proceed to do that here.
-			bucketsToBeUpdated.Add(bucket.Name)
-			continue
+			if !globalDomainIPs.Intersection(set.CreateStringSet(getHostsSlice(r)...)).IsEmpty() {
+				if globalDomainIPs.Difference(set.CreateStringSet(getHostsSlice(r)...)).IsEmpty() {
+					// No difference in terms of domainIPs and nothing
+					// has changed so we don't change anything on the etcd.
+					continue
+				}
+				// if domain IPs intersect then it won't be an empty set.
+				// such an intersection means that bucket exists on etcd.
+				// but if we do see a difference with local domain IPs with
+				// hostSlice from etcd then we should update with newer
+				// domainIPs, we proceed to do that here.
+				bucketsToBeUpdated.Add(bucket.Name)
+				continue
+			}
+			// No IPs seem to intersect, this means that bucket exists but has
+			// different IP addresses perhaps from a different deployment.
+			// bucket names are globally unique in federation at a given
+			// path prefix, name collision is not allowed. We simply log
+			// an error and continue.
+			bucketsInConflict.Add(bucket.Name)
 		}
-		// No IPs seem to intersect, this means that bucket exists but has
-		// different IP addresses perhaps from a different deployment.
-		// bucket names are globally unique in federation at a given
-		// path prefix, name collision is not allowed. We simply log
-		// an error and continue.
-		bucketsInConflict.Add(bucket.Name)
 	}
 
 	// Add/update buckets that are not registered with the DNS
@@ -449,14 +451,15 @@ func (api objectAPIHandlers) DeleteMultipleObjectsHandler(w http.ResponseWriter,
 
 	deleteList := toNames(objectsToDelete)
 	dObjects, errs := deleteObjectsFn(ctx, bucket, deleteList, ObjectOptions{
-		Versioned: globalBucketVersioningSys.Enabled(bucket),
+		Versioned:        globalBucketVersioningSys.Enabled(bucket),
+		VersionSuspended: globalBucketVersioningSys.Suspended(bucket),
 	})
 
 	deletedObjects := make([]DeletedObject, len(deleteObjects.Objects))
 	for i := range errs {
 		dindex := objectsToDelete[deleteList[i]]
 		apiErr := toAPIError(ctx, errs[i])
-		if apiErr.Code == "" || apiErr.Code == "NoSuchKey" {
+		if apiErr.Code == "" || apiErr.Code == "NoSuchKey" || apiErr.Code == "InvalidArgument" {
 			deletedObjects[dindex] = dObjects[i]
 			continue
 		}
@@ -561,7 +564,9 @@ func (api objectAPIHandlers) PutBucketHandler(w http.ResponseWriter, r *http.Req
 	if globalDNSConfig != nil {
 		sr, err := globalDNSConfig.Get(bucket)
 		if err != nil {
-			if err == dns.ErrNoEntriesFound {
+			// ErrNotImplemented indicates a DNS backend that doesn't need to check if bucket already
+			// exists elsewhere
+			if err == dns.ErrNoEntriesFound || err == dns.ErrNotImplemented {
 				// Proceed to creating a bucket.
 				if err = objectAPI.MakeBucketWithLocation(ctx, bucket, opts); err != nil {
 					writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
@@ -654,16 +659,14 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 		writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL, guessIsBrowserReq(r))
 		return
 	}
-	if !api.EncryptionEnabled() && crypto.IsRequested(r.Header) {
+
+	if !objectAPI.IsEncryptionSupported() && crypto.IsRequested(r.Header) {
 		writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL, guessIsBrowserReq(r))
 		return
 	}
 
 	bucket := mux.Vars(r)["bucket"]
 
-	// To detect if the client has disconnected.
-	r.Body = &contextReader{r.Body, r.Context()}
-
 	// Require Content-Length to be set in the request
 	size := r.ContentLength
 	if size < 0 {
@@ -762,7 +765,7 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 
 		// Make sure formValues adhere to policy restrictions.
 		if err = checkPostPolicy(formValues, postPolicyForm); err != nil {
-			writeCustomErrorResponseXML(ctx, w, errorCodes.ToAPIErr(ErrAccessDenied), err.Error(), r.URL, guessIsBrowserReq(r))
+			writeErrorResponse(ctx, w, errorCodes.ToAPIErrWithErr(ErrAccessDenied, err), r.URL, guessIsBrowserReq(r))
 			return
 		}
 
@@ -804,7 +807,7 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 	if _, err = globalBucketSSEConfigSys.Get(bucket); err == nil || globalAutoEncryption {
 		// This request header needs to be set prior to setting ObjectOptions
 		if !crypto.SSEC.IsRequested(r.Header) {
-			r.Header.Add(crypto.SSEHeader, crypto.SSEAlgorithmAES256)
+			r.Header.Set(crypto.SSEHeader, crypto.SSEAlgorithmAES256)
 		}
 	}
 
@@ -999,7 +1002,7 @@ func (api objectAPIHandlers) DeleteBucketHandler(w http.ResponseWriter, r *http.
 
 	if globalDNSConfig != nil {
 		if err := globalDNSConfig.Delete(bucket); err != nil {
-			logger.LogIf(ctx, fmt.Errorf("Unable to delete bucket DNS entry %w, please delete it manually using etcdctl", err))
+			logger.LogIf(ctx, fmt.Errorf("Unable to delete bucket DNS entry %w, please delete it manually", err))
 			writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
 			return
 		}
@@ -1271,7 +1274,7 @@ func (api objectAPIHandlers) PutBucketReplicationConfigHandler(w http.ResponseWr
 		writeErrorResponse(ctx, w, apiErr, r.URL, guessIsBrowserReq(r))
 		return
 	}
-	sameTarget, err := globalBucketReplicationSys.validateDestination(ctx, bucket, replicationConfig)
+	sameTarget, err := validateReplicationDestination(ctx, bucket, replicationConfig)
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
 		return

cmd/bucket-listobjects-handlers.go

@@ -24,12 +24,36 @@ import (
 	"strings"
 
 	"github.com/gorilla/mux"
-	"github.com/minio/minio/cmd/crypto"
 	"github.com/minio/minio/cmd/logger"
 
 	"github.com/minio/minio/pkg/bucket/policy"
+	"github.com/minio/minio/pkg/sync/errgroup"
 )
 
+func concurrentDecryptETag(ctx context.Context, objects []ObjectInfo) {
+	inParallel := func(objects []ObjectInfo) {
+		g := errgroup.WithNErrs(len(objects))
+		for index := range objects {
+			index := index
+			g.Go(func() error {
+				objects[index].ETag = objects[index].GetActualETag(nil)
+				objects[index].Size, _ = objects[index].GetActualSize()
+				return nil
+			}, index)
+		}
+		g.Wait()
+	}
+	const maxConcurrent = 500
+	for {
+		if len(objects) < maxConcurrent {
+			inParallel(objects)
+			return
+		}
+		inParallel(objects[:maxConcurrent])
+		objects = objects[maxConcurrent:]
+	}
+}
+
 // 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
@@ -69,7 +93,7 @@ func (api objectAPIHandlers) ListObjectVersionsHandler(w http.ResponseWriter, r
 		return
 	}
 
-	if s3Error := checkRequestAuthType(ctx, r, policy.ListBucketAction, bucket, ""); s3Error != ErrNone {
+	if s3Error := checkRequestAuthType(ctx, r, policy.ListBucketVersionsAction, bucket, ""); s3Error != ErrNone {
 		writeErrorResponse(ctx, w, errorCodes.ToAPIErr(s3Error), r.URL, guessIsBrowserReq(r))
 		return
 	}
@@ -89,6 +113,10 @@ func (api objectAPIHandlers) ListObjectVersionsHandler(w http.ResponseWriter, r
 		return
 	}
 
+	if proxyRequestByBucket(ctx, w, r, bucket) {
+		return
+	}
+
 	listObjectVersions := objectAPI.ListObjectVersions
 
 	// Inititate a list object versions operation based on the input params.
@@ -100,16 +128,7 @@ func (api objectAPIHandlers) ListObjectVersionsHandler(w http.ResponseWriter, r
 		return
 	}
 
-	for i := range listObjectVersionsInfo.Objects {
-		if crypto.IsEncrypted(listObjectVersionsInfo.Objects[i].UserDefined) {
-			listObjectVersionsInfo.Objects[i].ETag = getDecryptedETag(r.Header, listObjectVersionsInfo.Objects[i], false)
-		}
-		listObjectVersionsInfo.Objects[i].Size, err = listObjectVersionsInfo.Objects[i].GetActualSize()
-		if err != nil {
-			writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
-			return
-		}
-	}
+	concurrentDecryptETag(ctx, listObjectVersionsInfo.Objects)
 
 	response := generateListVersionsResponse(bucket, prefix, marker, versionIDMarker, delimiter, encodingType, maxkeys, listObjectVersionsInfo)
 
@@ -178,16 +197,7 @@ func (api objectAPIHandlers) ListObjectsV2MHandler(w http.ResponseWriter, r *htt
 		return
 	}
 
-	for i := range listObjectsV2Info.Objects {
-		if crypto.IsEncrypted(listObjectsV2Info.Objects[i].UserDefined) {
-			listObjectsV2Info.Objects[i].ETag = getDecryptedETag(r.Header, listObjectsV2Info.Objects[i], false)
-		}
-		listObjectsV2Info.Objects[i].Size, err = listObjectsV2Info.Objects[i].GetActualSize()
-		if err != nil {
-			writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
-			return
-		}
-	}
+	concurrentDecryptETag(ctx, listObjectsV2Info.Objects)
 
 	// The next continuation token has id@node_index format to optimize paginated listing
 	nextContinuationToken := listObjectsV2Info.NextContinuationToken
@@ -264,16 +274,7 @@ func (api objectAPIHandlers) ListObjectsV2Handler(w http.ResponseWriter, r *http
 		return
 	}
 
-	for i := range listObjectsV2Info.Objects {
-		if crypto.IsEncrypted(listObjectsV2Info.Objects[i].UserDefined) {
-			listObjectsV2Info.Objects[i].ETag = getDecryptedETag(r.Header, listObjectsV2Info.Objects[i], false)
-		}
-		listObjectsV2Info.Objects[i].Size, err = listObjectsV2Info.Objects[i].GetActualSize()
-		if err != nil {
-			writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
-			return
-		}
-	}
+	concurrentDecryptETag(ctx, listObjectsV2Info.Objects)
 
 	// The next continuation token has id@node_index format to optimize paginated listing
 	nextContinuationToken := listObjectsV2Info.NextContinuationToken
@@ -396,16 +397,7 @@ func (api objectAPIHandlers) ListObjectsV1Handler(w http.ResponseWriter, r *http
 		return
 	}
 
-	for i := range listObjectsInfo.Objects {
-		if crypto.IsEncrypted(listObjectsInfo.Objects[i].UserDefined) {
-			listObjectsInfo.Objects[i].ETag = getDecryptedETag(r.Header, listObjectsInfo.Objects[i], false)
-		}
-		listObjectsInfo.Objects[i].Size, err = listObjectsInfo.Objects[i].GetActualSize()
-		if err != nil {
-			writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL, guessIsBrowserReq(r))
-			return
-		}
-	}
+	concurrentDecryptETag(ctx, listObjectsInfo.Objects)
 
 	response := generateListObjectsV1Response(bucket, prefix, marker, delimiter, encodingType, maxKeys, listObjectsInfo)
 

cmd/bucket-metadata-sys.go

@@ -24,6 +24,7 @@ import (
 	"sync"
 
 	"github.com/minio/minio-go/v7/pkg/tags"
+	"github.com/minio/minio/cmd/logger"
 	bucketsse "github.com/minio/minio/pkg/bucket/encryption"
 	"github.com/minio/minio/pkg/bucket/lifecycle"
 	objectlock "github.com/minio/minio/pkg/bucket/object/lock"
@@ -80,7 +81,7 @@ func (sys *BucketMetadataSys) Update(bucket string, configFile string, configDat
 		// This code is needed only for gateway implementations.
 		switch configFile {
 		case bucketSSEConfig:
-			if globalGatewayName == "nas" {
+			if globalGatewayName == NASBackendGateway {
 				meta, err := loadBucketMetadata(GlobalContext, objAPI, bucket)
 				if err != nil {
 					return err
@@ -89,7 +90,7 @@ func (sys *BucketMetadataSys) Update(bucket string, configFile string, configDat
 				return meta.Save(GlobalContext, objAPI)
 			}
 		case bucketLifecycleConfig:
-			if globalGatewayName == "nas" {
+			if globalGatewayName == NASBackendGateway {
 				meta, err := loadBucketMetadata(GlobalContext, objAPI, bucket)
 				if err != nil {
 					return err
@@ -98,7 +99,7 @@ func (sys *BucketMetadataSys) Update(bucket string, configFile string, configDat
 				return meta.Save(GlobalContext, objAPI)
 			}
 		case bucketTaggingConfig:
-			if globalGatewayName == "nas" {
+			if globalGatewayName == NASBackendGateway {
 				meta, err := loadBucketMetadata(GlobalContext, objAPI, bucket)
 				if err != nil {
 					return err
@@ -107,7 +108,7 @@ func (sys *BucketMetadataSys) Update(bucket string, configFile string, configDat
 				return meta.Save(GlobalContext, objAPI)
 			}
 		case bucketNotificationConfig:
-			if globalGatewayName == "nas" {
+			if globalGatewayName == NASBackendGateway {
 				meta, err := loadBucketMetadata(GlobalContext, objAPI, bucket)
 				if err != nil {
 					return err
@@ -148,16 +149,28 @@ func (sys *BucketMetadataSys) Update(bucket string, configFile string, configDat
 		meta.EncryptionConfigXML = configData
 	case bucketTaggingConfig:
 		meta.TaggingConfigXML = configData
-	case objectLockConfig:
-		meta.ObjectLockConfigXML = configData
-	case bucketVersioningConfig:
-		meta.VersioningConfigXML = configData
 	case bucketQuotaConfigFile:
 		meta.QuotaConfigJSON = configData
+	case objectLockConfig:
+		if !globalIsErasure && !globalIsDistErasure {
+			return NotImplemented{}
+		}
+		meta.ObjectLockConfigXML = configData
+	case bucketVersioningConfig:
+		if !globalIsErasure && !globalIsDistErasure {
+			return NotImplemented{}
+		}
+		meta.VersioningConfigXML = configData
 	case bucketReplicationConfig:
+		if !globalIsErasure && !globalIsDistErasure {
+			return NotImplemented{}
+		}
 		meta.ReplicationConfigXML = configData
-	case bucketReplicationTargetsFile:
-		meta.ReplicationTargetsConfigJSON = configData
+	case bucketTargetsFile:
+		if !globalIsErasure && !globalIsDistErasure {
+			return NotImplemented{}
+		}
+		meta.BucketTargetsConfigJSON = configData
 	default:
 		return fmt.Errorf("Unknown bucket %s metadata update requested %s", bucket, configFile)
 	}
@@ -176,6 +189,13 @@ func (sys *BucketMetadataSys) Update(bucket string, configFile string, configDat
 // If no metadata exists errConfigNotFound is returned and a new metadata is returned.
 // Only a shallow copy is returned, so referenced data should not be modified,
 // but can be replaced atomically.
+//
+// This function should only be used with
+// - GetBucketInfo
+// - ListBuckets
+// - ListBucketsHeal (only in case of erasure coding mode)
+// For all other bucket specific metadata, use the relevant
+// calls implemented specifically for each of those features.
 func (sys *BucketMetadataSys) Get(bucket string) (BucketMetadata, error) {
 	if globalIsGateway || bucket == minioMetaBucket {
 		return newBucketMetadata(bucket), errConfigNotFound
@@ -253,7 +273,7 @@ func (sys *BucketMetadataSys) GetLifecycleConfig(bucket string) (*lifecycle.Life
 // GetNotificationConfig returns configured notification config
 // The returned object may not be modified.
 func (sys *BucketMetadataSys) GetNotificationConfig(bucket string) (*event.Config, error) {
-	if globalIsGateway && globalGatewayName == "nas" {
+	if globalIsGateway && globalGatewayName == NASBackendGateway {
 		// Only needed in case of NAS gateway.
 		objAPI := newObjectLayerWithoutSafeModeFn()
 		if objAPI == nil {
@@ -340,17 +360,17 @@ func (sys *BucketMetadataSys) GetReplicationConfig(ctx context.Context, bucket s
 	return meta.replicationConfig, nil
 }
 
-// GetReplicationTargetConfig returns configured bucket replication target for this bucket
+// GetBucketTargetsConfig returns configured bucket targets for this bucket
 // The returned object may not be modified.
-func (sys *BucketMetadataSys) GetReplicationTargetConfig(bucket string) (*madmin.BucketReplicationTarget, error) {
+func (sys *BucketMetadataSys) GetBucketTargetsConfig(bucket string) (*madmin.BucketTargets, error) {
 	meta, err := sys.GetConfig(bucket)
 	if err != nil {
 		return nil, err
 	}
-	if meta.replicationTargetConfig == nil {
-		return nil, BucketReplicationTargetNotFound{Bucket: bucket}
+	if meta.bucketTargetConfig == nil {
+		return nil, BucketRemoteTargetNotFound{Bucket: bucket}
 	}
-	return meta.replicationTargetConfig, nil
+	return meta.bucketTargetConfig, nil
 }
 
 // GetConfig returns a specific configuration from the bucket metadata.
@@ -397,8 +417,9 @@ func (sys *BucketMetadataSys) Init(ctx context.Context, buckets []BucketInfo, ob
 		return nil
 	}
 
-	// Load PolicySys once during boot.
-	return sys.load(ctx, buckets, objAPI)
+	// Load bucket metadata sys in background
+	go logger.LogIf(ctx, sys.load(ctx, buckets, objAPI))
+	return nil
 }
 
 // concurrently load bucket metadata to speed up loading bucket metadata.

cmd/bucket-metadata.go

@@ -61,31 +61,31 @@ var (
 // bucketMetadataFormat refers to the format.
 // bucketMetadataVersion can be used to track a rolling upgrade of a field.
 type BucketMetadata struct {
-	Name                         string
-	Created                      time.Time
-	LockEnabled                  bool // legacy not used anymore.
-	PolicyConfigJSON             []byte
-	NotificationConfigXML        []byte
-	LifecycleConfigXML           []byte
-	ObjectLockConfigXML          []byte
-	VersioningConfigXML          []byte
-	EncryptionConfigXML          []byte
-	TaggingConfigXML             []byte
-	QuotaConfigJSON              []byte
-	ReplicationConfigXML         []byte
-	ReplicationTargetsConfigJSON []byte
+	Name                    string
+	Created                 time.Time
+	LockEnabled             bool // legacy not used anymore.
+	PolicyConfigJSON        []byte
+	NotificationConfigXML   []byte
+	LifecycleConfigXML      []byte
+	ObjectLockConfigXML     []byte
+	VersioningConfigXML     []byte
+	EncryptionConfigXML     []byte
+	TaggingConfigXML        []byte
+	QuotaConfigJSON         []byte
+	ReplicationConfigXML    []byte
+	BucketTargetsConfigJSON []byte
 
 	// Unexported fields. Must be updated atomically.
-	policyConfig            *policy.Policy
-	notificationConfig      *event.Config
-	lifecycleConfig         *lifecycle.Lifecycle
-	objectLockConfig        *objectlock.Config
-	versioningConfig        *versioning.Versioning
-	sseConfig               *bucketsse.BucketSSEConfig
-	taggingConfig           *tags.Tags
-	quotaConfig             *madmin.BucketQuota
-	replicationConfig       *replication.Config
-	replicationTargetConfig *madmin.BucketReplicationTarget
+	policyConfig       *policy.Policy
+	notificationConfig *event.Config
+	lifecycleConfig    *lifecycle.Lifecycle
+	objectLockConfig   *objectlock.Config
+	versioningConfig   *versioning.Versioning
+	sseConfig          *bucketsse.BucketSSEConfig
+	taggingConfig      *tags.Tags
+	quotaConfig        *madmin.BucketQuota
+	replicationConfig  *replication.Config
+	bucketTargetConfig *madmin.BucketTargets
 }
 
 // newBucketMetadata creates BucketMetadata with the supplied name and Created to Now.
@@ -100,7 +100,7 @@ func newBucketMetadata(name string) BucketMetadata {
 		versioningConfig: &versioning.Versioning{
 			XMLNS: "http://s3.amazonaws.com/doc/2006-03-01/",
 		},
-		replicationTargetConfig: &madmin.BucketReplicationTarget{},
+		bucketTargetConfig: &madmin.BucketTargets{},
 	}
 }
 
@@ -227,12 +227,12 @@ func (b *BucketMetadata) parseAllConfigs(ctx context.Context, objectAPI ObjectLa
 		b.replicationConfig = nil
 	}
 
-	if len(b.ReplicationTargetsConfigJSON) != 0 {
-		if err = json.Unmarshal(b.ReplicationTargetsConfigJSON, b.replicationTargetConfig); err != nil {
+	if len(b.BucketTargetsConfigJSON) != 0 {
+		if err = json.Unmarshal(b.BucketTargetsConfigJSON, b.bucketTargetConfig); err != nil {
 			return err
 		}
 	} else {
-		b.replicationTargetConfig = &madmin.BucketReplicationTarget{}
+		b.bucketTargetConfig = &madmin.BucketTargets{}
 	}
 	return nil
 }
@@ -247,7 +247,7 @@ func (b *BucketMetadata) convertLegacyConfigs(ctx context.Context, objectAPI Obj
 		bucketSSEConfig,
 		bucketTaggingConfig,
 		bucketReplicationConfig,
-		bucketReplicationTargetsFile,
+		bucketTargetsFile,
 		objectLockConfig,
 	}
 
@@ -265,6 +265,12 @@ func (b *BucketMetadata) convertLegacyConfigs(ctx context.Context, objectAPI Obj
 
 		configData, err := readConfig(ctx, objectAPI, configFile)
 		if err != nil {
+			switch err.(type) {
+			case ObjectExistsAsDirectory:
+				// in FS mode it possible that we have actual
+				// files in this folder with `.minio.sys/buckets/bucket/configFile`
+				continue
+			}
 			if errors.Is(err, errConfigNotFound) {
 				// legacy file config not found, proceed to look for new metadata.
 				continue
@@ -306,8 +312,8 @@ func (b *BucketMetadata) convertLegacyConfigs(ctx context.Context, objectAPI Obj
 			b.QuotaConfigJSON = configData
 		case bucketReplicationConfig:
 			b.ReplicationConfigXML = configData
-		case bucketReplicationTargetsFile:
-			b.ReplicationTargetsConfigJSON = configData
+		case bucketTargetsFile:
+			b.BucketTargetsConfigJSON = configData
 		}
 	}
 
@@ -348,7 +354,7 @@ func (b *BucketMetadata) Save(ctx context.Context, api ObjectLayer) error {
 
 // deleteBucketMetadata deletes bucket metadata
 // If config does not exist no error is returned.
-func deleteBucketMetadata(ctx context.Context, obj ObjectLayer, bucket string) error {
+func deleteBucketMetadata(ctx context.Context, obj objectDeleter, bucket string) error {
 	metadataFiles := []string{
 		dataUsageCacheName,
 		bucketMetadataFile,

cmd/bucket-metadata_gen.go

@@ -96,10 +96,10 @@ func (z *BucketMetadata) DecodeMsg(dc *msgp.Reader) (err error) {
 				err = msgp.WrapError(err, "ReplicationConfigXML")
 				return
 			}
-		case "ReplicationTargetsConfigJSON":
-			z.ReplicationTargetsConfigJSON, err = dc.ReadBytes(z.ReplicationTargetsConfigJSON)
+		case "BucketTargetsConfigJSON":
+			z.BucketTargetsConfigJSON, err = dc.ReadBytes(z.BucketTargetsConfigJSON)
 			if err != nil {
-				err = msgp.WrapError(err, "ReplicationTargetsConfigJSON")
+				err = msgp.WrapError(err, "BucketTargetsConfigJSON")
 				return
 			}
 		default:
@@ -236,14 +236,14 @@ func (z *BucketMetadata) EncodeMsg(en *msgp.Writer) (err error) {
 		err = msgp.WrapError(err, "ReplicationConfigXML")
 		return
 	}
-	// write "ReplicationTargetsConfigJSON"
-	err = en.Append(0xbc, 0x52, 0x65, 0x70, 0x6c, 0x69, 0x63, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x54, 0x61, 0x72, 0x67, 0x65, 0x74, 0x73, 0x43, 0x6f, 0x6e, 0x66, 0x69, 0x67, 0x4a, 0x53, 0x4f, 0x4e)
+	// write "BucketTargetsConfigJSON"
+	err = en.Append(0xb7, 0x42, 0x75, 0x63, 0x6b, 0x65, 0x74, 0x54, 0x61, 0x72, 0x67, 0x65, 0x74, 0x73, 0x43, 0x6f, 0x6e, 0x66, 0x69, 0x67, 0x4a, 0x53, 0x4f, 0x4e)
 	if err != nil {
 		return
 	}
-	err = en.WriteBytes(z.ReplicationTargetsConfigJSON)
+	err = en.WriteBytes(z.BucketTargetsConfigJSON)
 	if err != nil {
-		err = msgp.WrapError(err, "ReplicationTargetsConfigJSON")
+		err = msgp.WrapError(err, "BucketTargetsConfigJSON")
 		return
 	}
 	return
@@ -289,9 +289,9 @@ func (z *BucketMetadata) MarshalMsg(b []byte) (o []byte, err error) {
 	// string "ReplicationConfigXML"
 	o = append(o, 0xb4, 0x52, 0x65, 0x70, 0x6c, 0x69, 0x63, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x43, 0x6f, 0x6e, 0x66, 0x69, 0x67, 0x58, 0x4d, 0x4c)
 	o = msgp.AppendBytes(o, z.ReplicationConfigXML)
-	// string "ReplicationTargetsConfigJSON"
-	o = append(o, 0xbc, 0x52, 0x65, 0x70, 0x6c, 0x69, 0x63, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x54, 0x61, 0x72, 0x67, 0x65, 0x74, 0x73, 0x43, 0x6f, 0x6e, 0x66, 0x69, 0x67, 0x4a, 0x53, 0x4f, 0x4e)
-	o = msgp.AppendBytes(o, z.ReplicationTargetsConfigJSON)
+	// string "BucketTargetsConfigJSON"
+	o = append(o, 0xb7, 0x42, 0x75, 0x63, 0x6b, 0x65, 0x74, 0x54, 0x61, 0x72, 0x67, 0x65, 0x74, 0x73, 0x43, 0x6f, 0x6e, 0x66, 0x69, 0x67, 0x4a, 0x53, 0x4f, 0x4e)
+	o = msgp.AppendBytes(o, z.BucketTargetsConfigJSON)
 	return
 }
 
@@ -385,10 +385,10 @@ func (z *BucketMetadata) UnmarshalMsg(bts []byte) (o []byte, err error) {
 				err = msgp.WrapError(err, "ReplicationConfigXML")
 				return
 			}
-		case "ReplicationTargetsConfigJSON":
-			z.ReplicationTargetsConfigJSON, bts, err = msgp.ReadBytesBytes(bts, z.ReplicationTargetsConfigJSON)
+		case "BucketTargetsConfigJSON":
+			z.BucketTargetsConfigJSON, bts, err = msgp.ReadBytesBytes(bts, z.BucketTargetsConfigJSON)
 			if err != nil {
-				err = msgp.WrapError(err, "ReplicationTargetsConfigJSON")
+				err = msgp.WrapError(err, "BucketTargetsConfigJSON")
 				return
 			}
 		default:
@@ -405,6 +405,6 @@ func (z *BucketMetadata) UnmarshalMsg(bts []byte) (o []byte, err error) {
 
 // Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
 func (z *BucketMetadata) Msgsize() (s int) {
-	s = 1 + 5 + msgp.StringPrefixSize + len(z.Name) + 8 + msgp.TimeSize + 12 + msgp.BoolSize + 17 + msgp.BytesPrefixSize + len(z.PolicyConfigJSON) + 22 + msgp.BytesPrefixSize + len(z.NotificationConfigXML) + 19 + msgp.BytesPrefixSize + len(z.LifecycleConfigXML) + 20 + msgp.BytesPrefixSize + len(z.ObjectLockConfigXML) + 20 + msgp.BytesPrefixSize + len(z.VersioningConfigXML) + 20 + msgp.BytesPrefixSize + len(z.EncryptionConfigXML) + 17 + msgp.BytesPrefixSize + len(z.TaggingConfigXML) + 16 + msgp.BytesPrefixSize + len(z.QuotaConfigJSON) + 21 + msgp.BytesPrefixSize + len(z.ReplicationConfigXML) + 29 + msgp.BytesPrefixSize + len(z.ReplicationTargetsConfigJSON)
+	s = 1 + 5 + msgp.StringPrefixSize + len(z.Name) + 8 + msgp.TimeSize + 12 + msgp.BoolSize + 17 + msgp.BytesPrefixSize + len(z.PolicyConfigJSON) + 22 + msgp.BytesPrefixSize + len(z.NotificationConfigXML) + 19 + msgp.BytesPrefixSize + len(z.LifecycleConfigXML) + 20 + msgp.BytesPrefixSize + len(z.ObjectLockConfigXML) + 20 + msgp.BytesPrefixSize + len(z.VersioningConfigXML) + 20 + msgp.BytesPrefixSize + len(z.EncryptionConfigXML) + 17 + msgp.BytesPrefixSize + len(z.TaggingConfigXML) + 16 + msgp.BytesPrefixSize + len(z.QuotaConfigJSON) + 21 + msgp.BytesPrefixSize + len(z.ReplicationConfigXML) + 24 + msgp.BytesPrefixSize + len(z.BucketTargetsConfigJSON)
 	return
 }

cmd/bucket-policy.go

@@ -68,6 +68,12 @@ func getConditionValues(r *http.Request, lc string, username string, claims map[
 	principalType := "Anonymous"
 	if username != "" {
 		principalType = "User"
+		if len(claims) > 0 {
+			principalType = "AssumedRole"
+		}
+		if username == globalActiveCred.AccessKey {
+			principalType = "Account"
+		}
 	}
 
 	vid := r.URL.Query().Get("versionId")
@@ -143,7 +149,12 @@ func getConditionValues(r *http.Request, lc string, username string, claims map[
 	for k, v := range claims {
 		vStr, ok := v.(string)
 		if ok {
-			args[k] = []string{vStr}
+			// Special case for AD/LDAP STS users
+			if k == ldapUser {
+				args["user"] = []string{vStr}
+			} else {
+				args[k] = []string{vStr}
+			}
 		}
 	}
 

cmd/bucket-replication.go

@@ -21,32 +21,21 @@ import (
 	"fmt"
 	"net/http"
 	"strings"
-	"sync"
 	"time"
 
 	miniogo "github.com/minio/minio-go/v7"
-	"github.com/minio/minio-go/v7/pkg/credentials"
 	"github.com/minio/minio-go/v7/pkg/encrypt"
 	"github.com/minio/minio-go/v7/pkg/tags"
 	"github.com/minio/minio/cmd/crypto"
 	xhttp "github.com/minio/minio/cmd/http"
 	"github.com/minio/minio/cmd/logger"
 	"github.com/minio/minio/pkg/bucket/replication"
-	"github.com/minio/minio/pkg/bucket/versioning"
 	"github.com/minio/minio/pkg/event"
 	iampolicy "github.com/minio/minio/pkg/iam/policy"
-	"github.com/minio/minio/pkg/madmin"
 )
 
-// BucketReplicationSys represents replication subsystem
-type BucketReplicationSys struct {
-	sync.RWMutex
-	targetsMap    map[string]*miniogo.Core
-	targetsARNMap map[string]string
-}
-
-// GetConfig - gets replication config associated to a given bucket name.
-func (sys *BucketReplicationSys) GetConfig(ctx context.Context, bucketName string) (rc *replication.Config, err error) {
+// gets replication config associated to a given bucket name.
+func getReplicationConfig(ctx context.Context, bucketName string) (rc *replication.Config, err error) {
 	if globalIsGateway {
 		objAPI := newObjectLayerWithoutSafeModeFn()
 		if objAPI == nil {
@@ -59,153 +48,52 @@ func (sys *BucketReplicationSys) GetConfig(ctx context.Context, bucketName strin
 	return globalBucketMetadataSys.GetReplicationConfig(ctx, bucketName)
 }
 
-// SetTarget - sets a new minio-go client replication target for this bucket.
-func (sys *BucketReplicationSys) SetTarget(ctx context.Context, bucket string, tgt *madmin.BucketReplicationTarget) error {
-	if globalIsGateway {
-		return nil
-	}
-	// delete replication targets that were removed
-	if tgt.Empty() {
-		sys.Lock()
-		if currTgt, ok := sys.targetsMap[bucket]; ok {
-			delete(sys.targetsARNMap, currTgt.EndpointURL().String())
-		}
-		delete(sys.targetsMap, bucket)
-		sys.Unlock()
-		return nil
-	}
-	clnt, err := getReplicationTargetClient(tgt)
-	if err != nil {
-		return BucketReplicationTargetNotFound{Bucket: tgt.TargetBucket}
-	}
-	ok, err := clnt.BucketExists(ctx, tgt.TargetBucket)
-	if err != nil {
-		return err
-	}
-	if !ok {
-		return BucketReplicationDestinationNotFound{Bucket: tgt.TargetBucket}
-	}
-	vcfg, err := clnt.GetBucketVersioning(ctx, tgt.TargetBucket)
-	if err != nil || vcfg.Status != string(versioning.Enabled) {
-		return BucketReplicationTargetNotVersioned{Bucket: tgt.TargetBucket}
-	}
-	sys.Lock()
-	sys.targetsMap[bucket] = clnt
-	sys.targetsARNMap[tgt.URL()] = tgt.Arn
-	sys.Unlock()
-	return nil
-}
-
-// GetTargetClient returns minio-go client for target instance
-func (sys *BucketReplicationSys) GetTargetClient(ctx context.Context, bucket string) *miniogo.Core {
-	var clnt *miniogo.Core
-	sys.RLock()
-	if c, ok := sys.targetsMap[bucket]; ok {
-		clnt = c
-	}
-	sys.RUnlock()
-	return clnt
-}
-
-// validateDestination returns error if replication destination bucket missing or not configured
+// validateReplicationDestination returns error if replication destination bucket missing or not configured
 // It also returns true if replication destination is same as this server.
-func (sys *BucketReplicationSys) validateDestination(ctx context.Context, bucket string, rCfg *replication.Config) (bool, error) {
-	clnt := sys.GetTargetClient(ctx, bucket)
+func validateReplicationDestination(ctx context.Context, bucket string, rCfg *replication.Config) (bool, error) {
+	clnt := globalBucketTargetSys.GetReplicationTargetClient(ctx, rCfg.RoleArn)
 	if clnt == nil {
-		return false, BucketReplicationTargetNotFound{Bucket: bucket}
+		return false, BucketRemoteTargetNotFound{Bucket: bucket}
 	}
 	if found, _ := clnt.BucketExists(ctx, rCfg.GetDestination().Bucket); !found {
 		return false, BucketReplicationDestinationNotFound{Bucket: rCfg.GetDestination().Bucket}
 	}
-	// validate replication ARN against target endpoint
-	for k, v := range sys.targetsARNMap {
-		if v == rCfg.ReplicationArn {
-			if k == clnt.EndpointURL().String() {
-				sameTarget, _ := isLocalHost(clnt.EndpointURL().Hostname(), clnt.EndpointURL().Port(), globalMinioPort)
-				return sameTarget, nil
+	if ret, err := globalBucketObjectLockSys.Get(bucket); err == nil {
+		if ret.LockEnabled {
+			lock, _, _, _, err := clnt.GetObjectLockConfig(ctx, rCfg.GetDestination().Bucket)
+			if err != nil || lock != "Enabled" {
+				return false, BucketReplicationDestinationMissingLock{Bucket: rCfg.GetDestination().Bucket}
 			}
 		}
 	}
-	return false, BucketReplicationTargetNotFound{Bucket: bucket}
-}
-
-// NewBucketReplicationSys - creates new replication system.
-func NewBucketReplicationSys() *BucketReplicationSys {
-	return &BucketReplicationSys{
-		targetsMap:    make(map[string]*miniogo.Core),
-		targetsARNMap: make(map[string]string),
-	}
-}
-
-// Init initializes the bucket replication subsystem for buckets with replication config
-func (sys *BucketReplicationSys) Init(ctx context.Context, buckets []BucketInfo, objAPI ObjectLayer) error {
-	if objAPI == nil {
-		return errServerNotInitialized
-	}
-
-	// In gateway mode, replication is not supported.
-	if globalIsGateway {
-		return nil
-	}
-
-	// Load bucket replication targets once during boot.
-	sys.load(ctx, buckets, objAPI)
-	return nil
-}
-
-// create minio-go clients for buckets having replication targets
-func (sys *BucketReplicationSys) load(ctx context.Context, buckets []BucketInfo, objAPI ObjectLayer) {
-	for _, bucket := range buckets {
-		tgt, err := globalBucketMetadataSys.GetReplicationTargetConfig(bucket.Name)
-		if err != nil {
-			continue
+	// validate replication ARN against target endpoint
+	c, ok := globalBucketTargetSys.arnRemotesMap[rCfg.RoleArn]
+	if ok {
+		if c.EndpointURL().String() == clnt.EndpointURL().String() {
+			sameTarget, _ := isLocalHost(clnt.EndpointURL().Hostname(), clnt.EndpointURL().Port(), globalMinioPort)
+			return sameTarget, nil
 		}
-		if tgt == nil || tgt.Empty() {
-			continue
-		}
-		tgtClient, err := getReplicationTargetClient(tgt)
-		if err != nil {
-			continue
-		}
-		sys.Lock()
-		sys.targetsMap[bucket.Name] = tgtClient
-		sys.targetsARNMap[tgt.URL()] = tgt.Arn
-		sys.Unlock()
 	}
+	return false, BucketRemoteTargetNotFound{Bucket: bucket}
 }
 
-// GetARN returns the ARN associated with replication target URL
-func (sys *BucketReplicationSys) getARN(endpoint string) string {
-	return sys.targetsARNMap[endpoint]
-}
-
-// getReplicationTargetInstanceTransport contains a singleton roundtripper.
-var getReplicationTargetInstanceTransport http.RoundTripper
-var getReplicationTargetInstanceTransportOnce sync.Once
-
-// Returns a minio-go Client configured to access remote host described in replication target config.
-var getReplicationTargetClient = func(tcfg *madmin.BucketReplicationTarget) (*miniogo.Core, error) {
-	config := tcfg.Credentials
-	// if Signature version '4' use NewV4 directly.
-	creds := credentials.NewStaticV4(config.AccessKey, config.SecretKey, "")
-	// if Signature version '2' use NewV2 directly.
-	if strings.ToUpper(tcfg.API) == "S3V2" {
-		creds = credentials.NewStaticV2(config.AccessKey, config.SecretKey, "")
+func mustReplicateWeb(ctx context.Context, r *http.Request, bucket, object string, meta map[string]string, replStatus string, permErr APIErrorCode) bool {
+	if permErr != ErrNone {
+		return false
 	}
-
-	getReplicationTargetInstanceTransportOnce.Do(func() {
-		getReplicationTargetInstanceTransport = NewGatewayHTTPTransport()
-	})
-	core, err := miniogo.NewCore(tcfg.Endpoint, &miniogo.Options{
-		Creds:     creds,
-		Secure:    tcfg.IsSSL,
-		Transport: getReplicationTargetInstanceTransport,
-	})
-	return core, err
+	return mustReplicater(ctx, r, bucket, object, meta, replStatus)
 }
 
 // mustReplicate returns true if object meets replication criteria.
-func (sys *BucketReplicationSys) mustReplicate(ctx context.Context, r *http.Request, bucket, object string, meta map[string]string, replStatus string) bool {
+func mustReplicate(ctx context.Context, r *http.Request, bucket, object string, meta map[string]string, replStatus string) bool {
+	if s3Err := isPutActionAllowed(getRequestAuthType(r), bucket, object, r, iampolicy.GetReplicationConfigurationAction); s3Err != ErrNone {
+		return false
+	}
+	return mustReplicater(ctx, r, bucket, object, meta, replStatus)
+}
+
+// mustReplicater returns true if object meets replication criteria.
+func mustReplicater(ctx context.Context, r *http.Request, bucket, object string, meta map[string]string, replStatus string) bool {
 	if globalIsGateway {
 		return false
 	}
@@ -215,10 +103,7 @@ func (sys *BucketReplicationSys) mustReplicate(ctx context.Context, r *http.Requ
 	if replication.StatusType(replStatus) == replication.Replica {
 		return false
 	}
-	if s3Err := isPutActionAllowed(getRequestAuthType(r), bucket, object, r, iampolicy.GetReplicationConfigurationAction); s3Err != ErrNone {
-		return false
-	}
-	cfg, err := globalBucketReplicationSys.GetConfig(ctx, bucket)
+	cfg, err := getReplicationConfig(ctx, bucket)
 	if err != nil {
 		return false
 	}
@@ -239,22 +124,29 @@ func putReplicationOpts(dest replication.Destination, objInfo ObjectInfo) (putOp
 		if k == xhttp.AmzBucketReplicationStatus {
 			continue
 		}
+		if strings.HasPrefix(strings.ToLower(k), ReservedMetadataPrefixLower) {
+			continue
+		}
 		meta[k] = v
 	}
-
 	tag, err := tags.ParseObjectTags(objInfo.UserTags)
 	if err != nil {
 		return
 	}
+	sc := dest.StorageClass
+	if sc == "" {
+		sc = objInfo.StorageClass
+	}
 	putOpts = miniogo.PutObjectOptions{
 		UserMetadata:         meta,
 		UserTags:             tag.ToMap(),
 		ContentType:          objInfo.ContentType,
 		ContentEncoding:      objInfo.ContentEncoding,
-		StorageClass:         dest.StorageClass,
+		StorageClass:         sc,
 		ReplicationVersionID: objInfo.VersionID,
 		ReplicationStatus:    miniogo.ReplicationStatusReplica,
 		ReplicationMTime:     objInfo.ModTime,
+		ReplicationETag:      objInfo.ETag,
 	}
 	if mode, ok := objInfo.UserDefined[xhttp.AmzObjectLockMode]; ok {
 		rmode := miniogo.RetentionMode(mode)
@@ -278,38 +170,53 @@ func putReplicationOpts(dest replication.Destination, objInfo ObjectInfo) (putOp
 
 // 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, bucket, object, versionID string, objectAPI ObjectLayer, eventArg *eventArgs, healPending bool) {
-	cfg, err := globalBucketReplicationSys.GetConfig(ctx, bucket)
-	if err != nil {
-		logger.LogIf(ctx, err)
-		return
-	}
-	tgt := globalBucketReplicationSys.GetTargetClient(ctx, bucket)
-	if tgt == nil {
-		return
-	}
-	gr, err := objectAPI.GetObjectNInfo(ctx, bucket, object, nil, http.Header{}, readLock, ObjectOptions{})
-	if err != nil {
-		return
-	}
-	defer gr.Close()
-	objInfo := gr.ObjInfo
-	size, err := objInfo.GetActualSize()
+func replicateObject(ctx context.Context, objInfo ObjectInfo, objectAPI ObjectLayer) {
+	bucket := objInfo.Bucket
+	object := objInfo.Name
+
+	cfg, err := getReplicationConfig(ctx, bucket)
 	if err != nil {
 		logger.LogIf(ctx, err)
 		return
 	}
 
-	dest := cfg.GetDestination()
-	if dest.Bucket == "" {
+	tgt := globalBucketTargetSys.GetReplicationTargetClient(ctx, cfg.RoleArn)
+	if tgt == nil {
 		return
 	}
+
+	gr, err := objectAPI.GetObjectNInfo(ctx, bucket, object, nil, http.Header{}, readLock, ObjectOptions{
+		VersionID: objInfo.VersionID,
+	})
+	if err != nil {
+		return
+	}
+
+	objInfo = gr.ObjInfo
+	size, err := objInfo.GetActualSize()
+	if err != nil {
+		logger.LogIf(ctx, err)
+		gr.Close()
+		return
+	}
+
+	dest := cfg.GetDestination()
+	if dest.Bucket == "" {
+		gr.Close()
+		return
+	}
+
+	// if heal encounters a pending replication status, either replication
+	// has failed due to server shutdown or crawler and PutObject replication are in contention.
+	healPending := objInfo.ReplicationStatus == replication.Pending
+
 	// In the rare event that replication is in pending state either due to
 	// server shut down/crash before replication completed or healing and PutObject
 	// race - do an additional stat to see if the version ID exists
 	if healPending {
 		_, err := tgt.StatObject(ctx, dest.Bucket, object, miniogo.StatObjectOptions{VersionID: objInfo.VersionID})
 		if err == nil {
+			gr.Close()
 			// object with same VersionID already exists, replication kicked off by
 			// PutObject might have completed.
 			return
@@ -319,37 +226,96 @@ func replicateObject(ctx context.Context, bucket, object, versionID string, obje
 
 	replicationStatus := replication.Complete
 	_, err = tgt.PutObject(ctx, dest.Bucket, object, gr, size, "", "", putOpts)
+	gr.Close()
 	if err != nil {
 		replicationStatus = replication.Failed
-		// Notify replication failure  event.
-		if eventArg == nil {
-			eventArg = &eventArgs{
-				BucketName: bucket,
-				Object:     objInfo,
-				Host:       "Internal: [Replication]",
-			}
-		}
-		eventArg.EventName = event.OperationReplicationFailed
-		eventArg.Object.UserDefined[xhttp.AmzBucketReplicationStatus] = replicationStatus.String()
-		sendEvent(*eventArg)
 	}
 	objInfo.UserDefined[xhttp.AmzBucketReplicationStatus] = replicationStatus.String()
 	if objInfo.UserTags != "" {
 		objInfo.UserDefined[xhttp.AmzObjectTagging] = objInfo.UserTags
 	}
+
+	// FIXME: add support for missing replication events
+	// - event.ObjectReplicationNotTracked
+	// - event.ObjectReplicationMissedThreshold
+	// - event.ObjectReplicationReplicatedAfterThreshold
+	if replicationStatus == replication.Failed {
+		sendEvent(eventArgs{
+			EventName:  event.ObjectReplicationFailed,
+			BucketName: bucket,
+			Object:     objInfo,
+			Host:       "Internal: [Replication]",
+		})
+	}
+
 	objInfo.metadataOnly = true // Perform only metadata updates.
 	if _, err = objectAPI.CopyObject(ctx, bucket, object, bucket, object, objInfo, ObjectOptions{
 		VersionID: objInfo.VersionID,
-	}, ObjectOptions{VersionID: objInfo.VersionID}); err != nil {
-		logger.LogIf(ctx, err)
+	}, ObjectOptions{
+		VersionID: objInfo.VersionID,
+	}); err != nil {
+		logger.LogIf(ctx, fmt.Errorf("Unable to update replication metadata for %s: %s", objInfo.VersionID, err))
 	}
 }
 
-// getReplicationARN gets existing ARN for an endpoint or generates a new one.
-func (sys *BucketReplicationSys) getReplicationARN(endpoint string) string {
-	arn, ok := sys.targetsARNMap[endpoint]
-	if ok {
-		return arn
+// filterReplicationStatusMetadata filters replication status metadata for COPY
+func filterReplicationStatusMetadata(metadata map[string]string) map[string]string {
+	// Copy on write
+	dst := metadata
+	var copied bool
+	delKey := func(key string) {
+		if _, ok := metadata[key]; !ok {
+			return
+		}
+		if !copied {
+			dst = make(map[string]string, len(metadata))
+			for k, v := range metadata {
+				dst[k] = v
+			}
+			copied = true
+		}
+		delete(dst, key)
 	}
-	return fmt.Sprintf("arn:minio:s3::%s:*", mustGetUUID())
+
+	delKey(xhttp.AmzBucketReplicationStatus)
+	return dst
+}
+
+type replicationState struct {
+	// add future metrics here
+	replicaCh chan ObjectInfo
+}
+
+func (r *replicationState) queueReplicaTask(oi ObjectInfo) {
+	select {
+	case r.replicaCh <- oi:
+	default:
+	}
+}
+
+var globalReplicationState *replicationState
+
+func newReplicationState() *replicationState {
+	return &replicationState{
+		// TODO: currently keeping it conservative
+		// but eventually can be tuned in future
+		replicaCh: make(chan ObjectInfo, 100),
+	}
+}
+
+func initBackgroundReplication(ctx context.Context, objectAPI ObjectLayer) {
+	if globalReplicationState == nil {
+		return
+	}
+	go func() {
+		defer close(globalReplicationState.replicaCh)
+		for {
+			select {
+			case <-ctx.Done():
+				return
+			case oi := <-globalReplicationState.replicaCh:
+				replicateObject(ctx, oi, objectAPI)
+			}
+		}
+	}()
 }

cmd/bucket-targets.go

@@ -0,0 +1,313 @@
+/*
+ * MinIO Cloud Storage, (C) 2020 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 cmd
+
+import (
+	"context"
+	"net/http"
+	"sync"
+
+	minio "github.com/minio/minio-go/v7"
+	miniogo "github.com/minio/minio-go/v7"
+	"github.com/minio/minio-go/v7/pkg/credentials"
+	"github.com/minio/minio/pkg/bucket/versioning"
+	"github.com/minio/minio/pkg/madmin"
+)
+
+// BucketTargetSys represents bucket targets subsystem
+type BucketTargetSys struct {
+	sync.RWMutex
+	arnRemotesMap map[string]*miniogo.Core
+	targetsMap    map[string][]madmin.BucketTarget
+	clientsCache  map[string]*miniogo.Core
+}
+
+// ListTargets lists bucket targets across tenant or for individual bucket, and returns
+// results filtered by arnType
+func (sys *BucketTargetSys) ListTargets(ctx context.Context, bucket, arnType string) (targets []madmin.BucketTarget) {
+	if bucket != "" {
+		if ts, err := sys.ListBucketTargets(ctx, bucket); err == nil {
+			for _, t := range ts.Targets {
+				if string(t.Type) == arnType || arnType == "" {
+					targets = append(targets, t.Clone())
+				}
+			}
+		}
+		return targets
+	}
+	sys.RLock()
+	defer sys.RUnlock()
+	for _, tgts := range sys.targetsMap {
+		for _, t := range tgts {
+			if string(t.Type) == arnType || arnType == "" {
+				targets = append(targets, t.Clone())
+			}
+		}
+	}
+	return
+}
+
+// ListBucketTargets - gets list of bucket targets for this bucket.
+func (sys *BucketTargetSys) ListBucketTargets(ctx context.Context, bucket string) (*madmin.BucketTargets, error) {
+
+	sys.RLock()
+	defer sys.RUnlock()
+
+	tgts, ok := sys.targetsMap[bucket]
+	if ok {
+		return &madmin.BucketTargets{Targets: tgts}, nil
+	}
+	return nil, BucketRemoteTargetNotFound{Bucket: bucket}
+}
+
+// SetTarget - sets a new minio-go client target for this bucket.
+func (sys *BucketTargetSys) SetTarget(ctx context.Context, bucket string, tgt *madmin.BucketTarget) error {
+	if globalIsGateway {
+		return nil
+	}
+	if !tgt.Type.IsValid() {
+		return BucketRemoteArnTypeInvalid{Bucket: bucket}
+	}
+	clnt, err := sys.getRemoteTargetClient(tgt)
+	if err != nil {
+		return BucketRemoteTargetNotFound{Bucket: tgt.TargetBucket}
+	}
+	if tgt.Type == madmin.ReplicationService {
+		if !globalBucketVersioningSys.Enabled(bucket) {
+			return BucketReplicationSourceNotVersioned{Bucket: bucket}
+		}
+		vcfg, err := clnt.GetBucketVersioning(ctx, tgt.TargetBucket)
+		if err != nil {
+			if minio.ToErrorResponse(err).Code == "NoSuchBucket" {
+				return BucketRemoteTargetNotFound{Bucket: tgt.TargetBucket}
+			}
+			return BucketRemoteConnectionErr{Bucket: tgt.TargetBucket}
+		}
+		if vcfg.Status != string(versioning.Enabled) {
+			return BucketReplicationTargetNotVersioned{Bucket: tgt.TargetBucket}
+		}
+	}
+
+	sys.Lock()
+	defer sys.Unlock()
+
+	tgts := sys.targetsMap[bucket]
+	newtgts := make([]madmin.BucketTarget, len(tgts))
+	found := false
+	for idx, t := range tgts {
+		if t.Type == tgt.Type {
+			if t.Arn == tgt.Arn {
+				return BucketRemoteAlreadyExists{Bucket: t.TargetBucket}
+			}
+			newtgts[idx] = *tgt
+			found = true
+			continue
+		}
+		newtgts[idx] = t
+	}
+	if !found {
+		newtgts = append(newtgts, *tgt)
+	}
+
+	sys.targetsMap[bucket] = newtgts
+	sys.arnRemotesMap[tgt.Arn] = clnt
+	if _, ok := sys.clientsCache[clnt.EndpointURL().String()]; !ok {
+		sys.clientsCache[clnt.EndpointURL().String()] = clnt
+	}
+	return nil
+}
+
+// RemoveTarget - removes a remote bucket target for this source bucket.
+func (sys *BucketTargetSys) RemoveTarget(ctx context.Context, bucket, arnStr string) error {
+	if globalIsGateway {
+		return nil
+	}
+	if arnStr == "" {
+		return BucketRemoteArnInvalid{Bucket: bucket}
+	}
+	arn, err := madmin.ParseARN(arnStr)
+	if err != nil {
+		return BucketRemoteArnInvalid{Bucket: bucket}
+	}
+	if arn.Type == madmin.ReplicationService {
+		// reject removal of remote target if replication configuration is present
+		rcfg, err := getReplicationConfig(ctx, bucket)
+		if err == nil && rcfg.RoleArn == arnStr {
+			if _, ok := sys.arnRemotesMap[arnStr]; ok {
+				return BucketRemoteRemoveDisallowed{Bucket: bucket}
+			}
+		}
+	}
+	// delete ARN type from list of matching targets
+	sys.Lock()
+	defer sys.Unlock()
+	targets := make([]madmin.BucketTarget, 0)
+	found := false
+	tgts := sys.targetsMap[bucket]
+	for _, tgt := range tgts {
+		if tgt.Arn != arnStr {
+			targets = append(targets, tgt)
+			continue
+		}
+		found = true
+	}
+	if !found {
+		return BucketRemoteTargetNotFound{Bucket: bucket}
+	}
+	sys.targetsMap[bucket] = targets
+	delete(sys.arnRemotesMap, arnStr)
+	return nil
+}
+
+// GetReplicationTargetClient returns minio-go client for replication target instance
+func (sys *BucketTargetSys) GetReplicationTargetClient(ctx context.Context, arn string) *miniogo.Core {
+	sys.RLock()
+	defer sys.RUnlock()
+	return sys.arnRemotesMap[arn]
+}
+
+// NewBucketTargetSys - creates new replication system.
+func NewBucketTargetSys() *BucketTargetSys {
+	return &BucketTargetSys{
+		arnRemotesMap: make(map[string]*miniogo.Core),
+		targetsMap:    make(map[string][]madmin.BucketTarget),
+		clientsCache:  make(map[string]*miniogo.Core),
+	}
+}
+
+// Init initializes the bucket targets subsystem for buckets which have targets configured.
+func (sys *BucketTargetSys) Init(ctx context.Context, buckets []BucketInfo, objAPI ObjectLayer) error {
+	if objAPI == nil {
+		return errServerNotInitialized
+	}
+
+	// In gateway mode, bucket targets is not supported.
+	if globalIsGateway {
+		return nil
+	}
+
+	// Load bucket targets once during boot in background.
+	go sys.load(ctx, buckets, objAPI)
+	return nil
+}
+
+// UpdateTarget updates target to reflect metadata updates
+func (sys *BucketTargetSys) UpdateTarget(bucket string, cfg *madmin.BucketTargets) {
+	if sys == nil {
+		return
+	}
+	sys.Lock()
+	defer sys.Unlock()
+	if cfg == nil || cfg.Empty() {
+		// remove target and arn association
+		if tgts, ok := sys.targetsMap[bucket]; ok {
+			for _, t := range tgts {
+				delete(sys.arnRemotesMap, t.Arn)
+			}
+		}
+		delete(sys.targetsMap, bucket)
+		return
+	}
+
+	if len(cfg.Targets) > 0 {
+		sys.targetsMap[bucket] = cfg.Targets
+	}
+	for _, tgt := range cfg.Targets {
+		tgtClient, err := sys.getRemoteTargetClient(&tgt)
+		if err != nil {
+			continue
+		}
+		sys.arnRemotesMap[tgt.Arn] = tgtClient
+		if _, ok := sys.clientsCache[tgtClient.EndpointURL().String()]; !ok {
+			sys.clientsCache[tgtClient.EndpointURL().String()] = tgtClient
+		}
+	}
+	sys.targetsMap[bucket] = cfg.Targets
+}
+
+// create minio-go clients for buckets having remote targets
+func (sys *BucketTargetSys) load(ctx context.Context, buckets []BucketInfo, objAPI ObjectLayer) {
+	for _, bucket := range buckets {
+		cfg, err := globalBucketMetadataSys.GetBucketTargetsConfig(bucket.Name)
+		if err != nil {
+			continue
+		}
+		if cfg == nil || cfg.Empty() {
+			continue
+		}
+		if len(cfg.Targets) > 0 {
+			sys.targetsMap[bucket.Name] = cfg.Targets
+		}
+		for _, tgt := range cfg.Targets {
+			tgtClient, err := sys.getRemoteTargetClient(&tgt)
+			if err != nil {
+				continue
+			}
+			sys.arnRemotesMap[tgt.Arn] = tgtClient
+			if _, ok := sys.clientsCache[tgtClient.EndpointURL().String()]; !ok {
+				sys.clientsCache[tgtClient.EndpointURL().String()] = tgtClient
+			}
+		}
+		sys.targetsMap[bucket.Name] = cfg.Targets
+	}
+}
+
+// getRemoteTargetInstanceTransport contains a singleton roundtripper.
+var getRemoteTargetInstanceTransport http.RoundTripper
+var getRemoteTargetInstanceTransportOnce sync.Once
+
+// Returns a minio-go Client configured to access remote host described in replication target config.
+func (sys *BucketTargetSys) getRemoteTargetClient(tcfg *madmin.BucketTarget) (*miniogo.Core, error) {
+	if clnt, ok := sys.clientsCache[tcfg.Endpoint]; ok {
+		return clnt, nil
+	}
+	config := tcfg.Credentials
+	creds := credentials.NewStaticV4(config.AccessKey, config.SecretKey, "")
+
+	getRemoteTargetInstanceTransportOnce.Do(func() {
+		getRemoteTargetInstanceTransport = NewGatewayHTTPTransport()
+	})
+	core, err := miniogo.NewCore(tcfg.Endpoint, &miniogo.Options{
+		Creds:     creds,
+		Secure:    tcfg.Secure,
+		Transport: getRemoteTargetInstanceTransport,
+	})
+	return core, err
+}
+
+// getRemoteARN gets existing ARN for an endpoint or generates a new one.
+func (sys *BucketTargetSys) getRemoteARN(bucket string, target *madmin.BucketTarget) string {
+	if target == nil {
+		return ""
+	}
+	tgts := sys.targetsMap[bucket]
+	for _, tgt := range tgts {
+		if tgt.Type == target.Type && tgt.TargetBucket == target.TargetBucket && target.URL() == tgt.URL() {
+			return tgt.Arn
+		}
+	}
+	if !madmin.ServiceType(target.Type).IsValid() {
+		return ""
+	}
+	arn := madmin.ARN{
+		Type:   target.Type,
+		ID:     mustGetUUID(),
+		Region: target.Region,
+		Bucket: target.TargetBucket,
+	}
+	return arn.String()
+}

cmd/common-main.go

@@ -20,6 +20,7 @@ import (
 	"crypto/x509"
 	"encoding/gob"
 	"errors"
+	"fmt"
 	"net"
 	"net/url"
 	"os"
@@ -42,15 +43,11 @@ func init() {
 	logger.Init(GOPATH, GOROOT)
 	logger.RegisterError(config.FmtError)
 
