Support to store browser config settings (#18631 )

* csp_policy * hsts_seconds * hsts_include_subdomains * hsts_preload * referrer_policy
doc: fix the name of RoleArn param in AssumeRoleWithWebIdentity requests (#18711 )
2026-08-08 07:13:31 +03:00 · 2024-01-01 08:36:33 -08:00 · 2023-12-30 13:25:15 -08:00 · 2023-12-29 23:56:50 -08:00 · 2023-12-29 15:52:41 -08:00 · 2023-12-29 09:30:10 -08:00
299 changed files with 31277 additions and 6265 deletions
@@ -1,6 +1,6 @@
 ## Community Contribution License
 All community contributions in this pull request are licensed to the project maintainers
-under the terms of the [Apache 2 license] (https://www.apache.org/licenses/LICENSE-2.0). 
+under the terms of the [Apache 2 license](https://www.apache.org/licenses/LICENSE-2.0). 
 By creating this pull request I represent that I have the right to license the 
 contributions to the project maintainers under the Apache 2 license.

@@ -16,10 +16,10 @@ x-minio-common: &minio-common
    MINIO_COMPRESSION_ALLOW_ENCRYPTION: "on"
    MINIO_KMS_SECRET_KEY: "my-minio-key:OSMM+vkKUTCvQs9YL/CVMIMt43HFhkUpqJxTmGl6rYw="
  healthcheck:
-    test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
-    interval: 30s
-    timeout: 20s
-    retries: 3
+    test: ["CMD", "mc", "ready", "local"]
+    interval: 5s
+    timeout: 5s
+    retries: 5

 # starts 4 docker containers running minio server instances.
 # using nginx reverse proxy, load balancing, you can access
@@ -13,10 +13,10 @@ x-minio-common: &minio-common
    MINIO_ROOT_PASSWORD: "minio123"
    MINIO_KMS_SECRET_KEY: "my-minio-key:OSMM+vkKUTCvQs9YL/CVMIMt43HFhkUpqJxTmGl6rYw="
  healthcheck:
-    test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
-    interval: 30s
-    timeout: 20s
-    retries: 3
+    test: ["CMD", "mc", "ready", "local"]
+    interval: 5s
+    timeout: 5s
+    retries: 5

 # starts 4 docker containers running minio server instances.
 # using nginx reverse proxy, load balancing, you can access
@@ -13,10 +13,10 @@ x-minio-common: &minio-common
    MINIO_ROOT_PASSWORD: "minio123"
    MINIO_KMS_SECRET_KEY: "my-minio-key:OSMM+vkKUTCvQs9YL/CVMIMt43HFhkUpqJxTmGl6rYw="
  healthcheck:
-    test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
-    interval: 30s
-    timeout: 20s
-    retries: 3
+    test: ["CMD", "mc", "ready", "local"]
+    interval: 5s
+    timeout: 5s
+    retries: 5

 # starts 4 docker containers running minio server instances.
 # using nginx reverse proxy, load balancing, you can access
@@ -5,11 +5,19 @@ set -x
 ## change working directory
 cd .github/workflows/multipart/

-docker-compose -f docker-compose-site1.yaml rm -s -f
-docker-compose -f docker-compose-site2.yaml rm -s -f
-for volume in $(docker volume ls -q | grep minio); do
-	docker volume rm ${volume}
-done
+function cleanup() {
+	docker-compose -f docker-compose-site1.yaml rm -s -f || true
+	docker-compose -f docker-compose-site2.yaml rm -s -f || true
+	for volume in $(docker volume ls -q | grep minio); do
+		docker volume rm ${volume} || true
+	done
+
+	docker system prune -f || true
+	docker volume prune -f || true
+	docker volume rm $(docker volume ls -q -f dangling=true) || true
+}
+
+cleanup

 if [ ! -f ./mc ]; then
 	wget --quiet -O mc https://dl.minio.io/client/mc/release/linux-amd64/mc &&
@@ -101,15 +109,7 @@ if [ $failed_count_site2 -ne 0 ]; then
 	exit 1
 fi

-docker-compose -f docker-compose-site1.yaml rm -s -f
-docker-compose -f docker-compose-site2.yaml rm -s -f
-for volume in $(docker volume ls -q | grep minio); do
-	docker volume rm ${volume}
-done
-
-docker system prune -f || true
-docker volume prune -f || true
-docker volume rm $(docker volume ls -q -f dangling=true) || true
+cleanup

 ## change working directory
 cd ../../../
@@ -36,6 +36,12 @@ jobs:
          sudo sysctl net.ipv6.conf.default.disable_ipv6=0
          make test-decom

+      - name: Test Config File
+        run: |
+          sudo sysctl net.ipv6.conf.all.disable_ipv6=0
+          sudo sysctl net.ipv6.conf.default.disable_ipv6=0
+          make test-configfile
+
      - name: Test Replication
        run: |
          sudo sysctl net.ipv6.conf.all.disable_ipv6=0
@@ -3,11 +3,10 @@ on:
  pull_request:
    branches:
      - master
-      - next
+
  push:
    branches:
      - master
-      - next

 permissions:
  contents: read # to fetch code (actions/checkout)
@@ -22,7 +21,7 @@ jobs:
      - name: Set up Go
        uses: actions/setup-go@v3
        with:
-          go-version: 1.21.3
+          go-version: 1.21.5
          check-latest: true
      - name: Get official govulncheck
        run: go install golang.org/x/vuln/cmd/govulncheck@latest
@@ -42,3 +42,5 @@ docs/debugging/healing-bin/healing-bin
 docs/debugging/inspect/inspect
 docs/debugging/pprofgoparser/pprofgoparser
 docs/debugging/reorder-disks/reorder-disks
+docs/debugging/populate-hard-links/populate-hardlinks
+docs/debugging/xattr/xattr
@@ -29,5 +29,8 @@ linters:
 issues:
  exclude-use-default: false
  exclude:
+    - "empty-block:"
+    - "unused-parameter:"
+    - "dot-imports:"
    - should have a package comment
    - error strings should not be capitalized or end with punctuation or a newline
@@ -9,7 +9,7 @@ ENV CGO_ENABLED 0
 # Install curl and minisign
 RUN apk add -U --no-cache ca-certificates && \
    apk add -U --no-cache curl && \
-    go install aead.dev/minisign/cmd/minisign@v0.2.0
+    go install aead.dev/minisign/cmd/minisign@v0.2.1

 # Download minio binary and signature file
 RUN curl -s -q https://dl.min.io/server/minio/hotfixes/linux-${TARGETARCH}/archive/minio.${RELEASE} -o /go/bin/minio && \
@@ -25,7 +25,7 @@ RUN curl -s -q https://dl.min.io/client/mc/release/linux-${TARGETARCH}/mc -o /go
 RUN minisign -Vqm /go/bin/minio -x /go/bin/minio.minisig -P RWTx5Zr1tiHQLwG9keckT0c45M3AGeHD6IvimQHpyRywVWGbP1aVSGav && \
    minisign -Vqm /go/bin/mc -x /go/bin/mc.minisig -P RWTx5Zr1tiHQLwG9keckT0c45M3AGeHD6IvimQHpyRywVWGbP1aVSGav

-FROM registry.access.redhat.com/ubi9/ubi-micro:9.2
+FROM registry.access.redhat.com/ubi9/ubi-micro:latest

 ARG RELEASE

@@ -9,7 +9,7 @@ ENV CGO_ENABLED 0
 # Install curl and minisign
 RUN apk add -U --no-cache ca-certificates && \
    apk add -U --no-cache curl && \
-    go install aead.dev/minisign/cmd/minisign@v0.2.0
+    go install aead.dev/minisign/cmd/minisign@v0.2.1

 # Download minio binary and signature file
 RUN curl -s -q https://dl.min.io/server/minio/release/linux-${TARGETARCH}/archive/minio.${RELEASE} -o /go/bin/minio && \
@@ -9,7 +9,7 @@ ENV CGO_ENABLED 0
 # Install curl and minisign
 RUN apk add -U --no-cache ca-certificates && \
    apk add -U --no-cache curl && \
-    go install aead.dev/minisign/cmd/minisign@v0.2.0
+    go install aead.dev/minisign/cmd/minisign@v0.2.1

 # Download minio binary and signature file
 RUN curl -s -q https://dl.min.io/server/minio/release/linux-${TARGETARCH}/archive/minio.${RELEASE}.fips -o /go/bin/minio && \
@@ -19,7 +19,7 @@ RUN curl -s -q https://dl.min.io/server/minio/release/linux-${TARGETARCH}/archiv
 # Verify binary signature using public key "RWTx5Zr1tiHQLwG9keckT0c45M3AGeHD6IvimQHpyRywVWGbP1aVSGavRUN"
 RUN minisign -Vqm /go/bin/minio -x /go/bin/minio.minisig -P RWTx5Zr1tiHQLwG9keckT0c45M3AGeHD6IvimQHpyRywVWGbP1aVSGav

-FROM registry.access.redhat.com/ubi9/ubi-micro:9.2
+FROM registry.access.redhat.com/ubi9/ubi-micro:latest

 ARG RELEASE

@@ -9,7 +9,7 @@ ENV CGO_ENABLED 0
 # Install curl and minisign
 RUN apk add -U --no-cache ca-certificates && \
    apk add -U --no-cache curl && \
-    go install aead.dev/minisign/cmd/minisign@v0.2.0
+    go install aead.dev/minisign/cmd/minisign@v0.2.1

 # Download minio binary and signature file
 RUN curl -s -q https://dl.min.io/server/minio/release/linux-${TARGETARCH}/archive/minio.${RELEASE} -o /go/bin/minio && \
@@ -8,7 +8,6 @@ GOOS := $(shell go env GOOS)
 VERSION ?= $(shell git describe --tags)
 TAG ?= "quay.io/minio/minio:$(VERSION)"

-GOLANGCI_VERSION = v1.51.2
 GOLANGCI_DIR = .bin/golangci/$(GOLANGCI_VERSION)
 GOLANGCI = $(GOLANGCI_DIR)/golangci-lint

@@ -23,8 +22,8 @@ help: ## print this help

 getdeps: ## fetch necessary dependencies
 	@mkdir -p ${GOPATH}/bin
-	@echo "Installing golangci-lint" && curl -sSfL https://raw.githubusercontent.com/golangci/golangci-lint/master/install.sh | sh -s -- -b $(GOLANGCI_DIR) $(GOLANGCI_VERSION)
-	@echo "Installing msgp" && go install -v github.com/tinylib/msgp@v1.1.7
+	@echo "Installing golangci-lint" && curl -sSfL https://raw.githubusercontent.com/golangci/golangci-lint/master/install.sh | sh -s -- -b $(GOLANGCI_DIR)
+	@echo "Installing msgp" && go install -v github.com/tinylib/msgp@6ac204f0b4d48d17ab4fa442134c7fba13127a4e
 	@echo "Installing stringer" && go install -v golang.org/x/tools/cmd/stringer@latest

 crosscompile: ## cross compile minio
@@ -60,6 +59,9 @@ test-decom: install-race
 	@env bash $(PWD)/docs/distributed/decom-encrypted-sse-s3.sh
 	@env bash $(PWD)/docs/distributed/decom-compressed-sse-s3.sh

+test-configfile: install-race
+	@env bash $(PWD)/docs/distributed/distributed-from-config-file.sh
+
 test-upgrade: install-race
 	@echo "Running minio upgrade tests"
 	@(env bash $(PWD)/buildscripts/minio-upgrade.sh)
@@ -141,14 +143,23 @@ hotfix-vars:
 	$(eval VERSION := $(shell git describe --tags --abbrev=0).hotfix.$(shell git rev-parse --short HEAD))
 	$(eval TAG := "minio/minio:$(VERSION)")

-hotfix: hotfix-vars install ## builds minio binary with hotfix tags
-	@mv -f ./minio ./minio.$(VERSION)
-	@minisign -qQSm ./minio.$(VERSION) -s "${CRED_DIR}/minisign.key" < "${CRED_DIR}/minisign-passphrase"
-	@sha256sum < ./minio.$(VERSION) | sed 's, -,minio.$(VERSION),g' > minio.$(VERSION).sha256sum
+hotfix: hotfix-vars clean install ## builds minio binary with hotfix tags
+	@wget -q -c https://github.com/minio/pkger/releases/download/v2.2.0/pkger_2.2.0_linux_amd64.deb
+	@wget -q -c https://raw.githubusercontent.com/minio/minio-service/v1.0.0/linux-systemd/distributed/minio.service
+	@sudo apt install ./pkger_2.2.0_linux_amd64.deb --yes
+	@mkdir -p minio-release/$(GOOS)-$(GOARCH)/archive
+	@cp -af ./minio minio-release/$(GOOS)-$(GOARCH)/minio
+	@cp -af ./minio minio-release/$(GOOS)-$(GOARCH)/minio.$(VERSION)
+	@minisign -qQSm minio-release/$(GOOS)-$(GOARCH)/minio.$(VERSION) -s "${CRED_DIR}/minisign.key" < "${CRED_DIR}/minisign-passphrase"
+	@sha256sum < minio-release/$(GOOS)-$(GOARCH)/minio.$(VERSION) | sed 's, -,minio.$(VERSION),g' > minio-release/$(GOOS)-$(GOARCH)/minio.$(VERSION).sha256sum
+	@cp -af minio-release/$(GOOS)-$(GOARCH)/minio.$(VERSION)* minio-release/$(GOOS)-$(GOARCH)/archive/
+	@pkger -r $(VERSION) --ignore

 hotfix-push: hotfix
-	@scp -q -r minio.$(VERSION)* minio@dl-0.minio.io:~/releases/server/minio/hotfixes/linux-amd64/archive/
-	@scp -q -r minio.$(VERSION)* minio@dl-1.minio.io:~/releases/server/minio/hotfixes/linux-amd64/archive/
+	@scp -q -r minio-release/$(GOOS)-$(GOARCH)/* minio@dl-0.minio.io:~/releases/server/minio/hotfixes/linux-amd64/
+	@scp -q -r minio-release/$(GOOS)-$(GOARCH)/* minio@dl-0.minio.io:~/releases/server/minio/hotfixes/linux-amd64/archive
+	@scp -q -r minio-release/$(GOOS)-$(GOARCH)/* minio@dl-1.minio.io:~/releases/server/minio/hotfixes/linux-amd64/
+	@scp -q -r minio-release/$(GOOS)-$(GOARCH)/* minio@dl-1.minio.io:~/releases/server/minio/hotfixes/linux-amd64/archive
 	@echo "Published new hotfix binaries at https://dl.min.io/server/minio/hotfixes/linux-amd64/archive/minio.$(VERSION)"

 docker-hotfix-push: docker-hotfix
@@ -183,3 +194,6 @@ clean: ## cleanup all generated assets
 	@rm -rvf build
 	@rm -rvf release
 	@rm -rvf .verify*
+	@rm -rvf minio-release
+	@rm -rvf minio.RELEASE*.hotfix.*
+	@rm -rvf pkger_*.deb
@@ -9,7 +9,7 @@ function _init() {
 	export CGO_ENABLED=0

 	## List of architectures and OS to test coss compilation.
-	SUPPORTED_OSARCH="linux/ppc64le linux/mips64 linux/amd64 linux/arm64 linux/s390x darwin/arm64 darwin/amd64 freebsd/amd64 windows/amd64 linux/arm linux/386 netbsd/amd64 linux/mips"
+	SUPPORTED_OSARCH="linux/ppc64le linux/mips64 linux/amd64 linux/arm64 linux/s390x darwin/arm64 darwin/amd64 freebsd/amd64 windows/amd64 linux/arm linux/386 netbsd/amd64 linux/mips openbsd/amd64"
 }

 function _build() {
@@ -9,11 +9,6 @@ x-minio-common: &minio-common
  expose:
    - "9000"
    - "9001"
-  healthcheck:
-    test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
-    interval: 30s
-    timeout: 20s
-    retries: 3

 # starts 4 docker containers running minio server instances.
 # using nginx reverse proxy, load balancing, you can access
@@ -94,7 +94,7 @@ func (a adminAPIHandlers) PutBucketQuotaConfigHandler(w http.ResponseWriter, r *
 		Quota:     data,
 		UpdatedAt: updatedAt,
 	}
-	if quotaConfig.Size == 0 || quotaConfig.Quota == 0 {
+	if quotaConfig.Size == 0 && quotaConfig.Quota == 0 {
 		bucketMeta.Quota = nil
 	}

@@ -19,10 +19,14 @@ package cmd

 import (
 	"encoding/json"
+	"errors"
+	"fmt"
 	"io"
 	"net/http"
+	"strings"

 	"github.com/minio/madmin-go/v3"
+	"github.com/minio/minio/internal/auth"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/mux"
 	"github.com/minio/pkg/v2/policy"
@@ -175,3 +179,267 @@ func (a adminAPIHandlers) AttachDetachPolicyLDAP(w http.ResponseWriter, r *http.