-	// Initialize globalConsoleSys system
-	globalConsoleSys = NewConsoleLogger(GlobalContext)
-	logger.AddTarget(globalConsoleSys)
-
 	gob.Register(StorageErr(""))
 }
 
 func verifyObjectLayerFeatures(name string, objAPI ObjectLayer) {
-	if (globalAutoEncryption || GlobalKMS != nil) && !objAPI.IsEncryptionSupported() {
+	if (GlobalKMS != nil) && !objAPI.IsEncryptionSupported() {
 		logger.Fatal(errInvalidArgument,
 			"Encryption support is requested but '%s' does not support encryption", name)
 	}
@@ -293,7 +290,7 @@ func logStartupMessage(msg string) {
 	logger.StartupMessage(msg)
 }
 
-func getTLSConfig() (x509Certs []*x509.Certificate, c *certs.Certs, secureConn bool, err error) {
+func getTLSConfig() (x509Certs []*x509.Certificate, manager *certs.Manager, secureConn bool, err error) {
 	if !(isFile(getPublicCertFile()) && isFile(getPrivateKeyFile())) {
 		return nil, nil, false, nil
 	}
@@ -302,11 +299,62 @@ func getTLSConfig() (x509Certs []*x509.Certificate, c *certs.Certs, secureConn b
 		return nil, nil, false, err
 	}
 
-	c, err = certs.New(getPublicCertFile(), getPrivateKeyFile(), config.LoadX509KeyPair)
+	manager, err = certs.NewManager(GlobalContext, getPublicCertFile(), getPrivateKeyFile(), config.LoadX509KeyPair)
 	if err != nil {
 		return nil, nil, false, err
 	}
 
+	// MinIO has support for multiple certificates. It expects the following structure:
+	//  certs/
+	//   │
+	//   ├─ public.crt
+	//   ├─ private.key
+	//   │
+	//   ├─ example.com/
+	//   │   │
+	//   │   ├─ public.crt
+	//   │   └─ private.key
+	//   └─ foobar.org/
+	//      │
+	//      ├─ public.crt
+	//      └─ private.key
+	//   ...
+	//
+	// 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())
+	if err != nil {
+		return nil, nil, false, err
+	}
+	defer root.Close()
+
+	files, err := root.Readdir(-1)
+	if err != nil {
+		return nil, nil, false, err
+	}
+	for _, file := range files {
+		// We exclude any regular file and the "CAs/" directory.
+		// The "CAs/" directory contains (root) CA certificates
+		// that MinIO adds to its list of trusted roots (tls.Config.RootCAs).
+		// Therefore, "CAs/" does not contain X.509 certificates that
+		// are meant to be served by MinIO.
+		if !file.IsDir() || file.Name() == "CAs" {
+			continue
+		}
+
+		var (
+			certFile = filepath.Join(root.Name(), file.Name(), publicCertFile)
+			keyFile  = filepath.Join(root.Name(), file.Name(), privateKeyFile)
+		)
+		if !isFile(certFile) || !isFile(keyFile) {
+			continue
+		}
+		if err := manager.AddCertificate(certFile, keyFile); err != nil {
+			err = fmt.Errorf("Failed to load TLS certificate '%s': %v", certFile, err)
+			logger.LogIf(GlobalContext, err, logger.Minio)
+		}
+	}
 	secureConn = true
-	return x509Certs, c, secureConn, nil
+	return x509Certs, manager, secureConn, nil
 }

cmd/config-common.go

@@ -21,7 +21,6 @@ import (
 	"context"
 	"errors"
 
-	"github.com/minio/minio/cmd/logger"
 	"github.com/minio/minio/pkg/hash"
 )
 
@@ -36,8 +35,6 @@ func readConfig(ctx context.Context, objAPI ObjectLayer, configFile string) ([]b
 			return nil, errConfigNotFound
 		}
 
-		logger.GetReqInfo(ctx).AppendTags("configFile", configFile)
-		logger.LogIf(ctx, err)
 		return nil, err
 	}
 
@@ -49,7 +46,11 @@ func readConfig(ctx context.Context, objAPI ObjectLayer, configFile string) ([]b
 	return buffer.Bytes(), nil
 }
 
-func deleteConfig(ctx context.Context, objAPI ObjectLayer, configFile string) error {
+type objectDeleter interface {
+	DeleteObject(ctx context.Context, bucket, object string, opts ObjectOptions) (ObjectInfo, error)
+}
+
+func deleteConfig(ctx context.Context, objAPI objectDeleter, configFile string) error {
 	_, err := objAPI.DeleteObject(ctx, minioMetaBucket, configFile, ObjectOptions{})
 	if err != nil && isErrObjectNotFound(err) {
 		return errConfigNotFound
@@ -74,8 +75,6 @@ func checkConfig(ctx context.Context, objAPI ObjectLayer, configFile string) err
 			return errConfigNotFound
 		}
 
-		logger.GetReqInfo(ctx).AppendTags("configFile", configFile)
-		logger.LogIf(ctx, err)
 		return err
 	}
 	return nil

cmd/config-current.go

@@ -25,8 +25,9 @@ import (
 	"github.com/minio/minio/cmd/config/api"
 	"github.com/minio/minio/cmd/config/cache"
 	"github.com/minio/minio/cmd/config/compress"
+	"github.com/minio/minio/cmd/config/crawler"
+	"github.com/minio/minio/cmd/config/dns"
 	"github.com/minio/minio/cmd/config/etcd"
-	"github.com/minio/minio/cmd/config/etcd/dns"
 	xldap "github.com/minio/minio/cmd/config/identity/ldap"
 	"github.com/minio/minio/cmd/config/identity/openid"
 	"github.com/minio/minio/cmd/config/notify"
@@ -55,6 +56,7 @@ func initHelp() {
 		config.KmsKesSubSys:         crypto.DefaultKesKVS,
 		config.LoggerWebhookSubSys:  logger.DefaultKVS,
 		config.AuditWebhookSubSys:   logger.DefaultAuditKVS,
+		config.CrawlerSubSys:        crawler.DefaultKVS,
 	}
 	for k, v := range notify.DefaultNotificationKVS {
 		kvs[k] = v
@@ -92,7 +94,7 @@ func initHelp() {
 		},
 		config.HelpKV{
 			Key:         config.PolicyOPASubSys,
-			Description: "enable external OPA for policy enforcement",
+			Description: "[DEPRECATED] enable external OPA for policy enforcement",
 		},
 		config.HelpKV{
 			Key:         config.KmsVaultSubSys,
@@ -106,6 +108,10 @@ func initHelp() {
 			Key:         config.APISubSys,
 			Description: "manage global HTTP API call specific features, such as throttling, authentication types, etc.",
 		},
+		config.HelpKV{
+			Key:         config.CrawlerSubSys,
+			Description: "manage continuous disk crawling for bucket disk usage, lifecycle, quota and data integrity checks",
+		},
 		config.HelpKV{
 			Key:             config.LoggerWebhookSubSys,
 			Description:     "send server logs to webhook endpoints",
@@ -185,6 +191,7 @@ func initHelp() {
 		config.EtcdSubSys:           etcd.Help,
 		config.CacheSubSys:          cache.Help,
 		config.CompressionSubSys:    compress.Help,
+		config.CrawlerSubSys:        crawler.Help,
 		config.IdentityOpenIDSubSys: openid.Help,
 		config.IdentityLDAPSubSys:   xldap.Help,
 		config.PolicyOPASubSys:      opa.Help,
@@ -213,7 +220,7 @@ var (
 	globalServerConfigMu sync.RWMutex
 )
 
-func validateConfig(s config.Config) error {
+func validateConfig(s config.Config, setDriveCount int) error {
 	// Disable merging env values with config for validation.
 	env.SetEnvOff()
 
@@ -233,8 +240,7 @@ func validateConfig(s config.Config) error {
 	}
 
 	if globalIsErasure {
-		if _, err := storageclass.LookupConfig(s[config.StorageClassSubSys][config.Default],
-			globalErasureSetDriveCount); err != nil {
+		if _, err := storageclass.LookupConfig(s[config.StorageClassSubSys][config.Default], setDriveCount); err != nil {
 			return err
 		}
 	}
@@ -247,6 +253,10 @@ func validateConfig(s config.Config) error {
 		return err
 	}
 
+	if _, err := crawler.LookupConfig(s[config.CrawlerSubSys][config.Default]); err != nil {
+		return err
+	}
+
 	{
 		etcdCfg, err := etcd.LookupConfig(s[config.EtcdSubSys][config.Default], globalRootCAs)
 		if err != nil {
@@ -307,11 +317,10 @@ func validateConfig(s config.Config) error {
 		return err
 	}
 
-	return notify.TestNotificationTargets(s, GlobalContext.Done(), NewGatewayHTTPTransport(),
-		globalNotificationSys.ConfiguredTargetIDs())
+	return notify.TestNotificationTargets(GlobalContext, s, NewGatewayHTTPTransport(), globalNotificationSys.ConfiguredTargetIDs())
 }
 
-func lookupConfigs(s config.Config) {
+func lookupConfigs(s config.Config, setDriveCount int) {
 	ctx := GlobalContext
 
 	var err error
@@ -323,6 +332,19 @@ func lookupConfigs(s config.Config) {
 		}
 	}
 
+	if dnsURL, dnsUser, dnsPass, ok := env.LookupEnv(config.EnvDNSWebhook); ok {
+		globalDNSConfig, err = dns.NewOperatorDNS(dnsURL,
+			dns.Authentication(dnsUser, dnsPass),
+			dns.RootCAs(globalRootCAs))
+		if err != nil {
+			if globalIsGateway {
+				logger.FatalIf(err, "Unable to initialize remote webhook DNS config")
+			} else {
+				logger.LogIf(ctx, fmt.Errorf("Unable to initialize remote webhook DNS config %w", err))
+			}
+		}
+	}
+
 	etcdCfg, err := etcd.LookupConfig(s[config.EtcdSubSys][config.Default], globalRootCAs)
 	if err != nil {
 		if globalIsGateway {
@@ -344,19 +366,25 @@ func lookupConfigs(s config.Config) {
 			}
 		}
 
-		if len(globalDomainNames) != 0 && !globalDomainIPs.IsEmpty() && globalEtcdClient != nil && globalDNSConfig == nil {
-			globalDNSConfig, err = dns.NewCoreDNS(etcdCfg.Config,
-				dns.DomainNames(globalDomainNames),
-				dns.DomainIPs(globalDomainIPs),
-				dns.DomainPort(globalMinioPort),
-				dns.CoreDNSPath(etcdCfg.CoreDNSPath),
-			)
-			if err != nil {
-				if globalIsGateway {
-					logger.FatalIf(err, "Unable to initialize DNS config")
-				} else {
-					logger.LogIf(ctx, fmt.Errorf("Unable to initialize DNS config for %s: %w",
-						globalDomainNames, err))
+		if len(globalDomainNames) != 0 && !globalDomainIPs.IsEmpty() && globalEtcdClient != nil {
+			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))
+			} else {
+				globalDNSConfig, err = dns.NewCoreDNS(etcdCfg.Config,
+					dns.DomainNames(globalDomainNames),
+					dns.DomainIPs(globalDomainIPs),
+					dns.DomainPort(globalMinioPort),
+					dns.CoreDNSPath(etcdCfg.CoreDNSPath),
+				)
+				if err != nil {
+					if globalIsGateway {
+						logger.FatalIf(err, "Unable to initialize DNS config")
+					} else {
+						logger.LogIf(ctx, fmt.Errorf("Unable to initialize DNS config for %s: %w",
+							globalDomainNames, err))
+					}
 				}
 			}
 		}
@@ -379,11 +407,15 @@ func lookupConfigs(s config.Config) {
 		logger.LogIf(ctx, fmt.Errorf("Invalid api configuration: %w", err))
 	}
 
-	globalAPIConfig.init(apiConfig)
+	globalAPIConfig.init(apiConfig, setDriveCount)
+
+	// Initialize remote instance transport once.
+	getRemoteInstanceTransportOnce.Do(func() {
+		getRemoteInstanceTransport = newGatewayHTTPTransport(apiConfig.RemoteTransportDeadline)
+	})
 
 	if globalIsErasure {
-		globalStorageClass, err = storageclass.LookupConfig(s[config.StorageClassSubSys][config.Default],
-			globalErasureSetDriveCount)
+		globalStorageClass, err = storageclass.LookupConfig(s[config.StorageClassSubSys][config.Default], setDriveCount)
 		if err != nil {
 			logger.LogIf(ctx, fmt.Errorf("Unable to initialize storage class config: %w", err))
 		}
@@ -406,6 +438,10 @@ func lookupConfigs(s config.Config) {
 			}
 		}
 	}
+	globalCrawlerConfig, err = crawler.LookupConfig(s[config.CrawlerSubSys][config.Default])
+	if err != nil {
+		logger.LogIf(ctx, fmt.Errorf("Unable to read crawler config: %w", err))
+	}
 
 	kmsCfg, err := crypto.LookupConfig(s, globalCertsCADir.Get(), NewGatewayHTTPTransport())
 	if err != nil {
@@ -419,6 +455,9 @@ func lookupConfigs(s config.Config) {
 
 	// Enable auto-encryption if enabled
 	globalAutoEncryption = kmsCfg.AutoEncryption
+	if globalAutoEncryption && !globalIsGateway {
+		logger.LogIf(ctx, fmt.Errorf("%s env is deprecated please migrate to using `mc encrypt` at bucket level", crypto.EnvKMSAutoEncryption))
+	}
 
 	globalCompressConfig, err = compress.LookupConfig(s[config.CompressionSubSys][config.Default])
 	if err != nil {
@@ -457,37 +496,41 @@ func lookupConfigs(s config.Config) {
 	for _, l := range loggerCfg.HTTP {
 		if l.Enabled {
 			// Enable http logging
-			logger.AddTarget(
+			if err = logger.AddTarget(
 				http.New(http.WithEndpoint(l.Endpoint),
 					http.WithAuthToken(l.AuthToken),
 					http.WithUserAgent(loggerUserAgent),
 					http.WithLogKind(string(logger.All)),
 					http.WithTransport(NewGatewayHTTPTransport()),
 				),
-			)
+			); err != nil {
+				logger.LogIf(ctx, fmt.Errorf("Unable to initialize console HTTP target: %w", err))
+			}
 		}
 	}
 
 	for _, l := range loggerCfg.Audit {
 		if l.Enabled {
 			// Enable http audit logging
-			logger.AddAuditTarget(
+			if err = logger.AddAuditTarget(
 				http.New(http.WithEndpoint(l.Endpoint),
 					http.WithAuthToken(l.AuthToken),
 					http.WithUserAgent(loggerUserAgent),
 					http.WithLogKind(string(logger.All)),
 					http.WithTransport(NewGatewayHTTPTransport()),
 				),
-			)
+			); err != nil {
+				logger.LogIf(ctx, fmt.Errorf("Unable to initialize audit HTTP target: %w", err))
+			}
 		}
 	}
 
-	globalConfigTargetList, err = notify.GetNotificationTargets(s, GlobalContext.Done(), NewGatewayHTTPTransport(), false)
+	globalConfigTargetList, err = notify.GetNotificationTargets(GlobalContext, s, NewGatewayHTTPTransport(), false)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("Unable to initialize notification target(s): %w", err))
 	}
 
-	globalEnvTargetList, err = notify.GetNotificationTargets(newServerConfig(), GlobalContext.Done(), NewGatewayHTTPTransport(), true)
+	globalEnvTargetList, err = notify.GetNotificationTargets(GlobalContext, newServerConfig(), NewGatewayHTTPTransport(), true)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("Unable to initialize notification target(s): %w", err))
 	}
@@ -601,7 +644,7 @@ func loadConfig(objAPI ObjectLayer) error {
 	}
 
 	// Override any values from ENVs.
-	lookupConfigs(srvCfg)
+	lookupConfigs(srvCfg, objAPI.SetDriveCount())
 
 	// hold the mutex lock before a new config is assigned.
 	globalServerConfigMu.Lock()

cmd/config/api/api.go

@@ -29,15 +29,17 @@ import (
 
 // API sub-system constants
 const (
-	apiRequestsMax      = "requests_max"
-	apiRequestsDeadline = "requests_deadline"
-	apiReadyDeadline    = "ready_deadline"
-	apiCorsAllowOrigin  = "cors_allow_origin"
+	apiRequestsMax             = "requests_max"
+	apiRequestsDeadline        = "requests_deadline"
+	apiReadyDeadline           = "ready_deadline"
+	apiCorsAllowOrigin         = "cors_allow_origin"
+	apiRemoteTransportDeadline = "remote_transport_deadline"
 
-	EnvAPIRequestsMax      = "MINIO_API_REQUESTS_MAX"
-	EnvAPIRequestsDeadline = "MINIO_API_REQUESTS_DEADLINE"
-	EnvAPIReadyDeadline    = "MINIO_API_READY_DEADLINE"
-	EnvAPICorsAllowOrigin  = "MINIO_API_CORS_ALLOW_ORIGIN"
+	EnvAPIRequestsMax             = "MINIO_API_REQUESTS_MAX"
+	EnvAPIRequestsDeadline        = "MINIO_API_REQUESTS_DEADLINE"
+	EnvAPIReadyDeadline           = "MINIO_API_READY_DEADLINE"
+	EnvAPICorsAllowOrigin         = "MINIO_API_CORS_ALLOW_ORIGIN"
+	EnvAPIRemoteTransportDeadline = "MINIO_API_REMOTE_TRANSPORT_DEADLINE"
 )
 
 // DefaultKVS - default storage class config
@@ -59,15 +61,20 @@ var (
 			Key:   apiCorsAllowOrigin,
 			Value: "*",
 		},
+		config.KV{
+			Key:   apiRemoteTransportDeadline,
+			Value: "2h",
+		},
 	}
 )
 
 // Config storage class configuration
 type Config struct {
-	APIRequestsMax      int           `json:"requests_max"`
-	APIRequestsDeadline time.Duration `json:"requests_deadline"`
-	APIReadyDeadline    time.Duration `json:"ready_deadline"`
-	APICorsAllowOrigin  []string      `json:"cors_allow_origin"`
+	RequestsMax             int           `json:"requests_max"`
+	RequestsDeadline        time.Duration `json:"requests_deadline"`
+	ReadyDeadline           time.Duration `json:"ready_deadline"`
+	CorsAllowOrigin         []string      `json:"cors_allow_origin"`
+	RemoteTransportDeadline time.Duration `json:"remote_transport_deadline"`
 }
 
 // UnmarshalJSON - Validate SS and RRS parity when unmarshalling JSON.
@@ -108,10 +115,17 @@ func LookupConfig(kvs config.KVS) (cfg Config, err error) {
 	}
 
 	corsAllowOrigin := strings.Split(env.Get(EnvAPICorsAllowOrigin, kvs.Get(apiCorsAllowOrigin)), ",")
+
+	remoteTransportDeadline, err := time.ParseDuration(env.Get(EnvAPIRemoteTransportDeadline, kvs.Get(apiRemoteTransportDeadline)))
+	if err != nil {
+		return cfg, err
+	}
+
 	return Config{
-		APIRequestsMax:      requestsMax,
-		APIRequestsDeadline: requestsDeadline,
-		APIReadyDeadline:    readyDeadline,
-		APICorsAllowOrigin:  corsAllowOrigin,
+		RequestsMax:             requestsMax,
+		RequestsDeadline:        requestsDeadline,
+		ReadyDeadline:           readyDeadline,
+		CorsAllowOrigin:         corsAllowOrigin,
+		RemoteTransportDeadline: remoteTransportDeadline,
 	}, nil
 }

cmd/config/api/help.go

@@ -39,5 +39,11 @@ var (
 			Optional:    true,
 			Type:        "csv",
 		},
+		config.HelpKV{
+			Key:         apiRemoteTransportDeadline,
+			Description: `set the deadline for API requests on remote transports while proxying between federated instances e.g. "2h"`,
+			Optional:    true,
+			Type:        "duration",
+		},
 	}
 )

cmd/config/certs.go

@@ -24,8 +24,6 @@ import (
 	"crypto/x509"
 	"encoding/pem"
 	"io/ioutil"
-	"os"
-	"path"
 
 	"github.com/minio/minio/pkg/env"
 )
@@ -69,38 +67,6 @@ func ParsePublicCertFile(certFile string) (x509Certs []*x509.Certificate, err er
 	return x509Certs, nil
 }
 
-// GetRootCAs - returns all the root CAs into certPool
-// at the input certsCADir
-func GetRootCAs(certsCAsDir string) (*x509.CertPool, error) {
-	rootCAs, _ := x509.SystemCertPool()
-	if rootCAs == nil {
-		// In some systems (like Windows) system cert pool is
-		// not supported or no certificates are present on the
-		// system - so we create a new cert pool.
-		rootCAs = x509.NewCertPool()
-	}
-
-	fis, err := ioutil.ReadDir(certsCAsDir)
-	if err != nil {
-		if os.IsNotExist(err) || os.IsPermission(err) {
-			// Return success if CA's directory is missing or permission denied.
-			err = nil
-		}
-		return rootCAs, err
-	}
-
-	// Load all custom CA files.
-	for _, fi := range fis {
-		caCert, err := ioutil.ReadFile(path.Join(certsCAsDir, fi.Name()))
-		if err != nil {
-			// ignore files which are not readable.
-			continue
-		}
-		rootCAs.AppendCertsFromPEM(caCert)
-	}
-	return rootCAs, nil
-}
-
 // LoadX509KeyPair - load an X509 key pair (private key , certificate)
 // from the provided paths. The private key may be encrypted and is
 // decrypted using the ENV_VAR: MINIO_CERT_PASSWD.

cmd/config/certs_test.go

@@ -20,7 +20,6 @@ import (
 	"fmt"
 	"io/ioutil"
 	"os"
-	"path/filepath"
 	"runtime"
 	"testing"
 )
@@ -194,60 +193,6 @@ M9ofSEt/bdRD
 	}
 }
 
-func TestGetRootCAs(t *testing.T) {
-	emptydir, err := ioutil.TempDir("", "test-get-root-cas")
-	if err != nil {
-		t.Fatalf("Unable create temp directory. %v", emptydir)
-	}
-	defer os.RemoveAll(emptydir)
-
-	dir1, err := ioutil.TempDir("", "test-get-root-cas")
-	if err != nil {
-		t.Fatalf("Unable create temp directory. %v", dir1)
-	}
-	defer os.RemoveAll(dir1)
-	if err = os.Mkdir(filepath.Join(dir1, "empty-dir"), 0755); err != nil {
-		t.Fatalf("Unable create empty dir. %v", err)
-	}
-
-	dir2, err := ioutil.TempDir("", "test-get-root-cas")
-	if err != nil {
-		t.Fatalf("Unable create temp directory. %v", dir2)
-	}
-	defer os.RemoveAll(dir2)
-	if err = ioutil.WriteFile(filepath.Join(dir2, "empty-file"), []byte{}, 0644); err != nil {
-		t.Fatalf("Unable create test file. %v", err)
-	}
-
-	testCases := []struct {
-		certCAsDir  string
-		expectedErr error
-	}{
-		// ignores non-existent directories.
-		{"nonexistent-dir", nil},
-		// Ignores directories.
-		{dir1, nil},
-		// Ignore empty directory.
-		{emptydir, nil},
-		// Loads the cert properly.
-		{dir2, nil},
-	}
-
-	for _, testCase := range testCases {
-		_, err := GetRootCAs(testCase.certCAsDir)
-
-		if testCase.expectedErr == nil {
-			if err != nil {
-				t.Fatalf("error: expected = <nil>, got = %v", err)
-			}
-		} else if err == nil {
-			t.Fatalf("error: expected = %v, got = <nil>", testCase.expectedErr)
-		} else if testCase.expectedErr.Error() != err.Error() {
-			t.Fatalf("error: expected = %v, got = %v", testCase.expectedErr, err)
-		}
-	}
-}
-
 func TestLoadX509KeyPair(t *testing.T) {
 	for i, testCase := range loadX509KeyPairTests {
 		privateKey, err := createTempFile("private.key", testCase.privateKey)

cmd/config/config.go

@@ -76,6 +76,7 @@ const (
 	KmsKesSubSys         = "kms_kes"
 	LoggerWebhookSubSys  = "logger_webhook"
 	AuditWebhookSubSys   = "audit_webhook"
+	CrawlerSubSys        = "crawler"
 
 	// Add new constants here if you add new fields to config.
 )
@@ -112,6 +113,7 @@ var SubSystems = set.CreateStringSet([]string{
 	PolicyOPASubSys,
 	IdentityLDAPSubSys,
 	IdentityOpenIDSubSys,
+	CrawlerSubSys,
 	NotifyAMQPSubSys,
 	NotifyESSubSys,
 	NotifyKafkaSubSys,
@@ -138,6 +140,7 @@ var SubSystemsSingleTargets = set.CreateStringSet([]string{
 	PolicyOPASubSys,
 	IdentityLDAPSubSys,
 	IdentityOpenIDSubSys,
+	CrawlerSubSys,
 }...)
 