 	writeSuccessResponseJSON(w, encryptedData)
 }
+
+// AddServiceAccountLDAP adds a new service account for provided LDAP username or DN
+//
+// PUT /minio/admin/v3/idp/ldap/add-service-account
+func (a adminAPIHandlers) AddServiceAccountLDAP(w http.ResponseWriter, r *http.Request) {
+	ctx, cred, opts, createReq, targetUser, APIError := commonAddServiceAccount(r)
+	if APIError.Code != "" {
+		writeErrorResponseJSON(ctx, w, APIError, r.URL)
+		return
+	}
+
+	// fail if ldap is not enabled
+	if !globalIAMSys.LDAPConfig.Enabled() {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errors.New("LDAP not enabled")), r.URL)
+	}
+
+	// Find the user for the request sender (as it may be sent via a service
+	// account or STS account):
+	requestorUser := cred.AccessKey
+	requestorParentUser := cred.AccessKey
+	requestorGroups := cred.Groups
+	requestorIsDerivedCredential := false
+	if cred.IsServiceAccount() || cred.IsTemp() {
+		requestorParentUser = cred.ParentUser
+		requestorIsDerivedCredential = true
+	}
+
+	// Check if we are creating svc account for request sender.
+	isSvcAccForRequestor := false
+	if targetUser == requestorUser || targetUser == requestorParentUser {
+		isSvcAccForRequestor = true
+	}
+
+	var (
+		targetGroups []string
+		err          error
+	)
+
+	// If we are creating svc account for request sender, ensure
+	// that targetUser is a real user (i.e. not derived
+	// credentials).
+	if isSvcAccForRequestor {
+		if requestorIsDerivedCredential {
+			if requestorParentUser == "" {
+				writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx,
+					errors.New("service accounts cannot be generated for temporary credentials without parent")), r.URL)
+				return
+			}
+			targetUser = requestorParentUser
+		}
+		targetGroups = requestorGroups
+
+		// Deny if the target user is not LDAP
+		isLDAP, err := globalIAMSys.LDAPConfig.DoesUsernameExist(targetUser)
+		if err != nil {
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+		if isLDAP == "" {
+			err := errors.New("Specified user does not exist on LDAP server")
+			APIErr := errorCodes.ToAPIErrWithErr(ErrAdminNoSuchUser, err)
+			writeErrorResponseJSON(ctx, w, APIErr, r.URL)
+			return
+		}
+
+		// In case of LDAP/OIDC we need to set `opts.claims` to ensure
+		// it is associated with the LDAP/OIDC user properly.
+		for k, v := range cred.Claims {
+			if k == expClaim {
+				continue
+			}
+			opts.claims[k] = v
+		}
+	} else {
+		isDN := globalIAMSys.LDAPConfig.IsLDAPUserDN(targetUser)
+
+		opts.claims[ldapUserN] = targetUser // simple username
+		targetUser, targetGroups, err = globalIAMSys.LDAPConfig.LookupUserDN(targetUser)
+		if err != nil {
+			// if not found, check if DN
+			if strings.Contains(err.Error(), "not found") && isDN {
+				// warn user that DNs are not allowed
+				err = fmt.Errorf("Must use short username to add service account. %w", err)
+			}
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+		opts.claims[ldapUser] = targetUser // DN
+	}
+
+	newCred, updatedAt, err := globalIAMSys.NewServiceAccount(ctx, targetUser, targetGroups, opts)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	createResp := madmin.AddServiceAccountResp{
+		Credentials: madmin.Credentials{
+			AccessKey:  newCred.AccessKey,
+			SecretKey:  newCred.SecretKey,
+			Expiration: newCred.Expiration,
+		},
+	}
+
+	data, err := json.Marshal(createResp)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	encryptedData, err := madmin.EncryptData(cred.SecretKey, data)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	writeSuccessResponseJSON(w, encryptedData)
+
+	// Call hook for cluster-replication if the service account is not for a
+	// root user.
+	if newCred.ParentUser != globalActiveCred.AccessKey {
+		logger.LogIf(ctx, globalSiteReplicationSys.IAMChangeHook(ctx, madmin.SRIAMItem{
+			Type: madmin.SRIAMItemSvcAcc,
+			SvcAccChange: &madmin.SRSvcAccChange{
+				Create: &madmin.SRSvcAccCreate{
+					Parent:        newCred.ParentUser,
+					AccessKey:     newCred.AccessKey,
+					SecretKey:     newCred.SecretKey,
+					Groups:        newCred.Groups,
+					Name:          newCred.Name,
+					Description:   newCred.Description,
+					Claims:        opts.claims,
+					SessionPolicy: createReq.Policy,
+					Status:        auth.AccountOn,
+					Expiration:    createReq.Expiration,
+				},
+			},
+			UpdatedAt: updatedAt,
+		}))
+	}
+}
+
+// ListAccessKeysLDAP - GET /minio/admin/v3/idp/ldap/list-access-keys
+func (a adminAPIHandlers) ListAccessKeysLDAP(w http.ResponseWriter, r *http.Request) {
+	ctx := r.Context()
+
+	// Get current object layer instance.
+	objectAPI := newObjectLayerFn()
+	if objectAPI == nil || globalNotificationSys == nil {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
+		return
+	}
+
+	cred, owner, s3Err := validateAdminSignature(ctx, r, "")
+	if s3Err != ErrNone {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(s3Err), r.URL)
+		return
+	}
+
+	userDN := r.Form.Get("userDN")
+
+	// If listing is requested for a specific user (who is not the request
+	// sender), check that the user has permissions.
+	if userDN != "" && userDN != cred.ParentUser {
+		if !globalIAMSys.IsAllowed(policy.Args{
+			AccountName:     cred.AccessKey,
+			Groups:          cred.Groups,
+			Action:          policy.ListServiceAccountsAdminAction,
+			ConditionValues: getConditionValues(r, "", cred),
+			IsOwner:         owner,
+			Claims:          cred.Claims,
+		}) {
+			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAccessDenied), r.URL)
+			return
+		}
+	} else {
+		if !globalIAMSys.IsAllowed(policy.Args{
+			AccountName:     cred.AccessKey,
+			Groups:          cred.Groups,
+			Action:          policy.ListServiceAccountsAdminAction,
+			ConditionValues: getConditionValues(r, "", cred),
+			IsOwner:         owner,
+			Claims:          cred.Claims,
+			DenyOnly:        true,
+		}) {
+			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAccessDenied), r.URL)
+			return
+		}
+		userDN = cred.AccessKey
+		if cred.ParentUser != "" {
+			userDN = cred.ParentUser
+		}
+	}
+
+	targetAccount, err := globalIAMSys.LDAPConfig.DoesUsernameExist(userDN)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	} else if userDN == "" {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errNoSuchUser), r.URL)
+		return
+	}
+
+	listType := r.Form.Get("listType")
+	if listType != "sts-only" && listType != "svcacc-only" && listType != "" {
+		// default to both
+		listType = ""
+	}
+
+	var serviceAccounts []auth.Credentials
+	var stsKeys []auth.Credentials
+
+	if listType == "" || listType == "sts-only" {
+		stsKeys, err = globalIAMSys.ListSTSAccounts(ctx, targetAccount)
+		if err != nil {
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+	}
+	if listType == "" || listType == "svcacc-only" {
+		serviceAccounts, err = globalIAMSys.ListServiceAccounts(ctx, targetAccount)
+		if err != nil {
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+	}
+
+	var serviceAccountList []madmin.ServiceAccountInfo
+	var stsKeyList []madmin.ServiceAccountInfo
+
+	for _, svc := range serviceAccounts {
+		expiryTime := svc.Expiration
+		serviceAccountList = append(serviceAccountList, madmin.ServiceAccountInfo{
+			AccessKey:  svc.AccessKey,
+			Expiration: &expiryTime,
+		})
+	}
+	for _, sts := range stsKeys {
+		expiryTime := sts.Expiration
+		stsKeyList = append(stsKeyList, madmin.ServiceAccountInfo{
+			AccessKey:  sts.AccessKey,
+			Expiration: &expiryTime,
+		})
+	}
+
+	listResp := madmin.ListAccessKeysLDAPResp{
+		ServiceAccounts: serviceAccountList,
+		STSKeys:         stsKeyList,
+	}
+
+	data, err := json.Marshal(listResp)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	encryptedData, err := madmin.EncryptData(cred.SecretKey, data)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	writeSuccessResponseJSON(w, encryptedData)
+}
@@ -22,6 +22,7 @@ import (
 	"errors"
 	"fmt"
 	"net/http"
+	"strconv"
 	"strings"

 	"github.com/minio/minio/internal/logger"
@@ -66,16 +67,28 @@ func (a adminAPIHandlers) StartDecommission(w http.ResponseWriter, r *http.Reque

 	vars := mux.Vars(r)
 	v := vars["pool"]
+	byID := vars["by-id"] == "true"

 	pools := strings.Split(v, ",")
 	poolIndices := make([]int, 0, len(pools))

 	for _, pool := range pools {
-		idx := globalEndpoints.GetPoolIdx(pool)
-		if idx == -1 {
-			// We didn't find any matching pools, invalid input
-			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errInvalidArgument), r.URL)
-			return
+		var idx int
+		if byID {
+			var err error
+			idx, err = strconv.Atoi(pool)
+			if err != nil {
+				// We didn't find any matching pools, invalid input
+				writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errInvalidArgument), r.URL)
+				return
+			}
+		} else {
+			idx = globalEndpoints.GetPoolIdx(pool)
+			if idx == -1 {
+				// We didn't find any matching pools, invalid input
+				writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errInvalidArgument), r.URL)
+				return
+			}
 		}
 		var pool *erasureSets
 		for pidx := range z.serverPools {
@@ -132,8 +145,17 @@ func (a adminAPIHandlers) CancelDecommission(w http.ResponseWriter, r *http.Requ

 	vars := mux.Vars(r)
 	v := vars["pool"]
+	byID := vars["by-id"] == "true"
+	idx := -1
+
+	if byID {
+		if i, err := strconv.Atoi(v); err == nil && i >= 0 && i < len(globalEndpoints) {
+			idx = i
+		}
+	} else {
+		idx = globalEndpoints.GetPoolIdx(v)
+	}

-	idx := globalEndpoints.GetPoolIdx(v)
 	if idx == -1 {
 		// We didn't find any matching pools, invalid input
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errInvalidArgument), r.URL)
@@ -178,8 +200,17 @@ func (a adminAPIHandlers) StatusPool(w http.ResponseWriter, r *http.Request) {

 	vars := mux.Vars(r)
 	v := vars["pool"]
+	byID := vars["by-id"] == "true"
+	idx := -1
+
+	if byID {
+		if i, err := strconv.Atoi(v); err == nil && i >= 0 && i < len(globalEndpoints) {
+			idx = i
+		}
+	} else {
+		idx = globalEndpoints.GetPoolIdx(v)
+	}

-	idx := globalEndpoints.GetPoolIdx(v)
 	if idx == -1 {
 		apiErr := toAdminAPIErr(ctx, errInvalidArgument)
 		apiErr.Description = fmt.Sprintf("specified pool '%s' not found, please specify a valid pool", v)
@@ -31,6 +31,7 @@ import (

 	"github.com/dustin/go-humanize"
 	"github.com/minio/madmin-go/v3"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/mux"
 	"github.com/minio/pkg/v2/policy"
@@ -51,7 +52,8 @@ func (a adminAPIHandlers) SiteReplicationAdd(w http.ResponseWriter, r *http.Requ
 		return
 	}

-	status, err := globalSiteReplicationSys.AddPeerClusters(ctx, sites)
+	opts := getSRAddOptions(r)
+	status, err := globalSiteReplicationSys.AddPeerClusters(ctx, sites, opts)
 	if err != nil {
 		logger.LogIf(ctx, err)
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
@@ -67,6 +69,11 @@ func (a adminAPIHandlers) SiteReplicationAdd(w http.ResponseWriter, r *http.Requ
 	writeSuccessResponseJSON(w, body)
 }

+func getSRAddOptions(r *http.Request) (opts madmin.SRAddOptions) {
+	opts.ReplicateILMExpiry = r.Form.Get("replicateILMExpiry") == "true"
+	return
+}
+
 // SRPeerJoin - PUT /minio/admin/v3/site-replication/join
 //
 // used internally to tell current cluster to enable SR with
@@ -191,7 +198,7 @@ func (a adminAPIHandlers) SRPeerReplicateIAMItem(w http.ResponseWriter, r *http.
 	}
 }

-// SRPeerReplicateBucketItem - PUT /minio/admin/v3/site-replication/bucket-meta
+// SRPeerReplicateBucketItem - PUT /minio/admin/v3/site-replication/peer/bucket-meta
 func (a adminAPIHandlers) SRPeerReplicateBucketItem(w http.ResponseWriter, r *http.Request) {
 	ctx := r.Context()

@@ -252,6 +259,8 @@ func (a adminAPIHandlers) SRPeerReplicateBucketItem(w http.ResponseWriter, r *ht
 		err = globalSiteReplicationSys.PeerBucketObjectLockConfigHandler(ctx, item.Bucket, item.ObjectLockConfig, item.UpdatedAt)
 	case madmin.SRBucketMetaTypeSSEConfig:
 		err = globalSiteReplicationSys.PeerBucketSSEConfigHandler(ctx, item.Bucket, item.SSEConfig, item.UpdatedAt)
+	case madmin.SRBucketMetaLCConfig:
+		err = globalSiteReplicationSys.PeerBucketLCConfigHandler(ctx, item.Bucket, item.ExpiryLCConfig, item.UpdatedAt)
 	}
 	if err != nil {
 		logger.LogIf(ctx, err)
@@ -333,6 +342,7 @@ func (a adminAPIHandlers) SiteReplicationStatus(w http.ResponseWriter, r *http.R
 		opts.Users = true
 		opts.Policies = true
 		opts.Groups = true
+		opts.ILMExpiryRules = true
 	}
 	info, err := globalSiteReplicationSys.SiteReplicationStatus(ctx, objectAPI, opts)
 	if err != nil {
@@ -382,7 +392,9 @@ func (a adminAPIHandlers) SiteReplicationEdit(w http.ResponseWriter, r *http.Req
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
 	}
-	status, err := globalSiteReplicationSys.EditPeerCluster(ctx, site)
+
+	opts := getSREditOptions(r)
+	status, err := globalSiteReplicationSys.EditPeerCluster(ctx, site, opts)
 	if err != nil {
 		logger.LogIf(ctx, err)
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
@@ -397,6 +409,12 @@ func (a adminAPIHandlers) SiteReplicationEdit(w http.ResponseWriter, r *http.Req
 	writeSuccessResponseJSON(w, body)
 }

+func getSREditOptions(r *http.Request) (opts madmin.SREditOptions) {
+	opts.DisableILMExpiryReplication = r.Form.Get("disableILMExpiryReplication") == "true"
+	opts.EnableILMExpiryReplication = r.Form.Get("enableILMExpiryReplication") == "true"
+	return
+}
+
 // SRPeerEdit - PUT /minio/admin/v3/site-replication/peer/edit
 //
 // used internally to tell current cluster to update endpoint for peer
@@ -421,12 +439,37 @@ func (a adminAPIHandlers) SRPeerEdit(w http.ResponseWriter, r *http.Request) {
 	}
 }

+// SRStateEdit - PUT /minio/admin/v3/site-replication/state/edit
+//
+// used internally to tell current cluster to update site replication state
+func (a adminAPIHandlers) SRStateEdit(w http.ResponseWriter, r *http.Request) {
+	ctx := r.Context()
+
+	objectAPI, _ := validateAdminReq(ctx, w, r, policy.SiteReplicationOperationAction)
+	if objectAPI == nil {
+		return
+	}
+
+	var state madmin.SRStateEditReq
+	if err := parseJSONBody(ctx, r.Body, &state, ""); err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+	if err := globalSiteReplicationSys.PeerStateEditReq(ctx, state); err != nil {
+		logger.LogIf(ctx, err)
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+}
+
 func getSRStatusOptions(r *http.Request) (opts madmin.SRStatusOptions) {
 	q := r.Form
 	opts.Buckets = q.Get("buckets") == "true"
 	opts.Policies = q.Get("policies") == "true"
 	opts.Groups = q.Get("groups") == "true"
 	opts.Users = q.Get("users") == "true"
+	opts.ILMExpiryRules = q.Get("ilm-expiry-rules") == "true"
+	opts.PeerState = q.Get("peer-state") == "true"
 	opts.Entity = madmin.GetSREntityType(q.Get("entity"))
 	opts.EntityValue = q.Get("entityvalue")
 	opts.ShowDeleted = q.Get("showDeleted") == "true"
@@ -537,7 +580,7 @@ func (a adminAPIHandlers) SiteReplicationDevNull(w http.ResponseWriter, r *http.

 	connectTime := time.Now()
 	for {
-		n, err := io.CopyN(io.Discard, r.Body, 128*humanize.KiByte)
+		n, err := io.CopyN(xioutil.Discard, r.Body, 128*humanize.KiByte)
 		atomic.AddUint64(&globalSiteNetPerfRX.RX, uint64(n))
 		if err != nil && err != io.EOF && err != io.ErrUnexpectedEOF {
 			// If there is a disconnection before globalNetPerfMinDuration (we give a margin of error of 1 sec)
@@ -545,14 +545,16 @@ func (a adminAPIHandlers) TemporaryAccountInfo(w http.ResponseWriter, r *http.Re
 		return
 	}

-	if !globalIAMSys.IsAllowed(policy.Args{
+	args := policy.Args{
 		AccountName:     cred.AccessKey,
 		Groups:          cred.Groups,
 		Action:          policy.ListTemporaryAccountsAdminAction,
 		ConditionValues: getConditionValues(r, "", cred),
 		IsOwner:         owner,
 		Claims:          cred.Claims,
-	}) {
+	}
+
+	if !globalIAMSys.IsAllowed(args) {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAccessDenied), r.URL)
 		return
 	}
@@ -568,11 +570,16 @@ func (a adminAPIHandlers) TemporaryAccountInfo(w http.ResponseWriter, r *http.Re
 	if sessionPolicy != nil {
 		stsAccountPolicy = *sessionPolicy
 	} else {
-		policiesNames, err := globalIAMSys.PolicyDBGet(stsAccount.ParentUser, false)
+		policiesNames, err := globalIAMSys.PolicyDBGet(stsAccount.ParentUser, stsAccount.Groups...)
 		if err != nil {
 			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 			return
 		}
+		if len(policiesNames) == 0 {
+			policySet, _ := args.GetPolicies(iamPolicyClaimNameOpenID())
+			policiesNames = policySet.ToSlice()
+		}
+
 		stsAccountPolicy = globalIAMSys.GetCombinedPolicy(policiesNames...)
 	}

@@ -607,72 +614,17 @@ func (a adminAPIHandlers) TemporaryAccountInfo(w http.ResponseWriter, r *http.Re

 // AddServiceAccount - PUT /minio/admin/v3/add-service-account
 func (a adminAPIHandlers) AddServiceAccount(w http.ResponseWriter, r *http.Request) {
-	ctx := r.Context()
-
-	// Get current object layer instance.
-	objectAPI := newObjectLayerFn()
-	if objectAPI == nil || globalNotificationSys == nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrServerNotInitialized), r.URL)
-		return
-	}
-
-	cred, owner, 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))
-	if err != nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err), r.URL)
-		return
-	}
-
-	var createReq madmin.AddServiceAccountReq
-	if err = json.Unmarshal(reqBytes, &createReq); err != nil {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err), r.URL)
-		return
-	}
-
-	// service account access key cannot have space characters beginning and end of the string.
-	if hasSpaceBE(createReq.AccessKey) {
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAdminResourceInvalidArgument), r.URL)
-		return
-	}
-
-	if err := createReq.Validate(); err != nil {
-		// Since this validation would happen client side as well, we only send
-		// a generic error message here.
-		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAdminResourceInvalidArgument), r.URL)
+	ctx, cred, opts, createReq, targetUser, APIError := commonAddServiceAccount(r)
+	if APIError.Code != "" {
+		writeErrorResponseJSON(ctx, w, APIError, r.URL)
 		return
 	}

 	var (
-		targetUser   string
 		targetGroups []string
+		err          error
 	)

-	// If the request did not set a TargetUser, the service account is
-	// created for the request sender.
-	targetUser = createReq.TargetUser
-	if targetUser == "" {
-		targetUser = cred.AccessKey
-	}
-
-	description := createReq.Description
-	if description == "" {
-		description = createReq.Comment
-	}
-	opts := newServiceAccountOpts{
-		accessKey:   createReq.AccessKey,
-		secretKey:   createReq.SecretKey,
-		name:        createReq.Name,
-		description: description,
-		expiration:  createReq.Expiration,
-		claims:      make(map[string]interface{}),
-	}
-
 	// Find the user for the request sender (as it may be sent via a service
 	// account or STS account):
 	requestorUser := cred.AccessKey
@@ -706,23 +658,6 @@ func (a adminAPIHandlers) AddServiceAccount(w http.ResponseWriter, r *http.Reque
 	// that targetUser is a real user (i.e. not derived
 	// credentials).
 	if isSvcAccForRequestor {
-		// Check if adding service account is explicitly denied.
-		//
-		// This allows turning off service accounts for request sender,
-		// if there is no deny statement this call is implicitly enabled.
-		if !globalIAMSys.IsAllowed(policy.Args{
-			AccountName:     requestorUser,
-			Groups:          requestorGroups,
-			Action:          policy.CreateServiceAccountAdminAction,
-			ConditionValues: getConditionValues(r, "", cred),
-			IsOwner:         owner,
-			Claims:          cred.Claims,
-			DenyOnly:        true,
-		}) {
-			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAccessDenied), r.URL)
-			return
-		}
-
 		if requestorIsDerivedCredential {
 			if requestorParentUser == "" {
 				writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx,
@@ -741,32 +676,16 @@ func (a adminAPIHandlers) AddServiceAccount(w http.ResponseWriter, r *http.Reque
 			}
 			opts.claims[k] = v
 		}
-	} else {
-		// Need permission if we are creating a service account for a
-		// user <> to the request sender
-		if !globalIAMSys.IsAllowed(policy.Args{
-			AccountName:     requestorUser,
-			Groups:          requestorGroups,
-			Action:          policy.CreateServiceAccountAdminAction,
-			ConditionValues: getConditionValues(r, "", cred),
-			IsOwner:         owner,
-			Claims:          cred.Claims,
-		}) {
-			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAccessDenied), r.URL)
-			return
-		}
-
+	} else if globalIAMSys.LDAPConfig.Enabled() {
 		// In case of LDAP we need to resolve the targetUser to a DN and
 		// query their groups:
-		if globalIAMSys.LDAPConfig.Enabled() {
-			opts.claims[ldapUserN] = targetUser // simple username
-			targetUser, targetGroups, err = globalIAMSys.LDAPConfig.LookupUserDN(targetUser)
-			if err != nil {
-				writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
-				return
-			}
-			opts.claims[ldapUser] = targetUser // username DN
+		opts.claims[ldapUserN] = targetUser // simple username
+		targetUser, targetGroups, err = globalIAMSys.LDAPConfig.LookupUserDN(targetUser)
+		if err != nil {
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
 		}
+		opts.claims[ldapUser] = targetUser // username DN

 		// NOTE: if not using LDAP, then internal IDP or open ID is
 		// being used - in the former, group info is enforced when
@@ -774,19 +693,6 @@ func (a adminAPIHandlers) AddServiceAccount(w http.ResponseWriter, r *http.Reque
 		// latter, a group notion is not supported.
 	}

-	var sp *policy.Policy
-	if len(createReq.Policy) > 0 {
-		sp, err = policy.ParseConfig(bytes.NewReader(createReq.Policy))
-		if err != nil {
-			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
-			return
-		}
-		if sp.Version == "" && len(sp.Statements) == 0 {
-			sp = nil
-		}
-	}
-
-	opts.sessionPolicy = sp
 	newCred, updatedAt, err := globalIAMSys.NewServiceAccount(ctx, targetUser, targetGroups, opts)
 	if err != nil {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
@@ -807,7 +713,7 @@ func (a adminAPIHandlers) AddServiceAccount(w http.ResponseWriter, r *http.Reque
 		return
 	}

-	encryptedData, err := madmin.EncryptData(password, data)
+	encryptedData, err := madmin.EncryptData(cred.SecretKey, data)
 	if err != nil {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
@@ -1010,7 +916,7 @@ func (a adminAPIHandlers) InfoServiceAccount(w http.ResponseWriter, r *http.Requ
 	if !impliedPolicy {
 		svcAccountPolicy = *sessionPolicy
 	} else {
-		policiesNames, err := globalIAMSys.PolicyDBGet(svcAccount.ParentUser, false)
+		policiesNames, err := globalIAMSys.PolicyDBGet(svcAccount.ParentUser, svcAccount.Groups...)
 		if err != nil {
 			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 			return
@@ -1357,7 +1263,7 @@ func (a adminAPIHandlers) AccountInfoHandler(w http.ResponseWriter, r *http.Requ
 		effectivePolicy = globalIAMSys.GetCombinedPolicy(policySetFromClaims.ToSlice()...)

 	default:
-		policies, err := globalIAMSys.PolicyDBGet(accountName, false, cred.Groups...)
+		policies, err := globalIAMSys.PolicyDBGet(accountName, cred.Groups...)
 		if err != nil {
 			logger.LogIf(ctx, err)
 			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
@@ -2507,3 +2413,85 @@ func (a adminAPIHandlers) ImportIAM(w http.ResponseWriter, r *http.Request) {
 		}
 	}
 }
+
+func commonAddServiceAccount(r *http.Request) (context.Context, auth.Credentials, newServiceAccountOpts, madmin.AddServiceAccountReq, string, APIError) {
+	ctx := r.Context()
+
+	// Get current object layer instance.
+	objectAPI := newObjectLayerFn()
+	if objectAPI == nil || globalNotificationSys == nil {
+		return ctx, auth.Credentials{}, newServiceAccountOpts{}, madmin.AddServiceAccountReq{}, "", errorCodes.ToAPIErr(ErrServerNotInitialized)
+	}
+
+	cred, owner, s3Err := validateAdminSignature(ctx, r, "")
+	if s3Err != ErrNone {
+		return ctx, auth.Credentials{}, newServiceAccountOpts{}, madmin.AddServiceAccountReq{}, "", errorCodes.ToAPIErr(s3Err)
+	}
+
+	password := cred.SecretKey
+	reqBytes, err := madmin.DecryptData(password, io.LimitReader(r.Body, r.ContentLength))
+	if err != nil {
+		return ctx, auth.Credentials{}, newServiceAccountOpts{}, madmin.AddServiceAccountReq{}, "", errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err)
+	}
+
+	var createReq madmin.AddServiceAccountReq
+	if err = json.Unmarshal(reqBytes, &createReq); err != nil {
+		return ctx, auth.Credentials{}, newServiceAccountOpts{}, madmin.AddServiceAccountReq{}, "", errorCodes.ToAPIErrWithErr(ErrAdminConfigBadJSON, err)
+	}
+
+	// service account access key cannot have space characters beginning and end of the string.
+	if hasSpaceBE(createReq.AccessKey) {
+		return ctx, auth.Credentials{}, newServiceAccountOpts{}, madmin.AddServiceAccountReq{}, "", errorCodes.ToAPIErr(ErrAdminResourceInvalidArgument)
+	}
+
+	if err := createReq.Validate(); err != nil {
+		// Since this validation would happen client side as well, we only send
+		// a generic error message here.
+		return ctx, auth.Credentials{}, newServiceAccountOpts{}, madmin.AddServiceAccountReq{}, "", errorCodes.ToAPIErr(ErrAdminResourceInvalidArgument)
+	}
+	// If the request did not set a TargetUser, the service account is
+	// created for the request sender.
+	targetUser := createReq.TargetUser
+	if targetUser == "" {
+		targetUser = cred.AccessKey
+	}
+
+	description := createReq.Description
+	if description == "" {
+		description = createReq.Comment
+	}
+	opts := newServiceAccountOpts{
+		accessKey:   createReq.AccessKey,
+		secretKey:   createReq.SecretKey,
+		name:        createReq.Name,
+		description: description,
+		expiration:  createReq.Expiration,
+		claims:      make(map[string]interface{}),
+	}
+
+	// Check if action is allowed if creating access key for another user
+	// Check if action is explicitly denied if for self
+	if !globalIAMSys.IsAllowed(policy.Args{
+		AccountName:     cred.AccessKey,
+		Groups:          cred.Groups,
+		Action:          policy.CreateServiceAccountAdminAction,
+		ConditionValues: getConditionValues(r, "", cred),
+		IsOwner:         owner,
+		Claims:          cred.Claims,
+		DenyOnly:        (targetUser == cred.AccessKey || targetUser == cred.ParentUser),
+	}) {
+		return ctx, auth.Credentials{}, newServiceAccountOpts{}, madmin.AddServiceAccountReq{}, "", errorCodes.ToAPIErr(ErrAccessDenied)
+	}
+
+	var sp *policy.Policy
+	if len(createReq.Policy) > 0 {
+		sp, err = policy.ParseConfig(bytes.NewReader(createReq.Policy))
+		if err != nil {
+			return ctx, auth.Credentials{}, newServiceAccountOpts{}, madmin.AddServiceAccountReq{}, "", toAdminAPIErr(ctx, err)
+		}
+	}
+
+	opts.sessionPolicy = sp
+
+	return ctx, cred, opts, createReq, targetUser, APIError{}
+}
@@ -52,6 +52,7 @@ import (
 	"github.com/minio/minio/internal/dsync"
 	"github.com/minio/minio/internal/handlers"
 	xhttp "github.com/minio/minio/internal/http"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/kms"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/mux"
@@ -59,11 +60,14 @@ import (
 	xnet "github.com/minio/pkg/v2/net"
 	"github.com/minio/pkg/v2/policy"
 	"github.com/secure-io/sio-go"
+	"github.com/zeebo/xxh3"
 )

 const (
 	maxEConfigJSONSize        = 262272
 	kubernetesVersionEndpoint = "https://kubernetes.default.svc/version"
+	anonymizeParam            = "anonymize"
+	anonymizeStrict           = "strict"
 )

 // Only valid query params for mgmt admin APIs.
@@ -336,7 +340,7 @@ func (a adminAPIHandlers) StorageInfoHandler(w http.ResponseWriter, r *http.Requ
 		return
 	}

-	storageInfo := objectAPI.StorageInfo(ctx)
+	storageInfo := objectAPI.StorageInfo(ctx, true)

 	// Collect any disk healing.
 	healing, _ := getAggregatedBackgroundHealState(ctx, nil)
@@ -754,11 +758,8 @@ func (a adminAPIHandlers) ProfileHandler(w http.ResponseWriter, r *http.Request)
 			return
 		}
 	}
-	// read request body
-	io.CopyN(io.Discard, r.Body, 1)

 	globalProfilerMu.Lock()
-
 	if globalProfiler == nil {
 		globalProfiler = make(map[string]minioProfiler, 10)
 	}
@@ -1218,7 +1219,7 @@ func (a adminAPIHandlers) ClientDevNull(w http.ResponseWriter, r *http.Request)
 	totalRx := int64(0)
 	connectTime := time.Now()
 	for {
-		n, err := io.CopyN(io.Discard, r.Body, 128*humanize.KiByte)
+		n, err := io.CopyN(xioutil.Discard, r.Body, 128*humanize.KiByte)
 		if err != nil && err != io.EOF && err != io.ErrUnexpectedEOF {
 			// would mean the network is not stable. Logging here will help in debugging network issues.
 			if time.Since(connectTime) < (globalNetPerfMinDuration - time.Second) {
@@ -1296,6 +1297,7 @@ func (a adminAPIHandlers) ObjectSpeedTestHandler(w http.ResponseWriter, r *http.
 	customBucket := strings.TrimSpace(r.Form.Get(peerRESTBucket))
 	autotune := r.Form.Get("autotune") == "true"
 	noClear := r.Form.Get("noclear") == "true"
+	enableSha256 := r.Form.Get("enableSha256") == "true"

 	size, err := strconv.Atoi(sizeStr)
 	if err != nil {
@@ -1312,7 +1314,7 @@ func (a adminAPIHandlers) ObjectSpeedTestHandler(w http.ResponseWriter, r *http.
 		duration = time.Second * 10
 	}

-	storageInfo := objectAPI.StorageInfo(ctx)
+	storageInfo := objectAPI.StorageInfo(ctx, true)

 	sufficientCapacity, canAutotune, capacityErrMsg := validateObjPerfOptions(ctx, storageInfo, concurrent, size, autotune)
 	if !sufficientCapacity {
@@ -1365,6 +1367,7 @@ func (a adminAPIHandlers) ObjectSpeedTestHandler(w http.ResponseWriter, r *http.
 		autotune:         autotune,
 		storageClass:     storageClass,
 		bucketName:       customBucket,
+		enableSha256:     enableSha256,
 	})
 	var prevResult madmin.SpeedTestResult
 	for {
@@ -1400,13 +1403,21 @@ func (a adminAPIHandlers) ObjectSpeedTestHandler(w http.ResponseWriter, r *http.
 }

 func makeObjectPerfBucket(ctx context.Context, objectAPI ObjectLayer, bucketName string) (bucketExists bool, err error) {
-	if err = objectAPI.MakeBucket(ctx, bucketName, MakeBucketOptions{}); err != nil {
+	if err = objectAPI.MakeBucket(ctx, bucketName, MakeBucketOptions{VersioningEnabled: globalSiteReplicationSys.isEnabled()}); err != nil {
 		if _, ok := err.(BucketExists); !ok {
 			// Only BucketExists error can be ignored.
 			return false, err
 		}
 		bucketExists = true
 	}
+
+	if globalSiteReplicationSys.isEnabled() {
+		configData := []byte(`<VersioningConfiguration><Status>Enabled</Status><ExcludedPrefixes><Prefix>speedtest/*</Prefix></ExcludedPrefixes></VersioningConfiguration>`)
+		if _, err = globalBucketMetadataSys.Update(ctx, bucketName, bucketVersioningConfig, configData); err != nil {
+			return false, err
+		}
+	}
+
 	return bucketExists, nil
 }

@@ -1902,7 +1913,7 @@ func getServerInfo(ctx context.Context, poolsInfoEnabled bool, r *http.Request)
 		}
 	}

-	log, audit := fetchLoggerInfo()
+	log, audit := fetchLoggerInfo(ctx)

 	// Get the notification target info
 	notifyTarget := fetchLambdaInfo()
@@ -2027,8 +2038,14 @@ func getKubernetesInfo(dctx context.Context) madmin.KubernetesInfo {

 func fetchHealthInfo(healthCtx context.Context, objectAPI ObjectLayer, query *url.Values, healthInfoCh chan madmin.HealthInfo, healthInfo madmin.HealthInfo) {
 	hostAnonymizer := createHostAnonymizer()
-	// anonAddr - Anonymizes hosts in given input string.
+
+	anonParam := query.Get(anonymizeParam)
+	// anonAddr - Anonymizes hosts in given input string
+	// (only if the anonymize param is set to srict).
 	anonAddr := func(addr string) string {
+		if anonParam != anonymizeStrict {
+			return addr
+		}
 		newAddr, found := hostAnonymizer[addr]
 		if found {
 			return newAddr
@@ -2089,6 +2106,20 @@ func fetchHealthInfo(healthCtx context.Context, objectAPI ObjectLayer, query *ur
 		}
 	}

+	getAndWriteNetInfo := func() {
+		if query.Get(string(madmin.HealthDataTypeSysNet)) == "true" {
+			localNetInfo := madmin.GetNetInfo(globalLocalNodeName, globalInternodeInterface)
+			healthInfo.Sys.NetInfo = append(healthInfo.Sys.NetInfo, localNetInfo)
+
+			peerNetInfos := globalNotificationSys.GetNetInfo(healthCtx)
+			for _, n := range peerNetInfos {
+				anonymizeAddr(&n)
+				healthInfo.Sys.NetInfo = append(healthInfo.Sys.NetInfo, n)
+			}
+			partialWrite(healthInfo)
+		}
+	}
+
 	getAndWriteOSInfo := func() {
 		if query.Get("sysosinfo") == "true" {
 			localOSInfo := madmin.GetOSInfo(healthCtx, globalLocalNodeName)
@@ -2189,6 +2220,10 @@ func fetchHealthInfo(healthCtx context.Context, objectAPI ObjectLayer, query *ur
 	}

 	anonymizeCmdLine := func(cmdLine string) string {
+		if anonParam != anonymizeStrict {
+			return cmdLine
+		}
+
 		if !globalIsDistErasure {
 			// FS mode - single server - hard code to `server1`
 			anonCmdLine := strings.ReplaceAll(cmdLine, globalLocalNodeName, "server1")
@@ -2310,6 +2345,7 @@ func fetchHealthInfo(healthCtx context.Context, objectAPI ObjectLayer, query *ur
 		getAndWritePlatformInfo()
 		getAndWriteCPUs()
 		getAndWritePartitions()
+		getAndWriteNetInfo()
 		getAndWriteOSInfo()
 		getAndWriteMemInfo()
 		getAndWriteProcInfo()
@@ -2467,11 +2503,27 @@ func getTLSInfo() madmin.TLSInfo {

 	if globalIsTLS {
 		for _, c := range globalPublicCerts {
+			check := xxh3.Hash(c.RawIssuer)
+			check ^= xxh3.Hash(c.RawSubjectPublicKeyInfo)
+			// We XOR, so order doesn't matter.
+			for _, v := range c.DNSNames {
+				check ^= xxh3.HashString(v)
+			}
+			for _, v := range c.EmailAddresses {
+				check ^= xxh3.HashString(v)
+			}
+			for _, v := range c.IPAddresses {
+				check ^= xxh3.HashString(v.String())
+			}
+			for _, v := range c.URIs {
+				check ^= xxh3.HashString(v.String())
+			}
 			tlsInfo.Certs = append(tlsInfo.Certs, madmin.TLSCert{
 				PubKeyAlgo:    c.PublicKeyAlgorithm.String(),
 				SignatureAlgo: c.SignatureAlgorithm.String(),
 				NotBefore:     c.NotBefore,
 				NotAfter:      c.NotAfter,
+				Checksum:      strconv.FormatUint(check, 16),
 			})
 		}
 	}
@@ -2522,7 +2574,7 @@ func assignPoolNumbers(servers []madmin.ServerProperties) {
 func fetchLambdaInfo() []map[string][]madmin.TargetIDStatus {
 	lambdaMap := make(map[string][]madmin.TargetIDStatus)

-	for _, tgt := range globalNotifyTargetList.Targets() {
+	for _, tgt := range globalEventNotifier.Targets() {
 		targetIDStatus := make(map[string]madmin.Status)
 		active, _ := tgt.IsActive()
 		targetID := tgt.ID()
@@ -2610,19 +2662,26 @@ func fetchKMSStatusV2(ctx context.Context) []madmin.KMS {
 	return stats
 }

+func targetStatus(ctx context.Context, h logger.Target) madmin.Status {
+	if h.IsOnline(ctx) {
+		return madmin.Status{Status: string(madmin.ItemOnline)}
+	}
+	return madmin.Status{Status: string(madmin.ItemOffline)}
+}
+
 // fetchLoggerDetails return log info
-func fetchLoggerInfo() ([]madmin.Logger, []madmin.Audit) {
+func fetchLoggerInfo(ctx context.Context) ([]madmin.Logger, []madmin.Audit) {
 	var loggerInfo []madmin.Logger
 	var auditloggerInfo []madmin.Audit
 	for _, tgt := range logger.SystemTargets() {
 		if tgt.Endpoint() != "" {
-			loggerInfo = append(loggerInfo, madmin.Logger{tgt.String(): logger.TargetStatus(GlobalContext, tgt)})
+			loggerInfo = append(loggerInfo, madmin.Logger{tgt.String(): targetStatus(ctx, tgt)})
 		}
 	}

 	for _, tgt := range logger.AuditTargets() {
 		if tgt.Endpoint() != "" {
-			auditloggerInfo = append(auditloggerInfo, madmin.Audit{tgt.String(): logger.TargetStatus(GlobalContext, tgt)})
+			auditloggerInfo = append(auditloggerInfo, madmin.Audit{tgt.String(): targetStatus(ctx, tgt)})
 		}
 	}

@@ -2676,7 +2735,7 @@ func getClusterMetaInfo(ctx context.Context) []byte {
 		ci.Info.NoOfServers = len(globalEndpoints.Hostnames())
 		ci.Info.MinioVersion = Version

-		si := objectAPI.StorageInfo(ctx)
+		si := objectAPI.StorageInfo(ctx, true)

 		ci.Info.NoOfDrives = len(si.Disks)
 		for _, disk := range si.Disks {
@@ -810,7 +810,7 @@ func (h *healSequence) healMinioSysMeta(objAPI ObjectLayer, metaPrefix string) f
 		// NOTE: Healing on meta is run regardless
 		// of any bucket being selected, this is to ensure that
 		// meta are always upto date and correct.
-		return objAPI.HealObjects(h.ctx, minioMetaBucket, metaPrefix, h.settings, func(bucket, object, versionID string) error {
+		return objAPI.HealObjects(h.ctx, minioMetaBucket, metaPrefix, h.settings, func(bucket, object, versionID string, scanMode madmin.HealScanMode) error {
 			if h.isQuitting() {
 				return errHealStopSignalled
 			}
@@ -867,7 +867,7 @@ func (h *healSequence) healBucket(objAPI ObjectLayer, bucket string, bucketsOnly

 	if !h.settings.Recursive {
 		if h.object != "" {
-			if err := h.healObject(bucket, h.object, ""); err != nil {
+			if err := h.healObject(bucket, h.object, "", h.settings.ScanMode); err != nil {
 				return err
 			}
 		}
@@ -882,7 +882,7 @@ func (h *healSequence) healBucket(objAPI ObjectLayer, bucket string, bucketsOnly
 }

 // healObject - heal the given object and record result
-func (h *healSequence) healObject(bucket, object, versionID string) error {
+func (h *healSequence) healObject(bucket, object, versionID string, scanMode madmin.HealScanMode) error {
 	if h.isQuitting() {
 		return errHealStopSignalled
 	}
@@ -300,6 +300,12 @@ func registerAdminRouter(router *mux.Router, enableConfigOps bool) {
 		adminRouter.Methods(http.MethodGet).Path(adminVersion + "/idp-config/{type}/{name}").HandlerFunc(adminMiddleware(adminAPI.GetIdentityProviderCfg))
 		adminRouter.Methods(http.MethodDelete).Path(adminVersion + "/idp-config/{type}/{name}").HandlerFunc(adminMiddleware(adminAPI.DeleteIdentityProviderCfg))