 // Constant separators
@@ -448,6 +451,11 @@ func (c Config) Merge() Config {
 					ckvs.Set(kv.Key, kv.Value)
 				}
 			}
+			if _, ok := cp[subSys]; !ok {
+				// A config subsystem was removed or server was downgraded.
+				Logger.Info("config: ignoring unknown subsystem config %q\n", subSys)
+				continue
+			}
 			cp[subSys][tgt] = ckvs
 		}
 	}

cmd/config/constants.go

@@ -31,11 +31,13 @@ const (
 	EnvDomain       = "MINIO_DOMAIN"
 	EnvRegionName   = "MINIO_REGION_NAME"
 	EnvPublicIPs    = "MINIO_PUBLIC_IPS"
-	EnvEndpoints    = "MINIO_ENDPOINTS"
 	EnvFSOSync      = "MINIO_FS_OSYNC"
+	EnvArgs         = "MINIO_ARGS"
+	EnvDNSWebhook   = "MINIO_DNS_WEBHOOK_ENDPOINT"
 
 	EnvUpdate = "MINIO_UPDATE"
 
-	EnvWorm   = "MINIO_WORM"   // legacy
-	EnvRegion = "MINIO_REGION" // legacy
+	EnvEndpoints = "MINIO_ENDPOINTS" // legacy
+	EnvWorm      = "MINIO_WORM"      // legacy
+	EnvRegion    = "MINIO_REGION"    // legacy
 )

cmd/config/crawler/crawler.go

@@ -0,0 +1,67 @@
+/*
+ * MinIO Cloud Storage, (C) 2020 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 crawler
+
+import (
+	"errors"
+
+	"github.com/minio/minio/cmd/config"
+)
+
+// Compression environment variables
+const (
+	BitrotScan = "bitrotscan"
+)
+
+// Config represents the crawler settings.
+type Config struct {
+	// Bitrot will perform bitrot scan on local disk when checking objects.
+	Bitrot bool `json:"bitrotscan"`
+}
+
+var (
+	// DefaultKVS - default KV config for crawler settings
+	DefaultKVS = config.KVS{
+		config.KV{
+			Key:   BitrotScan,
+			Value: config.EnableOff,
+		},
+	}
+
+	// Help provides help for config values
+	Help = config.HelpKVS{
+		config.HelpKV{
+			Key:         BitrotScan,
+			Description: `perform bitrot scan on disks when checking objects during crawl`,
+			Optional:    true,
+			Type:        "on|off",
+		},
+	}
+)
+
+// LookupConfig - lookup config and override with valid environment settings if any.
+func LookupConfig(kvs config.KVS) (cfg Config, err error) {
+	if err = config.CheckValidKeys(config.CrawlerSubSys, kvs, DefaultKVS); err != nil {
+		return cfg, err
+	}
+	bitrot := kvs.Get(BitrotScan)
+	if bitrot != config.EnableOn && bitrot != config.EnableOff {
+		return cfg, errors.New(BitrotScan + ": must be 'on' or 'off'")
+	}
+	cfg.Bitrot = bitrot == config.EnableOn
+	return cfg, nil
+}

cmd/config/dns/etcd_dns.go

@@ -26,9 +26,8 @@ import (
 	"strings"
 	"time"
 
-	"github.com/minio/minio-go/v7/pkg/set"
-
 	"github.com/coredns/coredns/plugin/etcd/msg"
+	"github.com/minio/minio-go/v7/pkg/set"
 	"go.etcd.io/etcd/v3/clientv3"
 )
 
@@ -214,6 +213,11 @@ func (c *CoreDNS) DeleteRecord(record SrvRecord) error {
 	return nil
 }
 
+// String stringer name for this implementation of dns.Store
+func (c *CoreDNS) String() string {
+	return "etcdDNS"
+}
+
 // CoreDNS - represents dns config for coredns server.
 type CoreDNS struct {
 	domainNames []string
@@ -223,13 +227,13 @@ type CoreDNS struct {
 	etcdClient  *clientv3.Client
 }
 
-// Option - functional options pattern style
-type Option func(*CoreDNS)
+// EtcdOption - functional options pattern style
+type EtcdOption func(*CoreDNS)
 
 // DomainNames set a list of domain names used by this CoreDNS
 // client setting, note this will fail if set to empty when
 // constructor initializes.
-func DomainNames(domainNames []string) Option {
+func DomainNames(domainNames []string) EtcdOption {
 	return func(args *CoreDNS) {
 		args.domainNames = domainNames
 	}
@@ -237,14 +241,14 @@ func DomainNames(domainNames []string) Option {
 
 // DomainIPs set a list of custom domain IPs, note this will
 // fail if set to empty when constructor initializes.
-func DomainIPs(domainIPs set.StringSet) Option {
+func DomainIPs(domainIPs set.StringSet) EtcdOption {
 	return func(args *CoreDNS) {
 		args.domainIPs = domainIPs
 	}
 }
 
 // DomainPort - is a string version of server port
-func DomainPort(domainPort string) Option {
+func DomainPort(domainPort string) EtcdOption {
 	return func(args *CoreDNS) {
 		args.domainPort = domainPort
 	}
@@ -253,14 +257,14 @@ func DomainPort(domainPort string) Option {
 // CoreDNSPath - custom prefix on etcd to populate DNS
 // service records, optional and can be empty.
 // if empty then c.prefixPath is used i.e "/skydns"
-func CoreDNSPath(prefix string) Option {
+func CoreDNSPath(prefix string) EtcdOption {
 	return func(args *CoreDNS) {
 		args.prefixPath = prefix
 	}
 }
 
 // NewCoreDNS - initialize a new coreDNS set/unset values.
-func NewCoreDNS(cfg clientv3.Config, setters ...Option) (Store, error) {
+func NewCoreDNS(cfg clientv3.Config, setters ...EtcdOption) (Store, error) {
 	etcdClient, err := clientv3.New(cfg)
 	if err != nil {
 		return nil, err

cmd/config/dns/operator_dns.go

@@ -0,0 +1,221 @@
+/*
+ * MinIO Cloud Storage, (C) 2020 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 dns
+
+import (
+	"context"
+	"crypto/tls"
+	"crypto/x509"
+	"errors"
+	"fmt"
+	"net"
+	"net/http"
+	"net/url"
+	"strconv"
+	"time"
+
+	"github.com/dgrijalva/jwt-go"
+	"github.com/minio/minio/cmd/config"
+	xhttp "github.com/minio/minio/cmd/http"
+)
+
+var (
+	defaultOperatorContextTimeout = 10 * time.Second
+	// ErrNotImplemented - Indicates the functionality which is not implemented
+	ErrNotImplemented = errors.New("The method is not implemented")
+)
+
+func (c *OperatorDNS) addAuthHeader(r *http.Request) error {
+	if c.username == "" || c.password == "" {
+		return nil
+	}
+
+	claims := &jwt.StandardClaims{
+		ExpiresAt: int64(15 * time.Minute),
+		Issuer:    c.username,
+		Subject:   config.EnvDNSWebhook,
+	}
+
+	token := jwt.NewWithClaims(jwt.SigningMethodHS512, claims)
+	ss, err := token.SignedString([]byte(c.password))
+	if err != nil {
+		return err
+	}
+
+	r.Header.Set("Authorization", "Bearer "+ss)
+	return nil
+}
+
+func (c *OperatorDNS) endpoint(bucket string, delete bool) (string, error) {
+	u, err := url.Parse(c.Endpoint)
+	if err != nil {
+		return "", err
+	}
+	q := u.Query()
+	q.Add("bucket", bucket)
+	q.Add("delete", strconv.FormatBool(delete))
+	u.RawQuery = q.Encode()
+	return u.String(), nil
+}
+
+// Put - Adds DNS entries into operator webhook server
+func (c *OperatorDNS) Put(bucket string) error {
+	ctx, cancel := context.WithTimeout(context.Background(), defaultOperatorContextTimeout)
+	defer cancel()
+	e, err := c.endpoint(bucket, false)
+	if err != nil {
+		return err
+	}
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, e, nil)
+	if err != nil {
+		return err
+	}
+	if err = c.addAuthHeader(req); err != nil {
+		return err
+	}
+	resp, err := c.httpClient.Do(req)
+	if err != nil {
+		if derr := c.Delete(bucket); derr != nil {
+			return derr
+		}
+	}
+	xhttp.DrainBody(resp.Body)
+	if resp.StatusCode != http.StatusOK {
+		return fmt.Errorf("request to create the service for bucket %s, failed with status %s", bucket, resp.Status)
+	}
+	return nil
+}
+
+// Delete - Removes DNS entries added in Put().
+func (c *OperatorDNS) Delete(bucket string) error {
+	ctx, cancel := context.WithTimeout(context.Background(), defaultOperatorContextTimeout)
+	defer cancel()
+	e, err := c.endpoint(bucket, true)
+	if err != nil {
+		return err
+	}
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, e, nil)
+	if err != nil {
+		return err
+	}
+	if err = c.addAuthHeader(req); err != nil {
+		return err
+	}
+	resp, err := c.httpClient.Do(req)
+	if err != nil {
+		return err
+	}
+	xhttp.DrainBody(resp.Body)
+	if resp.StatusCode != http.StatusOK {
+		return fmt.Errorf("request to delete the service for bucket %s, failed with status %s", bucket, resp.Status)
+	}
+	return nil
+}
+
+// DeleteRecord - Removes a specific DNS entry
+// No Op for Operator because operator deals on with bucket entries
+func (c *OperatorDNS) DeleteRecord(record SrvRecord) error {
+	return ErrNotImplemented
+}
+
+// Close closes the internal http client
+func (c *OperatorDNS) Close() error {
+	c.httpClient.CloseIdleConnections()
+	return nil
+}
+
+// List - Retrieves list of DNS entries for the domain.
+// This is a No Op for Operator because, there is no intent to enforce global
+// namespace at MinIO level with this DNS entry. The global namespace in
+// enforced by the Kubernetes Operator
+func (c *OperatorDNS) List() (srvRecords map[string][]SrvRecord, err error) {
+	return nil, ErrNotImplemented
+}
+
+// Get - Retrieves DNS records for a bucket.
+// This is a No Op for Operator because, there is no intent to enforce global
+// namespace at MinIO level with this DNS entry. The global namespace in
+// enforced by the Kubernetes Operator
+func (c *OperatorDNS) Get(bucket string) (srvRecords []SrvRecord, err error) {
+	return nil, ErrNotImplemented
+}
+
+// String stringer name for this implementation of dns.Store
+func (c *OperatorDNS) String() string {
+	return "webhookDNS"
+}
+
+// OperatorDNS - represents dns config for MinIO k8s operator.
+type OperatorDNS struct {
+	httpClient *http.Client
+	Endpoint   string
+	rootCAs    *x509.CertPool
+	username   string
+	password   string
+}
+
+// OperatorOption - functional options pattern style for OperatorDNS
+type OperatorOption func(*OperatorDNS)
+
+// Authentication - custom username and password for authenticating at the endpoint
+func Authentication(username, password string) OperatorOption {
+	return func(args *OperatorDNS) {
+		args.username = username
+		args.password = password
+	}
+}
+
+// RootCAs - add custom trust certs pool
+func RootCAs(CAs *x509.CertPool) OperatorOption {
+	return func(args *OperatorDNS) {
+		args.rootCAs = CAs
+	}
+}
+
+// NewOperatorDNS - initialize a new K8S Operator DNS set/unset values.
+func NewOperatorDNS(endpoint string, setters ...OperatorOption) (Store, error) {
+	if endpoint == "" {
+		return nil, errors.New("invalid argument")
+	}
+
+	args := &OperatorDNS{
+		Endpoint: endpoint,
+	}
+	for _, setter := range setters {
+		setter(args)
+	}
+	args.httpClient = &http.Client{
+		Transport: &http.Transport{
+			Proxy: http.ProxyFromEnvironment,
+			DialContext: (&net.Dialer{
+				Timeout:   3 * time.Second,
+				KeepAlive: 5 * time.Second,
+			}).DialContext,
+			ResponseHeaderTimeout: 3 * time.Second,
+			TLSHandshakeTimeout:   3 * time.Second,
+			ExpectContinueTimeout: 3 * time.Second,
+			TLSClientConfig: &tls.Config{
+				RootCAs: args.rootCAs,
+			},
+			// Go net/http automatically unzip if content-type is
+			// gzip disable this feature, as we are always interested
+			// in raw stream.
+			DisableCompression: true,
+		},
+	}
+	return args, nil
+}

cmd/config/dns/store.go

@@ -24,4 +24,5 @@ type Store interface {
 	List() (map[string][]SrvRecord, error)
 	DeleteRecord(record SrvRecord) error
 	Close() error
+	String() string
 }

cmd/config/errors.go

@@ -187,12 +187,6 @@ Example 1:
 		"",
 	)
 
-	ErrCorruptedBackend = newErrFn(
-		"Unable to use the specified backend, pre-existing content detected",
-		"Please ensure your disk mount does not have any pre-existing content",
-		"",
-	)
-
 	ErrUnableToWriteInBackend = newErrFn(
 		"Unable to write to the backend",
 		"Please ensure MinIO binary has write permissions for the backend",

cmd/config/identity/openid/jwt.go

@@ -25,6 +25,7 @@ import (
 	"net/http"
 	"strconv"
 	"strings"
+	"sync"
 	"time"
 
 	jwtgo "github.com/dgrijalva/jwt-go"
@@ -49,10 +50,14 @@ type Config struct {
 	publicKeys   map[string]crypto.PublicKey
 	transport    *http.Transport
 	closeRespFn  func(io.ReadCloser)
+	mutex        *sync.Mutex
 }
 
 // PopulatePublicKey - populates a new publickey from the JWKS URL.
 func (r *Config) PopulatePublicKey() error {
+	r.mutex.Lock()
+	defer r.mutex.Unlock()
+
 	if r.JWKS.URL == nil || r.JWKS.URL.String() == "" {
 		return nil
 	}
@@ -185,7 +190,15 @@ func (p *JWT) Validate(token, dsecs string) (map[string]interface{}, error) {
 	var claims jwtgo.MapClaims
 	jwtToken, err := jp.ParseWithClaims(token, &claims, keyFuncCallback)
 	if err != nil {
-		return nil, err
+		// Re-populate the public key in-case the JWKS
+		// pubkeys are refreshed
+		if err = p.PopulatePublicKey(); err != nil {
+			return nil, err
+		}
+		jwtToken, err = jwtgo.ParseWithClaims(token, &claims, keyFuncCallback)
+		if err != nil {
+			return nil, err
+		}
 	}
 
 	if !jwtToken.Valid {
@@ -317,6 +330,7 @@ func LookupConfig(kvs config.KVS, transport *http.Transport, closeRespFn func(io
 		ClientID:    env.Get(EnvIdentityOpenIDClientID, kvs.Get(ClientID)),
 		transport:   transport,
 		closeRespFn: closeRespFn,
+		mutex:       &sync.Mutex{}, // allocate for copying
 	}
 
 	configURL := env.Get(EnvIdentityOpenIDURL, kvs.Get(ConfigURL))

cmd/config/identity/openid/jwt_test.go

@@ -20,6 +20,7 @@ import (
 	"crypto"
 	"encoding/json"
 	"net/url"
+	"sync"
 	"testing"
 	"time"
 
@@ -89,6 +90,7 @@ func TestJWTAzureFail(t *testing.T) {
 	}
 
 	cfg := Config{}
+	cfg.mutex = &sync.Mutex{}
 	cfg.JWKS.URL = u1
 	cfg.publicKeys = keys
 	jwt := NewJWT(cfg)
@@ -136,6 +138,7 @@ func TestJWT(t *testing.T) {
 	}
 
 	cfg := Config{}
+	cfg.mutex = &sync.Mutex{}
 	cfg.JWKS.URL = u1
 	cfg.publicKeys = keys
 	jwt := NewJWT(cfg)

cmd/config/logger.go

@@ -0,0 +1,27 @@
+/*
+ * MinIO Cloud Storage, (C) 2020 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 config
+
+import "context"
+
+// Logger contains injected logger methods.
+var Logger = struct {
+	Info  func(msg string, data ...interface{})
+	LogIf func(ctx context.Context, err error, errKind ...interface{})
+}{
+	// Initialized via injection.
+}

cmd/config/notify/help.go

@@ -559,6 +559,18 @@ var (
 			Optional:    true,
 			Type:        "number",
 		},
+		config.HelpKV{
+			Key:         target.ElasticUsername,
+			Description: "username for Elasticsearch basic-auth",
+			Optional:    true,
+			Type:        "string",
+		},
+		config.HelpKV{
+			Key:         target.ElasticPassword,
+			Description: "password for Elasticsearch basic-auth",
+			Optional:    true,
+			Type:        "string",
+		},
 		config.HelpKV{
 			Key:         config.Comment,
 			Description: config.DefaultComment,

cmd/config/notify/legacy.go

@@ -201,6 +201,14 @@ func SetNotifyES(s config.Config, esName string, cfg target.ElasticsearchArgs) e
 			Key:   target.ElasticQueueLimit,
 			Value: strconv.Itoa(int(cfg.QueueLimit)),
 		},
+		config.KV{
+			Key:   target.ElasticUsername,
+			Value: cfg.Username,
+		},
+		config.KV{
+			Key:   target.ElasticPassword,
+			Value: cfg.Password,
+		},
 	}
 
 	return nil
@@ -331,7 +339,7 @@ func SetNotifyPostgres(s config.Config, psqName string, cfg target.PostgreSQLArg
 		},
 		config.KV{
 			Key:   target.PostgresUsername,
-			Value: cfg.User,
+			Value: cfg.Username,
 		},
 		config.KV{
 			Key:   target.PostgresPassword,

cmd/config/notify/parse.go

@@ -43,11 +43,10 @@ var ErrTargetsOffline = errors.New("one or more targets are offline. Please use
 
 // TestNotificationTargets is similar to GetNotificationTargets()
 // avoids explicit registration.
-func TestNotificationTargets(cfg config.Config, doneCh <-chan struct{}, transport *http.Transport,
-	targetIDs []event.TargetID) error {
+func TestNotificationTargets(ctx context.Context, cfg config.Config, transport *http.Transport, targetIDs []event.TargetID) error {
 	test := true
 	returnOnTargetError := true
-	targets, err := RegisterNotificationTargets(cfg, doneCh, transport, targetIDs, test, returnOnTargetError)
+	targets, err := RegisterNotificationTargets(ctx, cfg, transport, targetIDs, test, returnOnTargetError)
 	if err == nil {
 		// Close all targets since we are only testing connections.
 		for _, t := range targets.TargetMap() {
@@ -60,9 +59,9 @@ func TestNotificationTargets(cfg config.Config, doneCh <-chan struct{}, transpor
 
 // GetNotificationTargets registers and initializes all notification
 // targets, returns error if any.
-func GetNotificationTargets(cfg config.Config, doneCh <-chan struct{}, transport *http.Transport, test bool) (*event.TargetList, error) {
+func GetNotificationTargets(ctx context.Context, cfg config.Config, transport *http.Transport, test bool) (*event.TargetList, error) {
 	returnOnTargetError := false
-	return RegisterNotificationTargets(cfg, doneCh, transport, nil, test, returnOnTargetError)
+	return RegisterNotificationTargets(ctx, cfg, transport, nil, test, returnOnTargetError)
 }
 
 // RegisterNotificationTargets - returns TargetList which contains enabled targets in serverConfig.
@@ -70,8 +69,8 @@ func GetNotificationTargets(cfg config.Config, doneCh <-chan struct{}, transport
 // * Add a new target in pkg/event/target package.
 // * Add newly added target configuration to serverConfig.Notify.<TARGET_NAME>.
 // * Handle the configuration in this function to create/add into TargetList.
-func RegisterNotificationTargets(cfg config.Config, doneCh <-chan struct{}, transport *http.Transport, targetIDs []event.TargetID, test bool, returnOnTargetError bool) (*event.TargetList, error) {
-	targetList, err := FetchRegisteredTargets(cfg, doneCh, transport, test, returnOnTargetError)
+func RegisterNotificationTargets(ctx context.Context, cfg config.Config, transport *http.Transport, targetIDs []event.TargetID, test bool, returnOnTargetError bool) (*event.TargetList, error) {
+	targetList, err := FetchRegisteredTargets(ctx, cfg, transport, test, returnOnTargetError)
 	if err != nil {
 		return targetList, err
 	}
@@ -94,7 +93,7 @@ func RegisterNotificationTargets(cfg config.Config, doneCh <-chan struct{}, tran
 // FetchRegisteredTargets - Returns a set of configured TargetList
 // If `returnOnTargetError` is set to true, The function returns when a target initialization fails
 // Else, the function will return a complete TargetList irrespective of errors
-func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport *http.Transport, test bool, returnOnTargetError bool) (_ *event.TargetList, err error) {
+func FetchRegisteredTargets(ctx context.Context, cfg config.Config, transport *http.Transport, test bool, returnOnTargetError bool) (_ *event.TargetList, err error) {
 	targetList := event.NewTargetList()
 	var targetsOffline bool
 
@@ -118,7 +117,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 		return nil, err
 	}
 
-	esTargets, err := GetNotifyES(cfg[config.NotifyESSubSys])
+	esTargets, err := GetNotifyES(cfg[config.NotifyESSubSys], transport)
 	if err != nil {
 		return nil, err
 	}
@@ -167,7 +166,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 		if !args.Enable {
 			continue
 		}
-		newTarget, err := target.NewAMQPTarget(id, args, doneCh, logger.LogOnceIf, test)
+		newTarget, err := target.NewAMQPTarget(id, args, ctx.Done(), logger.LogOnceIf, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -188,7 +187,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 		if !args.Enable {
 			continue
 		}
-		newTarget, err := target.NewElasticsearchTarget(id, args, doneCh, logger.LogOnceIf, test)
+		newTarget, err := target.NewElasticsearchTarget(id, args, ctx.Done(), logger.LogOnceIf, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -209,7 +208,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 			continue
 		}
 		args.TLS.RootCAs = transport.TLSClientConfig.RootCAs
-		newTarget, err := target.NewKafkaTarget(id, args, doneCh, logger.LogOnceIf, test)
+		newTarget, err := target.NewKafkaTarget(id, args, ctx.Done(), logger.LogOnceIf, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -230,7 +229,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 			continue
 		}
 		args.RootCAs = transport.TLSClientConfig.RootCAs
-		newTarget, err := target.NewMQTTTarget(id, args, doneCh, logger.LogOnceIf, test)
+		newTarget, err := target.NewMQTTTarget(id, args, ctx.Done(), logger.LogOnceIf, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -250,7 +249,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 		if !args.Enable {
 			continue
 		}
-		newTarget, err := target.NewMySQLTarget(id, args, doneCh, logger.LogOnceIf, test)
+		newTarget, err := target.NewMySQLTarget(id, args, ctx.Done(), logger.LogOnceIf, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -270,7 +269,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 		if !args.Enable {
 			continue
 		}
-		newTarget, err := target.NewNATSTarget(id, args, doneCh, logger.LogOnceIf, test)
+		newTarget, err := target.NewNATSTarget(id, args, ctx.Done(), logger.LogOnceIf, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -290,7 +289,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 		if !args.Enable {
 			continue
 		}
-		newTarget, err := target.NewNSQTarget(id, args, doneCh, logger.LogOnceIf, test)
+		newTarget, err := target.NewNSQTarget(id, args, ctx.Done(), logger.LogOnceIf, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -310,7 +309,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 		if !args.Enable {
 			continue
 		}
-		newTarget, err := target.NewPostgreSQLTarget(id, args, doneCh, logger.LogOnceIf, test)
+		newTarget, err := target.NewPostgreSQLTarget(id, args, ctx.Done(), logger.LogOnceIf, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -330,7 +329,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 		if !args.Enable {
 			continue
 		}
-		newTarget, err := target.NewRedisTarget(id, args, doneCh, logger.LogOnceIf, test)
+		newTarget, err := target.NewRedisTarget(id, args, ctx.Done(), logger.LogOnceIf, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -350,7 +349,7 @@ func FetchRegisteredTargets(cfg config.Config, doneCh <-chan struct{}, transport
 		if !args.Enable {
 			continue
 		}
-		newTarget, err := target.NewWebhookTarget(id, args, doneCh, logger.LogOnceIf, transport, test)
+		newTarget, err := target.NewWebhookTarget(ctx, id, args, logger.LogOnceIf, transport, test)
 		if err != nil {
 			targetsOffline = true
 			if returnOnTargetError {
@@ -1532,11 +1531,19 @@ var (
 			Key:   target.ElasticQueueLimit,
 			Value: "0",
 		},
+		config.KV{
+			Key:   target.ElasticUsername,
+			Value: "",
+		},
+		config.KV{
+			Key:   target.ElasticPassword,
+			Value: "",
+		},
 	}
 )
 
 // GetNotifyES - returns a map of registered notification 'elasticsearch' targets
-func GetNotifyES(esKVS map[string]config.KVS) (map[string]target.ElasticsearchArgs, error) {
+func GetNotifyES(esKVS map[string]config.KVS, transport *http.Transport) (map[string]target.ElasticsearchArgs, error) {
 	esTargets := make(map[string]target.ElasticsearchArgs)
 	for k, kv := range mergeTargets(esKVS, target.EnvElasticEnable, DefaultESKVS) {
 		enableEnv := target.EnvElasticEnable
@@ -1586,6 +1593,16 @@ func GetNotifyES(esKVS map[string]config.KVS) (map[string]target.ElasticsearchAr
 			queueDirEnv = queueDirEnv + config.Default + k
 		}
 
+		usernameEnv := target.EnvElasticUsername
+		if k != config.Default {
+			usernameEnv = usernameEnv + config.Default + k
+		}
+
+		passwordEnv := target.EnvElasticPassword
+		if k != config.Default {
+			passwordEnv = passwordEnv + config.Default + k
+		}
+
 		esArgs := target.ElasticsearchArgs{
 			Enable:     enabled,
 			Format:     env.Get(formatEnv, kv.Get(target.ElasticFormat)),
@@ -1593,6 +1610,9 @@ func GetNotifyES(esKVS map[string]config.KVS) (map[string]target.ElasticsearchAr
 			Index:      env.Get(indexEnv, kv.Get(target.ElasticIndex)),
 			QueueDir:   env.Get(queueDirEnv, kv.Get(target.ElasticQueueDir)),
 			QueueLimit: uint64(queueLimit),
+			Transport:  transport,
+			Username:   env.Get(usernameEnv, kv.Get(target.ElasticUsername)),
+			Password:   env.Get(passwordEnv, kv.Get(target.ElasticPassword)),
 		}
 		if err = esArgs.Validate(); err != nil {
 			return nil, err

cmd/config/policy/opa/help.go

@@ -1,5 +1,5 @@
 /*
- * MinIO Cloud Storage, (C) 2019 MinIO, Inc.
+ * MinIO Cloud Storage, (C) 2019,2020 MinIO, Inc.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -23,12 +23,12 @@ var (
 	Help = config.HelpKVS{
 		config.HelpKV{
 			Key:         URL,
-			Description: `OPA HTTP(s) endpoint e.g. "http://localhost:8181/v1/data/httpapi/authz/allow"`,
+			Description: `[DEPRECATED] OPA HTTP(s) endpoint e.g. "http://localhost:8181/v1/data/httpapi/authz/allow"`,
 			Type:        "url",
 		},
 		config.HelpKV{
 			Key:         AuthToken,
-			Description: "authorization token for OPA endpoint",
+			Description: "[DEPRECATED] authorization token for OPA endpoint",
 			Optional:    true,
 			Type:        "string",
 		},

cmd/config/storageclass/storage-class.go

@@ -156,7 +156,7 @@ func parseStorageClass(storageClassEnv string) (sc StorageClass, err error) {
 }
 
 // Validates the parity disks.
-func validateParity(ssParity, rrsParity, drivesPerSet int) (err error) {
+func validateParity(ssParity, rrsParity, setDriveCount int) (err error) {
 	if ssParity == 0 && rrsParity == 0 {
 		return nil
 	}
@@ -174,12 +174,12 @@ func validateParity(ssParity, rrsParity, drivesPerSet int) (err error) {
 		return fmt.Errorf("Reduced redundancy storage class parity %d should be greater than or equal to %d", rrsParity, minParityDisks)
 	}
 
-	if ssParity > drivesPerSet/2 {
-		return fmt.Errorf("Standard storage class parity %d should be less than or equal to %d", ssParity, drivesPerSet/2)
+	if ssParity > setDriveCount/2 {
+		return fmt.Errorf("Standard storage class parity %d should be less than or equal to %d", ssParity, setDriveCount/2)
 	}
 
-	if rrsParity > drivesPerSet/2 {
-		return fmt.Errorf("Reduced redundancy storage class parity %d should be less than  or equal to %d", rrsParity, drivesPerSet/2)
+	if rrsParity > setDriveCount/2 {
+		return fmt.Errorf("Reduced redundancy storage class parity %d should be less than  or equal to %d", rrsParity, setDriveCount/2)
 	}
 
 	if ssParity > 0 && rrsParity > 0 {
@@ -220,13 +220,13 @@ func Enabled(kvs config.KVS) bool {
 }
 
 // LookupConfig - lookup storage class config and override with valid environment settings if any.
-func LookupConfig(kvs config.KVS, drivesPerSet int) (cfg Config, err error) {
+func LookupConfig(kvs config.KVS, setDriveCount int) (cfg Config, err error) {
 	cfg = Config{}
-	cfg.Standard.Parity = drivesPerSet / 2
+	cfg.Standard.Parity = setDriveCount / 2
 	cfg.RRS.Parity = defaultRRSParity
 
 	if err = config.CheckValidKeys(config.StorageClassSubSys, kvs, DefaultKVS); err != nil {
-		return cfg, err
+		return Config{}, err
 	}
 
 	ssc := env.Get(StandardEnv, kvs.Get(ClassStandard))
@@ -235,17 +235,17 @@ func LookupConfig(kvs config.KVS, drivesPerSet int) (cfg Config, err error) {
 	if ssc != "" {
 		cfg.Standard, err = parseStorageClass(ssc)
 		if err != nil {
-			return cfg, err
+			return Config{}, err
 		}
 	}
 	if cfg.Standard.Parity == 0 {
-		cfg.Standard.Parity = drivesPerSet / 2
+		cfg.Standard.Parity = setDriveCount / 2
 	}
 
 	if rrsc != "" {
 		cfg.RRS, err = parseStorageClass(rrsc)
 		if err != nil {
-			return cfg, err
+			return Config{}, err
 		}
 	}
 	if cfg.RRS.Parity == 0 {
@@ -254,8 +254,8 @@ func LookupConfig(kvs config.KVS, drivesPerSet int) (cfg Config, err error) {
 
 	// Validation is done after parsing both the storage classes. This is needed because we need one
 	// storage class value to deduce the correct value of the other storage class.
-	if err = validateParity(cfg.Standard.Parity, cfg.RRS.Parity, drivesPerSet); err != nil {
-		return cfg, err
+	if err = validateParity(cfg.Standard.Parity, cfg.RRS.Parity, setDriveCount); err != nil {
+		return Config{}, err
 	}
 
 	return cfg, nil

cmd/config/storageclass/storage-class_test.go

@@ -69,10 +69,10 @@ func TestParseStorageClass(t *testing.T) {
 
 func TestValidateParity(t *testing.T) {
 	tests := []struct {
-		rrsParity    int
-		ssParity     int
-		success      bool
-		drivesPerSet int
+		rrsParity     int
+		ssParity      int
+		success       bool
+		setDriveCount int
 	}{
 		{2, 4, true, 16},
 		{3, 3, true, 16},
@@ -85,7 +85,7 @@ func TestValidateParity(t *testing.T) {
 		{9, 2, false, 16},
 	}
 	for i, tt := range tests {
-		err := validateParity(tt.ssParity, tt.rrsParity, tt.drivesPerSet)
+		err := validateParity(tt.ssParity, tt.rrsParity, tt.setDriveCount)
 		if err != nil && tt.success {
 			t.Errorf("Test %d, Expected success, got %s", i+1, err)
 		}

cmd/consolelogger.go

@@ -117,6 +117,29 @@ func (sys *HTTPConsoleLoggerSys) Subscribe(subCh chan interface{}, doneCh <-chan
 	sys.pubsub.Subscribe(subCh, doneCh, filter)
 }
 
+// Validate if HTTPConsoleLoggerSys is valid, always returns nil right now
+func (sys *HTTPConsoleLoggerSys) Validate() error {
+	return nil
+}
+
+// Content returns the console stdout log
+func (sys *HTTPConsoleLoggerSys) Content() (logs []log.Entry) {
+	sys.RLock()
+	sys.logBuf.Do(func(p interface{}) {
+		if p != nil {
+			lg, ok := p.(log.Info)
+			if ok {
+				if (lg.Entry != log.Entry{}) {
+					logs = append(logs, lg.Entry)
+				}
+			}
+		}
+	})
+	sys.RUnlock()
+
+	return
+}
+
 // Send log message 'e' to console and publish to console
 // log pubsub system
 func (sys *HTTPConsoleLoggerSys) Send(e interface{}, logKind string) error {

cmd/crypto/config.go

@@ -16,11 +16,14 @@ package crypto
 
 import (
 	"errors"
+	"math/rand"
 	"net/http"
 	"reflect"
 	"strconv"
+	"strings"
 
 	"github.com/minio/minio/cmd/config"
+	"github.com/minio/minio/pkg/ellipses"
 	"github.com/minio/minio/pkg/env"
 	xnet "github.com/minio/minio/pkg/net"
 )
@@ -167,7 +170,8 @@ const (
 
 const (
 	// EnvKMSKesEndpoint is the environment variable used to specify
-	// the kes server HTTPS endpoint.
+	// one or multiple KES server HTTPS endpoints. The individual
+	// endpoints should be separated by ','.
 	EnvKMSKesEndpoint = "MINIO_KMS_KES_ENDPOINT"
 
 	// EnvKMSKesKeyFile is the environment variable used to specify
@@ -216,16 +220,36 @@ func LookupKesConfig(kvs config.KVS) (KesConfig, error) {
 	kesCfg := KesConfig{}
 
 	endpointStr := env.Get(EnvKMSKesEndpoint, kvs.Get(KMSKesEndpoint))
-	if endpointStr != "" {
-		// Lookup kes configuration & overwrite config entry if ENV var is present
-		endpoint, err := xnet.ParseHTTPURL(endpointStr)
+	var endpoints []string
+	for _, endpoint := range strings.Split(endpointStr, ",") {
+		if strings.TrimSpace(endpoint) == "" {
+			continue
+		}
+		if !ellipses.HasEllipses(endpoint) {
+			endpoints = append(endpoints, endpoint)
+			continue
+		}
+		pattern, err := ellipses.FindEllipsesPatterns(endpoint)
 		if err != nil {
 			return kesCfg, err
 		}
-		endpointStr = endpoint.String()
+		for _, p := range pattern {
+			endpoints = append(endpoints, p.Expand()...)
+		}
+	}
+	if len(endpoints) == 0 {
+		return kesCfg, nil
 	}
 
-	kesCfg.Endpoint = endpointStr
+	randNum := rand.Intn(len(endpoints) + 1) // We add 1 b/c len(endpoints) may be 0: See: rand.Intn docs
+	kesCfg.Endpoint = make([]string, len(endpoints))
+	for i, endpoint := range endpoints {
+		endpoint, err := xnet.ParseHTTPURL(endpoint)
+		if err != nil {
+			return kesCfg, err
+		}
+		kesCfg.Endpoint[(randNum+i)%len(endpoints)] = endpoint.String()
+	}
 	kesCfg.KeyFile = env.Get(EnvKMSKesKeyFile, kvs.Get(KMSKesKeyFile))
 	kesCfg.CertFile = env.Get(EnvKMSKesCertFile, kvs.Get(KMSKesCertFile))
 	kesCfg.CAPath = env.Get(EnvKMSKesCAPath, kvs.Get(KMSKesCAPath))

cmd/crypto/header.go

@@ -18,9 +18,11 @@ import (
 	"bytes"
 	"crypto/md5"
 	"encoding/base64"
-	"encoding/json"
 	"net/http"
 	"strings"
+
+	jsoniter "github.com/json-iterator/go"
+	xhttp "github.com/minio/minio/cmd/http"
 )
 
 // SSEHeader is the general AWS SSE HTTP header key.
@@ -81,6 +83,8 @@ const (
 func RemoveSensitiveHeaders(h http.Header) {
 	h.Del(SSECKey)
 	h.Del(SSECopyKey)
+	h.Del(xhttp.AmzMetaUnencryptedContentLength)
+	h.Del(xhttp.AmzMetaUnencryptedContentMD5)
 }
 
 // IsRequested returns true if the HTTP headers indicates
@@ -144,6 +148,7 @@ func (s3KMS) ParseHTTP(h http.Header) (string, interface{}, error) {
 	contextStr, ok := h[SSEKmsContext]
 	if ok {
 		var context map[string]interface{}
+		var json = jsoniter.ConfigCompatibleWithStandardLibrary
 		if err := json.Unmarshal([]byte(contextStr[0]), &context); err != nil {
 			return "", nil, err
 		}

cmd/crypto/header_test.go

@@ -457,6 +457,16 @@ var removeSensitiveHeadersTests = []struct {
 			"X-Amz-Meta-Test-1": []string{"Test-1"},
 		},
 	},
+	{ // https://github.com/google/security-research/security/advisories/GHSA-76wf-9vgp-pj7w
+		Header: http.Header{
+			"X-Amz-Meta-X-Amz-Unencrypted-Content-Md5":    []string{"value"},
+			"X-Amz-Meta-X-Amz-Unencrypted-Content-Length": []string{"value"},
+			"X-Amz-Meta-Test-1":                           []string{"Test-1"},
+		},
+		ExpectedHeader: http.Header{
+			"X-Amz-Meta-Test-1": []string{"Test-1"},
+		},
+	},
 }
 
 func TestRemoveSensitiveHeaders(t *testing.T) {

cmd/crypto/kes.go

@@ -16,9 +16,9 @@ package crypto
 
 import (
 	"bytes"
+	"context"
 	"crypto/tls"
 	"crypto/x509"
-	"encoding/json"
 	"errors"
 	"fmt"
 	"io"
@@ -30,10 +30,13 @@ import (
 	"strings"
 	"time"
 
+	jsoniter "github.com/json-iterator/go"
 	xhttp "github.com/minio/minio/cmd/http"
 	xnet "github.com/minio/minio/pkg/net"
 )
 
+var json = jsoniter.ConfigCompatibleWithStandardLibrary
+
 // ErrKESKeyExists is the error returned a KES server
 // when a master key does exist.
 var ErrKESKeyExists = NewKESError(http.StatusBadRequest, "key does already exist")
@@ -43,8 +46,8 @@ var ErrKESKeyExists = NewKESError(http.StatusBadRequest, "key does already exist
 type KesConfig struct {
 	Enabled bool
 
-	// The kes server endpoint.
-	Endpoint string
+	// The KES server endpoints.
+	Endpoint []string
 
 	// The path to the TLS private key used
 	// by MinIO to authenticate to the kes
@@ -83,7 +86,7 @@ type KesConfig struct {
 // Verify verifies if the kes configuration is correct
 func (k KesConfig) Verify() (err error) {
 	switch {
-	case k.Endpoint == "":
+	case len(k.Endpoint) == 0:
 		err = Errorf("crypto: missing kes endpoint")
 	case k.CertFile == "":
 		err = Errorf("crypto: missing cert file")
@@ -98,7 +101,7 @@ func (k KesConfig) Verify() (err error) {
 type kesService struct {
 	client *kesClient
 
-	endpoint     string
+	endpoints    []string
 	defaultKeyID string
 }
 
@@ -113,23 +116,37 @@ func NewKes(cfg KesConfig) (KMS, error) {
 	if err != nil {
 		return nil, err
 	}
-	certPool, err := loadCACertificates(cfg.CAPath)
-	if err != nil {
-		return nil, err
+	if cfg.Transport.TLSClientConfig != nil {
+		if err = loadCACertificates(cfg.CAPath,
+			cfg.Transport.TLSClientConfig.RootCAs); err != nil {
+			return nil, err
+		}
+	} else {
+		rootCAs, _ := x509.SystemCertPool()
+		if rootCAs == nil {
+			// In some systems (like Windows) system cert pool is
+			// not supported or no certificates are present on the
+			// system - so we create a new cert pool.
+			rootCAs = x509.NewCertPool()
+		}
+		if err = loadCACertificates(cfg.CAPath, rootCAs); err != nil {
+			return nil, err
+		}
+		cfg.Transport.TLSClientConfig = &tls.Config{
+			RootCAs: rootCAs,
+		}
 	}
-	cfg.Transport.TLSClientConfig = &tls.Config{
-		Certificates: []tls.Certificate{cert},
-		RootCAs:      certPool,
-	}
-	cfg.Transport.ForceAttemptHTTP2 = true
+	cfg.Transport.TLSClientConfig.Certificates = []tls.Certificate{cert}
+	cfg.Transport.TLSClientConfig.NextProtos = []string{"h2"}
+
 	return &kesService{
 		client: &kesClient{
-			addr: cfg.Endpoint,
+			endpoints: cfg.Endpoint,
 			httpClient: http.Client{
 				Transport: cfg.Transport,
 			},
 		},
-		endpoint:     cfg.Endpoint,
+		endpoints:    cfg.Endpoint,
 		defaultKeyID: cfg.DefaultKeyID,
 	}, nil
 }
@@ -146,9 +163,9 @@ func (kes *kesService) DefaultKeyID() string {
 // method.
 func (kes *kesService) Info() KMSInfo {
 	return KMSInfo{
-		Endpoint: kes.endpoint,
-		Name:     kes.DefaultKeyID(),
-		AuthType: "TLS",
+		Endpoints: kes.endpoints,
+		Name:      kes.DefaultKeyID(),
+		AuthType:  "TLS",
 	}
 }
 
@@ -204,7 +221,7 @@ func (kes *kesService) UnsealKey(keyID string, sealedKey []byte, ctx Context) (k
 //   • GenerateDataKey (API: /v1/key/generate/)
 //   • DecryptDataKey  (API: /v1/key/decrypt/)
 type kesClient struct {
-	addr       string
+	endpoints  []string
 	httpClient http.Client
 }
 
@@ -215,8 +232,8 @@ type kesClient struct {
 // application does not have the cryptographic key at
 // any point in time.
 func (c *kesClient) CreateKey(name string) error {
-	url := fmt.Sprintf("%s/v1/key/create/%s", c.addr, url.PathEscape(name))
-	_, err := c.postRetry(url, nil, 0) // No request body and no response expected
+	path := fmt.Sprintf("/v1/key/create/%s", url.PathEscape(name))
+	_, err := c.postRetry(path, nil, 0) // No request body and no response expected
 	if err != nil {
 		return err
 	}
@@ -248,8 +265,8 @@ func (c *kesClient) GenerateDataKey(name string, context []byte) ([]byte, []byte
 	}
 
 	const limit = 1 << 20 // A plaintext/ciphertext key pair will never be larger than 1 MB
-	url := fmt.Sprintf("%s/v1/key/generate/%s", c.addr, url.PathEscape(name))
-	resp, err := c.postRetry(url, bytes.NewReader(body), limit)
+	path := fmt.Sprintf("/v1/key/generate/%s", url.PathEscape(name))
+	resp, err := c.postRetry(path, bytes.NewReader(body), limit)
 	if err != nil {
 		return nil, nil, err
 	}
@@ -285,8 +302,8 @@ func (c *kesClient) DecryptDataKey(name string, ciphertext, context []byte) ([]b
 	}
 
 	const limit = 1 << 20 // A data key will never be larger than 1 MiB
-	url := fmt.Sprintf("%s/v1/key/decrypt/%s", c.addr, url.PathEscape(name))
-	resp, err := c.postRetry(url, bytes.NewReader(body), limit)
+	path := fmt.Sprintf("/v1/key/decrypt/%s", url.PathEscape(name))
+	resp, err := c.postRetry(path, bytes.NewReader(body), limit)
 	if err != nil {
 		return nil, err
 	}
@@ -357,7 +374,16 @@ func parseErrorResponse(resp *http.Response) error {
 }
 
 func (c *kesClient) post(url string, body io.Reader, limit int64) (io.Reader, error) {
-	resp, err := c.httpClient.Post(url, "application/json", body)
+	ctx, cancel := context.WithTimeout(context.Background(), 3*time.Second)
+	defer cancel()
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, body)
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Content-Type", "application/json")
+
+	resp, err := c.httpClient.Do(req)
 	if err != nil {
 		return nil, err
 	}
@@ -376,26 +402,34 @@ func (c *kesClient) post(url string, body io.Reader, limit int64) (io.Reader, er
 	return &respBody, nil
 }
 
-func (c *kesClient) postRetry(url string, body io.ReadSeeker, limit int64) (io.Reader, error) {
+func (c *kesClient) postRetry(path string, body io.ReadSeeker, limit int64) (io.Reader, error) {
+	retryMax := 1 + len(c.endpoints)
 	for i := 0; ; i++ {
 		if body != nil {
 			body.Seek(0, io.SeekStart) // seek to the beginning of the body.
 		}
-		response, err := c.post(url, body, limit)
+
+		response, err := c.post(c.endpoints[i%len(c.endpoints)]+path, body, limit)
 		if err == nil {
 			return response, nil
 		}
 
-		if !xnet.IsNetworkOrHostDown(err) && !errors.Is(err, io.EOF) && !errors.Is(err, io.ErrUnexpectedEOF) {
+		// If the error is not temp. / retryable => fail the request immediately.
+		if !xnet.IsNetworkOrHostDown(err) &&
+			!errors.Is(err, io.EOF) &&
+			!errors.Is(err, io.ErrUnexpectedEOF) &&
+			!errors.Is(err, context.DeadlineExceeded) {
 			return nil, err
 		}
-
-		// retriable network errors.
-		remain := retryMax - i
-		if remain <= 0 {
+		if remain := retryMax - i; remain <= 0 { // Fail if we exceeded our retry limit.
 			return response, err
 		}
 
+		// If there are more KES instances then skip waiting and
+		// try the next endpoint directly.
+		if i < len(c.endpoints) {
+			continue
+		}
 		<-time.After(LinearJitterBackoff(retryWaitMin, retryWaitMax, i))
 	}
 }
@@ -413,24 +447,17 @@ func (c *kesClient) postRetry(url string, body io.ReadSeeker, limit int64) (io.R
 // file as PEM-encoded certificate and add it to
 // the CertPool. If a file is not a PEM certificate
 // it will be ignored.
-func loadCACertificates(path string) (*x509.CertPool, error) {
-	rootCAs, _ := x509.SystemCertPool()
-	if rootCAs == nil {
-		// In some systems (like Windows) system cert pool is
-		// not supported or no certificates are present on the
-		// system - so we create a new cert pool.
-		rootCAs = x509.NewCertPool()
-	}
+func loadCACertificates(path string, rootCAs *x509.CertPool) error {
 	if path == "" {
-		return rootCAs, nil
+		return nil
 	}
 
 	stat, err := os.Stat(path)
 	if err != nil {
 		if os.IsNotExist(err) || os.IsPermission(err) {
-			return rootCAs, nil
+			return nil
 		}
-		return nil, Errorf("crypto: cannot open '%s': %v", path, err)
+		return Errorf("crypto: cannot open '%s': %v", path, err)
 	}
 
 	// If path is a file, parse as PEM-encoded certifcate
@@ -439,12 +466,12 @@ func loadCACertificates(path string) (*x509.CertPool, error) {
 	if !stat.IsDir() {
 		cert, err := ioutil.ReadFile(path)
 		if err != nil {
-			return nil, err
+			return err
 		}
 		if !rootCAs.AppendCertsFromPEM(cert) {
-			return nil, Errorf("crypto: '%s' is not a valid PEM-encoded certificate", path)
+			return Errorf("crypto: '%s' is not a valid PEM-encoded certificate", path)
 		}
-		return rootCAs, nil
+		return nil
 	}
 
 	// If path is a directory then try
@@ -454,7 +481,7 @@ func loadCACertificates(path string) (*x509.CertPool, error) {
 	// we ignore it.
 	files, err := ioutil.ReadDir(path)
 	if err != nil {
-		return nil, err
+		return err
 	}
 	for _, file := range files {
 		cert, err := ioutil.ReadFile(filepath.Join(path, file.Name()))
@@ -463,6 +490,6 @@ func loadCACertificates(path string) (*x509.CertPool, error) {
 		}
 		rootCAs.AppendCertsFromPEM(cert) // ignore files which are not PEM certtificates
 	}
-	return rootCAs, nil
+	return nil
 
 }

cmd/crypto/kms.go

@@ -109,9 +109,9 @@ type masterKeyKMS struct {
 // KMSInfo contains some describing information about
 // the KMS.
 type KMSInfo struct {
-	Endpoint string
-	Name     string
-	AuthType string
+	Endpoints []string
+	Name      string
+	AuthType  string
 }
 
 // NewMasterKey returns a basic KMS implementation from a single 256 bit master key.
@@ -147,9 +147,9 @@ func (kms *masterKeyKMS) GenerateKey(keyID string, ctx Context) (key [32]byte, s
 // KMS is configured directly using master key
 func (kms *masterKeyKMS) Info() (info KMSInfo) {
 	return KMSInfo{
-		Endpoint: "",
-		Name:     "",
-		AuthType: "master-key",
+		Endpoints: []string{},
+		Name:      "",
+		AuthType:  "master-key",
 	}
 }
 

cmd/crypto/metadata.go

@@ -19,6 +19,7 @@ import (
 	"encoding/base64"
 	"errors"
 
+	xhttp "github.com/minio/minio/cmd/http"
 	"github.com/minio/minio/cmd/logger"
 )
 
@@ -38,6 +39,8 @@ func IsMultiPart(metadata map[string]string) bool {
 func RemoveSensitiveEntries(metadata map[string]string) { // The functions is tested in TestRemoveSensitiveHeaders for compatibility reasons
 	delete(metadata, SSECKey)
 	delete(metadata, SSECopyKey)
+	delete(metadata, xhttp.AmzMetaUnencryptedContentLength)
+	delete(metadata, xhttp.AmzMetaUnencryptedContentMD5)
 }
 
 // RemoveSSEHeaders removes all crypto-specific SSE
@@ -62,6 +65,17 @@ func RemoveInternalEntries(metadata map[string]string) {
 	delete(metadata, S3KMSSealedKey)
 }
 
+// IsSourceEncrypted returns true if the source is encrypted
+func IsSourceEncrypted(metadata map[string]string) bool {
+	if _, ok := metadata[SSECAlgorithm]; ok {
+		return true
+	}
+	if _, ok := metadata[SSEHeader]; ok {
+		return true
+	}
+	return false
+}
+
 // IsEncrypted returns true if the object metadata indicates
 // that it was uploaded using some form of server-side-encryption.
 //
@@ -115,7 +129,7 @@ func (ssec) IsEncrypted(metadata map[string]string) bool {
 // metadata is nil.
 func CreateMultipartMetadata(metadata map[string]string) map[string]string {
 	if metadata == nil {
-		metadata = map[string]string{}
+		return map[string]string{SSEMultipart: ""}
 	}
 	metadata[SSEMultipart] = ""
 	return metadata
@@ -142,7 +156,7 @@ func (s3) CreateMetadata(metadata map[string]string, keyID string, kmsKey []byte
 	}
 
 	if metadata == nil {
-		metadata = map[string]string{}
+		metadata = make(map[string]string, 5)
 	}
 
 	metadata[SSESealAlgorithm] = sealedKey.Algorithm
@@ -222,7 +236,7 @@ func (ssec) CreateMetadata(metadata map[string]string, sealedKey SealedKey) map[
 	}
 
 	if metadata == nil {
-		metadata = map[string]string{}
+		metadata = make(map[string]string, 3)
 	}
 	metadata[SSESealAlgorithm] = SealAlgorithm
 	metadata[SSEIV] = base64.StdEncoding.EncodeToString(sealedKey.IV[:])

cmd/crypto/retry.go

@@ -21,9 +21,8 @@ import (
 
 // default retry configuration
 const (
-	retryWaitMin = 500 * time.Millisecond // minimum retry limit.
-	retryWaitMax = 3 * time.Second        // 3 secs worth of max retry.
-	retryMax     = 2
+	retryWaitMin = 100 * time.Millisecond  // minimum retry limit.
+	retryWaitMax = 1500 * time.Millisecond // 1.5 secs worth of max retry.
 )
 
 // LinearJitterBackoff provides the time.Duration for a caller to

cmd/crypto/vault.go

@@ -199,13 +199,13 @@ func (v *vaultService) DefaultKeyID() string {
 }
 
 // Info returns some information about the Vault,
-// configuration - like the endpoint or authentication
+// configuration - like the endpoints or authentication
 // method.
 func (v *vaultService) Info() KMSInfo {
 	return KMSInfo{
-		Endpoint: v.config.Endpoint,
-		Name:     v.DefaultKeyID(),
-		AuthType: v.config.Auth.Type,
+		Endpoints: []string{v.config.Endpoint},
+		Name:      v.DefaultKeyID(),
+		AuthType:  v.config.Auth.Type,
 	}
 }
 

cmd/data-crawler.go

@@ -28,6 +28,7 @@ import (
 	"time"
 
 	"github.com/minio/minio/cmd/config"
+	"github.com/minio/minio/cmd/config/crawler"
 	"github.com/minio/minio/cmd/logger"
 	"github.com/minio/minio/pkg/bucket/lifecycle"
 	"github.com/minio/minio/pkg/bucket/replication"
@@ -35,6 +36,7 @@ import (
 	"github.com/minio/minio/pkg/env"
 	"github.com/minio/minio/pkg/event"
 	"github.com/minio/minio/pkg/hash"
+	"github.com/minio/minio/pkg/madmin"
 	"github.com/willf/bloom"
 )
 
@@ -44,6 +46,14 @@ const (
 	dataCrawlStartDelay      = 5 * time.Minute  // Time to wait on startup and between cycles.
 	dataUsageUpdateDirCycles = 16               // Visit all folders every n cycles.
 
+	healDeleteDangling    = true
+	healFolderIncludeProb = 32  // Include a clean folder one in n cycles.
+	healObjectSelectProb  = 512 // Overall probability of a file being scanned; one in n.
+
+)
+
+var (
+	globalCrawlerConfig crawler.Config
 )
 
 // initDataCrawler will start the crawler unless disabled.
@@ -87,12 +97,6 @@ func runDataCrawler(ctx context.Context, objAPI ObjectLayer) {
 			if err == nil {
 				// Store new cycle...
 				nextBloomCycle++
-				if nextBloomCycle%dataUpdateTrackerResetEvery == 0 {
-					if intDataUpdateTracker.debug {
-						logger.Info(color.Green("runDataCrawler:") + " Resetting bloom filter for next runs.")
-					}
-					nextBloomCycle++
-				}
 				var tmp [8]byte
 				binary.LittleEndian.PutUint64(tmp[:], nextBloomCycle)
 				r, err := hash.NewReader(bytes.NewReader(tmp[:]), int64(len(tmp)), "", "", int64(len(tmp)), false)
@@ -111,8 +115,9 @@ func runDataCrawler(ctx context.Context, objAPI ObjectLayer) {
 }
 
 type cachedFolder struct {
-	name   string
-	parent *dataUsageHash
+	name              string
+	parent            *dataUsageHash
+	objectHealProbDiv uint32
 }
 
 type folderScanner struct {
@@ -125,6 +130,8 @@ type folderScanner struct {
 
 	dataUsageCrawlMult  float64
 	dataUsageCrawlDebug bool
+	healFolderInclude   uint32 // Include a clean folder one in n cycles.
+	healObjectSelect    uint32 // Do a heal check on an object once every n cycles. Must divide into healFolderInclude
 
 	newFolders      []cachedFolder
 	existingFolders []cachedFolder
@@ -167,8 +174,17 @@ func crawlDataFolder(ctx context.Context, basePath string, cache dataUsageCache,
 		existingFolders:     nil,
 		dataUsageCrawlMult:  delayMult,
 		dataUsageCrawlDebug: intDataUpdateTracker.debug,
+		healFolderInclude:   0,
+		healObjectSelect:    0,
 	}
 
+	// Enable healing in XL mode.
+	if globalIsErasure {
+		// Include a clean folder one in n cycles.
+		s.healFolderInclude = healFolderIncludeProb
+		// Do a heal check on an object once every n cycles. Must divide into healFolderInclude
+		s.healObjectSelect = healObjectSelectProb
+	}
 	if len(cache.Info.BloomFilter) > 0 {
 		s.withFilter = &bloomFilter{BloomFilter: &bloom.BloomFilter{}}
 		_, err := s.withFilter.ReadFrom(bytes.NewBuffer(cache.Info.BloomFilter))
@@ -189,7 +205,7 @@ func crawlDataFolder(ctx context.Context, basePath string, cache dataUsageCache,
 	}
 
 	// Always scan flattenLevels deep. Cache root is level 0.
-	todo := []cachedFolder{{name: cache.Info.Name}}
+	todo := []cachedFolder{{name: cache.Info.Name, objectHealProbDiv: 1}}
 	for i := 0; i < flattenLevels; i++ {
 		if s.dataUsageCrawlDebug {
 			logger.Info(logPrefix+"Level %v, scanning %v directories."+logSuffix, i, len(todo))
@@ -218,7 +234,7 @@ func crawlDataFolder(ctx context.Context, basePath string, cache dataUsageCache,
 			return s.newCache, ctx.Err()
 		default:
 		}
-		du, err := s.deepScanFolder(ctx, folder.name)
+		du, err := s.deepScanFolder(ctx, folder)
 		if err != nil {
 			logger.LogIf(ctx, err)
 			continue
@@ -249,26 +265,38 @@ func crawlDataFolder(ctx context.Context, basePath string, cache dataUsageCache,
 		}
 		h := hashPath(folder.name)
 		if !h.mod(s.oldCache.Info.NextCycle, dataUsageUpdateDirCycles) {
-			s.newCache.replaceHashed(h, folder.parent, s.oldCache.Cache[h.Key()])
-			continue
+			if !h.mod(s.oldCache.Info.NextCycle, s.healFolderInclude/folder.objectHealProbDiv) {
+				s.newCache.replaceHashed(h, folder.parent, s.oldCache.Cache[h.Key()])
+				continue
+			} else {
+				folder.objectHealProbDiv = s.healFolderInclude
+			}
+			folder.objectHealProbDiv = dataUsageUpdateDirCycles
 		}
-
 		if s.withFilter != nil {
 			_, prefix := path2BucketObjectWithBasePath(basePath, folder.name)
 			if s.oldCache.Info.lifeCycle == nil || !s.oldCache.Info.lifeCycle.HasActiveRules(prefix, true) {
 				// If folder isn't in filter, skip it completely.
 				if !s.withFilter.containsDir(folder.name) {
-					if s.dataUsageCrawlDebug {
-						logger.Info(logPrefix+"Skipping non-updated folder: %v"+logSuffix, folder)
+					if !h.mod(s.oldCache.Info.NextCycle, s.healFolderInclude/folder.objectHealProbDiv) {
+						if s.dataUsageCrawlDebug {
+							logger.Info(logPrefix+"Skipping non-updated folder: %v"+logSuffix, folder)
+						}
+						s.newCache.replaceHashed(h, folder.parent, s.oldCache.Cache[h.Key()])
+						continue
+					} else {
+						if s.dataUsageCrawlDebug {
+							logger.Info(logPrefix+"Adding non-updated folder to heal check: %v"+logSuffix, folder.name)
+						}
+						// Update probability of including objects
+						folder.objectHealProbDiv = s.healFolderInclude
 					}
-					s.newCache.replaceHashed(h, folder.parent, s.oldCache.Cache[h.Key()])
-					continue
 				}
 			}
 		}
 