+		// LDAP specific service accounts ops
+		adminRouter.Methods(http.MethodPut).Path(adminVersion + "/idp/ldap/add-service-account").HandlerFunc(adminMiddleware(adminAPI.AddServiceAccountLDAP))
+		adminRouter.Methods(http.MethodGet).Path(adminVersion+"/idp/ldap/list-access-keys").
+			HandlerFunc(adminMiddleware(adminAPI.ListAccessKeysLDAP)).
+			Queries("userDN", "{userDN:.*}", "listType", "{listType:.*}")
+
 		// LDAP IAM operations
 		adminRouter.Methods(http.MethodGet).Path(adminVersion + "/idp/ldap/policy-entities").HandlerFunc(adminMiddleware(adminAPI.ListLDAPPolicyMappingEntities))
 		adminRouter.Methods(http.MethodPost).Path(adminVersion + "/idp/ldap/policy/{operation}").HandlerFunc(adminMiddleware(adminAPI.AttachDetachPolicyLDAP))
@@ -376,6 +382,7 @@ func registerAdminRouter(router *mux.Router, enableConfigOps bool) {
 		adminRouter.Methods(http.MethodPut).Path(adminVersion + "/site-replication/peer/edit").HandlerFunc(adminMiddleware(adminAPI.SRPeerEdit))
 		adminRouter.Methods(http.MethodPut).Path(adminVersion + "/site-replication/peer/remove").HandlerFunc(adminMiddleware(adminAPI.SRPeerRemove))
 		adminRouter.Methods(http.MethodPut).Path(adminVersion+"/site-replication/resync/op").HandlerFunc(adminMiddleware(adminAPI.SiteReplicationResyncOp)).Queries("operation", "{operation:.*}")
+		adminRouter.Methods(http.MethodPut).Path(adminVersion + "/site-replication/state/edit").HandlerFunc(adminMiddleware(adminAPI.SRStateEdit))

 		if globalIsDistErasure {
 			// Top locks
@@ -141,7 +141,7 @@ func getLocalServerProperty(endpointServerPools EndpointServerPools, r *http.Req

 	objLayer := newObjectLayerFn()
 	if objLayer != nil {
-		storageInfo := objLayer.LocalStorageInfo(GlobalContext)
+		storageInfo := objLayer.LocalStorageInfo(GlobalContext, true)
 		props.State = string(madmin.ItemOnline)
 		props.Disks = storageInfo.Disks
 	} else {
@@ -502,6 +502,47 @@ func generateListBucketsResponse(buckets []BucketInfo) ListBucketsResponse {
 	return data
 }

+func cleanReservedKeys(metadata map[string]string) map[string]string {
+	m := cloneMSS(metadata)
+
+	switch kind, _ := crypto.IsEncrypted(metadata); kind {
+	case crypto.S3:
+		m[xhttp.AmzServerSideEncryption] = xhttp.AmzEncryptionAES
+	case crypto.S3KMS:
+		m[xhttp.AmzServerSideEncryption] = xhttp.AmzEncryptionKMS
+		m[xhttp.AmzServerSideEncryptionKmsID] = kmsKeyIDFromMetadata(metadata)
+		if kmsCtx, ok := metadata[crypto.MetaContext]; ok {
+			m[xhttp.AmzServerSideEncryptionKmsContext] = kmsCtx
+		}
+	case crypto.SSEC:
+		m[xhttp.AmzServerSideEncryptionCustomerAlgorithm] = xhttp.AmzEncryptionAES
+
+	}
+
+	var toRemove []string
+	for k := range cleanMinioInternalMetadataKeys(m) {
+		if stringsHasPrefixFold(k, ReservedMetadataPrefixLower) {
+			// Do not need to send any internal metadata
+			// values to client.
+			toRemove = append(toRemove, k)
+			continue
+		}
+
+		// https://github.com/google/security-research/security/advisories/GHSA-76wf-9vgp-pj7w
+		if equals(k, xhttp.AmzMetaUnencryptedContentLength, xhttp.AmzMetaUnencryptedContentMD5) {
+			toRemove = append(toRemove, k)
+			continue
+		}
+	}
+
+	for _, k := range toRemove {
+		delete(m, k)
+		delete(m, strings.ToLower(k))
+	}
+
+	return m
+}
+
 // 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, metadata metaCheckFn) ListVersionsResponse {
 	versions := make([]ObjectVersion, 0, len(resp.Objects))
@@ -549,18 +590,11 @@ func generateListVersionsResponse(bucket, prefix, marker, versionIDMarker, delim
 			case crypto.SSEC:
 				content.UserMetadata.Set(xhttp.AmzServerSideEncryptionCustomerAlgorithm, xhttp.AmzEncryptionAES)
 			}
-			for k, v := range cleanMinioInternalMetadataKeys(object.UserDefined) {
-				if stringsHasPrefixFold(k, ReservedMetadataPrefixLower) {
-					// Do not need to send any internal metadata
-					// values to client.
-					continue
-				}
-				// https://github.com/google/security-research/security/advisories/GHSA-76wf-9vgp-pj7w
-				if equals(k, xhttp.AmzMetaUnencryptedContentLength, xhttp.AmzMetaUnencryptedContentMD5) {
-					continue
-				}
+			for k, v := range cleanReservedKeys(object.UserDefined) {
 				content.UserMetadata.Set(k, v)
 			}
+
+			content.UserMetadata.Set("expires", object.Expires.Format(http.TimeFormat))
 			content.Internal = &ObjectInternalInfo{
 				K: object.DataBlocks,
 				M: object.ParityBlocks,
@@ -692,18 +726,10 @@ func generateListObjectsV2Response(bucket, prefix, token, nextToken, startAfter,
 				case crypto.SSEC:
 					content.UserMetadata.Set(xhttp.AmzServerSideEncryptionCustomerAlgorithm, xhttp.AmzEncryptionAES)
 				}
-				for k, v := range cleanMinioInternalMetadataKeys(object.UserDefined) {
-					if stringsHasPrefixFold(k, ReservedMetadataPrefixLower) {
-						// Do not need to send any internal metadata
-						// values to client.
-						continue
-					}
-					// https://github.com/google/security-research/security/advisories/GHSA-76wf-9vgp-pj7w
-					if equals(k, xhttp.AmzMetaUnencryptedContentLength, xhttp.AmzMetaUnencryptedContentMD5) {
-						continue
-					}
+				for k, v := range cleanReservedKeys(object.UserDefined) {
 					content.UserMetadata.Set(k, v)
 				}
+				content.UserMetadata.Set("expires", object.Expires.Format(http.TimeFormat))
 				content.Internal = &ObjectInternalInfo{
 					K: object.DataBlocks,
 					M: object.ParityBlocks,
@@ -23,7 +23,7 @@ import (
 	"net/http"

 	"github.com/klauspost/compress/gzhttp"
-	"github.com/minio/console/restapi"
+	consoleapi "github.com/minio/console/api"
 	xhttp "github.com/minio/minio/internal/http"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/mux"
@@ -43,13 +43,13 @@ func setHTTPServer(h *xhttp.Server) {
 	globalObjLayerMutex.Unlock()
 }

-func newConsoleServerFn() *restapi.Server {
+func newConsoleServerFn() *consoleapi.Server {
 	globalObjLayerMutex.RLock()
 	defer globalObjLayerMutex.RUnlock()
 	return globalConsoleSrv
 }

-func setConsoleSrv(srv *restapi.Server) {
+func setConsoleSrv(srv *consoleapi.Server) {
 	globalObjLayerMutex.Lock()
 	globalConsoleSrv = srv
 	globalObjLayerMutex.Unlock()
@@ -232,7 +232,7 @@ func (h *healingTracker) delete(ctx context.Context) error {
 		pathJoin(bucketMetaPrefix, healingTrackerFilename),
 		DeleteOptions{
 			Recursive: false,
-			Force:     false,
+			Immediate: false,
 		},
 	)
 }
@@ -345,18 +345,17 @@ func initAutoHeal(ctx context.Context, objAPI ObjectLayer) {

 	initBackgroundHealing(ctx, objAPI) // start quick background healing

-	globalBackgroundHealState.pushHealLocalDisks(getLocalDisksToHeal()...)
-
 	if env.Get("_MINIO_AUTO_DRIVE_HEALING", config.EnableOn) == config.EnableOn || env.Get("_MINIO_AUTO_DISK_HEALING", config.EnableOn) == config.EnableOn {
+		globalBackgroundHealState.pushHealLocalDisks(getLocalDisksToHeal()...)
 		go monitorLocalDisksAndHeal(ctx, z)
 	}
 }

 func getLocalDisksToHeal() (disksToHeal Endpoints) {
 	globalLocalDrivesMu.RLock()
-	globalLocalDrives := globalLocalDrives
+	localDrives := globalLocalDrives
 	globalLocalDrivesMu.RUnlock()
-	for _, disk := range globalLocalDrives {
+	for _, disk := range localDrives {
 		_, err := disk.GetDiskID()
 		if errors.Is(err, errUnformattedDisk) {
 			disksToHeal = append(disksToHeal, disk.Endpoint())
@@ -0,0 +1,677 @@
+// Copyright (c) 2015-2023 MinIO, Inc.
+//
+// This file is part of MinIO Object Storage stack
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package cmd
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"runtime"
+	"strconv"
+	"time"
+
+	"github.com/minio/minio-go/v7/pkg/tags"
+	"github.com/minio/minio/internal/bucket/versioning"
+	xhttp "github.com/minio/minio/internal/http"
+	"github.com/minio/minio/internal/logger"
+	"github.com/minio/pkg/v2/env"
+	"github.com/minio/pkg/v2/wildcard"
+	"github.com/minio/pkg/v2/workers"
+)
+
+// expire: # Expire objects that match a condition
+//   apiVersion: v1
+//   bucket: mybucket # Bucket where this batch job will expire matching objects from
+//   prefix: myprefix # (Optional) Prefix under which this job will expire objects matching the rules below.
+//   rules:
+//     - type: object  # regular objects with zero ore more older versions
+//       name: NAME # match object names that satisfy the wildcard expression.
+//       olderThan: 70h # match objects older than this value
+//       createdBefore: "2006-01-02T15:04:05.00Z" # match objects created before "date"
+//       tags:
+//         - key: name
+//           value: pick* # match objects with tag 'name', all values starting with 'pick'
+//       metadata:
+//         - key: content-type
+//           value: image/* # match objects with 'content-type', all values starting with 'image/'
+//       size:
+//         lessThan: "10MiB" # match objects with size less than this value (e.g. 10MiB)
+//         greaterThan: 1MiB # match objects with size greater than this value (e.g. 1MiB)
+//       purge:
+//           # retainVersions: 0 # (default) delete all versions of the object. This option is the fastest.
+//           # retainVersions: 5 # keep the latest 5 versions of the object.
+//
+//     - type: deleted # objects with delete marker as their latest version
+//       name: NAME # match object names that satisfy the wildcard expression.
+//       olderThan: 10h # match objects older than this value (e.g. 7d10h31s)
+//       createdBefore: "2006-01-02T15:04:05.00Z" # match objects created before "date"
+//       purge:
+//           # retainVersions: 0 # (default) delete all versions of the object. This option is the fastest.
+//           # retainVersions: 5 # keep the latest 5 versions of the object including delete markers.
+//
+//   notify:
+//     endpoint: https://notify.endpoint # notification endpoint to receive job completion status
+//     token: Bearer xxxxx # optional authentication token for the notification endpoint
+//
+//   retry:
+//     attempts: 10 # number of retries for the job before giving up
+//     delay: 500ms # least amount of delay between each retry
+
+//go:generate msgp -file $GOFILE
+
+// BatchJobExpirePurge type accepts non-negative versions to be retained
+type BatchJobExpirePurge struct {
+	RetainVersions int `yaml:"retainVersions" json:"retainVersions"`
+}
+
+// Validate returns nil if value is valid, ie > 0.
+func (p BatchJobExpirePurge) Validate() error {
+	if p.RetainVersions < 0 {
+		return errors.New("retainVersions must be >= 0")
+	}
+	return nil
+}
+
+// BatchJobExpireFilter holds all the filters currently supported for batch replication
+type BatchJobExpireFilter struct {
+	OlderThan     time.Duration       `yaml:"olderThan,omitempty" json:"olderThan"`
+	CreatedBefore *time.Time          `yaml:"createdBefore,omitempty" json:"createdBefore"`
+	Tags          []BatchJobKV        `yaml:"tags,omitempty" json:"tags"`
+	Metadata      []BatchJobKV        `yaml:"metadata,omitempty" json:"metadata"`
+	Size          BatchJobSizeFilter  `yaml:"size" json:"size"`
+	Type          string              `yaml:"type" json:"type"`
+	Name          string              `yaml:"name" json:"name"`
+	Purge         BatchJobExpirePurge `yaml:"purge" json:"purge"`
+}
+
+// Matches returns true if obj matches the filter conditions specified in ef.
+func (ef BatchJobExpireFilter) Matches(obj ObjectInfo, now time.Time) bool {
+	switch ef.Type {
+	case BatchJobExpireObject:
+		if obj.DeleteMarker {
+			return false
+		}
+	case BatchJobExpireDeleted:
+		if !obj.DeleteMarker {
+			return false
+		}
+	default:
+		// we should never come here, Validate should have caught this.
+		logger.LogOnceIf(context.Background(), fmt.Errorf("invalid filter type: %s", ef.Type), ef.Type)
+		return false
+	}
+
+	if len(ef.Name) > 0 && !wildcard.Match(ef.Name, obj.Name) {
+		return false
+	}
+	if ef.OlderThan > 0 && now.Sub(obj.ModTime) <= ef.OlderThan {
+		return false
+	}
+
+	if ef.CreatedBefore != nil && !obj.ModTime.Before(*ef.CreatedBefore) {
+		return false
+	}
+
+	if len(ef.Tags) > 0 && !obj.DeleteMarker {
+		// Only parse object tags if tags filter is specified.
+		var tagMap map[string]string
+		if len(obj.UserTags) != 0 {
+			t, err := tags.ParseObjectTags(obj.UserTags)
+			if err != nil {
+				return false
+			}
+			tagMap = t.ToMap()
+		}
+
+		for _, kv := range ef.Tags {
+			// Object (version) must match all tags specified in
+			// the filter
+			var match bool
+			for t, v := range tagMap {
+				if kv.Match(BatchJobKV{Key: t, Value: v}) {
+					match = true
+				}
+			}
+			if !match {
+				return false
+			}
+		}
+
+	}
+	if len(ef.Metadata) > 0 && !obj.DeleteMarker {
+		for _, kv := range ef.Metadata {
+			// Object (version) must match all x-amz-meta and
+			// standard metadata headers
+			// specified in the filter
+			var match bool
+			for k, v := range obj.UserDefined {
+				if !stringsHasPrefixFold(k, "x-amz-meta-") && !isStandardHeader(k) {
+					continue
+				}
+				// We only need to match x-amz-meta or standardHeaders
+				if kv.Match(BatchJobKV{Key: k, Value: v}) {
+					match = true
+				}
+			}
+			if !match {
+				return false
+			}
+		}
+	}
+
+	return ef.Size.InRange(obj.Size)
+}
+
+const (
+	// BatchJobExpireObject - object type
+	BatchJobExpireObject string = "object"
+	// BatchJobExpireDeleted - delete marker type
+	BatchJobExpireDeleted string = "deleted"
+)
+
+// Validate returns nil if ef has valid fields, validation error otherwise.
+func (ef BatchJobExpireFilter) Validate() error {
+	switch ef.Type {
+	case BatchJobExpireObject:
+	case BatchJobExpireDeleted:
+		if len(ef.Tags) > 0 || len(ef.Metadata) > 0 {
+			return errors.New("invalid batch-expire rule filter")
+		}
+	default:
+		return errors.New("invalid batch-expire type")
+	}
+
+	for _, tag := range ef.Tags {
+		if err := tag.Validate(); err != nil {
+			return err
+		}
+	}
+
+	for _, meta := range ef.Metadata {
+		if err := meta.Validate(); err != nil {
+			return err
+		}
+	}
+	if err := ef.Purge.Validate(); err != nil {
+		return err
+	}
+	if err := ef.Size.Validate(); err != nil {
+		return err
+	}
+	if ef.CreatedBefore != nil && !ef.CreatedBefore.Before(time.Now()) {
+		return errors.New("CreatedBefore is in the future")
+	}
+	return nil
+}
+
+// BatchJobExpire represents configuration parameters for a batch expiration
+// job typically supplied in yaml form
+type BatchJobExpire struct {
+	APIVersion      string                 `yaml:"apiVersion" json:"apiVersion"`
+	Bucket          string                 `yaml:"bucket" json:"bucket"`
+	Prefix          string                 `yaml:"prefix" json:"prefix"`
+	NotificationCfg BatchJobNotification   `yaml:"notify" json:"notify"`
+	Retry           BatchJobRetry          `yaml:"retry" json:"retry"`
+	Rules           []BatchJobExpireFilter `yaml:"rules" json:"rules"`
+}
+
+// Notify notifies notification endpoint if configured regarding job failure or success.
+func (r BatchJobExpire) Notify(ctx context.Context, body io.Reader) error {
+	if r.NotificationCfg.Endpoint == "" {
+		return nil
+	}
+
+	ctx, cancel := context.WithTimeout(ctx, 10*time.Second)
+	defer cancel()
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, r.NotificationCfg.Endpoint, body)
+	if err != nil {
+		return err
+	}
+
+	if r.NotificationCfg.Token != "" {
+		req.Header.Set("Authorization", r.NotificationCfg.Token)
+	}
+
+	clnt := http.Client{Transport: getRemoteInstanceTransport}
+	resp, err := clnt.Do(req)
+	if err != nil {
+		return err
+	}
+
+	xhttp.DrainBody(resp.Body)
+	if resp.StatusCode != http.StatusOK {
+		return errors.New(resp.Status)
+	}
+
+	return nil
+}
+
+// Expire expires object versions which have already matched supplied filter conditions
+func (r *BatchJobExpire) Expire(ctx context.Context, api ObjectLayer, vc *versioning.Versioning, objsToDel []ObjectToDelete) []error {
+	opts := ObjectOptions{
+		PrefixEnabledFn:  vc.PrefixEnabled,
+		VersionSuspended: vc.Suspended(),
+	}
+	_, errs := api.DeleteObjects(ctx, r.Bucket, objsToDel, opts)
+	return errs
+}
+
+const (
+	batchExpireName                 = "batch-expire.bin"
+	batchExpireFormat               = 1
+	batchExpireVersionV1            = 1
+	batchExpireVersion              = batchExpireVersionV1
+	batchExpireAPIVersion           = "v1"
+	batchExpireJobDefaultRetries    = 3
+	batchExpireJobDefaultRetryDelay = 250 * time.Millisecond
+)
+
+type objInfoCache map[string]*ObjectInfo
+
+func newObjInfoCache() objInfoCache {
+	return objInfoCache(make(map[string]*ObjectInfo))
+}
+
+func (oiCache objInfoCache) Add(toDel ObjectToDelete, oi *ObjectInfo) {
+	oiCache[fmt.Sprintf("%s-%s", toDel.ObjectName, toDel.VersionID)] = oi
+}
+
+func (oiCache objInfoCache) Get(toDel ObjectToDelete) (*ObjectInfo, bool) {
+	oi, ok := oiCache[fmt.Sprintf("%s-%s", toDel.ObjectName, toDel.VersionID)]
+	return oi, ok
+}
+
+func batchObjsForDelete(ctx context.Context, r *BatchJobExpire, ri *batchJobInfo, job BatchJobRequest, api ObjectLayer, wk *workers.Workers, expireCh <-chan []expireObjInfo) {
+	vc, _ := globalBucketVersioningSys.Get(r.Bucket)
+	retryAttempts := r.Retry.Attempts
+	delay := job.Expire.Retry.Delay
+	if delay == 0 {
+		delay = batchExpireJobDefaultRetryDelay
+	}
+
+	var i int
+	for toExpire := range expireCh {
+		select {
+		case <-ctx.Done():
+			return
+		default:
+		}
+		if i > 0 {
+			if wait := globalBatchConfig.ExpirationWait(); wait > 0 {
+				time.Sleep(wait)
+			}
+		}
+		i++
+		wk.Take()
+		go func(toExpire []expireObjInfo) {
+			defer wk.Give()
+
+			toExpireAll := make([]ObjectInfo, 0, len(toExpire))
+			toDel := make([]ObjectToDelete, 0, len(toExpire))
+			oiCache := newObjInfoCache()
+			for _, exp := range toExpire {
+				if exp.ExpireAll {
+					toExpireAll = append(toExpireAll, exp.ObjectInfo)
+					continue
+				}
+				// Cache ObjectInfo value via pointers for
+				// subsequent use to track objects which
+				// couldn't be deleted.
+				od := ObjectToDelete{
+					ObjectV: ObjectV{
+						ObjectName: exp.Name,
+						VersionID:  exp.VersionID,
+					},
+				}
+				toDel = append(toDel, od)
+				oiCache.Add(od, &exp.ObjectInfo)
+			}
+
+			var done bool
+			// DeleteObject(deletePrefix: true) to expire all versions of an object
+			for _, exp := range toExpireAll {
+				var success bool
+				for attempts := 1; attempts <= retryAttempts; attempts++ {
+					select {
+					case <-ctx.Done():
+						done = true
+					default:
+					}
+					stopFn := globalBatchJobsMetrics.trace(batchJobMetricExpire, ri.JobID, attempts)
+					_, err := api.DeleteObject(ctx, exp.Bucket, exp.Name, ObjectOptions{
+						DeletePrefix: true,
+					})
+					if err != nil {
+						stopFn(exp, err)
+						logger.LogIf(ctx, fmt.Errorf("Failed to expire %s/%s versionID=%s due to %v (attempts=%d)", toExpire[i].Bucket, toExpire[i].Name, toExpire[i].VersionID, err, attempts))
+					} else {
+						stopFn(exp, err)
+						success = true
+						break
+					}
+				}
+				ri.trackMultipleObjectVersions(r.Bucket, exp, success)
+				if done {
+					break
+				}
+			}
+
+			if done {
+				return
+			}
+
+			// DeleteMultiple objects
+			toDelCopy := make([]ObjectToDelete, len(toDel))
+			for attempts := 1; attempts <= retryAttempts; attempts++ {
+				select {
+				case <-ctx.Done():
+					return
+				default:
+				}
+
+				stopFn := globalBatchJobsMetrics.trace(batchJobMetricExpire, ri.JobID, attempts)
+				// Copying toDel to select from objects whose
+				// deletion failed
+				copy(toDelCopy, toDel)
+				var failed int
+				errs := r.Expire(ctx, api, vc, toDel)
+				// reslice toDel in preparation for next retry
+				// attempt
+				toDel = toDel[:0]
+				for i, err := range errs {
+					if err != nil {
+						stopFn(toDelCopy[i], err)
+						logger.LogIf(ctx, fmt.Errorf("Failed to expire %s/%s versionID=%s due to %v (attempts=%d)", ri.Bucket, toDelCopy[i].ObjectName, toDelCopy[i].VersionID, err, attempts))
+						failed++
+						if attempts == retryAttempts { // all retry attempts failed, record failure
+							if oi, ok := oiCache.Get(toDelCopy[i]); ok {
+								ri.trackCurrentBucketObject(r.Bucket, *oi, false)
+							}
+						} else {
+							toDel = append(toDel, toDelCopy[i])
+						}
+					} else {
+						stopFn(toDelCopy[i], nil)
+						if oi, ok := oiCache.Get(toDelCopy[i]); ok {
+							ri.trackCurrentBucketObject(r.Bucket, *oi, true)
+						}
+					}
+				}
+
+				globalBatchJobsMetrics.save(ri.JobID, ri)
+
+				if failed == 0 {
+					break
+				}
+
+				// Add a delay between retry attempts
+				if attempts < retryAttempts {
+					time.Sleep(delay)
+				}
+			}
+		}(toExpire)
+	}
+}
+
+type expireObjInfo struct {
+	ObjectInfo
+	ExpireAll bool
+}
+
+// Start the batch expiration job, resumes if there was a pending job via "job.ID"
+func (r *BatchJobExpire) Start(ctx context.Context, api ObjectLayer, job BatchJobRequest) error {
+	ri := &batchJobInfo{
+		JobID:     job.ID,
+		JobType:   string(job.Type()),
+		StartTime: job.Started,
+	}
+	if err := ri.load(ctx, api, job); err != nil {
+		return err
+	}
+
+	globalBatchJobsMetrics.save(job.ID, ri)
+	lastObject := ri.Object
+
+	now := time.Now().UTC()
+
+	workerSize, err := strconv.Atoi(env.Get("_MINIO_BATCH_EXPIRATION_WORKERS", strconv.Itoa(runtime.GOMAXPROCS(0)/2)))
+	if err != nil {
+		return err
+	}
+
+	wk, err := workers.New(workerSize)
+	if err != nil {
+		// invalid worker size.
+		return err
+	}
+
+	ctx, cancel := context.WithCancel(ctx)
+	defer cancel()
+
+	results := make(chan ObjectInfo, workerSize)
+	if err := api.Walk(ctx, r.Bucket, r.Prefix, results, WalkOptions{
+		Marker:       lastObject,
+		LatestOnly:   false, // we need to visit all versions of the object to implement purge: retainVersions
+		VersionsSort: WalkVersionsSortDesc,
+	}); err != nil {
+		// Do not need to retry if we can't list objects on source.
+		return err
+	}
+
+	// Goroutine to periodically save batch-expire job's in-memory state
+	saverQuitCh := make(chan struct{})
+	go func() {
+		saveTicker := time.NewTicker(10 * time.Second)
+		defer saveTicker.Stop()
+		for {
+			select {
+			case <-saveTicker.C:
+				// persist in-memory state to disk after every 10secs.
+				logger.LogIf(ctx, ri.updateAfter(ctx, api, 10*time.Second, job))
+
+			case <-ctx.Done():
+				// persist in-memory state immediately before exiting due to context cancelation.
+				logger.LogIf(ctx, ri.updateAfter(ctx, api, 0, job))
+				return
+
+			case <-saverQuitCh:
+				// persist in-memory state immediately to disk.
+				logger.LogIf(ctx, ri.updateAfter(ctx, api, 0, job))
+				return
+			}
+		}
+	}()
+
+	expireCh := make(chan []expireObjInfo, workerSize)
+	go batchObjsForDelete(ctx, r, ri, job, api, wk, expireCh)
+
+	var (
+		prevObj       ObjectInfo
+		matchedFilter BatchJobExpireFilter
+		versionsCount int
+		toDel         []expireObjInfo
+	)
+	for result := range results {
+		// Apply filter to find the matching rule to apply expiry
+		// actions accordingly.
+		// nolint:gocritic
+		if result.IsLatest {
+			// send down filtered entries to be deleted using
+			// DeleteObjects method
+			if len(toDel) > 10 { // batch up to 10 objects/versions to be expired simultaneously.
+				xfer := make([]expireObjInfo, len(toDel))
+				copy(xfer, toDel)
+
+				var done bool
+				select {
+				case <-ctx.Done():
+					done = true
+				case expireCh <- xfer:
+					toDel = toDel[:0] // resetting toDel
+				}
+				if done {
+					break
+				}
+			}
+			var match BatchJobExpireFilter
+			var found bool
+			for _, rule := range r.Rules {
+				if rule.Matches(result, now) {
+					match = rule
+					found = true
+					break
+				}
+			}
+			if !found {
+				continue
+			}
+
+			prevObj = result
+			matchedFilter = match
+			versionsCount = 1
+			// Include the latest version
+			if matchedFilter.Purge.RetainVersions == 0 {
+				toDel = append(toDel, expireObjInfo{
+					ObjectInfo: result,
+					ExpireAll:  true,
+				})
+				continue
+			}
+		} else if prevObj.Name == result.Name {
+			if matchedFilter.Purge.RetainVersions == 0 {
+				continue // including latest version in toDel suffices, skipping other versions
+			}
+			versionsCount++
+		} else {
+			continue
+		}
+
+		if versionsCount <= matchedFilter.Purge.RetainVersions {
+			continue // retain versions
+		}
+		toDel = append(toDel, expireObjInfo{
+			ObjectInfo: result,
+		})
+	}
+	// Send any remaining objects downstream
+	if len(toDel) > 0 {
+		select {
+		case <-ctx.Done():
+		case expireCh <- toDel:
+		}
+	}
+	close(expireCh)
+
+	wk.Wait() // waits for all expire goroutines to complete
+
+	ri.Complete = ri.ObjectsFailed == 0
+	ri.Failed = ri.ObjectsFailed > 0
+	globalBatchJobsMetrics.save(job.ID, ri)
+
+	// Close the saverQuitCh - this also triggers saving in-memory state
+	// immediately one last time before we exit this method.
+	close(saverQuitCh)
+
+	// Notify expire jobs final status to the configured endpoint
+	buf, _ := json.Marshal(ri)
+	if err := r.Notify(context.Background(), bytes.NewReader(buf)); err != nil {
+		logger.LogIf(context.Background(), fmt.Errorf("unable to notify %v", err))
+	}
+
+	return nil
+}
+
+//msgp:ignore batchExpireJobError
+type batchExpireJobError struct {
+	Code           string
+	Description    string
+	HTTPStatusCode int
+}
+
+func (e batchExpireJobError) Error() string {
+	return e.Description
+}
+
+// maxBatchRules maximum number of rules a batch-expiry job supports
+const maxBatchRules = 50
+
+// Validate validates the job definition input
+func (r *BatchJobExpire) Validate(ctx context.Context, job BatchJobRequest, o ObjectLayer) error {
+	if r == nil {
+		return nil
+	}
+
+	if r.APIVersion != batchExpireAPIVersion {
+		return batchExpireJobError{
+			Code:           "InvalidArgument",
+			Description:    "Unsupported batch expire API version",
+			HTTPStatusCode: http.StatusBadRequest,
+		}
+	}
+
+	if r.Bucket == "" {
+		return batchExpireJobError{
+			Code:           "InvalidArgument",
+			Description:    "Bucket argument missing",
+			HTTPStatusCode: http.StatusBadRequest,
+		}
+	}
+
+	if _, err := o.GetBucketInfo(ctx, r.Bucket, BucketOptions{}); err != nil {
+		if isErrBucketNotFound(err) {
+			return batchExpireJobError{
+				Code:           "NoSuchSourceBucket",
+				Description:    "The specified source bucket does not exist",
+				HTTPStatusCode: http.StatusNotFound,
+			}
+		}
+		return err
+	}
+
+	if len(r.Rules) > maxBatchRules {
+		return batchExpireJobError{
+			Code:           "InvalidArgument",
+			Description:    "Too many rules. Batch expire job can't have more than 100 rules",
+			HTTPStatusCode: http.StatusBadRequest,
+		}
+	}
+
+	for _, rule := range r.Rules {
+		if err := rule.Validate(); err != nil {
+			return batchExpireJobError{
+				Code:           "InvalidArgument",
+				Description:    fmt.Sprintf("Invalid batch expire rule: %s", err),
+				HTTPStatusCode: http.StatusBadRequest,
+			}
+		}
+	}
+
+	if err := r.Retry.Validate(); err != nil {
+		return batchExpireJobError{
+			Code:           "InvalidArgument",
+			Description:    fmt.Sprintf("Invalid batch expire retry configuration: %s", err),
+			HTTPStatusCode: http.StatusBadRequest,
+		}
+	}
+	return nil
+}
@@ -0,0 +1,856 @@
+package cmd
+
+// Code generated by github.com/tinylib/msgp DO NOT EDIT.
+
+import (
+	"time"
+
+	"github.com/tinylib/msgp/msgp"
+)
+
+// DecodeMsg implements msgp.Decodable
+func (z *BatchJobExpire) DecodeMsg(dc *msgp.Reader) (err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, err = dc.ReadMapHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, err = dc.ReadMapKeyPtr()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "APIVersion":
+			z.APIVersion, err = dc.ReadString()
+			if err != nil {
+				err = msgp.WrapError(err, "APIVersion")
+				return
+			}
+		case "Bucket":
+			z.Bucket, err = dc.ReadString()
+			if err != nil {
+				err = msgp.WrapError(err, "Bucket")
+				return
+			}
+		case "Prefix":
+			z.Prefix, err = dc.ReadString()
+			if err != nil {
+				err = msgp.WrapError(err, "Prefix")
+				return
+			}
+		case "NotificationCfg":
+			err = z.NotificationCfg.DecodeMsg(dc)
+			if err != nil {
+				err = msgp.WrapError(err, "NotificationCfg")
+				return
+			}
+		case "Retry":
+			err = z.Retry.DecodeMsg(dc)
+			if err != nil {
+				err = msgp.WrapError(err, "Retry")
+				return
+			}
+		case "Rules":
+			var zb0002 uint32
+			zb0002, err = dc.ReadArrayHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "Rules")
+				return
+			}
+			if cap(z.Rules) >= int(zb0002) {
+				z.Rules = (z.Rules)[:zb0002]
+			} else {
+				z.Rules = make([]BatchJobExpireFilter, zb0002)
+			}
+			for za0001 := range z.Rules {
+				err = z.Rules[za0001].DecodeMsg(dc)
+				if err != nil {
+					err = msgp.WrapError(err, "Rules", za0001)
+					return
+				}
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z *BatchJobExpire) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 6
+	// write "APIVersion"
+	err = en.Append(0x86, 0xaa, 0x41, 0x50, 0x49, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e)
+	if err != nil {
+		return
+	}
+	err = en.WriteString(z.APIVersion)
+	if err != nil {
+		err = msgp.WrapError(err, "APIVersion")
+		return
+	}
+	// write "Bucket"
+	err = en.Append(0xa6, 0x42, 0x75, 0x63, 0x6b, 0x65, 0x74)
+	if err != nil {
+		return
+	}
+	err = en.WriteString(z.Bucket)
+	if err != nil {
+		err = msgp.WrapError(err, "Bucket")
+		return
+	}
+	// write "Prefix"
+	err = en.Append(0xa6, 0x50, 0x72, 0x65, 0x66, 0x69, 0x78)
+	if err != nil {
+		return
+	}
+	err = en.WriteString(z.Prefix)
+	if err != nil {
+		err = msgp.WrapError(err, "Prefix")
+		return
+	}
+	// write "NotificationCfg"
+	err = en.Append(0xaf, 0x4e, 0x6f, 0x74, 0x69, 0x66, 0x69, 0x63, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x43, 0x66, 0x67)
+	if err != nil {
+		return
+	}
+	err = z.NotificationCfg.EncodeMsg(en)
+	if err != nil {
+		err = msgp.WrapError(err, "NotificationCfg")
+		return
+	}
+	// write "Retry"
+	err = en.Append(0xa5, 0x52, 0x65, 0x74, 0x72, 0x79)
+	if err != nil {
+		return
+	}
+	err = z.Retry.EncodeMsg(en)
+	if err != nil {
+		err = msgp.WrapError(err, "Retry")
+		return
+	}
+	// write "Rules"
+	err = en.Append(0xa5, 0x52, 0x75, 0x6c, 0x65, 0x73)
+	if err != nil {
+		return
+	}
+	err = en.WriteArrayHeader(uint32(len(z.Rules)))
+	if err != nil {
+		err = msgp.WrapError(err, "Rules")
+		return
+	}
+	for za0001 := range z.Rules {
+		err = z.Rules[za0001].EncodeMsg(en)
+		if err != nil {
+			err = msgp.WrapError(err, "Rules", za0001)
+			return
+		}
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z *BatchJobExpire) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 6
+	// string "APIVersion"
+	o = append(o, 0x86, 0xaa, 0x41, 0x50, 0x49, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e)
+	o = msgp.AppendString(o, z.APIVersion)
+	// string "Bucket"
+	o = append(o, 0xa6, 0x42, 0x75, 0x63, 0x6b, 0x65, 0x74)
+	o = msgp.AppendString(o, z.Bucket)
+	// string "Prefix"
+	o = append(o, 0xa6, 0x50, 0x72, 0x65, 0x66, 0x69, 0x78)
+	o = msgp.AppendString(o, z.Prefix)
+	// string "NotificationCfg"
+	o = append(o, 0xaf, 0x4e, 0x6f, 0x74, 0x69, 0x66, 0x69, 0x63, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x43, 0x66, 0x67)