 		// Update on this cycle...
-		du, err := s.deepScanFolder(ctx, folder.name)
+		du, err := s.deepScanFolder(ctx, folder)
 		if err != nil {
 			logger.LogIf(ctx, err)
 			continue
@@ -301,6 +329,7 @@ func (f *folderScanner) scanQueuedLevels(ctx context.Context, folders []cachedFo
 		default:
 		}
 		thisHash := hashPath(folder.name)
+		existing := f.oldCache.findChildrenCopy(thisHash)
 
 		// If there are lifecycle rules for the prefix, remove the filter.
 		filter := f.withFilter
@@ -309,7 +338,7 @@ func (f *folderScanner) scanQueuedLevels(ctx context.Context, folders []cachedFo
 			_, prefix := path2BucketObjectWithBasePath(f.root, folder.name)
 			if f.oldCache.Info.lifeCycle.HasActiveRules(prefix, true) {
 				if f.dataUsageCrawlDebug {
-					logger.Info(color.Green("data-usage:")+" Prefix %q has active rules", prefix)
+					logger.Info(color.Green("folder-scanner:")+" Prefix %q has active rules", prefix)
 				}
 				activeLifeCycle = f.oldCache.Info.lifeCycle
 				filter = nil
@@ -318,11 +347,19 @@ func (f *folderScanner) scanQueuedLevels(ctx context.Context, folders []cachedFo
 		if _, ok := f.oldCache.Cache[thisHash.Key()]; filter != nil && ok {
 			// If folder isn't in filter and we have data, skip it completely.
 			if folder.name != dataUsageRoot && !filter.containsDir(folder.name) {
-				f.newCache.copyWithChildren(&f.oldCache, thisHash, folder.parent)
-				if f.dataUsageCrawlDebug {
-					logger.Info(color.Green("data-usage:")+" Skipping non-updated folder: %v", folder.name)
+				if !thisHash.mod(f.oldCache.Info.NextCycle, f.healFolderInclude/folder.objectHealProbDiv) {
+					f.newCache.copyWithChildren(&f.oldCache, thisHash, folder.parent)
+					if f.dataUsageCrawlDebug {
+						logger.Info(color.Green("folder-scanner:")+" Skipping non-updated folder: %v", folder.name)
+					}
+					continue
+				} else {
+					if f.dataUsageCrawlDebug {
+						logger.Info(color.Green("folder-scanner:")+" Adding non-updated folder to heal check: %v", folder.name)
+					}
+					// If probability was already crawlerHealFolderInclude, keep it.
+					folder.objectHealProbDiv = f.healFolderInclude
 				}
-				continue
 			}
 		}
 		f.waitForLowActiveIO()
@@ -336,14 +373,14 @@ func (f *folderScanner) scanQueuedLevels(ctx context.Context, folders []cachedFo
 			bucket, prefix := path2BucketObjectWithBasePath(f.root, entName)
 			if bucket == "" {
 				if f.dataUsageCrawlDebug {
-					logger.Info(color.Green("data-usage:")+" no bucket (%s,%s)", f.root, entName)
+					logger.Info(color.Green("folder-scanner:")+" no bucket (%s,%s)", f.root, entName)
 				}
 				return nil
 			}
 
 			if isReservedOrInvalidBucket(bucket, false) {
 				if f.dataUsageCrawlDebug {
-					logger.Info(color.Green("data-usage:")+" invalid bucket: %v, entry: %v", bucket, entName)
+					logger.Info(color.Green("folder-scanner:")+" invalid bucket: %v, entry: %v", bucket, entName)
 				}
 				return nil
 			}
@@ -359,7 +396,8 @@ func (f *folderScanner) scanQueuedLevels(ctx context.Context, folders []cachedFo
 				_, exists := f.oldCache.Cache[h.Key()]
 				cache.addChildString(entName)
 
-				this := cachedFolder{name: entName, parent: &thisHash}
+				this := cachedFolder{name: entName, parent: &thisHash, objectHealProbDiv: folder.objectHealProbDiv}
+				delete(existing, h.Key())
 				cache.addChild(h)
 				if final {
 					if exists {
@@ -385,11 +423,12 @@ func (f *folderScanner) scanQueuedLevels(ctx context.Context, folders []cachedFo
 				objectName: path.Base(entName),
 				debug:      f.dataUsageCrawlDebug,
 				lifeCycle:  activeLifeCycle,
+				heal:       thisHash.mod(f.oldCache.Info.NextCycle, f.healObjectSelect/folder.objectHealProbDiv),
 			}
 			size, err := f.getSize(item)
 
 			sleepDuration(time.Since(t), f.dataUsageCrawlMult)
-			if err == errSkipFile || err == errFileNotFound {
+			if err == errSkipFile {
 				return nil
 			}
 			logger.LogIf(ctx, err)
@@ -402,19 +441,78 @@ func (f *folderScanner) scanQueuedLevels(ctx context.Context, folders []cachedFo
 		if err != nil {
 			return nil, err
 		}
+
+		if f.healObjectSelect == 0 {
+			// If we are not scanning, return now.
+			f.newCache.replaceHashed(thisHash, folder.parent, cache)
+			continue
+		}
+
+		objAPI := newObjectLayerWithoutSafeModeFn()
+		if objAPI == nil {
+			continue
+		}
+
+		bgSeq, found := globalBackgroundHealState.getHealSequenceByToken(bgHealingUUID)
+		if !found {
+			continue
+		}
+
+		// Whatever remains in 'existing' are folders at this level
+		// that existed in the previous run but wasn't found now.
+		//
+		// This may be because of 2 reasons:
+		//
+		// 1) The folder/object was deleted.
+		// 2) We come from another disk and this disk missed the write.
+		//
+		// We therefore perform a heal check.
+		// If that doesn't bring it back we remove the folder and assume it was deleted.
+		// This means that the next run will not look for it.
+		for k := range existing {
+			f.waitForLowActiveIO()
+			bucket, prefix := path2BucketObject(k)
+			if f.dataUsageCrawlDebug {
+				logger.Info(color.Green("folder-scanner:")+" checking disappeared folder: %v/%v", bucket, prefix)
+			}
+
+			err = objAPI.HealObjects(ctx, bucket, prefix, madmin.HealOpts{Recursive: true, Remove: healDeleteDangling},
+				func(bucket, object, versionID string) error {
+					return bgSeq.queueHealTask(healSource{
+						bucket:    bucket,
+						object:    object,
+						versionID: versionID,
+					}, madmin.HealItemObject)
+				})
+
+			if f.dataUsageCrawlDebug && err != nil {
+				logger.Info(color.Green("healObjects:")+" checking returned value %v", err)
+			}
+
+			// Add unless healing returned an error.
+			if err == nil {
+				this := cachedFolder{name: k, parent: &thisHash, objectHealProbDiv: folder.objectHealProbDiv}
+				cache.addChild(hashPath(k))
+				if final {
+					f.existingFolders = append(f.existingFolders, this)
+				} else {
+					nextFolders = append(nextFolders, this)
+				}
+			}
+		}
 		f.newCache.replaceHashed(thisHash, folder.parent, cache)
 	}
 	return nextFolders, nil
 }
 
 // deepScanFolder will deep scan a folder and return the size if no error occurs.
-func (f *folderScanner) deepScanFolder(ctx context.Context, folder string) (*dataUsageEntry, error) {
+func (f *folderScanner) deepScanFolder(ctx context.Context, folder cachedFolder) (*dataUsageEntry, error) {
 	var cache dataUsageEntry
 
 	done := ctx.Done()
 
 	var addDir func(entName string, typ os.FileMode) error
-	var dirStack = []string{f.root, folder}
+	var dirStack = []string{f.root, folder.name}
 
 	addDir = func(entName string, typ os.FileMode) error {
 		select {
@@ -445,7 +543,7 @@ func (f *folderScanner) deepScanFolder(ctx context.Context, folder string) (*dat
 		if f.oldCache.Info.lifeCycle != nil {
 			if f.oldCache.Info.lifeCycle.HasActiveRules(prefix, false) {
 				if f.dataUsageCrawlDebug {
-					logger.Info(color.Green("data-usage:")+" Prefix %q has active rules", prefix)
+					logger.Info(color.Green("folder-scanner:")+" Prefix %q has active rules", prefix)
 				}
 				activeLifeCycle = f.oldCache.Info.lifeCycle
 			}
@@ -460,6 +558,7 @@ func (f *folderScanner) deepScanFolder(ctx context.Context, folder string) (*dat
 				objectName: path.Base(entName),
 				debug:      f.dataUsageCrawlDebug,
 				lifeCycle:  activeLifeCycle,
+				heal:       hashPath(path.Join(prefix, entName)).mod(f.oldCache.Info.NextCycle, f.healObjectSelect/folder.objectHealProbDiv),
 			})
 
 		// Don't sleep for really small amount of time
@@ -490,6 +589,7 @@ type crawlItem struct {
 	prefix     string // Only the prefix if any, does not have final object name.
 	objectName string // Only the object name without prefixes.
 	lifeCycle  *lifecycle.Lifecycle
+	heal       bool // Has the object been selected for heal check?
 	debug      bool
 }
 
@@ -509,9 +609,9 @@ func (i *crawlItem) transformMetaDir() {
 
 // actionMeta contains information used to apply actions.
 type actionMeta struct {
-	oi          ObjectInfo
-	trustOI     bool // Set true if oi can be trusted and has been read with quorum.
-	numVersions int  // The number of versions of this object
+	oi               ObjectInfo
+	successorModTime time.Time // The modtime of the successor version
+	numVersions      int       // The number of versions of this object
 }
 
 // applyActions will apply lifecycle checks on to a scanned item.
@@ -523,6 +623,20 @@ func (i *crawlItem) applyActions(ctx context.Context, o ObjectLayer, meta action
 	if i.debug {
 		logger.LogIf(ctx, err)
 	}
+	if i.heal {
+		if i.debug {
+			logger.Info(color.Green("applyActions:")+" heal checking: %v/%v v%s", i.bucket, i.objectPath(), meta.oi.VersionID)
+		}
+		res, err := o.HealObject(ctx, i.bucket, i.objectPath(), meta.oi.VersionID, madmin.HealOpts{Remove: healDeleteDangling})
+		if isErrObjectNotFound(err) || isErrVersionNotFound(err) {
+			return 0
+		}
+		if err != nil && !errors.Is(err, NotImplemented{}) {
+			logger.LogIf(ctx, err)
+			return 0
+		}
+		size = res.ObjectSize
+	}
 	if i.lifeCycle == nil {
 		return size
 	}
@@ -530,16 +644,17 @@ func (i *crawlItem) applyActions(ctx context.Context, o ObjectLayer, meta action
 	versionID := meta.oi.VersionID
 	action := i.lifeCycle.ComputeAction(
 		lifecycle.ObjectOpts{
-			Name:         i.objectPath(),
-			UserTags:     meta.oi.UserTags,
-			ModTime:      meta.oi.ModTime,
-			VersionID:    meta.oi.VersionID,
-			DeleteMarker: meta.oi.DeleteMarker,
-			IsLatest:     meta.oi.IsLatest,
-			NumVersions:  meta.numVersions,
+			Name:             i.objectPath(),
+			UserTags:         meta.oi.UserTags,
+			ModTime:          meta.oi.ModTime,
+			VersionID:        meta.oi.VersionID,
+			DeleteMarker:     meta.oi.DeleteMarker,
+			IsLatest:         meta.oi.IsLatest,
+			NumVersions:      meta.numVersions,
+			SuccessorModTime: meta.successorModTime,
 		})
 	if i.debug {
-		logger.Info(color.Green("applyActions:")+" lifecycle: %q, Initial scan: %v", i.objectPath(), action)
+		logger.Info(color.Green("applyActions:")+" lifecycle: %q (version-id=%s), Initial scan: %v", i.objectPath(), versionID, action)
 	}
 	switch action {
 	case lifecycle.DeleteAction, lifecycle.DeleteVersionAction:
@@ -548,63 +663,72 @@ func (i *crawlItem) applyActions(ctx context.Context, o ObjectLayer, meta action
 		return size
 	}
 
-	// These (expensive) operations should only run on items we are likely to delete.
-	// Load to ensure that we have the correct version and not an unsynced version.
-	if !meta.trustOI {
-		obj, err := o.GetObjectInfo(ctx, i.bucket, i.objectPath(), ObjectOptions{
-			VersionID: versionID,
-		})
-		if err != nil {
-			switch err.(type) {
-			case MethodNotAllowed: // This happens usually for a delete marker
-				if !obj.DeleteMarker { // if this is not a delete marker log and return
-					// Do nothing - heal in the future.
-					logger.LogIf(ctx, err)
-					return size
-				}
-			case ObjectNotFound:
-				// object not found return 0
-				return 0
-			default:
-				// All other errors proceed.
+	obj, err := o.GetObjectInfo(ctx, i.bucket, i.objectPath(), ObjectOptions{
+		VersionID: versionID,
+	})
+	if err != nil {
+		switch err.(type) {
+		case MethodNotAllowed: // This happens usually for a delete marker
+			if !obj.DeleteMarker { // if this is not a delete marker log and return
+				// Do nothing - heal in the future.
 				logger.LogIf(ctx, err)
 				return size
 			}
-		}
-		size = obj.Size
-
-		// Recalculate action.
-		action = i.lifeCycle.ComputeAction(
-			lifecycle.ObjectOpts{
-				Name:         i.objectPath(),
-				UserTags:     obj.UserTags,
-				ModTime:      obj.ModTime,
-				VersionID:    obj.VersionID,
-				DeleteMarker: obj.DeleteMarker,
-				IsLatest:     obj.IsLatest,
-				NumVersions:  meta.numVersions,
-			})
-		if i.debug {
-			logger.Info(color.Green("applyActions:")+" lifecycle: Secondary scan: %v", action)
-		}
-		versionID = obj.VersionID
-		switch action {
-		case lifecycle.DeleteAction, lifecycle.DeleteVersionAction:
+		case ObjectNotFound:
+			// object not found return 0
+			return 0
 		default:
-			// No action.
+			// All other errors proceed.
+			logger.LogIf(ctx, err)
 			return size
 		}
 	}
+	size = obj.Size
+
+	// Recalculate action.
+	action = i.lifeCycle.ComputeAction(
+		lifecycle.ObjectOpts{
+			Name:             i.objectPath(),
+			UserTags:         obj.UserTags,
+			ModTime:          obj.ModTime,
+			VersionID:        obj.VersionID,
+			DeleteMarker:     obj.DeleteMarker,
+			IsLatest:         obj.IsLatest,
+			NumVersions:      meta.numVersions,
+			SuccessorModTime: meta.successorModTime,
+		})
+	if i.debug {
+		logger.Info(color.Green("applyActions:")+" lifecycle: Secondary scan: %v", action)
+	}
+	switch action {
+	case lifecycle.DeleteAction, lifecycle.DeleteVersionAction:
+	default:
+		// No action.
+		return size
+	}
 
 	opts := ObjectOptions{}
 	switch action {
 	case lifecycle.DeleteVersionAction:
-		opts.VersionID = versionID
+		// Defensive code, should never happen
+		if obj.VersionID == "" {
+			return size
+		}
+		if rcfg, _ := globalBucketObjectLockSys.Get(i.bucket); rcfg.LockEnabled {
+			locked := enforceRetentionForDeletion(ctx, obj)
+			if locked {
+				if i.debug {
+					logger.Info(color.Green("applyActions:")+" lifecycle: %s is locked, not deleting", i.objectPath())
+				}
+				return size
+			}
+		}
+		opts.VersionID = obj.VersionID
 	case lifecycle.DeleteAction:
 		opts.Versioned = globalBucketVersioningSys.Enabled(i.bucket)
 	}
 
-	obj, err := o.DeleteObject(ctx, i.bucket, i.objectPath(), opts)
+	obj, err = o.DeleteObject(ctx, i.bucket, i.objectPath(), opts)
 	if err != nil {
 		// Assume it is still there.
 		logger.LogIf(ctx, err)
@@ -642,9 +766,6 @@ func sleepDuration(d time.Duration, x float64) {
 func (i *crawlItem) healReplication(ctx context.Context, o ObjectLayer, meta actionMeta) {
 	if meta.oi.ReplicationStatus == replication.Pending ||
 		meta.oi.ReplicationStatus == replication.Failed {
-		// if heal encounters a pending replication status, either replication
-		// has failed due to server shutdown or crawler and PutObject replication are in contention.
-		healPending := meta.oi.ReplicationStatus == replication.Pending
-		replicateObject(ctx, meta.oi.Bucket, meta.oi.Name, meta.oi.VersionID, o, nil, healPending)
+		globalReplicationState.queueReplicaTask(meta.oi)
 	}
 }

cmd/data-update-tracker.go

@@ -48,9 +48,6 @@ const (
 	dataUpdateTrackerFilename     = dataUsageBucket + SlashSeparator + ".tracker.bin"
 	dataUpdateTrackerVersion      = 2
 	dataUpdateTrackerSaveInterval = 5 * time.Minute
-
-	// Reset bloom filters every n cycle
-	dataUpdateTrackerResetEvery = 1000
 )
 
 var (
@@ -181,6 +178,8 @@ func (d *dataUpdateTracker) start(ctx context.Context, drives ...string) {
 	}
 	d.load(ctx, drives...)
 	go d.startCollector(ctx)
+	// startSaver will unlock.
+	d.mu.Lock()
 	go d.startSaver(ctx, dataUpdateTrackerSaveInterval, drives)
 }
 
@@ -214,17 +213,17 @@ func (d *dataUpdateTracker) load(ctx context.Context, drives ...string) {
 }
 
 // startSaver will start a saver that will write d to all supplied drives at specific intervals.
+// 'd' must be write locked when started and will be unlocked.
 // The saver will save and exit when supplied context is closed.
 func (d *dataUpdateTracker) startSaver(ctx context.Context, interval time.Duration, drives []string) {
-	t := time.NewTicker(interval)
-	defer t.Stop()
-	var buf bytes.Buffer
-	d.mu.Lock()
 	saveNow := d.save
 	exited := make(chan struct{})
 	d.saveExited = exited
 	d.mu.Unlock()
+	t := time.NewTicker(interval)
+	defer t.Stop()
 	defer close(exited)
+	var buf bytes.Buffer
 	for {
 		var exit bool
 		select {
@@ -427,6 +426,8 @@ func (d *dataUpdateTracker) deserialize(src io.Reader, newerThan time.Time) erro
 	}
 	// Ignore what remains on the stream.
 	// Update d:
+	d.mu.Lock()
+	defer d.mu.Unlock()
 	d.Current = dst.Current
 	d.History = dst.History
 	d.Saved = dst.Saved

cmd/data-update-tracker_test.go

@@ -42,6 +42,10 @@ type testingLogger struct {
 	t  testLoggerI
 }
 
+func (t *testingLogger) Validate() error {
+	return nil
+}
+
 func (t *testingLogger) Send(entry interface{}, errKind string) error {
 	t.mu.Lock()
 	defer t.mu.Unlock()

cmd/data-usage-cache.go

@@ -85,7 +85,12 @@ func (e *dataUsageEntry) merge(other dataUsageEntry) {
 }
 
 // mod returns true if the hash mod cycles == cycle.
+// If cycles is 0 false is always returned.
+// If cycles is 1 true is always returned (as expected).
 func (h dataUsageHash) mod(cycle uint32, cycles uint32) bool {
+	if cycles <= 1 {
+		return cycles == 1
+	}
 	return uint32(xxhash.Sum64String(string(h)))%cycles == cycle%cycles
 }
 
@@ -117,6 +122,16 @@ func (d *dataUsageCache) find(path string) *dataUsageEntry {
 	return &due
 }
 
+// findChildrenCopy returns a copy of the children of the supplied hash.
+func (d *dataUsageCache) findChildrenCopy(h dataUsageHash) dataUsageHashMap {
+	ch := d.Cache[h.String()].Children
+	res := make(dataUsageHashMap, len(ch))
+	for k := range ch {
+		res[k] = struct{}{}
+	}
+	return res
+}
+
 // Returns nil if not found.
 func (d *dataUsageCache) subCache(path string) dataUsageCache {
 	dst := dataUsageCache{Info: dataUsageCacheInfo{
@@ -328,7 +343,7 @@ func (d *dataUsageCache) bucketsUsageInfo(buckets []BucketInfo) map[string]Bucke
 		flat := d.flatten(*e)
 		dst[bucket.Name] = BucketUsageInfo{
 			Size:                 uint64(flat.Size),
-			ObjectsCount:         uint64(flat.Objects),
+			ObjectsCount:         flat.Objects,
 			ObjectSizesHistogram: flat.ObjSizes.toMap(),
 		}
 	}
@@ -345,7 +360,7 @@ func (d *dataUsageCache) bucketUsageInfo(bucket string) BucketUsageInfo {
 	flat := d.flatten(*e)
 	return BucketUsageInfo{
 		Size:                 uint64(flat.Size),
-		ObjectsCount:         uint64(flat.Objects),
+		ObjectsCount:         flat.Objects,
 		ObjectSizesHistogram: flat.ObjSizes.toMap(),
 	}
 }
@@ -413,10 +428,15 @@ func (d *dataUsageCache) merge(other dataUsageCache) {
 	}
 }
 
+type objectIO interface {
+	GetObject(ctx context.Context, bucket, object string, startOffset int64, length int64, writer io.Writer, etag string, opts ObjectOptions) (err error)
+	PutObject(ctx context.Context, bucket, object string, data *PutObjReader, opts ObjectOptions) (objInfo ObjectInfo, err error)
+}
+
 // load the cache content with name from minioMetaBackgroundOpsBucket.
 // Only backend errors are returned as errors.
 // If the object is not found or unable to deserialize d is cleared and nil error is returned.
-func (d *dataUsageCache) load(ctx context.Context, store ObjectLayer, name string) error {
+func (d *dataUsageCache) load(ctx context.Context, store objectIO, name string) error {
 	var buf bytes.Buffer
 	err := store.GetObject(ctx, dataUsageBucket, name, 0, -1, &buf, "", ObjectOptions{})
 	if err != nil {
@@ -435,7 +455,7 @@ func (d *dataUsageCache) load(ctx context.Context, store ObjectLayer, name strin
 }
 
 // save the content of the cache to minioMetaBackgroundOpsBucket with the provided name.
-func (d *dataUsageCache) save(ctx context.Context, store ObjectLayer, name string) error {
+func (d *dataUsageCache) save(ctx context.Context, store objectIO, name string) error {
 	b := d.serialize()
 	size := int64(len(b))
 	r, err := hash.NewReader(bytes.NewReader(b), size, "", "", size, false)

cmd/data-usage_test.go

@@ -240,7 +240,6 @@ func TestDataUsageUpdate(t *testing.T) {
 				t.Fatal("got nil result")
 			}
 			if w.flatten {
-				t.Log(e.Children)
 				*e = got.flatten(*e)
 			}
 			if e.Size != int64(w.size) {
@@ -360,6 +359,13 @@ func TestDataUsageUpdatePrefix(t *testing.T) {
 	if err != nil {
 		t.Fatal(err)
 	}
+	if got.root() == nil {
+		t.Log("cached folders:")
+		for folder := range got.Cache {
+			t.Log("folder:", folder)
+		}
+		t.Fatal("got nil root.")
+	}
 
 	// Test dirs
 	var want = []struct {

cmd/disk-cache-backend.go

@@ -46,7 +46,7 @@ const (
 	cacheMetaJSONFile = "cache.json"
 	cacheDataFile     = "part.1"
 	cacheMetaVersion  = "1.0.0"
-	cacheExpiryDays   = time.Duration(90 * time.Hour * 24) // defaults to 90 days
+	cacheExpiryDays   = 90 * time.Hour * 24 // defaults to 90 days
 	// SSECacheEncrypted is the metadata key indicating that the object
 	// is a cache entry encrypted with cache KMS master key in globalCacheKMS.
 	SSECacheEncrypted = "X-Minio-Internal-Encrypted-Cache"
@@ -157,7 +157,7 @@ func newDiskCache(ctx context.Context, dir string, config cache.Config) (*diskCa
 	}
 	cache := diskCache{
 		dir:           dir,
-		triggerGC:     make(chan struct{}),
+		triggerGC:     make(chan struct{}, 1),
 		stats:         CacheDiskStats{Dir: dir},
 		quotaPct:      quotaPct,
 		after:         config.After,
@@ -174,7 +174,7 @@ func newDiskCache(ctx context.Context, dir string, config cache.Config) (*diskCa
 		nsMutex: newNSLock(false),
 	}
 	go cache.purgeWait(ctx)
-	cache.diskUsageHigh() // update if cache usage is already high.
+	cache.diskSpaceAvailable(0) // update if cache usage is already high.
 	cache.NewNSLockFn = func(ctx context.Context, cachePath string) RWLocker {
 		return cache.nsMutex.NewNSLock(ctx, nil, cachePath, "")
 	}
@@ -203,9 +203,9 @@ func (c *diskCache) diskUsageLow() bool {
 	return low
 }
 
-// Returns if the disk usage reaches high water mark w.r.t the configured cache quota.
-// gc starts if high water mark reached.
-func (c *diskCache) diskUsageHigh() bool {
+// Returns if the disk usage reaches  or exceeds configured cache quota when size is added.
+// If current usage without size exceeds high watermark a GC is automatically queued.
+func (c *diskCache) diskSpaceAvailable(size int64) bool {
 	gcTriggerPct := c.quotaPct * c.highWatermark / 100
 	di, err := disk.GetInfo(c.dir)
 	if err != nil {
@@ -214,27 +214,30 @@ func (c *diskCache) diskUsageHigh() bool {
 		logger.LogIf(ctx, err)
 		return false
 	}
-	usedPercent := (di.Total - di.Free) * 100 / di.Total
-	high := int(usedPercent) >= gcTriggerPct
-	atomic.StoreUint64(&c.stats.UsagePercent, usedPercent)
-	if high {
-		atomic.StoreInt32(&c.stats.UsageState, 1)
-	}
-	return high
-}
-
-// Returns if size space can be allocated without exceeding
-// max disk usable for caching
-func (c *diskCache) diskAvailable(size int64) bool {
-	di, err := disk.GetInfo(c.dir)
-	if err != nil {
-		reqInfo := (&logger.ReqInfo{}).AppendTags("cachePath", c.dir)
-		ctx := logger.SetReqInfo(GlobalContext, reqInfo)
-		logger.LogIf(ctx, err)
+	if di.Total == 0 {
+		logger.Info("diskCache: Received 0 total disk size")
 		return false
 	}
-	usedPercent := (di.Total - (di.Free - uint64(size))) * 100 / di.Total
-	return int(usedPercent) < c.quotaPct
+	usedPercent := float64(di.Total-di.Free) * 100 / float64(di.Total)
+	if usedPercent >= float64(gcTriggerPct) {
+		atomic.StoreInt32(&c.stats.UsageState, 1)
+		c.queueGC()
+	}
+	atomic.StoreUint64(&c.stats.UsagePercent, uint64(usedPercent))
+
+	// Recalculate percentage with provided size added.
+	usedPercent = float64(di.Total-di.Free+uint64(size)) * 100 / float64(di.Total)
+
+	return usedPercent < float64(c.quotaPct)
+}
+
+// queueGC will queue a GC.
+// Calling this function is always non-blocking.
+func (c *diskCache) queueGC() {
+	select {
+	case c.triggerGC <- struct{}{}:
+	default:
+	}
 }
 
 // toClear returns how many bytes should be cleared to reach the low watermark quota.
@@ -247,7 +250,7 @@ func (c *diskCache) toClear() uint64 {
 		logger.LogIf(ctx, err)
 		return 0
 	}
-	return bytesToClear(int64(di.Total), int64(di.Free), uint64(c.quotaPct), uint64(c.lowWatermark))
+	return bytesToClear(int64(di.Total), int64(di.Free), uint64(c.quotaPct), uint64(c.lowWatermark), uint64(c.highWatermark))
 }
 
 var (
@@ -417,7 +420,7 @@ func (c *diskCache) Stat(ctx context.Context, bucket, object string) (oi ObjectI
 func (c *diskCache) statCachedMeta(ctx context.Context, cacheObjPath string) (meta *cacheMeta, partial bool, numHits int, err error) {
 
 	cLock := c.NewNSLockFn(ctx, cacheObjPath)
-	if err = cLock.GetRLock(globalObjectTimeout); err != nil {
+	if err = cLock.GetRLock(globalOperationTimeout); err != nil {
 		return
 	}
 
@@ -499,7 +502,7 @@ func (c *diskCache) statCache(ctx context.Context, cacheObjPath string) (meta *c
 func (c *diskCache) SaveMetadata(ctx context.Context, bucket, object string, meta map[string]string, actualSize int64, rs *HTTPRangeSpec, rsFileName string, incHitsOnly bool) error {
 	cachedPath := getCacheSHADir(c.dir, bucket, object)
 	cLock := c.NewNSLockFn(ctx, cachedPath)
-	if err := cLock.GetLock(globalObjectTimeout); err != nil {
+	if err := cLock.GetLock(globalOperationTimeout); err != nil {
 		return err
 	}
 	defer cLock.Unlock()
@@ -658,14 +661,13 @@ func newCacheEncryptMetadata(bucket, object string, metadata map[string]string)
 