+	o, err = z.NotificationCfg.MarshalMsg(o)
+	if err != nil {
+		err = msgp.WrapError(err, "NotificationCfg")
+		return
+	}
+	// string "Retry"
+	o = append(o, 0xa5, 0x52, 0x65, 0x74, 0x72, 0x79)
+	o, err = z.Retry.MarshalMsg(o)
+	if err != nil {
+		err = msgp.WrapError(err, "Retry")
+		return
+	}
+	// string "Rules"
+	o = append(o, 0xa5, 0x52, 0x75, 0x6c, 0x65, 0x73)
+	o = msgp.AppendArrayHeader(o, uint32(len(z.Rules)))
+	for za0001 := range z.Rules {
+		o, err = z.Rules[za0001].MarshalMsg(o)
+		if err != nil {
+			err = msgp.WrapError(err, "Rules", za0001)
+			return
+		}
+	}
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *BatchJobExpire) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadMapHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, bts, err = msgp.ReadMapKeyZC(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "APIVersion":
+			z.APIVersion, bts, err = msgp.ReadStringBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "APIVersion")
+				return
+			}
+		case "Bucket":
+			z.Bucket, bts, err = msgp.ReadStringBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Bucket")
+				return
+			}
+		case "Prefix":
+			z.Prefix, bts, err = msgp.ReadStringBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Prefix")
+				return
+			}
+		case "NotificationCfg":
+			bts, err = z.NotificationCfg.UnmarshalMsg(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "NotificationCfg")
+				return
+			}
+		case "Retry":
+			bts, err = z.Retry.UnmarshalMsg(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Retry")
+				return
+			}
+		case "Rules":
+			var zb0002 uint32
+			zb0002, bts, err = msgp.ReadArrayHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Rules")
+				return
+			}
+			if cap(z.Rules) >= int(zb0002) {
+				z.Rules = (z.Rules)[:zb0002]
+			} else {
+				z.Rules = make([]BatchJobExpireFilter, zb0002)
+			}
+			for za0001 := range z.Rules {
+				bts, err = z.Rules[za0001].UnmarshalMsg(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Rules", za0001)
+					return
+				}
+			}
+		default:
+			bts, err = msgp.Skip(bts)
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z *BatchJobExpire) Msgsize() (s int) {
+	s = 1 + 11 + msgp.StringPrefixSize + len(z.APIVersion) + 7 + msgp.StringPrefixSize + len(z.Bucket) + 7 + msgp.StringPrefixSize + len(z.Prefix) + 16 + z.NotificationCfg.Msgsize() + 6 + z.Retry.Msgsize() + 6 + msgp.ArrayHeaderSize
+	for za0001 := range z.Rules {
+		s += z.Rules[za0001].Msgsize()
+	}
+	return
+}
+
+// DecodeMsg implements msgp.Decodable
+func (z *BatchJobExpireFilter) DecodeMsg(dc *msgp.Reader) (err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, err = dc.ReadMapHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, err = dc.ReadMapKeyPtr()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "OlderThan":
+			z.OlderThan, err = dc.ReadDuration()
+			if err != nil {
+				err = msgp.WrapError(err, "OlderThan")
+				return
+			}
+		case "CreatedBefore":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "CreatedBefore")
+					return
+				}
+				z.CreatedBefore = nil
+			} else {
+				if z.CreatedBefore == nil {
+					z.CreatedBefore = new(time.Time)
+				}
+				*z.CreatedBefore, err = dc.ReadTime()
+				if err != nil {
+					err = msgp.WrapError(err, "CreatedBefore")
+					return
+				}
+			}
+		case "Tags":
+			var zb0002 uint32
+			zb0002, err = dc.ReadArrayHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "Tags")
+				return
+			}
+			if cap(z.Tags) >= int(zb0002) {
+				z.Tags = (z.Tags)[:zb0002]
+			} else {
+				z.Tags = make([]BatchJobKV, zb0002)
+			}
+			for za0001 := range z.Tags {
+				err = z.Tags[za0001].DecodeMsg(dc)
+				if err != nil {
+					err = msgp.WrapError(err, "Tags", za0001)
+					return
+				}
+			}
+		case "Metadata":
+			var zb0003 uint32
+			zb0003, err = dc.ReadArrayHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "Metadata")
+				return
+			}
+			if cap(z.Metadata) >= int(zb0003) {
+				z.Metadata = (z.Metadata)[:zb0003]
+			} else {
+				z.Metadata = make([]BatchJobKV, zb0003)
+			}
+			for za0002 := range z.Metadata {
+				err = z.Metadata[za0002].DecodeMsg(dc)
+				if err != nil {
+					err = msgp.WrapError(err, "Metadata", za0002)
+					return
+				}
+			}
+		case "Size":
+			err = z.Size.DecodeMsg(dc)
+			if err != nil {
+				err = msgp.WrapError(err, "Size")
+				return
+			}
+		case "Type":
+			z.Type, err = dc.ReadString()
+			if err != nil {
+				err = msgp.WrapError(err, "Type")
+				return
+			}
+		case "Name":
+			z.Name, err = dc.ReadString()
+			if err != nil {
+				err = msgp.WrapError(err, "Name")
+				return
+			}
+		case "Purge":
+			var zb0004 uint32
+			zb0004, err = dc.ReadMapHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "Purge")
+				return
+			}
+			for zb0004 > 0 {
+				zb0004--
+				field, err = dc.ReadMapKeyPtr()
+				if err != nil {
+					err = msgp.WrapError(err, "Purge")
+					return
+				}
+				switch msgp.UnsafeString(field) {
+				case "RetainVersions":
+					z.Purge.RetainVersions, err = dc.ReadInt()
+					if err != nil {
+						err = msgp.WrapError(err, "Purge", "RetainVersions")
+						return
+					}
+				default:
+					err = dc.Skip()
+					if err != nil {
+						err = msgp.WrapError(err, "Purge")
+						return
+					}
+				}
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z *BatchJobExpireFilter) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 8
+	// write "OlderThan"
+	err = en.Append(0x88, 0xa9, 0x4f, 0x6c, 0x64, 0x65, 0x72, 0x54, 0x68, 0x61, 0x6e)
+	if err != nil {
+		return
+	}
+	err = en.WriteDuration(z.OlderThan)
+	if err != nil {
+		err = msgp.WrapError(err, "OlderThan")
+		return
+	}
+	// write "CreatedBefore"
+	err = en.Append(0xad, 0x43, 0x72, 0x65, 0x61, 0x74, 0x65, 0x64, 0x42, 0x65, 0x66, 0x6f, 0x72, 0x65)
+	if err != nil {
+		return
+	}
+	if z.CreatedBefore == nil {
+		err = en.WriteNil()
+		if err != nil {
+			return
+		}
+	} else {
+		err = en.WriteTime(*z.CreatedBefore)
+		if err != nil {
+			err = msgp.WrapError(err, "CreatedBefore")
+			return
+		}
+	}
+	// write "Tags"
+	err = en.Append(0xa4, 0x54, 0x61, 0x67, 0x73)
+	if err != nil {
+		return
+	}
+	err = en.WriteArrayHeader(uint32(len(z.Tags)))
+	if err != nil {
+		err = msgp.WrapError(err, "Tags")
+		return
+	}
+	for za0001 := range z.Tags {
+		err = z.Tags[za0001].EncodeMsg(en)
+		if err != nil {
+			err = msgp.WrapError(err, "Tags", za0001)
+			return
+		}
+	}
+	// write "Metadata"
+	err = en.Append(0xa8, 0x4d, 0x65, 0x74, 0x61, 0x64, 0x61, 0x74, 0x61)
+	if err != nil {
+		return
+	}
+	err = en.WriteArrayHeader(uint32(len(z.Metadata)))
+	if err != nil {
+		err = msgp.WrapError(err, "Metadata")
+		return
+	}
+	for za0002 := range z.Metadata {
+		err = z.Metadata[za0002].EncodeMsg(en)
+		if err != nil {
+			err = msgp.WrapError(err, "Metadata", za0002)
+			return
+		}
+	}
+	// write "Size"
+	err = en.Append(0xa4, 0x53, 0x69, 0x7a, 0x65)
+	if err != nil {
+		return
+	}
+	err = z.Size.EncodeMsg(en)
+	if err != nil {
+		err = msgp.WrapError(err, "Size")
+		return
+	}
+	// write "Type"
+	err = en.Append(0xa4, 0x54, 0x79, 0x70, 0x65)
+	if err != nil {
+		return
+	}
+	err = en.WriteString(z.Type)
+	if err != nil {
+		err = msgp.WrapError(err, "Type")
+		return
+	}
+	// write "Name"
+	err = en.Append(0xa4, 0x4e, 0x61, 0x6d, 0x65)
+	if err != nil {
+		return
+	}
+	err = en.WriteString(z.Name)
+	if err != nil {
+		err = msgp.WrapError(err, "Name")
+		return
+	}
+	// write "Purge"
+	err = en.Append(0xa5, 0x50, 0x75, 0x72, 0x67, 0x65)
+	if err != nil {
+		return
+	}
+	// map header, size 1
+	// write "RetainVersions"
+	err = en.Append(0x81, 0xae, 0x52, 0x65, 0x74, 0x61, 0x69, 0x6e, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x73)
+	if err != nil {
+		return
+	}
+	err = en.WriteInt(z.Purge.RetainVersions)
+	if err != nil {
+		err = msgp.WrapError(err, "Purge", "RetainVersions")
+		return
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z *BatchJobExpireFilter) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 8
+	// string "OlderThan"
+	o = append(o, 0x88, 0xa9, 0x4f, 0x6c, 0x64, 0x65, 0x72, 0x54, 0x68, 0x61, 0x6e)
+	o = msgp.AppendDuration(o, z.OlderThan)
+	// string "CreatedBefore"
+	o = append(o, 0xad, 0x43, 0x72, 0x65, 0x61, 0x74, 0x65, 0x64, 0x42, 0x65, 0x66, 0x6f, 0x72, 0x65)
+	if z.CreatedBefore == nil {
+		o = msgp.AppendNil(o)
+	} else {
+		o = msgp.AppendTime(o, *z.CreatedBefore)
+	}
+	// string "Tags"
+	o = append(o, 0xa4, 0x54, 0x61, 0x67, 0x73)
+	o = msgp.AppendArrayHeader(o, uint32(len(z.Tags)))
+	for za0001 := range z.Tags {
+		o, err = z.Tags[za0001].MarshalMsg(o)
+		if err != nil {
+			err = msgp.WrapError(err, "Tags", za0001)
+			return
+		}
+	}
+	// string "Metadata"
+	o = append(o, 0xa8, 0x4d, 0x65, 0x74, 0x61, 0x64, 0x61, 0x74, 0x61)
+	o = msgp.AppendArrayHeader(o, uint32(len(z.Metadata)))
+	for za0002 := range z.Metadata {
+		o, err = z.Metadata[za0002].MarshalMsg(o)
+		if err != nil {
+			err = msgp.WrapError(err, "Metadata", za0002)
+			return
+		}
+	}
+	// string "Size"
+	o = append(o, 0xa4, 0x53, 0x69, 0x7a, 0x65)
+	o, err = z.Size.MarshalMsg(o)
+	if err != nil {
+		err = msgp.WrapError(err, "Size")
+		return
+	}
+	// string "Type"
+	o = append(o, 0xa4, 0x54, 0x79, 0x70, 0x65)
+	o = msgp.AppendString(o, z.Type)
+	// string "Name"
+	o = append(o, 0xa4, 0x4e, 0x61, 0x6d, 0x65)
+	o = msgp.AppendString(o, z.Name)
+	// string "Purge"
+	o = append(o, 0xa5, 0x50, 0x75, 0x72, 0x67, 0x65)
+	// map header, size 1
+	// string "RetainVersions"
+	o = append(o, 0x81, 0xae, 0x52, 0x65, 0x74, 0x61, 0x69, 0x6e, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x73)
+	o = msgp.AppendInt(o, z.Purge.RetainVersions)
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *BatchJobExpireFilter) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadMapHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, bts, err = msgp.ReadMapKeyZC(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "OlderThan":
+			z.OlderThan, bts, err = msgp.ReadDurationBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "OlderThan")
+				return
+			}
+		case "CreatedBefore":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.CreatedBefore = nil
+			} else {
+				if z.CreatedBefore == nil {
+					z.CreatedBefore = new(time.Time)
+				}
+				*z.CreatedBefore, bts, err = msgp.ReadTimeBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "CreatedBefore")
+					return
+				}
+			}
+		case "Tags":
+			var zb0002 uint32
+			zb0002, bts, err = msgp.ReadArrayHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Tags")
+				return
+			}
+			if cap(z.Tags) >= int(zb0002) {
+				z.Tags = (z.Tags)[:zb0002]
+			} else {
+				z.Tags = make([]BatchJobKV, zb0002)
+			}
+			for za0001 := range z.Tags {
+				bts, err = z.Tags[za0001].UnmarshalMsg(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Tags", za0001)
+					return
+				}
+			}
+		case "Metadata":
+			var zb0003 uint32
+			zb0003, bts, err = msgp.ReadArrayHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Metadata")
+				return
+			}
+			if cap(z.Metadata) >= int(zb0003) {
+				z.Metadata = (z.Metadata)[:zb0003]
+			} else {
+				z.Metadata = make([]BatchJobKV, zb0003)
+			}
+			for za0002 := range z.Metadata {
+				bts, err = z.Metadata[za0002].UnmarshalMsg(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Metadata", za0002)
+					return
+				}
+			}
+		case "Size":
+			bts, err = z.Size.UnmarshalMsg(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Size")
+				return
+			}
+		case "Type":
+			z.Type, bts, err = msgp.ReadStringBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Type")
+				return
+			}
+		case "Name":
+			z.Name, bts, err = msgp.ReadStringBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Name")
+				return
+			}
+		case "Purge":
+			var zb0004 uint32
+			zb0004, bts, err = msgp.ReadMapHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Purge")
+				return
+			}
+			for zb0004 > 0 {
+				zb0004--
+				field, bts, err = msgp.ReadMapKeyZC(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Purge")
+					return
+				}
+				switch msgp.UnsafeString(field) {
+				case "RetainVersions":
+					z.Purge.RetainVersions, bts, err = msgp.ReadIntBytes(bts)
+					if err != nil {
+						err = msgp.WrapError(err, "Purge", "RetainVersions")
+						return
+					}
+				default:
+					bts, err = msgp.Skip(bts)
+					if err != nil {
+						err = msgp.WrapError(err, "Purge")
+						return
+					}
+				}
+			}
+		default:
+			bts, err = msgp.Skip(bts)
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z *BatchJobExpireFilter) Msgsize() (s int) {
+	s = 1 + 10 + msgp.DurationSize + 14
+	if z.CreatedBefore == nil {
+		s += msgp.NilSize
+	} else {
+		s += msgp.TimeSize
+	}
+	s += 5 + msgp.ArrayHeaderSize
+	for za0001 := range z.Tags {
+		s += z.Tags[za0001].Msgsize()
+	}
+	s += 9 + msgp.ArrayHeaderSize
+	for za0002 := range z.Metadata {
+		s += z.Metadata[za0002].Msgsize()
+	}
+	s += 5 + z.Size.Msgsize() + 5 + msgp.StringPrefixSize + len(z.Type) + 5 + msgp.StringPrefixSize + len(z.Name) + 6 + 1 + 15 + msgp.IntSize
+	return
+}
+
+// DecodeMsg implements msgp.Decodable
+func (z *BatchJobExpirePurge) DecodeMsg(dc *msgp.Reader) (err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, err = dc.ReadMapHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, err = dc.ReadMapKeyPtr()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "RetainVersions":
+			z.RetainVersions, err = dc.ReadInt()
+			if err != nil {
+				err = msgp.WrapError(err, "RetainVersions")
+				return
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z BatchJobExpirePurge) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 1
+	// write "RetainVersions"
+	err = en.Append(0x81, 0xae, 0x52, 0x65, 0x74, 0x61, 0x69, 0x6e, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x73)
+	if err != nil {
+		return
+	}
+	err = en.WriteInt(z.RetainVersions)
+	if err != nil {
+		err = msgp.WrapError(err, "RetainVersions")
+		return
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z BatchJobExpirePurge) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 1
+	// string "RetainVersions"
+	o = append(o, 0x81, 0xae, 0x52, 0x65, 0x74, 0x61, 0x69, 0x6e, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x73)
+	o = msgp.AppendInt(o, z.RetainVersions)
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *BatchJobExpirePurge) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadMapHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, bts, err = msgp.ReadMapKeyZC(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "RetainVersions":
+			z.RetainVersions, bts, err = msgp.ReadIntBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "RetainVersions")
+				return
+			}
+		default:
+			bts, err = msgp.Skip(bts)
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z BatchJobExpirePurge) Msgsize() (s int) {
+	s = 1 + 15 + msgp.IntSize
+	return
+}
@@ -0,0 +1,349 @@
+package cmd
+
+// Code generated by github.com/tinylib/msgp DO NOT EDIT.
+
+import (
+	"bytes"
+	"testing"
+
+	"github.com/tinylib/msgp/msgp"
+)
+
+func TestMarshalUnmarshalBatchJobExpire(t *testing.T) {
+	v := BatchJobExpire{}
+	bts, err := v.MarshalMsg(nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	left, err := v.UnmarshalMsg(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after UnmarshalMsg(): %q", len(left), left)
+	}
+
+	left, err = msgp.Skip(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after Skip(): %q", len(left), left)
+	}
+}
+
+func BenchmarkMarshalMsgBatchJobExpire(b *testing.B) {
+	v := BatchJobExpire{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgBatchJobExpire(b *testing.B) {
+	v := BatchJobExpire{}
+	bts := make([]byte, 0, v.Msgsize())
+	bts, _ = v.MarshalMsg(bts[0:0])
+	b.SetBytes(int64(len(bts)))
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		bts, _ = v.MarshalMsg(bts[0:0])
+	}
+}
+
+func BenchmarkUnmarshalBatchJobExpire(b *testing.B) {
+	v := BatchJobExpire{}
+	bts, _ := v.MarshalMsg(nil)
+	b.ReportAllocs()
+	b.SetBytes(int64(len(bts)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		_, err := v.UnmarshalMsg(bts)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestEncodeDecodeBatchJobExpire(t *testing.T) {
+	v := BatchJobExpire{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeBatchJobExpire Msgsize() is inaccurate")
+	}
+
+	vn := BatchJobExpire{}
+	err := msgp.Decode(&buf, &vn)
+	if err != nil {
+		t.Error(err)
+	}
+
+	buf.Reset()
+	msgp.Encode(&buf, &v)
+	err = msgp.NewReader(&buf).Skip()
+	if err != nil {
+		t.Error(err)
+	}
+}
+
+func BenchmarkEncodeBatchJobExpire(b *testing.B) {
+	v := BatchJobExpire{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	en := msgp.NewWriter(msgp.Nowhere)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.EncodeMsg(en)
+	}
+	en.Flush()
+}
+
+func BenchmarkDecodeBatchJobExpire(b *testing.B) {
+	v := BatchJobExpire{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	rd := msgp.NewEndlessReader(buf.Bytes(), b)
+	dc := msgp.NewReader(rd)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		err := v.DecodeMsg(dc)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestMarshalUnmarshalBatchJobExpireFilter(t *testing.T) {
+	v := BatchJobExpireFilter{}
+	bts, err := v.MarshalMsg(nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	left, err := v.UnmarshalMsg(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after UnmarshalMsg(): %q", len(left), left)
+	}
+
+	left, err = msgp.Skip(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after Skip(): %q", len(left), left)
+	}
+}
+
+func BenchmarkMarshalMsgBatchJobExpireFilter(b *testing.B) {
+	v := BatchJobExpireFilter{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgBatchJobExpireFilter(b *testing.B) {
+	v := BatchJobExpireFilter{}
+	bts := make([]byte, 0, v.Msgsize())
+	bts, _ = v.MarshalMsg(bts[0:0])
+	b.SetBytes(int64(len(bts)))
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		bts, _ = v.MarshalMsg(bts[0:0])
+	}
+}
+
+func BenchmarkUnmarshalBatchJobExpireFilter(b *testing.B) {
+	v := BatchJobExpireFilter{}
+	bts, _ := v.MarshalMsg(nil)
+	b.ReportAllocs()
+	b.SetBytes(int64(len(bts)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		_, err := v.UnmarshalMsg(bts)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestEncodeDecodeBatchJobExpireFilter(t *testing.T) {
+	v := BatchJobExpireFilter{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeBatchJobExpireFilter Msgsize() is inaccurate")
+	}
+
+	vn := BatchJobExpireFilter{}
+	err := msgp.Decode(&buf, &vn)
+	if err != nil {
+		t.Error(err)
+	}
+
+	buf.Reset()
+	msgp.Encode(&buf, &v)
+	err = msgp.NewReader(&buf).Skip()
+	if err != nil {
+		t.Error(err)
+	}
+}
+
+func BenchmarkEncodeBatchJobExpireFilter(b *testing.B) {
+	v := BatchJobExpireFilter{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	en := msgp.NewWriter(msgp.Nowhere)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.EncodeMsg(en)
+	}
+	en.Flush()
+}
+
+func BenchmarkDecodeBatchJobExpireFilter(b *testing.B) {
+	v := BatchJobExpireFilter{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	rd := msgp.NewEndlessReader(buf.Bytes(), b)
+	dc := msgp.NewReader(rd)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		err := v.DecodeMsg(dc)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestMarshalUnmarshalBatchJobExpirePurge(t *testing.T) {
+	v := BatchJobExpirePurge{}
+	bts, err := v.MarshalMsg(nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	left, err := v.UnmarshalMsg(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after UnmarshalMsg(): %q", len(left), left)
+	}
+
+	left, err = msgp.Skip(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after Skip(): %q", len(left), left)
+	}
+}
+
+func BenchmarkMarshalMsgBatchJobExpirePurge(b *testing.B) {
+	v := BatchJobExpirePurge{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgBatchJobExpirePurge(b *testing.B) {
+	v := BatchJobExpirePurge{}
+	bts := make([]byte, 0, v.Msgsize())
+	bts, _ = v.MarshalMsg(bts[0:0])
+	b.SetBytes(int64(len(bts)))
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		bts, _ = v.MarshalMsg(bts[0:0])
+	}
+}
+
+func BenchmarkUnmarshalBatchJobExpirePurge(b *testing.B) {
+	v := BatchJobExpirePurge{}
+	bts, _ := v.MarshalMsg(nil)
+	b.ReportAllocs()
+	b.SetBytes(int64(len(bts)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		_, err := v.UnmarshalMsg(bts)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestEncodeDecodeBatchJobExpirePurge(t *testing.T) {
+	v := BatchJobExpirePurge{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeBatchJobExpirePurge Msgsize() is inaccurate")
+	}
+
+	vn := BatchJobExpirePurge{}
+	err := msgp.Decode(&buf, &vn)
+	if err != nil {
+		t.Error(err)
+	}
+
+	buf.Reset()
+	msgp.Encode(&buf, &v)
+	err = msgp.NewReader(&buf).Skip()
+	if err != nil {
+		t.Error(err)
+	}
+}
+
+func BenchmarkEncodeBatchJobExpirePurge(b *testing.B) {
+	v := BatchJobExpirePurge{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	en := msgp.NewWriter(msgp.Nowhere)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.EncodeMsg(en)
+	}
+	en.Flush()
+}
+
+func BenchmarkDecodeBatchJobExpirePurge(b *testing.B) {
+	v := BatchJobExpirePurge{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	rd := msgp.NewEndlessReader(buf.Bytes(), b)
+	dc := msgp.NewReader(rd)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		err := v.DecodeMsg(dc)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
@@ -0,0 +1,71 @@
+// Copyright (c) 2015-2023 MinIO, Inc.
+//
+// This file is part of MinIO Object Storage stack
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package cmd
+
+import (
+	"testing"
+
+	"gopkg.in/yaml.v2"
+)
+
+func TestParseBatchJobExpire(t *testing.T) {
+	expireYaml := `
+expire: # Expire objects that match a condition
+  apiVersion: v1
+  bucket: mybucket # Bucket where this batch job will expire matching objects from
+  prefix: myprefix # (Optional) Prefix under which this job will expire objects matching the rules below.
+  rules:
+    - type: object  # regular objects with zero ore more older versions
+      name: NAME # match object names that satisfy the wildcard expression.
+      olderThan: 70h # match objects older than this value
+      createdBefore: "2006-01-02T15:04:05.00Z" # match objects created before "date"
+      tags:
+        - key: name
+          value: pick* # match objects with tag 'name', all values starting with 'pick'
+      metadata:
+        - key: content-type
+          value: image/* # match objects with 'content-type', all values starting with 'image/'
+      size:
+        lessThan: "10MiB" # match objects with size less than this value (e.g. 10MiB)
+        greaterThan: 1MiB # match objects with size greater than this value (e.g. 1MiB)
+      purge:
+          # retainVersions: 0 # (default) delete all versions of the object. This option is the fastest.
+          # retainVersions: 5 # keep the latest 5 versions of the object.
+  
+    - type: deleted # objects with delete marker as their latest version
+      name: NAME # match object names that satisfy the wildcard expression.
+      olderThan: 10h # match objects older than this value (e.g. 7d10h31s)
+      createdBefore: "2006-01-02T15:04:05.00Z" # match objects created before "date"
+      purge:
+          # retainVersions: 0 # (default) delete all versions of the object. This option is the fastest.
+          # retainVersions: 5 # keep the latest 5 versions of the object including delete markers.
+
+  notify:
+    endpoint: https://notify.endpoint # notification endpoint to receive job completion status
+    token: Bearer xxxxx # optional authentication token for the notification endpoint
+  
+  retry:
+    attempts: 10 # number of retries for the job before giving up
+    delay: 500ms # least amount of delay between each retry
+`
+	var job BatchJobRequest
+	err := yaml.UnmarshalStrict([]byte(expireYaml), &job)
+	if err != nil {
+		t.Fatal("Failed to parse batch-job-expire yaml", err)
+	}
+}
@@ -42,6 +42,7 @@ import (
 	"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/internal/config/batch"
 	"github.com/minio/minio/internal/crypto"
 	"github.com/minio/minio/internal/hash"
 	xhttp "github.com/minio/minio/internal/http"
@@ -54,6 +55,8 @@ import (
 	"gopkg.in/yaml.v2"
 )