 // Caches the object to disk
 func (c *diskCache) Put(ctx context.Context, bucket, object string, data io.Reader, size int64, rs *HTTPRangeSpec, opts ObjectOptions, incHitsOnly bool) error {
-	if c.diskUsageHigh() {
-		c.triggerGC <- struct{}{}
+	if !c.diskSpaceAvailable(size) {
 		io.Copy(ioutil.Discard, data)
 		return errDiskFull
 	}
 	cachePath := getCacheSHADir(c.dir, bucket, object)
 	cLock := c.NewNSLockFn(ctx, cachePath)
-	if err := cLock.GetLock(globalObjectTimeout); err != nil {
+	if err := cLock.GetLock(globalOperationTimeout); err != nil {
 		return err
 	}
 	defer cLock.Unlock()
@@ -688,16 +690,13 @@ func (c *diskCache) Put(ctx context.Context, bucket, object string, data io.Read
 	if rs != nil {
 		return c.putRange(ctx, bucket, object, data, size, rs, opts)
 	}
-	if !c.diskAvailable(size) {
+	if !c.diskSpaceAvailable(size) {
 		return errDiskFull
 	}
 	if err := os.MkdirAll(cachePath, 0777); err != nil {
 		return err
 	}
-	var metadata = make(map[string]string)
-	for k, v := range opts.UserDefined {
-		metadata[k] = v
-	}
+	var metadata = cloneMSS(opts.UserDefined)
 	var reader = data
 	var actualSize = uint64(size)
 	if globalCacheKMS != nil {
@@ -719,7 +718,7 @@ func (c *diskCache) Put(ctx context.Context, bucket, object string, data io.Read
 
 	if actualSize != uint64(n) {
 		removeAll(cachePath)
-		return IncompleteBody{}
+		return IncompleteBody{Bucket: bucket, Object: object}
 	}
 	return c.saveMetadata(ctx, bucket, object, metadata, n, nil, "", incHitsOnly)
 }
@@ -730,17 +729,14 @@ func (c *diskCache) putRange(ctx context.Context, bucket, object string, data io
 	if err != nil {
 		return err
 	}
-	if !c.diskAvailable(rlen) {
+	if !c.diskSpaceAvailable(rlen) {
 		return errDiskFull
 	}
 	cachePath := getCacheSHADir(c.dir, bucket, object)
 	if err := os.MkdirAll(cachePath, 0777); err != nil {
 		return err
 	}
-	var metadata = make(map[string]string)
-	for k, v := range opts.UserDefined {
-		metadata[k] = v
-	}
+	var metadata = cloneMSS(opts.UserDefined)
 	var reader = data
 	var actualSize = uint64(rlen)
 	// objSize is the actual size of object (with encryption overhead if any)
@@ -766,7 +762,7 @@ func (c *diskCache) putRange(ctx context.Context, bucket, object string, data io
 	}
 	if actualSize != uint64(n) {
 		removeAll(cachePath)
-		return IncompleteBody{}
+		return IncompleteBody{Bucket: bucket, Object: object}
 	}
 	return c.saveMetadata(ctx, bucket, object, metadata, int64(objSize), rs, cacheFile, false)
 }
@@ -871,7 +867,7 @@ func (c *diskCache) bitrotReadFromCache(ctx context.Context, filePath string, of
 func (c *diskCache) Get(ctx context.Context, bucket, object string, rs *HTTPRangeSpec, h http.Header, opts ObjectOptions) (gr *GetObjectReader, numHits int, err error) {
 	cacheObjPath := getCacheSHADir(c.dir, bucket, object)
 	cLock := c.NewNSLockFn(ctx, cacheObjPath)
-	if err := cLock.GetRLock(globalObjectTimeout); err != nil {
+	if err := cLock.GetRLock(globalOperationTimeout); err != nil {
 		return nil, numHits, err
 	}
 
@@ -935,7 +931,7 @@ func (c *diskCache) Get(ctx context.Context, bucket, object string, rs *HTTPRang
 // Deletes the cached object
 func (c *diskCache) delete(ctx context.Context, cacheObjPath string) (err error) {
 	cLock := c.NewNSLockFn(ctx, cacheObjPath)
-	if err := cLock.GetLock(globalObjectTimeout); err != nil {
+	if err := cLock.GetLock(globalOperationTimeout); err != nil {
 		return err
 	}
 	defer cLock.Unlock()

cmd/disk-cache-utils.go

@@ -102,7 +102,7 @@ func (c *cacheControl) isStale(modTime time.Time) bool {
 func cacheControlOpts(o ObjectInfo) *cacheControl {
 	c := cacheControl{}
 	m := o.UserDefined
-	if o.Expires != timeSentinel {
+	if !o.Expires.Equal(timeSentinel) {
 		c.expiry = o.Expires
 	}
 
@@ -489,9 +489,15 @@ func (f *fileScorer) queueString() string {
 // bytesToClear() returns the number of bytes to clear to reach low watermark
 // w.r.t quota given disk total and free space, quota in % allocated to cache
 // and low watermark % w.r.t allowed quota.
-func bytesToClear(total, free int64, quotaPct, lowWatermark uint64) uint64 {
-	used := (total - free)
+// If the high watermark hasn't been reached 0 will be returned.
+func bytesToClear(total, free int64, quotaPct, lowWatermark, highWatermark uint64) uint64 {
+	used := total - free
 	quotaAllowed := total * (int64)(quotaPct) / 100
-	lowWMUsage := (total * (int64)(lowWatermark*quotaPct) / (100 * 100))
+	highWMUsage := total * (int64)(highWatermark*quotaPct) / (100 * 100)
+	if used < highWMUsage {
+		return 0
+	}
+	// Return bytes needed to reach low watermark.
+	lowWMUsage := total * (int64)(lowWatermark*quotaPct) / (100 * 100)
 	return (uint64)(math.Min(float64(quotaAllowed), math.Max(0.0, float64(used-lowWMUsage))))
 }

cmd/disk-cache-utils_test.go

@@ -45,7 +45,7 @@ func TestGetCacheControlOpts(t *testing.T) {
 		t.Run("", func(t *testing.T) {
 			m := make(map[string]string)
 			m["cache-control"] = testCase.cacheControlHeaderVal
-			if testCase.expiryHeaderVal != timeSentinel {
+			if !testCase.expiryHeaderVal.Equal(timeSentinel) {
 				m["expires"] = testCase.expiryHeaderVal.String()
 			}
 			c := cacheControlOpts(ObjectInfo{UserDefined: m, Expires: testCase.expiryHeaderVal})
@@ -149,22 +149,26 @@ func TestNewFileScorer(t *testing.T) {
 }
 func TestBytesToClear(t *testing.T) {
 	testCases := []struct {
-		total        int64
-		free         int64
-		quotaPct     uint64
-		watermarkLow uint64
-		expected     uint64
+		total         int64
+		free          int64
+		quotaPct      uint64
+		watermarkLow  uint64
+		watermarkHigh uint64
+		expected      uint64
 	}{
-		{1000, 800, 40, 90, 0},
-		{1000, 200, 40, 90, 400},
-		{1000, 400, 40, 90, 240},
-		{1000, 600, 40, 90, 40},
-		{1000, 600, 40, 70, 120},
-		{1000, 1000, 90, 70, 0},
-		{1000, 0, 90, 70, 370},
+		{total: 1000, free: 800, quotaPct: 40, watermarkLow: 90, watermarkHigh: 90, expected: 0},
+		{total: 1000, free: 200, quotaPct: 40, watermarkLow: 90, watermarkHigh: 90, expected: 400},
+		{total: 1000, free: 400, quotaPct: 40, watermarkLow: 90, watermarkHigh: 90, expected: 240},
+		{total: 1000, free: 600, quotaPct: 40, watermarkLow: 90, watermarkHigh: 90, expected: 40},
+		{total: 1000, free: 600, quotaPct: 40, watermarkLow: 70, watermarkHigh: 70, expected: 120},
+		{total: 1000, free: 1000, quotaPct: 90, watermarkLow: 70, watermarkHigh: 70, expected: 0},
+
+		// High not yet reached..
+		{total: 1000, free: 250, quotaPct: 100, watermarkLow: 50, watermarkHigh: 90, expected: 0},
+		{total: 1000, free: 250, quotaPct: 100, watermarkLow: 50, watermarkHigh: 90, expected: 0},
 	}
 	for i, tc := range testCases {
-		toClear := bytesToClear(tc.total, tc.free, tc.quotaPct, tc.watermarkLow)
+		toClear := bytesToClear(tc.total, tc.free, tc.quotaPct, tc.watermarkLow, tc.watermarkHigh)
 		if tc.expected != toClear {
 			t.Errorf("test %d expected %v, got %v", i, tc.expected, toClear)
 		}

cmd/disk-cache.go

@@ -85,16 +85,15 @@ type cacheObjects struct {
 }
 
 func (c *cacheObjects) incHitsToMeta(ctx context.Context, dcache *diskCache, bucket, object string, size int64, eTag string, rs *HTTPRangeSpec) error {
-	metadata := make(map[string]string)
-	metadata["etag"] = eTag
+	metadata := map[string]string{"etag": eTag}
 	return dcache.SaveMetadata(ctx, bucket, object, metadata, size, rs, "", true)
 }
 
 // Backend metadata could have changed through server side copy - reset cache metadata if that is the case
 func (c *cacheObjects) updateMetadataIfChanged(ctx context.Context, dcache *diskCache, bucket, object string, bkObjectInfo, cacheObjInfo ObjectInfo, rs *HTTPRangeSpec) error {
 
-	bkMeta := make(map[string]string)
-	cacheMeta := make(map[string]string)
+	bkMeta := make(map[string]string, len(bkObjectInfo.UserDefined))
+	cacheMeta := make(map[string]string, len(cacheObjInfo.UserDefined))
 	for k, v := range bkObjectInfo.UserDefined {
 		if strings.HasPrefix(strings.ToLower(k), ReservedMetadataPrefixLower) {
 			// Do not need to send any internal metadata
@@ -166,13 +165,13 @@ func (c *cacheObjects) DeleteObjects(ctx context.Context, bucket string, objects
 
 // construct a metadata k-v map
 func getMetadata(objInfo ObjectInfo) map[string]string {
-	metadata := make(map[string]string)
+	metadata := make(map[string]string, len(objInfo.UserDefined)+4)
 	metadata["etag"] = objInfo.ETag
 	metadata["content-type"] = objInfo.ContentType
 	if objInfo.ContentEncoding != "" {
 		metadata["content-encoding"] = objInfo.ContentEncoding
 	}
-	if objInfo.Expires != timeSentinel {
+	if !objInfo.Expires.Equal(timeSentinel) {
 		metadata["expires"] = objInfo.Expires.Format(http.TimeFormat)
 	}
 	for k, v := range objInfo.UserDefined {
@@ -284,12 +283,6 @@ func (c *cacheObjects) GetObjectNInfo(ctx context.Context, bucket, object string
 	// Reaching here implies cache miss
 	c.cacheStats.incMiss()
 
-	// Since we got here, we are serving the request from backend,
-	// and also adding the object to the cache.
-	if dcache.diskUsageHigh() {
-		dcache.triggerGC <- struct{}{} // this is non-blocking
-	}
-
 	bkReader, bkErr := c.GetObjectNInfoFn(ctx, bucket, object, rs, h, lockType, opts)
 
 	if bkErr != nil {
@@ -306,7 +299,9 @@ func (c *cacheObjects) GetObjectNInfo(ctx context.Context, bucket, object string
 	if cacheErr == nil {
 		bkReader.ObjInfo.CacheLookupStatus = CacheHit
 	}
-	if !dcache.diskAvailable(objInfo.Size) {
+
+	// Check if we can add it without exceeding total cache size.
+	if !dcache.diskSpaceAvailable(objInfo.Size) {
 		return bkReader, bkErr
 	}
 
@@ -317,16 +312,18 @@ func (c *cacheObjects) GetObjectNInfo(ctx context.Context, bucket, object string
 				rs = nil
 			}
 			// fill cache in the background for range GET requests
-			bReader, bErr := c.GetObjectNInfoFn(ctx, bucket, object, rs, h, lockType, opts)
+			bReader, bErr := c.GetObjectNInfoFn(GlobalContext, bucket, object, rs, h, lockType, opts)
 			if bErr != nil {
 				return
 			}
 			defer bReader.Close()
-			oi, _, _, err := dcache.statRange(ctx, bucket, object, rs)
+			oi, _, _, err := dcache.statRange(GlobalContext, bucket, object, rs)
 			// avoid cache overwrite if another background routine filled cache
 			if err != nil || oi.ETag != bReader.ObjInfo.ETag {
 				// use a new context to avoid locker prematurely timing out operation when the GetObjectNInfo returns.
-				dcache.Put(context.Background(), bucket, object, bReader, bReader.ObjInfo.Size, rs, ObjectOptions{UserDefined: getMetadata(bReader.ObjInfo)}, false)
+				dcache.Put(GlobalContext, bucket, object, bReader, bReader.ObjInfo.Size, rs, ObjectOptions{
+					UserDefined: getMetadata(bReader.ObjInfo),
+				}, false)
 				return
 			}
 		}()
@@ -336,8 +333,13 @@ func (c *cacheObjects) GetObjectNInfo(ctx context.Context, bucket, object string
 	// Initialize pipe.
 	pipeReader, pipeWriter := io.Pipe()
 	teeReader := io.TeeReader(bkReader, pipeWriter)
+	userDefined := getMetadata(bkReader.ObjInfo)
 	go func() {
-		putErr := dcache.Put(ctx, bucket, object, io.LimitReader(pipeReader, bkReader.ObjInfo.Size), bkReader.ObjInfo.Size, nil, ObjectOptions{UserDefined: getMetadata(bkReader.ObjInfo)}, false)
+		putErr := dcache.Put(ctx, bucket, object,
+			io.LimitReader(pipeReader, bkReader.ObjInfo.Size),
+			bkReader.ObjInfo.Size, nil, ObjectOptions{
+				UserDefined: userDefined,
+			}, false)
 		// close the write end of the pipe, so the error gets
 		// propagated to getObjReader
 		pipeWriter.CloseWithError(putErr)
@@ -606,9 +608,10 @@ func (c *cacheObjects) PutObject(ctx context.Context, bucket, object string, r *
 	}
 
 	// fetch from backend if there is no space on cache drive
-	if !dcache.diskAvailable(size) {
+	if !dcache.diskSpaceAvailable(size) {
 		return putObjectFn(ctx, bucket, object, r, opts)
 	}
+
 	if opts.ServerSideEncryption != nil {
 		dcache.Delete(ctx, bucket, object)
 		return putObjectFn(ctx, bucket, object, r, opts)
@@ -634,15 +637,15 @@ func (c *cacheObjects) PutObject(ctx context.Context, bucket, object string, r *
 	if err == nil {
 		go func() {
 			// fill cache in the background
-			bReader, bErr := c.GetObjectNInfoFn(ctx, bucket, object, nil, http.Header{}, readLock, ObjectOptions{})
+			bReader, bErr := c.GetObjectNInfoFn(GlobalContext, bucket, object, nil, http.Header{}, readLock, ObjectOptions{})
 			if bErr != nil {
 				return
 			}
 			defer bReader.Close()
-			oi, _, err := dcache.Stat(ctx, bucket, object)
+			oi, _, err := dcache.Stat(GlobalContext, bucket, object)
 			// avoid cache overwrite if another background routine filled cache
 			if err != nil || oi.ETag != bReader.ObjInfo.ETag {
-				dcache.Put(ctx, bucket, object, bReader, bReader.ObjInfo.Size, nil, ObjectOptions{UserDefined: getMetadata(bReader.ObjInfo)}, false)
+				dcache.Put(GlobalContext, bucket, object, bReader, bReader.ObjInfo.Size, nil, ObjectOptions{UserDefined: getMetadata(bReader.ObjInfo)}, false)
 			}
 		}()
 	}
@@ -684,11 +687,13 @@ func newServerCacheObjects(ctx context.Context, config cache.Config) (CacheObjec
 	c.cacheStats.GetDiskStats = func() []CacheDiskStats {
 		cacheDiskStats := make([]CacheDiskStats, len(c.cache))
 		for i := range c.cache {
-			cacheDiskStats[i] = CacheDiskStats{
-				Dir: c.cache[i].stats.Dir,
+			dcache := c.cache[i]
+			cacheDiskStats[i] = CacheDiskStats{}
+			if dcache != nil {
+				cacheDiskStats[i].Dir = dcache.stats.Dir
+				atomic.StoreInt32(&cacheDiskStats[i].UsageState, atomic.LoadInt32(&dcache.stats.UsageState))
+				atomic.StoreUint64(&cacheDiskStats[i].UsagePercent, atomic.LoadUint64(&dcache.stats.UsagePercent))
 			}
-			atomic.StoreInt32(&cacheDiskStats[i].UsageState, atomic.LoadInt32(&c.cache[i].stats.UsageState))
-			atomic.StoreUint64(&cacheDiskStats[i].UsagePercent, atomic.LoadUint64(&c.cache[i].stats.UsagePercent))
 		}
 		return cacheDiskStats
 	}
@@ -712,7 +717,11 @@ func (c *cacheObjects) gc(ctx context.Context) {
 				continue
 			}
 			for _, dcache := range c.cache {
-				dcache.triggerGC <- struct{}{}
+				if dcache != nil {
+					// Check if there is disk.
+					// Will queue a GC scan if at high watermark.
+					dcache.diskSpaceAvailable(0)
+				}
 			}
 		}
 	}

cmd/disk-cache_test.go

@@ -27,7 +27,7 @@ func TestCacheMetadataObjInfo(t *testing.T) {
 	if objInfo.Size != 0 {
 		t.Fatal("Unexpected object info value for Size", objInfo.Size)
 	}
-	if objInfo.ModTime != timeSentinel {
+	if !objInfo.ModTime.Equal(timeSentinel) {
 		t.Fatal("Unexpected object info value for ModTime ", objInfo.ModTime)
 	}
 	if objInfo.IsDir {

cmd/dynamic-timeouts.go

@@ -26,7 +26,7 @@ const (
 	dynamicTimeoutIncreaseThresholdPct = 0.33 // Upper threshold for failures in order to increase timeout
 	dynamicTimeoutDecreaseThresholdPct = 0.10 // Lower threshold for failures in order to decrease timeout
 	dynamicTimeoutLogSize              = 16
-	maxDuration                        = time.Duration(1<<63 - 1)
+	maxDuration                        = 1<<63 - 1
 )
 
 // timeouts that are dynamically adapted based on actual usage results
@@ -110,7 +110,7 @@ func (dt *dynamicTimeout) adjust(entries [dynamicTimeoutLogSize]time.Duration) {
 		// We are hitting the timeout relatively few times, so decrease the timeout
 		average = average * 125 / 100 // Add buffer of 25% on top of average
 
-		timeout := (atomic.LoadInt64(&dt.timeout) + int64(average)) / 2 // Middle between current timeout and average success
+		timeout := (atomic.LoadInt64(&dt.timeout) + average) / 2 // Middle between current timeout and average success
 		if timeout < dt.minimum {
 			timeout = dt.minimum
 		}

cmd/encryption-v1.go

@@ -388,12 +388,7 @@ func DecryptBlocksRequestR(inputReader io.Reader, h http.Header, offset,
 		object:            object,
 		customerKeyHeader: h.Get(crypto.SSECKey),
 		copySource:        copySource,
-	}
-
-	w.metadata = map[string]string{}
-	// Copy encryption metadata for internal use.
-	for k, v := range oi.UserDefined {
-		w.metadata[k] = v
+		metadata:          cloneMSS(oi.UserDefined),
 	}
 
 	if w.copySource {
@@ -432,10 +427,7 @@ type DecryptBlocksReader struct {
 }
 
 func (d *DecryptBlocksReader) buildDecrypter(partID int) error {
-	m := make(map[string]string)
-	for k, v := range d.metadata {
-		m[k] = v
-	}
+	m := cloneMSS(d.metadata)
 	// Initialize the first decrypter; new decrypters will be
 	// initialized in Read() operation as needed.
 	var key []byte
@@ -605,12 +597,14 @@ func getDecryptedETag(headers http.Header, objInfo ObjectInfo, copySource bool)
 	if crypto.IsMultiPart(objInfo.UserDefined) {
 		return objInfo.ETag
 	}
+
 	if crypto.SSECopy.IsRequested(headers) {
 		key, err = crypto.SSECopy.ParseHTTP(headers)
 		if err != nil {
 			return objInfo.ETag
 		}
 	}
+
 	// As per AWS S3 Spec, ETag for SSE-C encrypted objects need not be MD5Sum of the data.
 	// Since server side copy with same source and dest just replaces the ETag, we save
 	// encrypted content MD5Sum as ETag for both SSE-C and SSE-S3, we standardize the ETag
@@ -664,7 +658,7 @@ func (o *ObjectInfo) GetDecryptedRange(rs *HTTPRangeSpec) (encOff, encLength, sk
 
 	if rs == nil {
 		// No range, so offsets refer to the whole object.
-		return 0, int64(o.Size), 0, 0, 0, nil
+		return 0, o.Size, 0, 0, 0, nil
 	}
 
 	// Assemble slice of (decrypted) part sizes in `sizes`

cmd/encryption-v1_test.go

@@ -373,7 +373,7 @@ func TestGetDecryptedRange(t *testing.T) {
 			sum := int64(0)
 			for i, s := range sizes {
 				r[i].Number = i
-				r[i].Size = int64(getEncSize(s))
+				r[i].Size = getEncSize(s)
 				sum += r[i].Size
 			}
 			return ObjectInfo{

cmd/endpoint-ellipses.go

@@ -329,28 +329,28 @@ var (
 // CreateServerEndpoints - validates and creates new endpoints from input args, supports
 // both ellipses and without ellipses transparently.
 func createServerEndpoints(serverAddr string, args ...string) (
-	endpointZones EndpointZones, setDriveCount int,
-	setupType SetupType, err error) {
+	endpointZones EndpointZones, setupType SetupType, err error) {
 
 	if len(args) == 0 {
-		return nil, -1, -1, errInvalidArgument
+		return nil, -1, errInvalidArgument
 	}
 
+	var setDriveCount int
 	if v := env.Get(EnvErasureSetDriveCount, ""); v != "" {
 		setDriveCount, err = strconv.Atoi(v)
 		if err != nil {
-			return nil, -1, -1, config.ErrInvalidErasureSetSize(err)
+			return nil, -1, config.ErrInvalidErasureSetSize(err)
 		}
 	}
 
 	if !ellipses.HasEllipses(args...) {
 		setArgs, err := GetAllSets(uint64(setDriveCount), args...)
 		if err != nil {
-			return nil, -1, -1, err
+			return nil, -1, err
 		}
 		endpointList, newSetupType, err := CreateEndpoints(serverAddr, false, setArgs...)
 		if err != nil {
-			return nil, -1, -1, err
+			return nil, -1, err
 		}
 		endpointZones = append(endpointZones, ZoneEndpoints{
 			SetCount:     len(setArgs),
@@ -358,7 +358,7 @@ func createServerEndpoints(serverAddr string, args ...string) (
 			Endpoints:    endpointList,
 		})
 		setupType = newSetupType
-		return endpointZones, len(setArgs[0]), setupType, nil
+		return endpointZones, setupType, nil
 	}
 
 	var prevSetupType SetupType
@@ -366,25 +366,25 @@ func createServerEndpoints(serverAddr string, args ...string) (
 	for _, arg := range args {
 		setArgs, err := GetAllSets(uint64(setDriveCount), arg)
 		if err != nil {
-			return nil, -1, -1, err
+			return nil, -1, err
 		}
 		var endpointList Endpoints
 		endpointList, setupType, err = CreateEndpoints(serverAddr, foundPrevLocal, setArgs...)
 		if err != nil {
-			return nil, -1, -1, err
+			return nil, -1, err
 		}
 		if setDriveCount != 0 && setDriveCount != len(setArgs[0]) {
-			return nil, -1, -1, fmt.Errorf("All zones should have same drive per set ratio - expected %d, got %d", setDriveCount, len(setArgs[0]))
+			return nil, -1, fmt.Errorf("All zones should have same drive per set ratio - expected %d, got %d", setDriveCount, len(setArgs[0]))
 		}
 		if prevSetupType != UnknownSetupType && prevSetupType != setupType {
-			return nil, -1, -1, fmt.Errorf("All zones should be of the same setup-type to maintain the original SLA expectations - expected %s, got %s", prevSetupType, setupType)
+			return nil, -1, fmt.Errorf("All zones should be of the same setup-type to maintain the original SLA expectations - expected %s, got %s", prevSetupType, setupType)
 		}
 		if err = endpointZones.Add(ZoneEndpoints{
 			SetCount:     len(setArgs),
 			DrivesPerSet: len(setArgs[0]),
 			Endpoints:    endpointList,
 		}); err != nil {
-			return nil, -1, -1, err
+			return nil, -1, err
 		}
 		foundPrevLocal = endpointList.atleastOneEndpointLocal()
 		if setDriveCount == 0 {
@@ -393,5 +393,5 @@ func createServerEndpoints(serverAddr string, args ...string) (
 		prevSetupType = setupType
 	}
 
-	return endpointZones, setDriveCount, setupType, nil
+	return endpointZones, setupType, nil
 }

cmd/endpoint-ellipses_test.go

@@ -56,7 +56,7 @@ func TestCreateServerEndpoints(t *testing.T) {
 	for _, testCase := range testCases {
 		testCase := testCase
 		t.Run("", func(t *testing.T) {
-			_, _, _, err := createServerEndpoints(testCase.serverAddr, testCase.args...)
+			_, _, err := createServerEndpoints(testCase.serverAddr, testCase.args...)
 			if err != nil && testCase.success {
 				t.Errorf("Expected success but failed instead %s", err)
 			}

cmd/endpoint.go

@@ -24,6 +24,7 @@ import (
 	"net/url"
 	"path"
 	"path/filepath"
+	"reflect"
 	"runtime"
 	"strconv"
 	"strings"
@@ -193,7 +194,7 @@ func NewEndpoint(arg string) (ep Endpoint, e error) {
 }
 
 // ZoneEndpoints represent endpoints in a given zone
-// along with its setCount and drivesPerSet.
+// along with its setCount and setDriveCount.
 type ZoneEndpoints struct {
 	SetCount     int
 	DrivesPerSet int
@@ -203,6 +204,21 @@ type ZoneEndpoints struct {
 // EndpointZones - list of list of endpoints
 type EndpointZones []ZoneEndpoints
 
+// GetLocalZoneIdx returns the zone which endpoint belongs to locally.
+// if ep is remote this code will return -1 zoneIndex
+func (l EndpointZones) GetLocalZoneIdx(ep Endpoint) int {
+	for i, zep := range l {
+		for _, cep := range zep.Endpoints {
+			if cep.IsLocal && ep.IsLocal {
+				if reflect.DeepEqual(cep, ep) {
+					return i
+				}
+			}
+		}
+	}
+	return -1
+}
+
 // Add add zone endpoints
 func (l *EndpointZones) Add(zeps ZoneEndpoints) error {
 	existSet := set.NewStringSet()
@@ -753,9 +769,16 @@ func GetProxyEndpoints(endpointZones EndpointZones) ([]ProxyEndpoint, error) {
 					RootCAs:    globalRootCAs,
 				}
 			}
+
+			tr := newCustomHTTPTransport(tlsConfig, rest.DefaultRESTTimeout)()
+			// Allow more requests to be in flight with higher response header timeout.
+			tr.ResponseHeaderTimeout = 30 * time.Minute
+			tr.MaxIdleConns = 64
+			tr.MaxIdleConnsPerHost = 64
+
 			proxyEps = append(proxyEps, ProxyEndpoint{
 				Endpoint:  endpoint,
-				Transport: newCustomHTTPTransport(tlsConfig, rest.DefaultRESTTimeout)(),
+				Transport: tr,
 			})
 		}
 	}

cmd/erasure-bucket.go

@@ -49,7 +49,7 @@ func (er erasureObjects) MakeBucketWithLocation(ctx context.Context, bucket stri
 		index := index
 		g.Go(func() error {
 			if storageDisks[index] != nil {
-				if err := storageDisks[index].MakeVol(bucket); err != nil {
+				if err := storageDisks[index].MakeVol(ctx, bucket); err != nil {
 					if err != errVolumeExists {
 						logger.LogIf(ctx, err)
 					}
@@ -75,7 +75,7 @@ func undoDeleteBucket(storageDisks []StorageAPI, bucket string) {
 		}
 		index := index
 		g.Go(func() error {
-			_ = storageDisks[index].MakeVol(bucket)
+			_ = storageDisks[index].MakeVol(context.Background(), bucket)
 			return nil
 		}, index)
 	}
@@ -92,7 +92,7 @@ func (er erasureObjects) getBucketInfo(ctx context.Context, bucketName string) (
 			bucketErrs = append(bucketErrs, errDiskNotFound)
 			continue
 		}
-		volInfo, serr := disk.StatVol(bucketName)
+		volInfo, serr := disk.StatVol(ctx, bucketName)
 		if serr == nil {
 			return BucketInfo(volInfo), nil
 		}
@@ -129,12 +129,11 @@ func (er erasureObjects) listBuckets(ctx context.Context) (bucketsInfo []BucketI
 			continue
 		}
 		var volsInfo []VolInfo
-		volsInfo, err = disk.ListVols()
+		volsInfo, err = disk.ListVols(ctx)
 		if err == nil {
 			// NOTE: The assumption here is that volumes across all disks in
 			// readQuorum have consistent view i.e they all have same number
-			// of buckets. This is essentially not verified since healing
-			// should take care of this.
+			// of buckets.
 			var bucketsInfo []BucketInfo
 			for _, volInfo := range volsInfo {
 				if isReservedOrInvalidBucket(volInfo.Name, true) {
@@ -182,10 +181,10 @@ func deleteDanglingBucket(ctx context.Context, storageDisks []StorageAPI, dErrs
 	for index, err := range dErrs {
 		if err == errVolumeNotEmpty {
 			// Attempt to delete bucket again.
-			if derr := storageDisks[index].DeleteVol(bucket, false); derr == errVolumeNotEmpty {
+			if derr := storageDisks[index].DeleteVol(ctx, bucket, false); derr == errVolumeNotEmpty {
 				_ = cleanupDir(ctx, storageDisks[index], bucket, "")
 
-				_ = storageDisks[index].DeleteVol(bucket, false)
+				_ = storageDisks[index].DeleteVol(ctx, bucket, false)
 
 				// Cleanup all the previously incomplete multiparts.
 				_ = cleanupDir(ctx, storageDisks[index], minioMetaMultipartBucket, bucket)
@@ -205,7 +204,7 @@ func (er erasureObjects) DeleteBucket(ctx context.Context, bucket string, forceD
 		index := index
 		g.Go(func() error {
 			if storageDisks[index] != nil {
-				if err := storageDisks[index].DeleteVol(bucket, forceDelete); err != nil {
+				if err := storageDisks[index].DeleteVol(ctx, bucket, forceDelete); err != nil {
 					return err
 				}
 				err := cleanupDir(ctx, storageDisks[index], minioMetaMultipartBucket, bucket)
@@ -244,7 +243,10 @@ func (er erasureObjects) DeleteBucket(ctx context.Context, bucket string, forceD
 	// If we reduce quorum to nil, means we have deleted buckets properly
 	// on some servers in quorum, we should look for volumeNotEmpty errors
 	// and delete those buckets as well.
-	deleteDanglingBucket(ctx, storageDisks, dErrs, bucket)
+	//
+	// let this call succeed, even if client cancels the context
+	// this is to ensure that we don't leave any stale content
+	deleteDanglingBucket(context.Background(), storageDisks, dErrs, bucket)
 
 	return nil
 }

cmd/erasure-coding.go

@@ -125,7 +125,7 @@ func (e *Erasure) ShardFileSize(totalLength int64) int64 {
 		return -1
 	}
 	numShards := totalLength / e.blockSize
-	lastBlockSize := totalLength % int64(e.blockSize)
+	lastBlockSize := totalLength % e.blockSize
 	lastShardSize := ceilFrac(lastBlockSize, int64(e.dataBlocks))
 	return numShards*e.ShardSize() + lastShardSize
 }
@@ -134,7 +134,7 @@ func (e *Erasure) ShardFileSize(totalLength int64) int64 {
 func (e *Erasure) ShardFileOffset(startOffset, length, totalLength int64) int64 {
 	shardSize := e.ShardSize()
 	shardFileSize := e.ShardFileSize(totalLength)
-	endShard := (startOffset + int64(length)) / e.blockSize
+	endShard := (startOffset + length) / e.blockSize
 	tillOffset := endShard*shardSize + shardSize
 	if tillOffset > shardFileSize {
 		tillOffset = shardFileSize

cmd/erasure-common.go

@@ -23,6 +23,35 @@ import (
 	"github.com/minio/minio/pkg/sync/errgroup"
 )
 
+func (er erasureObjects) getLoadBalancedLocalDisks() (newDisks []StorageAPI) {
+	disks := er.getDisks()
+	// Based on the random shuffling return back randomized disks.
+	for _, i := range hashOrder(UTCNow().String(), len(disks)) {
+		if disks[i-1] != nil && disks[i-1].IsLocal() {
+			newDisks = append(newDisks, disks[i-1])
+		}
+	}
+	return newDisks
+}
+
+// getLoadBalancedNDisks - fetches load balanced (sufficiently randomized) disk slice
+// with N disks online. If ndisks is zero or negative, then it will returns all disks,
+// same if ndisks is greater than the number of all disks.
+func (er erasureObjects) getLoadBalancedNDisks(ndisks int) (newDisks []StorageAPI) {
+	disks := er.getLoadBalancedDisks()
+	for _, disk := range disks {
+		if disk == nil {
+			continue
+		}
+		newDisks = append(newDisks, disk)
+		ndisks--
+		if ndisks == 0 {
+			break
+		}
+	}
+	return
+}
+
 // getLoadBalancedDisks - fetches load balanced (sufficiently randomized) disk slice.
 func (er erasureObjects) getLoadBalancedDisks() (newDisks []StorageAPI) {
 	disks := er.getDisks()
@@ -66,7 +95,7 @@ func (er erasureObjects) isObject(ctx context.Context, bucket, prefix string) (o
 				return errDiskNotFound
 			}
 			// Check if 'prefix' is an object on this 'disk', else continue the check the next disk
-			return storageDisks[index].CheckFile(bucket, prefix)
+			return storageDisks[index].CheckFile(ctx, bucket, prefix)
 		}, index)
 	}
 

cmd/erasure-common_test.go

@@ -32,6 +32,7 @@ func TestErasureParentDirIsObject(t *testing.T) {
 	if err != nil {
 		t.Fatalf("Unable to initialize 'Erasure' object layer.")
 	}
+	defer obj.Shutdown(context.Background())
 
 	// Remove all disks.
 	for _, disk := range fsDisks {

cmd/erasure-decode_test.go

@@ -28,7 +28,7 @@ import (
 	humanize "github.com/dustin/go-humanize"
 )
 
-func (a badDisk) ReadFile(volume string, path string, offset int64, buf []byte, verifier *BitrotVerifier) (n int64, err error) {
+func (a badDisk) ReadFile(ctx context.Context, volume string, path string, offset int64, buf []byte, verifier *BitrotVerifier) (n int64, err error) {
 	return 0, errFaultyDisk
 }
 

cmd/erasure-encode_test.go

@@ -32,11 +32,11 @@ func (a badDisk) String() string {
 	return "bad-disk"
 }
 
-func (a badDisk) AppendFile(volume string, path string, buf []byte) error {
+func (a badDisk) AppendFile(ctx context.Context, volume string, path string, buf []byte) error {
 	return errFaultyDisk
 }
 
-func (a badDisk) ReadFileStream(volume, path string, offset, length int64) (io.ReadCloser, error) {
+func (a badDisk) ReadFileStream(ctx context.Context, volume, path string, offset, length int64) (io.ReadCloser, error) {
 	return nil, errFaultyDisk
 }
 
@@ -44,7 +44,7 @@ func (a badDisk) UpdateBloomFilter(ctx context.Context, oldest, current uint64)
 	return nil, errFaultyDisk
 }
 
-func (a badDisk) CreateFile(volume, path string, size int64, reader io.Reader) error {
+func (a badDisk) CreateFile(ctx context.Context, volume, path string, size int64, reader io.Reader) error {
 	return errFaultyDisk
 }
 
@@ -195,7 +195,7 @@ func benchmarkErasureEncode(data, parity, dataDown, parityDown int, size int64,
 			if disk == OfflineDisk {
 				continue
 			}
-			disk.DeleteFile("testbucket", "object")
+			disk.DeleteFile(context.Background(), "testbucket", "object")
 			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)

cmd/erasure-healing-common.go

@@ -175,9 +175,9 @@ func disksWithAllParts(ctx context.Context, onlineDisks []StorageAPI, partsMetad
 			// disk has a valid xl.meta but may not have all the
 			// parts. This is considered an outdated disk, since
 			// it needs healing too.
-			dataErrs[i] = onlineDisk.VerifyFile(bucket, object, partsMetadata[i])
+			dataErrs[i] = onlineDisk.VerifyFile(ctx, bucket, object, partsMetadata[i])
 		case madmin.HealNormalScan:
-			dataErrs[i] = onlineDisk.CheckParts(bucket, object, partsMetadata[i])
+			dataErrs[i] = onlineDisk.CheckParts(ctx, bucket, object, partsMetadata[i])
 		}
 
 		if dataErrs[i] == nil {

cmd/erasure-healing-common_test.go

@@ -83,7 +83,7 @@ func TestCommonTime(t *testing.T) {
 	for i, testCase := range testCases {
 		// Obtain a common mod time from modTimes slice.
 		ctime, _ := commonTime(testCase.times)
-		if testCase.time != ctime {
+		if !testCase.time.Equal(ctime) {
 			t.Fatalf("Test case %d, expect to pass but failed. Wanted modTime: %s, got modTime: %s\n", i+1, testCase.time, ctime)
 		}
 	}
@@ -99,6 +99,7 @@ func TestListOnlineDisks(t *testing.T) {
 	if err != nil {
 		t.Fatalf("Prepare Erasure backend failed - %v", err)
 	}
+	defer obj.Shutdown(context.Background())
 	defer removeRoots(disks)
 
 	type tamperKind int
@@ -210,7 +211,7 @@ func TestListOnlineDisks(t *testing.T) {
 				// and check if that disk
 				// appears in outDatedDisks.
 				tamperedIndex = index
-				dErr := erasureDisks[index].DeleteFile(bucket, pathJoin(object, fi.DataDir, "part.1"))
+				dErr := erasureDisks[index].DeleteFile(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"))
 				if dErr != nil {
 					t.Fatalf("Test %d: Failed to delete %s - %v", i+1,
 						filepath.Join(object, "part.1"), dErr)
@@ -265,6 +266,7 @@ func TestDisksWithAllParts(t *testing.T) {
 	if err != nil {
 		t.Fatalf("Prepare Erasure backend failed - %v", err)
 	}
+	defer obj.Shutdown(context.Background())
 	defer removeRoots(disks)
 
 	bucket := "bucket"

cmd/erasure-healing.go

@@ -18,6 +18,7 @@ package cmd
 
 import (
 	"context"
+	"errors"
 	"fmt"
 	"io"
 	"sync"
@@ -28,16 +29,6 @@ import (
 	"github.com/minio/minio/pkg/sync/errgroup"
 )
 
-func (er erasureObjects) ReloadFormat(ctx context.Context, dryRun bool) error {
-	logger.LogIf(ctx, NotImplemented{})
-	return NotImplemented{}
-}
-
-func (er erasureObjects) HealFormat(ctx context.Context, dryRun bool) (madmin.HealResultItem, error) {
-	logger.LogIf(ctx, NotImplemented{})
-	return madmin.HealResultItem{}, NotImplemented{}
-}
-
 // Heals a bucket if it doesn't exist on one of the disks, additionally
 // also heals the missing entries for bucket metadata files
 // `policy.json, notification.xml, listeners.json`.
@@ -77,7 +68,7 @@ func healBucket(ctx context.Context, storageDisks []StorageAPI, storageEndpoints
 				afterState[index] = madmin.DriveStateOffline
 				return errDiskNotFound
 			}
-			if _, serr := storageDisks[index].StatVol(bucket); serr != nil {
+			if _, serr := storageDisks[index].StatVol(ctx, bucket); serr != nil {
 				if serr == errDiskNotFound {
 					beforeState[index] = madmin.DriveStateOffline
 					afterState[index] = madmin.DriveStateOffline
@@ -135,7 +126,7 @@ func healBucket(ctx context.Context, storageDisks []StorageAPI, storageEndpoints
 		index := index
 		g.Go(func() error {
 			if beforeState[index] == madmin.DriveStateMissing {
-				makeErr := storageDisks[index].MakeVol(bucket)
+				makeErr := storageDisks[index].MakeVol(ctx, bucket)
 				if makeErr == nil {
 					afterState[index] = madmin.DriveStateOk
 				}
@@ -170,7 +161,7 @@ func listAllBuckets(storageDisks []StorageAPI, healBuckets map[string]VolInfo) (
 			continue
 		}
 		var volsInfo []VolInfo
-		volsInfo, err = disk.ListVols()
+		volsInfo, err = disk.ListVols(context.TODO())
 		if err != nil {
 			if IsErrIgnored(err, bucketMetadataOpIgnoredErrs...) {
 				continue
@@ -197,10 +188,10 @@ func listAllBuckets(storageDisks []StorageAPI, healBuckets map[string]VolInfo) (
 // 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, quorumModTime time.Time) bool {
-	switch erErr {
-	case errFileNotFound, errFileVersionNotFound:
+	switch {
+	case errors.Is(erErr, errFileNotFound) || errors.Is(erErr, errFileVersionNotFound):
 		return true
-	case errCorruptedFormat:
+	case errors.Is(erErr, errCorruptedFormat):
 		return true
 	}
 	if erErr == nil {
@@ -471,7 +462,7 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 		}
 
 		// Attempt a rename now from healed data to final location.
-		if err = disk.RenameData(minioMetaTmpBucket, tmpID, partsMetadata[i].DataDir, bucket, object); err != nil {
+		if err = disk.RenameData(ctx, minioMetaTmpBucket, tmpID, partsMetadata[i].DataDir, bucket, object); err != nil {
 			if err != errIsNotRegular && err != errFileNotFound {
 				logger.LogIf(ctx, err)
 			}
@@ -524,7 +515,7 @@ func (er erasureObjects) healObjectDir(ctx context.Context, bucket, object strin
 				wg.Add(1)
 				go func(index int, disk StorageAPI) {
 					defer wg.Done()
-					_ = disk.DeleteFile(bucket, object)
+					_ = disk.DeleteFile(ctx, bucket, object)
 				}(index, disk)
 			}
 			wg.Wait()
@@ -556,7 +547,7 @@ func (er erasureObjects) healObjectDir(ctx context.Context, bucket, object strin
 	for i, err := range errs {
 		if err == errVolumeNotFound || err == errFileNotFound {
 			// Bucket or prefix/directory not found
-			merr := storageDisks[i].MakeVol(pathJoin(bucket, object))
+			merr := storageDisks[i].MakeVol(ctx, pathJoin(bucket, object))
 			switch merr {
 			case nil, errVolumeExists:
 				hr.After.Drives[i].State = madmin.DriveStateOk
@@ -641,7 +632,7 @@ func statAllDirs(ctx context.Context, storageDisks []StorageAPI, bucket, prefix
 		}
 		index := index
 		g.Go(func() error {
-			entries, err := storageDisks[index].ListDir(bucket, prefix, 1)
+			entries, err := storageDisks[index].ListDir(ctx, bucket, prefix, 1)
 			if err != nil {
 				return err
 			}
@@ -686,9 +677,9 @@ func isObjectDangling(metaArr []FileInfo, errs []error, dataErrs []error) (valid
 	// or when er.meta is not readable in read quorum disks.
 	var notFoundErasureMeta, corruptedErasureMeta int
 	for _, readErr := range errs {
-		if readErr == errFileNotFound || readErr == errFileVersionNotFound {
+		if errors.Is(readErr, errFileNotFound) || errors.Is(readErr, errFileVersionNotFound) {
 			notFoundErasureMeta++
-		} else if readErr == errCorruptedFormat {
+		} else if errors.Is(readErr, errCorruptedFormat) {
 			corruptedErasureMeta++
 		}
 	}
@@ -699,7 +690,10 @@ func isObjectDangling(metaArr []FileInfo, errs []error, dataErrs []error) (valid
 		// double counting when both parts and er.meta
 		// are not available.
 		if errs[i] != dataErrs[i] {
-			if dataErrs[i] == errFileNotFound || dataErrs[i] == errFileVersionNotFound {
+			if IsErr(dataErrs[i], []error{
+				errFileNotFound,
+				errFileVersionNotFound,
+			}...) {
 				notFoundParts++
 			}
 		}

cmd/erasure-healing_test.go

@@ -39,6 +39,7 @@ func TestHealing(t *testing.T) {
 	if err != nil {
 		t.Fatal(err)
 	}
+	defer obj.Shutdown(context.Background())
 	defer removeRoots(fsDirs)
 
 	z := obj.(*erasureZones)
@@ -66,7 +67,7 @@ func TestHealing(t *testing.T) {
 	}
 
 	disk := er.getDisks()[0]
-	fileInfoPreHeal, err := disk.ReadVersion(bucket, object, "")
+	fileInfoPreHeal, err := disk.ReadVersion(context.Background(), bucket, object, "")
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -83,7 +84,7 @@ func TestHealing(t *testing.T) {
 		t.Fatal(err)
 	}
 
-	fileInfoPostHeal, err := disk.ReadVersion(bucket, object, "")
+	fileInfoPostHeal, err := disk.ReadVersion(context.Background(), bucket, object, "")
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -102,7 +103,7 @@ func TestHealing(t *testing.T) {
 	// gone down when an object was replaced by a new object.
 	fileInfoOutDated := fileInfoPreHeal
 	fileInfoOutDated.ModTime = time.Now()
-	err = disk.WriteMetadata(bucket, object, fileInfoOutDated)
+	err = disk.WriteMetadata(context.Background(), bucket, object, fileInfoOutDated)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -112,7 +113,7 @@ func TestHealing(t *testing.T) {
 		t.Fatal(err)
 	}
 
-	fileInfoPostHeal, err = disk.ReadVersion(bucket, object, "")
+	fileInfoPostHeal, err = disk.ReadVersion(context.Background(), bucket, object, "")
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -134,7 +135,7 @@ func TestHealing(t *testing.T) {
 		t.Fatal(err)
 	}
 	// Stat the bucket to make sure that it was created.
-	_, err = er.getDisks()[0].StatVol(bucket)
+	_, err = er.getDisks()[0].StatVol(context.Background(), bucket)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -200,7 +201,7 @@ func TestHealObjectCorrupted(t *testing.T) {
 	er := z.zones[0].sets[0]
 	erasureDisks := er.getDisks()
 	firstDisk := erasureDisks[0]
-	err = firstDisk.DeleteFile(bucket, pathJoin(object, xlStorageFormatFile))
+	err = firstDisk.DeleteFile(context.Background(), bucket, pathJoin(object, xlStorageFormatFile))
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
 	}
@@ -216,16 +217,16 @@ func TestHealObjectCorrupted(t *testing.T) {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
 	}
 
-	if err = firstDisk.CheckFile(bucket, object); err != nil {
+	if err = firstDisk.CheckFile(context.Background(), bucket, object); err != nil {
 		t.Errorf("Expected er.meta file to be present but stat failed - %v", err)
 	}
 
-	err = firstDisk.DeleteFile(bucket, pathJoin(object, fi.DataDir, "part.1"))
+	err = firstDisk.DeleteFile(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"))
 	if err != nil {
 		t.Errorf("Failure during deleting part.1 - %v", err)
 	}
 
-	err = firstDisk.WriteAll(bucket, pathJoin(object, fi.DataDir, "part.1"), bytes.NewReader([]byte{}))
+	err = firstDisk.WriteAll(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), bytes.NewReader([]byte{}))
 	if err != nil {
 		t.Errorf("Failure during creating part.1 - %v", err)
 	}
@@ -245,13 +246,13 @@ func TestHealObjectCorrupted(t *testing.T) {
 		t.Fatalf("FileInfo not equal after healing")
 	}
 
-	err = firstDisk.DeleteFile(bucket, pathJoin(object, fi.DataDir, "part.1"))
+	err = firstDisk.DeleteFile(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"))
 	if err != nil {
 		t.Errorf("Failure during deleting part.1 - %v", err)
 	}
 
 	bdata := bytes.Repeat([]byte("b"), int(nfi.Size))
-	err = firstDisk.WriteAll(bucket, pathJoin(object, fi.DataDir, "part.1"), bytes.NewReader(bdata))
+	err = firstDisk.WriteAll(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), bytes.NewReader(bdata))
 	if err != nil {
 		t.Errorf("Failure during creating part.1 - %v", err)
 	}
@@ -274,7 +275,7 @@ func TestHealObjectCorrupted(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 <= len(er.getDisks())/2; i++ {
-		er.getDisks()[i].DeleteFile(bucket, pathJoin(object, xlStorageFormatFile))
+		er.getDisks()[i].DeleteFile(context.Background(), bucket, pathJoin(object, xlStorageFormatFile))
 	}
 
 	// Try healing now, expect to receive errFileNotFound.
@@ -350,7 +351,7 @@ func TestHealObjectErasure(t *testing.T) {
 		t.Fatalf("Failed to complete multipart upload - %v", err)
 	}
 
-	err = firstDisk.DeleteFile(bucket, pathJoin(object, xlStorageFormatFile))
+	err = firstDisk.DeleteFile(context.Background(), bucket, pathJoin(object, xlStorageFormatFile))
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
 	}
@@ -360,7 +361,7 @@ func TestHealObjectErasure(t *testing.T) {
 		t.Fatalf("Failed to heal object - %v", err)
 	}
 
-	if err = firstDisk.CheckFile(bucket, object); err != nil {
+	if err = firstDisk.CheckFile(context.Background(), bucket, object); err != nil {
 		t.Errorf("Expected er.meta file to be present but stat failed - %v", err)
 	}
 
@@ -421,7 +422,7 @@ func TestHealEmptyDirectoryErasure(t *testing.T) {
 	z := obj.(*erasureZones)
 	er := z.zones[0].sets[0]
 	firstDisk := er.getDisks()[0]
-	err = firstDisk.DeleteFile(bucket, object)
+	err = firstDisk.DeleteFile(context.Background(), bucket, object)
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
 	}
@@ -433,7 +434,7 @@ func TestHealEmptyDirectoryErasure(t *testing.T) {
 	}
 
 	// Check if the empty directory is restored in the first disk
-	_, err = firstDisk.StatVol(pathJoin(bucket, object))
+	_, err = firstDisk.StatVol(context.Background(), pathJoin(bucket, object))
 	if err != nil {
 		t.Fatalf("Expected object to be present but stat failed - %v", err)
 	}

cmd/erasure-list-objects.go

@@ -1,58 +0,0 @@
-/*
- * MinIO Cloud Storage, (C) 2016-2020 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 cmd
-
-import (
-	"context"
-
-	"github.com/minio/minio/pkg/madmin"
-)
-
-// ListObjectVersions - This is not implemented, look for erasure-zones.ListObjectVersions()
-func (er erasureObjects) ListObjectVersions(ctx context.Context, bucket, prefix, marker, versionMarker, delimiter string, maxKeys int) (loi ListObjectVersionsInfo, e error) {
-	return loi, NotImplemented{}
-}
-
-// ListObjectsV2 - This is not implemented/needed anymore, look for erasure-zones.ListObjectsV2()
-func (er erasureObjects) ListObjectsV2(ctx context.Context, bucket, prefix, continuationToken, delimiter string, maxKeys int, fetchOwner bool, startAfter string) (loi ListObjectsV2Info, e error) {
-	return loi, NotImplemented{}
-}
-
-// ListObjects - This is not implemented/needed anymore, look for erasure-zones.ListObjects()
-func (er erasureObjects) ListObjects(ctx context.Context, bucket, prefix, marker, delimiter string, maxKeys int) (loi ListObjectsInfo, e error) {
-	return loi, NotImplemented{}
-}
-
-// ListBucketsHeal - This is not implemented/needed anymore, look for erasure-zones.ListBucketHeal()
-func (er erasureObjects) ListBucketsHeal(ctx context.Context) ([]BucketInfo, error) {
-	return nil, NotImplemented{}
-}
-
-// ListObjectsHeal - This is not implemented, look for erasure-zones.ListObjectsHeal()
-func (er erasureObjects) ListObjectsHeal(ctx context.Context, bucket, prefix, marker, delimiter string, maxKeys int) (loi ListObjectsInfo, err error) {
-	return ListObjectsInfo{}, NotImplemented{}
-}
-
-// HealObjects - This is not implemented/needed anymore, look for erasure-zones.HealObjects()
-func (er erasureObjects) HealObjects(ctx context.Context, bucket, prefix string, _ madmin.HealOpts, _ HealObjectFn) (e error) {
-	return NotImplemented{}
-}
-
-// Walk - This is not implemented/needed anymore, look for erasure-zones.Walk()
-func (er erasureObjects) Walk(ctx context.Context, bucket, prefix string, results chan<- ObjectInfo, _ ObjectOptions) error {
-	return NotImplemented{}
-}

cmd/erasure-metadata-utils.go

@@ -119,14 +119,14 @@ func readAllFileInfo(ctx context.Context, disks []StorageAPI, bucket, object, ve
 	metadataArray := make([]FileInfo, len(disks))
 
 	g := errgroup.WithNErrs(len(disks))
-	// Read `xl.meta` parallelly across disks.
+	// Read `xl.meta` in parallel across disks.
 	for index := range disks {
 		index := index
 		g.Go(func() (err error) {
 			if disks[index] == nil {
 				return errDiskNotFound
 			}
-			metadataArray[index], err = disks[index].ReadVersion(bucket, object, versionID)
+			metadataArray[index], err = disks[index].ReadVersion(ctx, bucket, object, versionID)
 			if err != nil {
 				if err != errFileNotFound && err != errVolumeNotFound && err != errFileVersionNotFound {
 					logger.GetReqInfo(ctx).AppendTags("disk", disks[index].String())

cmd/erasure-metadata.go

@@ -88,9 +88,10 @@ func (fi FileInfo) IsValid() bool {
 		// for erasure coding information
 		return true
 	}
-	data := fi.Erasure.DataBlocks
-	parity := fi.Erasure.ParityBlocks
-	return ((data >= parity) && (data != 0) && (parity != 0))
+	dataBlocks := fi.Erasure.DataBlocks
+	parityBlocks := fi.Erasure.ParityBlocks
+	return ((dataBlocks >= parityBlocks) &&
+		(dataBlocks != 0) && (parityBlocks != 0))
 }
 
 // ToObjectInfo - Converts metadata to object info.
@@ -102,15 +103,22 @@ func (fi FileInfo) ToObjectInfo(bucket, object string) ObjectInfo {
 			IsDir:  true,
 		}
 	}
+
+	versionID := fi.VersionID
+	if globalBucketVersioningSys.Enabled(bucket) && versionID == "" {
+		versionID = nullVersionID
+	}
+
 	objInfo := ObjectInfo{
 		IsDir:           false,
 		Bucket:          bucket,
 		Name:            object,
-		VersionID:       fi.VersionID,
+		VersionID:       versionID,
 		IsLatest:        fi.IsLatest,
 		DeleteMarker:    fi.Deleted,
 		Size:            fi.Size,
 		ModTime:         fi.ModTime,
+		Legacy:          fi.XLV1,
 		ContentType:     fi.Metadata["content-type"],
 		ContentEncoding: fi.Metadata["content-encoding"],
 	}
@@ -271,7 +279,7 @@ func renameFileInfo(ctx context.Context, disks []StorageAPI, srcBucket, srcEntry
 			if disks[index] == nil {
 				return errDiskNotFound
 			}
-			if err := disks[index].RenameData(srcBucket, srcEntry, "", dstBucket, dstEntry); err != nil {
+			if err := disks[index].RenameData(ctx, srcBucket, srcEntry, "", dstBucket, dstEntry); err != nil {
 				if !IsErrIgnored(err, ignoredErr...) {
 					return err
 				}
@@ -302,7 +310,7 @@ func writeUniqueFileInfo(ctx context.Context, disks []StorageAPI, bucket, prefix
 			}
 			// Pick one FileInfo for a disk at index.
 			files[index].Erasure.Index = index + 1
-			return disks[index].WriteMetadata(bucket, prefix, files[index])
+			return disks[index].WriteMetadata(ctx, bucket, prefix, files[index])
 		}, index)
 	}
 
@@ -323,7 +331,18 @@ func objectQuorumFromMeta(ctx context.Context, er erasureObjects, partsMetaData
 		return 0, 0, err
 	}
 
+	dataBlocks := latestFileInfo.Erasure.DataBlocks
+	parityBlocks := globalStorageClass.GetParityForSC(latestFileInfo.Metadata[xhttp.AmzStorageClass])
+	if parityBlocks == 0 {
+		parityBlocks = dataBlocks
+	}
+
+	writeQuorum := dataBlocks
+	if dataBlocks == parityBlocks {
+		writeQuorum = dataBlocks + 1
+	}
+
 	// Since all the valid erasure code meta updated at the same time are equivalent, pass dataBlocks
 	// from latestFileInfo to get the quorum
-	return latestFileInfo.Erasure.DataBlocks, latestFileInfo.Erasure.DataBlocks + 1, nil
+	return dataBlocks, writeQuorum, nil
 }