+var globalBatchConfig batch.Config
+
 // BatchJobRequest this is an internal data structure not for external consumption.
 type BatchJobRequest struct {
 	ID        string               `yaml:"-" json:"name"`
@@ -62,6 +65,7 @@ type BatchJobRequest struct {
 	Location  string               `yaml:"-" json:"location"`
 	Replicate *BatchJobReplicateV1 `yaml:"replicate" json:"replicate"`
 	KeyRotate *BatchJobKeyRotateV1 `yaml:"keyrotate" json:"keyrotate"`
+	Expire    *BatchJobExpire      `yaml:"expire" json:"expire"`
 	ctx       context.Context      `msg:"-"`
 }

@@ -163,7 +167,7 @@ func (r *BatchJobReplicateV1) ReplicateFromSource(ctx context.Context, api Objec
 	}
 	rd, objInfo, _, err := core.GetObject(ctx, srcBucket, srcObject, gopts)
 	if err != nil {
-		return err
+		return ErrorRespToObjectError(err, srcBucket, srcObject, srcObjInfo.VersionID)
 	}
 	defer rd.Close()

@@ -226,7 +230,7 @@ func (r *BatchJobReplicateV1) copyWithMultipartfromSource(ctx context.Context, a
 		}
 		rd, objInfo, _, err := c.GetObject(ctx, srcBucket, srcObject, gopts)
 		if err != nil {
-			return err
+			return ErrorRespToObjectError(err, srcBucket, srcObject, srcObjInfo.VersionID)
 		}
 		defer rd.Close()

@@ -262,6 +266,9 @@ func (r *BatchJobReplicateV1) StartFromSource(ctx context.Context, api ObjectLay
 	if err := ri.load(ctx, api, job); err != nil {
 		return err
 	}
+	if ri.Complete {
+		return nil
+	}
 	globalBatchJobsMetrics.save(job.ID, ri)

 	delay := job.Replicate.Flags.Retry.Delay
@@ -428,23 +435,27 @@ func (r *BatchJobReplicateV1) StartFromSource(ctx context.Context, api ObjectLay
 			wk.Take()
 			go func() {
 				defer wk.Give()
-				stopFn := globalBatchJobsMetrics.trace(batchReplicationMetricObject, job.ID, attempts, oi)
+				stopFn := globalBatchJobsMetrics.trace(batchJobMetricReplication, job.ID, attempts)
 				success := true
 				if err := r.ReplicateFromSource(ctx, api, core, oi, retry); err != nil {
 					// object must be deleted concurrently, allow these failures but do not count them
 					if isErrVersionNotFound(err) || isErrObjectNotFound(err) {
 						return
 					}
-					stopFn(err)
+					stopFn(oi, err)
 					logger.LogIf(ctx, err)
 					success = false
 				} else {
-					stopFn(nil)
+					stopFn(oi, nil)
 				}
 				ri.trackCurrentBucketObject(r.Target.Bucket, oi, success)
 				globalBatchJobsMetrics.save(job.ID, ri)
 				// persist in-memory state to disk after every 10secs.
 				logger.LogIf(ctx, ri.updateAfter(ctx, api, 10*time.Second, job))
+
+				if wait := globalBatchConfig.ReplicationWait(); wait > 0 {
+					time.Sleep(wait)
+				}
 			}()
 		}
 		wk.Wait()
@@ -517,9 +528,68 @@ func toObjectInfo(bucket, object string, objInfo miniogo.ObjectInfo) ObjectInfo
 		oi.UserDefined[xhttp.AmzStorageClass] = objInfo.StorageClass
 	}

+	for k, v := range objInfo.UserMetadata {
+		oi.UserDefined[k] = v
+	}
+
 	return oi
 }

+func (r BatchJobReplicateV1) writeAsArchive(ctx context.Context, objAPI ObjectLayer, remoteClnt *minio.Client, entries []ObjectInfo) error {
+	input := make(chan minio.SnowballObject, 1)
+	opts := minio.SnowballOptions{
+		Opts:     minio.PutObjectOptions{},
+		InMemory: *r.Source.Snowball.InMemory,
+		Compress: *r.Source.Snowball.Compress,
+		SkipErrs: *r.Source.Snowball.SkipErrs,
+	}
+
+	go func() {
+		defer close(input)
+
+		for _, entry := range entries {
+			gr, err := objAPI.GetObjectNInfo(ctx, r.Source.Bucket,
+				entry.Name, nil, nil, ObjectOptions{
+					VersionID: entry.VersionID,
+				})
+			if err != nil {
+				logger.LogIf(ctx, err)
+				continue
+			}
+
+			snowballObj := minio.SnowballObject{
+				// Create path to store objects within the bucket.
+				Key:       entry.Name,
+				Size:      entry.Size,
+				ModTime:   entry.ModTime,
+				VersionID: entry.VersionID,
+				Content:   gr,
+				Headers:   make(http.Header),
+				Close: func() {
+					gr.Close()
+				},
+			}
+
+			opts, err := batchReplicationOpts(ctx, "", gr.ObjInfo)
+			if err != nil {
+				logger.LogIf(ctx, err)
+				continue
+			}
+
+			for k, vals := range opts.Header() {
+				for _, v := range vals {
+					snowballObj.Headers.Add(k, v)
+				}
+			}
+
+			input <- snowballObj
+		}
+	}()
+
+	// Collect and upload all entries.
+	return remoteClnt.PutObjectsSnowball(ctx, r.Target.Bucket, opts, input)
+}
+
 // ReplicateToTarget read from source and replicate to configured target
 func (r *BatchJobReplicateV1) ReplicateToTarget(ctx context.Context, api ObjectLayer, c *miniogo.Core, srcObjInfo ObjectInfo, retry bool) error {
 	srcBucket := r.Source.Bucket
@@ -663,7 +733,12 @@ func (ri *batchJobInfo) load(ctx context.Context, api ObjectLayer, job BatchJobR
 		fileName = batchKeyRotationName
 		version = batchKeyRotateVersionV1
 		format = batchKeyRotationFormat
-
+	case job.Expire != nil:
+		fileName = batchExpireName
+		version = batchExpireVersionV1
+		format = batchExpireFormat
+	default:
+		return errors.New("no supported batch job request specified")
 	}
 	data, err := readConfig(ctx, api, pathJoin(job.Location, fileName))
 	if err != nil {
@@ -680,6 +755,11 @@ func (ri *batchJobInfo) load(ctx context.Context, api ObjectLayer, job BatchJobR
 				if job.KeyRotate.Flags.Retry.Attempts > 0 {
 					ri.RetryAttempts = job.KeyRotate.Flags.Retry.Attempts
 				}
+			case job.Expire != nil:
+				ri.RetryAttempts = batchExpireJobDefaultRetries
+				if job.Expire.Retry.Attempts > 0 {
+					ri.RetryAttempts = job.Expire.Retry.Attempts
+				}
 			}
 			return nil
 		}
@@ -789,6 +869,12 @@ func (ri *batchJobInfo) updateAfter(ctx context.Context, api ObjectLayer, durati
 			jobTyp = string(job.Type())
 			fileName = batchKeyRotationName
 			ri.Version = batchKeyRotateVersionV1
+		case madmin.BatchJobExpire:
+			format = batchExpireFormat
+			version = batchExpireVersion
+			jobTyp = string(job.Type())
+			fileName = batchExpireName
+			ri.Version = batchExpireVersionV1
 		default:
 			return errInvalidArgument
 		}
@@ -814,7 +900,18 @@ func (ri *batchJobInfo) updateAfter(ctx context.Context, api ObjectLayer, durati
 	return nil
 }

-func (ri *batchJobInfo) trackCurrentBucketObject(bucket string, info ObjectInfo, failed bool) {
+// Note: to be used only with batch jobs that affect multiple versions through
+// a single action. e.g batch-expire has an option to expire all versions of an
+// object which matches the given filters.
+func (ri *batchJobInfo) trackMultipleObjectVersions(bucket string, info ObjectInfo, success bool) {
+	if success {
+		ri.Objects += int64(info.NumVersions)
+	} else {
+		ri.ObjectsFailed += int64(info.NumVersions)
+	}
+}
+
+func (ri *batchJobInfo) trackCurrentBucketObject(bucket string, info ObjectInfo, success bool) {
 	if ri == nil {
 		return
 	}
@@ -824,7 +921,7 @@ func (ri *batchJobInfo) trackCurrentBucketObject(bucket string, info ObjectInfo,

 	ri.Bucket = bucket
 	ri.Object = info.Name
-	ri.countItem(info.Size, info.DeleteMarker, failed)
+	ri.countItem(info.Size, info.DeleteMarker, success)
 }

 // Start start the batch replication job, resumes if there was a pending job via "job.ID"
@@ -837,6 +934,9 @@ func (r *BatchJobReplicateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 	if err := ri.load(ctx, api, job); err != nil {
 		return err
 	}
+	if ri.Complete {
+		return nil
+	}
 	globalBatchJobsMetrics.save(job.ID, ri)
 	lastObject := ri.Object

@@ -907,12 +1007,9 @@ func (r *BatchJobReplicateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 			// None of the provided metadata filters match skip the object.
 			return false
 		}
-		// if one of source or target is non MinIO, just replicate the top most version like `mc mirror`
-		if (r.Target.Type == BatchJobReplicateResourceS3 || r.Source.Type == BatchJobReplicateResourceS3) && !info.IsLatest {
-			return false
-		}

-		return true
+		// if one of source or target is non MinIO, just replicate the top most version like `mc mirror`
+		return !((r.Target.Type == BatchJobReplicateResourceS3 || r.Source.Type == BatchJobReplicateResourceS3) && !info.IsLatest)
 	}

 	u, err := url.Parse(r.Target.Endpoint)
@@ -931,8 +1028,73 @@ func (r *BatchJobReplicateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 	if err != nil {
 		return err
 	}
+
 	c.SetAppInfo("minio-"+batchJobPrefix, r.APIVersion+" "+job.ID)

+	var (
+		walkCh = make(chan ObjectInfo, 100)
+		slowCh = make(chan ObjectInfo, 100)
+	)
+
+	if !*r.Source.Snowball.Disable && r.Source.Type.isMinio() && r.Target.Type.isMinio() {
+		go func() {
+			defer close(slowCh)
+
+			// Snowball currently needs the high level minio-go Client, not the Core one
+			cl, err := miniogo.New(u.Host, &miniogo.Options{
+				Creds:        credentials.NewStaticV4(cred.AccessKey, cred.SecretKey, cred.SessionToken),
+				Secure:       u.Scheme == "https",
+				Transport:    getRemoteInstanceTransport,
+				BucketLookup: lookupStyle(r.Target.Path),
+			})
+			if err != nil {
+				logger.LogIf(ctx, err)
+				return
+			}
+
+			// Already validated before arriving here
+			smallerThan, _ := humanize.ParseBytes(*r.Source.Snowball.SmallerThan)
+
+			var (
+				obj   = ObjectInfo{}
+				batch = make([]ObjectInfo, 0, *r.Source.Snowball.Batch)
+				valid = true
+			)
+
+			for valid {
+				obj, valid = <-walkCh
+
+				if !valid {
+					goto write
+				}
+
+				if obj.DeleteMarker || !obj.VersionPurgeStatus.Empty() || obj.Size >= int64(smallerThan) {
+					slowCh <- obj
+					continue
+				}
+
+				batch = append(batch, obj)
+
+				if len(batch) < *r.Source.Snowball.Batch {
+					continue
+				}
+
+			write:
+				if len(batch) > 0 {
+					if err := r.writeAsArchive(ctx, api, cl, batch); err != nil {
+						logger.LogIf(ctx, err)
+						for _, b := range batch {
+							slowCh <- b
+						}
+					}
+					batch = batch[:0]
+				}
+			}
+		}()
+	} else {
+		slowCh = walkCh
+	}
+
 	workerSize, err := strconv.Atoi(env.Get("_MINIO_BATCH_REPLICATION_WORKERS", strconv.Itoa(runtime.GOMAXPROCS(0)/2)))
 	if err != nil {
 		return err
@@ -944,6 +1106,11 @@ func (r *BatchJobReplicateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 		return err
 	}

+	walkQuorum := env.Get("_MINIO_BATCH_REPLICATION_WALK_QUORUM", "strict")
+	if walkQuorum == "" {
+		walkQuorum = "strict"
+	}
+
 	retryAttempts := ri.RetryAttempts
 	retry := false
 	for attempts := 1; attempts <= retryAttempts; attempts++ {
@@ -953,10 +1120,10 @@ func (r *BatchJobReplicateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 		// one of source/target is s3, skip delete marker and all versions under the same object name.
 		s3Type := r.Target.Type == BatchJobReplicateResourceS3 || r.Source.Type == BatchJobReplicateResourceS3

-		results := make(chan ObjectInfo, 100)
-		if err := api.Walk(ctx, r.Source.Bucket, r.Source.Prefix, results, ObjectOptions{
-			WalkMarker: lastObject,
-			WalkFilter: selectObj,
+		if err := api.Walk(ctx, r.Source.Bucket, r.Source.Prefix, walkCh, WalkOptions{
+			Marker:   lastObject,
+			Filter:   selectObj,
+			AskDisks: walkQuorum,
 		}); err != nil {
 			cancel()
 			// Do not need to retry if we can't list objects on source.
@@ -966,7 +1133,7 @@ func (r *BatchJobReplicateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 		prevObj := ""

 		skipReplicate := false
-		for result := range results {
+		for result := range slowCh {
 			result := result
 			if result.Name != prevObj {
 				prevObj = result.Name
@@ -979,7 +1146,7 @@ func (r *BatchJobReplicateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 			go func() {
 				defer wk.Give()

-				stopFn := globalBatchJobsMetrics.trace(batchReplicationMetricObject, job.ID, attempts, result)
+				stopFn := globalBatchJobsMetrics.trace(batchJobMetricReplication, job.ID, attempts)
 				success := true
 				if err := r.ReplicateToTarget(ctx, api, c, result, retry); err != nil {
 					if miniogo.ToErrorResponse(err).Code == "PreconditionFailed" {
@@ -990,16 +1157,20 @@ func (r *BatchJobReplicateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 					if isErrVersionNotFound(err) || isErrObjectNotFound(err) {
 						return
 					}
-					stopFn(err)
+					stopFn(result, err)
 					logger.LogIf(ctx, err)
 					success = false
 				} else {
-					stopFn(nil)
+					stopFn(result, nil)
 				}
 				ri.trackCurrentBucketObject(r.Source.Bucket, result, success)
 				globalBatchJobsMetrics.save(job.ID, ri)
 				// persist in-memory state to disk after every 10secs.
 				logger.LogIf(ctx, ri.updateAfter(ctx, api, 10*time.Second, job))
+
+				if wait := globalBatchConfig.ReplicationWait(); wait > 0 {
+					time.Sleep(wait)
+				}
 			}()
 		}
 		wk.Wait()
@@ -1080,6 +1251,9 @@ func (r *BatchJobReplicateV1) Validate(ctx context.Context, job BatchJobRequest,
 	if err := r.Source.Type.Validate(); err != nil {
 		return err
 	}
+	if err := r.Source.Snowball.Validate(); err != nil {
+		return err
+	}
 	if r.Source.Creds.Empty() && r.Target.Creds.Empty() {
 		return errInvalidArgument
 	}
@@ -1201,6 +1375,8 @@ func (j BatchJobRequest) Type() madmin.BatchJobType {
 		return madmin.BatchJobReplicate
 	case j.KeyRotate != nil:
 		return madmin.BatchJobKeyRotate
+	case j.Expire != nil:
+		return madmin.BatchJobExpire
 	}
 	return madmin.BatchJobType("unknown")
 }
@@ -1213,6 +1389,8 @@ func (j BatchJobRequest) Validate(ctx context.Context, o ObjectLayer) error {
 		return j.Replicate.Validate(ctx, j, o)
 	case j.KeyRotate != nil:
 		return j.KeyRotate.Validate(ctx, j, o)
+	case j.Expire != nil:
+		return j.Expire.Validate(ctx, j, o)
 	}
 	return errInvalidArgument
 }
@@ -1223,12 +1401,14 @@ func (j BatchJobRequest) delete(ctx context.Context, api ObjectLayer) {
 		deleteConfig(ctx, api, pathJoin(j.Location, batchReplName))
 	case j.KeyRotate != nil:
 		deleteConfig(ctx, api, pathJoin(j.Location, batchKeyRotationName))
+	case j.Expire != nil:
+		deleteConfig(ctx, api, pathJoin(j.Location, batchExpireName))
 	}
 	deleteConfig(ctx, api, j.Location)
 }

 func (j *BatchJobRequest) save(ctx context.Context, api ObjectLayer) error {
-	if j.Replicate == nil && j.KeyRotate == nil {
+	if j.Replicate == nil && j.KeyRotate == nil && j.Expire == nil {
 		return errInvalidArgument
 	}

@@ -1297,7 +1477,7 @@ func (a adminAPIHandlers) ListBatchJobs(w http.ResponseWriter, r *http.Request)
 	ctx, cancel := context.WithCancel(ctx)
 	defer cancel()

-	if err := objectAPI.Walk(ctx, minioMetaBucket, batchJobPrefix, resultCh, ObjectOptions{}); err != nil {
+	if err := objectAPI.Walk(ctx, minioMetaBucket, batchJobPrefix, resultCh, WalkOptions{}); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -1389,6 +1569,34 @@ func (a adminAPIHandlers) StartBatchJob(w http.ResponseWriter, r *http.Request)
 		return
 	}

+	// Fill with default values
+	if job.Replicate != nil {
+		if job.Replicate.Source.Snowball.Disable == nil {
+			job.Replicate.Source.Snowball.Disable = ptr(false)
+		}
+		if job.Replicate.Source.Snowball.Batch == nil {
+			job.Replicate.Source.Snowball.Batch = ptr(100)
+		}
+		if job.Replicate.Source.Snowball.InMemory == nil {
+			job.Replicate.Source.Snowball.InMemory = ptr(true)
+		}
+		if job.Replicate.Source.Snowball.Compress == nil {
+			job.Replicate.Source.Snowball.Compress = ptr(false)
+		}
+		if job.Replicate.Source.Snowball.SmallerThan == nil {
+			job.Replicate.Source.Snowball.SmallerThan = ptr("5MiB")
+		}
+		if job.Replicate.Source.Snowball.SkipErrs == nil {
+			job.Replicate.Source.Snowball.SkipErrs = ptr(true)
+		}
+	}
+
+	//  Validate the incoming job request
+	if err := job.Validate(ctx, objectAPI); err != nil {
+		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
+		return
+	}
+
 	job.ID = fmt.Sprintf("%s:%d", shortuuid.New(), GetProxyEndpointLocalIndex(globalProxyEndpoints))
 	job.User = user
 	job.Started = time.Now()
@@ -1486,7 +1694,7 @@ func (j *BatchJobPool) resume() {
 	results := make(chan ObjectInfo, 100)
 	ctx, cancel := context.WithCancel(j.ctx)
 	defer cancel()
-	if err := j.objLayer.Walk(ctx, minioMetaBucket, batchJobPrefix, results, ObjectOptions{}); err != nil {
+	if err := j.objLayer.Walk(ctx, minioMetaBucket, batchJobPrefix, results, WalkOptions{}); err != nil {
 		logger.LogIf(j.ctx, err)
 		return
 	}
@@ -1525,7 +1733,8 @@ func (j *BatchJobPool) AddWorker() {
 			if !ok {
 				return
 			}
-			if job.Replicate != nil {
+			switch {
+			case job.Replicate != nil:
 				if job.Replicate.RemoteToLocal() {
 					if err := job.Replicate.StartFromSource(job.ctx, j.objLayer, *job); err != nil {
 						if !isErrBucketNotFound(err) {
@@ -1545,14 +1754,20 @@ func (j *BatchJobPool) AddWorker() {
 						// Bucket not found proceed to delete such a job.
 					}
 				}
-			}
-			if job.KeyRotate != nil {
+			case job.KeyRotate != nil:
 				if err := job.KeyRotate.Start(job.ctx, j.objLayer, *job); err != nil {
 					if !isErrBucketNotFound(err) {
 						logger.LogIf(j.ctx, err)
 						continue
 					}
 				}
+			case job.Expire != nil:
+				if err := job.Expire.Start(job.ctx, j.objLayer, *job); err != nil {
+					if !isErrBucketNotFound(err) {
+						logger.LogIf(j.ctx, err)
+						continue
+					}
+				}
 			}
 			job.delete(j.ctx, j.objLayer)
 			j.canceler(job.ID, false)
@@ -1630,24 +1845,26 @@ type batchJobMetrics struct {
 type batchJobMetric uint8

 const (
-	batchReplicationMetricObject batchJobMetric = iota
-	batchKeyRotationMetricObject
+	batchJobMetricReplication batchJobMetric = iota
+	batchJobMetricKeyRotation
+	batchJobMetricExpire
 )

-func batchJobTrace(d batchJobMetric, job string, startTime time.Time, duration time.Duration, info ObjectInfo, attempts int, err error) madmin.TraceInfo {
+func batchJobTrace(d batchJobMetric, job string, startTime time.Time, duration time.Duration, info objTraceInfoer, attempts int, err error) madmin.TraceInfo {
 	var errStr string
 	if err != nil {
 		errStr = err.Error()
 	}
-	jobKind := "batchReplication"
 	traceType := madmin.TraceBatchReplication
-	if d == batchKeyRotationMetricObject {
-		jobKind = "batchKeyRotation"
+	switch d {
+	case batchJobMetricKeyRotation:
 		traceType = madmin.TraceBatchKeyRotation
+	case batchJobMetricExpire:
+		traceType = madmin.TraceBatchExpire
 	}
-	funcName := fmt.Sprintf("%s.%s (job-name=%s)", jobKind, d.String(), job)
+	funcName := fmt.Sprintf("%s() (job-name=%s)", d.String(), job)
 	if attempts > 0 {
-		funcName = fmt.Sprintf("%s.%s (job-name=%s,attempts=%s)", jobKind, d.String(), job, humanize.Ordinal(attempts))
+		funcName = fmt.Sprintf("%s() (job-name=%s,attempts=%s)", d.String(), job, humanize.Ordinal(attempts))
 	}
 	return madmin.TraceInfo{
 		TraceType: traceType,
@@ -1655,55 +1872,65 @@ func batchJobTrace(d batchJobMetric, job string, startTime time.Time, duration t
 		NodeName:  globalLocalNodeName,
 		FuncName:  funcName,
 		Duration:  duration,
-		Path:      info.Name,
+		Path:      fmt.Sprintf("%s (versionID=%s)", info.TraceObjName(), info.TraceVersionID()),
 		Error:     errStr,
 	}
 }

+func (ri *batchJobInfo) metric() madmin.JobMetric {
+	m := madmin.JobMetric{
+		JobID:         ri.JobID,
+		JobType:       ri.JobType,
+		StartTime:     ri.StartTime,
+		LastUpdate:    ri.LastUpdate,
+		RetryAttempts: ri.RetryAttempts,
+		Complete:      ri.Complete,
+		Failed:        ri.Failed,
+	}
+
+	switch ri.JobType {
+	case string(madmin.BatchJobReplicate):
+		m.Replicate = &madmin.ReplicateInfo{
+			Bucket:           ri.Bucket,
+			Object:           ri.Object,
+			Objects:          ri.Objects,
+			ObjectsFailed:    ri.ObjectsFailed,
+			BytesTransferred: ri.BytesTransferred,
+			BytesFailed:      ri.BytesFailed,
+		}
+	case string(madmin.BatchJobKeyRotate):
+		m.KeyRotate = &madmin.KeyRotationInfo{
+			Bucket:        ri.Bucket,
+			Object:        ri.Object,
+			Objects:       ri.Objects,
+			ObjectsFailed: ri.ObjectsFailed,
+		}
+	case string(madmin.BatchJobExpire):
+		m.Expired = &madmin.ExpirationInfo{
+			Bucket:        ri.Bucket,
+			Object:        ri.Object,
+			Objects:       ri.Objects,
+			ObjectsFailed: ri.ObjectsFailed,
+		}
+	}
+
+	return m
+}
+
 func (m *batchJobMetrics) report(jobID string) (metrics *madmin.BatchJobMetrics) {
 	metrics = &madmin.BatchJobMetrics{CollectedAt: time.Now(), Jobs: make(map[string]madmin.JobMetric)}
 	m.RLock()
 	defer m.RUnlock()

-	match := true
+	if jobID != "" {
+		if job, ok := m.metrics[jobID]; ok {
+			metrics.Jobs[jobID] = job.metric()
+		}
+		return metrics
+	}
+
 	for id, job := range m.metrics {
-		if jobID != "" {
-			match = id == jobID
-		}
-		if !match {
-			continue
-		}
-
-		m := madmin.JobMetric{
-			JobID:         job.JobID,
-			JobType:       job.JobType,
-			StartTime:     job.StartTime,
-			LastUpdate:    job.LastUpdate,
-			RetryAttempts: job.RetryAttempts,
-			Complete:      job.Complete,
-			Failed:        job.Failed,
-		}
-
-		switch job.JobType {
-		case string(madmin.BatchJobReplicate):
-			m.Replicate = &madmin.ReplicateInfo{
-				Bucket:           job.Bucket,
-				Object:           job.Object,
-				Objects:          job.Objects,
-				ObjectsFailed:    job.ObjectsFailed,
-				BytesTransferred: job.BytesTransferred,
-				BytesFailed:      job.BytesFailed,
-			}
-		case string(madmin.BatchJobKeyRotate):
-			m.KeyRotate = &madmin.KeyRotationInfo{
-				Bucket:        job.Bucket,
-				Object:        job.Object,
-				Objects:       job.Objects,
-				ObjectsFailed: job.ObjectsFailed,
-			}
-		}
-
-		metrics.Jobs[id] = m
+		metrics.Jobs[id] = job.metric()
 	}
 	return metrics
 }
@@ -1748,19 +1975,52 @@ func (m *batchJobMetrics) save(jobID string, ri *batchJobInfo) {
 	m.metrics[jobID] = ri.clone()
 }

-func (m *batchJobMetrics) trace(d batchJobMetric, job string, attempts int, info ObjectInfo) func(err error) {
+type objTraceInfoer interface {
+	TraceObjName() string
+	TraceVersionID() string
+}
+
+// TraceObjName returns name of object being traced
+func (td ObjectToDelete) TraceObjName() string {
+	return td.ObjectName
+}
+
+// TraceVersionID returns version-id of object being traced
+func (td ObjectToDelete) TraceVersionID() string {
+	return td.VersionID
+}
+
+// TraceObjName returns name of object being traced
+func (oi ObjectInfo) TraceObjName() string {
+	return oi.Name
+}
+
+// TraceVersionID returns version-id of object being traced
+func (oi ObjectInfo) TraceVersionID() string {
+	return oi.VersionID
+}
+
+func (m *batchJobMetrics) trace(d batchJobMetric, job string, attempts int) func(info objTraceInfoer, err error) {
 	startTime := time.Now()
-	return func(err error) {
+	return func(info objTraceInfoer, err error) {
 		duration := time.Since(startTime)
+		if globalTrace.NumSubscribers(madmin.TraceBatch) > 0 {
+			globalTrace.Publish(batchJobTrace(d, job, startTime, duration, info, attempts, err))
+			return
+		}
 		switch d {
-		case batchReplicationMetricObject:
+		case batchJobMetricReplication:
 			if globalTrace.NumSubscribers(madmin.TraceBatchReplication) > 0 {
 				globalTrace.Publish(batchJobTrace(d, job, startTime, duration, info, attempts, err))
 			}
-		case batchKeyRotationMetricObject:
+		case batchJobMetricKeyRotation:
 			if globalTrace.NumSubscribers(madmin.TraceBatchKeyRotation) > 0 {
 				globalTrace.Publish(batchJobTrace(d, job, startTime, duration, info, attempts, err))
 			}
+		case batchJobMetricExpire:
+			if globalTrace.NumSubscribers(madmin.TraceBatchExpire) > 0 {
+				globalTrace.Publish(batchJobTrace(d, job, startTime, duration, info, attempts, err))
+			}
 		}
 	}
 }
@@ -84,6 +84,24 @@ func (z *BatchJobRequest) DecodeMsg(dc *msgp.Reader) (err error) {
 					return
 				}
 			}
+		case "Expire":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "Expire")
+					return
+				}
+				z.Expire = nil
+			} else {
+				if z.Expire == nil {
+					z.Expire = new(BatchJobExpire)
+				}
+				err = z.Expire.DecodeMsg(dc)
+				if err != nil {
+					err = msgp.WrapError(err, "Expire")
+					return
+				}
+			}
 		default:
 			err = dc.Skip()
 			if err != nil {
@@ -97,9 +115,9 @@ func (z *BatchJobRequest) DecodeMsg(dc *msgp.Reader) (err error) {

 // EncodeMsg implements msgp.Encodable
 func (z *BatchJobRequest) EncodeMsg(en *msgp.Writer) (err error) {
-	// map header, size 6
+	// map header, size 7
 	// write "ID"
-	err = en.Append(0x86, 0xa2, 0x49, 0x44)
+	err = en.Append(0x87, 0xa2, 0x49, 0x44)
 	if err != nil {
 		return
 	}
@@ -172,15 +190,32 @@ func (z *BatchJobRequest) EncodeMsg(en *msgp.Writer) (err error) {
 			return
 		}
 	}
+	// write "Expire"
+	err = en.Append(0xa6, 0x45, 0x78, 0x70, 0x69, 0x72, 0x65)
+	if err != nil {
+		return
+	}
+	if z.Expire == nil {
+		err = en.WriteNil()
+		if err != nil {
+			return
+		}
+	} else {
+		err = z.Expire.EncodeMsg(en)
+		if err != nil {
+			err = msgp.WrapError(err, "Expire")
+			return
+		}
+	}
 	return
 }

 // MarshalMsg implements msgp.Marshaler
 func (z *BatchJobRequest) MarshalMsg(b []byte) (o []byte, err error) {
 	o = msgp.Require(b, z.Msgsize())
-	// map header, size 6
+	// map header, size 7
 	// string "ID"
-	o = append(o, 0x86, 0xa2, 0x49, 0x44)
+	o = append(o, 0x87, 0xa2, 0x49, 0x44)
 	o = msgp.AppendString(o, z.ID)
 	// string "User"
 	o = append(o, 0xa4, 0x55, 0x73, 0x65, 0x72)
@@ -213,6 +248,17 @@ func (z *BatchJobRequest) MarshalMsg(b []byte) (o []byte, err error) {
 			return
 		}
 	}
+	// string "Expire"
+	o = append(o, 0xa6, 0x45, 0x78, 0x70, 0x69, 0x72, 0x65)
+	if z.Expire == nil {
+		o = msgp.AppendNil(o)
+	} else {
+		o, err = z.Expire.MarshalMsg(o)
+		if err != nil {
+			err = msgp.WrapError(err, "Expire")
+			return
+		}
+	}
 	return
 }

@@ -292,6 +338,23 @@ func (z *BatchJobRequest) UnmarshalMsg(bts []byte) (o []byte, err error) {
 					return
 				}
 			}
+		case "Expire":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.Expire = nil
+			} else {
+				if z.Expire == nil {
+					z.Expire = new(BatchJobExpire)
+				}
+				bts, err = z.Expire.UnmarshalMsg(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Expire")
+					return
+				}
+			}
 		default:
 			bts, err = msgp.Skip(bts)
 			if err != nil {
@@ -318,6 +381,12 @@ func (z *BatchJobRequest) Msgsize() (s int) {
 	} else {
 		s += z.KeyRotate.Msgsize()
 	}
+	s += 7
+	if z.Expire == nil {
+		s += msgp.NilSize
+	} else {
+		s += z.Expire.Msgsize()
+	}
 	return
 }

@@ -18,9 +18,11 @@
 package cmd

 import (
+	"errors"
 	"strings"
 	"time"

+	"github.com/dustin/go-humanize"
 	"github.com/minio/pkg/v2/wildcard"
 )

@@ -81,3 +83,81 @@ func (r BatchJobRetry) Validate() error {

 	return nil
 }
+
+//   # snowball based archive transfer is by default enabled when source
+//   # is local and target is remote which is also minio.
+//   snowball:
+//     disable: false # optionally turn-off snowball archive transfer
+//     batch: 100 # upto this many objects per archive
+//     inmemory: true # indicates if the archive must be staged locally or in-memory
+//     compress: true # S2/Snappy compressed archive
+//     smallerThan: 5MiB # create archive for all objects smaller than 5MiB
+//     skipErrs: false # skips any source side read() errors
+
+// BatchJobSnowball describes the snowball feature when replicating objects from a local source to a remote target
+type BatchJobSnowball struct {
+	Disable     *bool   `yaml:"disable" json:"disable"`
+	Batch       *int    `yaml:"batch" json:"batch"`
+	InMemory    *bool   `yaml:"inmemory" json:"inmemory"`
+	Compress    *bool   `yaml:"compress" json:"compress"`
+	SmallerThan *string `yaml:"smallerThan" json:"smallerThan"`
+	SkipErrs    *bool   `yaml:"skipErrs" json:"skipErrs"`
+}
+
+// Validate the snowball parameters in the job description
+func (b BatchJobSnowball) Validate() error {
+	if *b.Batch <= 0 {
+		return errors.New("batch number should be non positive zero")
+	}
+	_, err := humanize.ParseBytes(*b.SmallerThan)
+	return err
+}
+
+// BatchJobSizeFilter supports size based filters - LesserThan and GreaterThan
+type BatchJobSizeFilter struct {
+	UpperBound BatchJobSize `yaml:"lessThan" json:"lessThan"`
+	LowerBound BatchJobSize `yaml:"greaterThan" json:"greaterThan"`
+}
+
+// InRange returns true in the following cases and false otherwise,
+// - sf.LowerBound < sz, when sf.LowerBound alone is specified
+// - sz < sf.UpperBound, when sf.UpperBound alone is specified
+// - sf.LowerBound < sz < sf.UpperBound when both are specified,
+func (sf BatchJobSizeFilter) InRange(sz int64) bool {
+	if sf.UpperBound > 0 && sz > int64(sf.UpperBound) {
+		return false
+	}
+
+	if sf.LowerBound > 0 && sz < int64(sf.LowerBound) {
+		return false
+	}
+	return true
+}
+
+var errInvalidBatchJobSizeFilter = errors.New("invalid batch-job size filter")
+
+// Validate checks if sf is a valid batch-job size filter
+func (sf BatchJobSizeFilter) Validate() error {
+	if sf.LowerBound > 0 && sf.UpperBound > 0 && sf.LowerBound >= sf.UpperBound {
+		return errInvalidBatchJobSizeFilter
+	}
+	return nil
+}
+
+// BatchJobSize supports humanized byte values in yaml files type BatchJobSize uint64
+type BatchJobSize int64
+
+// UnmarshalYAML to parse humanized byte values
+func (s *BatchJobSize) UnmarshalYAML(unmarshal func(interface{}) error) error {
+	var batchExpireSz string
+	err := unmarshal(&batchExpireSz)
+	if err != nil {
+		return err
+	}
+	sz, err := humanize.ParseBytes(batchExpireSz)
+	if err != nil {
+		return err
+	}
+	*s = BatchJobSize(sz)
+	return nil
+}
@@ -389,3 +389,666 @@ func (z BatchJobRetry) Msgsize() (s int) {
 	s = 1 + 9 + msgp.IntSize + 6 + msgp.DurationSize
 	return
 }
+
+// DecodeMsg implements msgp.Decodable
+func (z *BatchJobSize) DecodeMsg(dc *msgp.Reader) (err error) {
+	{
+		var zb0001 int64
+		zb0001, err = dc.ReadInt64()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		(*z) = BatchJobSize(zb0001)
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z BatchJobSize) EncodeMsg(en *msgp.Writer) (err error) {
+	err = en.WriteInt64(int64(z))
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z BatchJobSize) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	o = msgp.AppendInt64(o, int64(z))
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *BatchJobSize) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	{
+		var zb0001 int64
+		zb0001, bts, err = msgp.ReadInt64Bytes(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		(*z) = BatchJobSize(zb0001)
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z BatchJobSize) Msgsize() (s int) {
+	s = msgp.Int64Size
+	return
+}
+
+// DecodeMsg implements msgp.Decodable
+func (z *BatchJobSizeFilter) DecodeMsg(dc *msgp.Reader) (err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, err = dc.ReadMapHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, err = dc.ReadMapKeyPtr()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "UpperBound":
+			{
+				var zb0002 int64
+				zb0002, err = dc.ReadInt64()
+				if err != nil {
+					err = msgp.WrapError(err, "UpperBound")
+					return
+				}
+				z.UpperBound = BatchJobSize(zb0002)
+			}
+		case "LowerBound":
+			{
+				var zb0003 int64
+				zb0003, err = dc.ReadInt64()
+				if err != nil {
+					err = msgp.WrapError(err, "LowerBound")
+					return
+				}
+				z.LowerBound = BatchJobSize(zb0003)
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z BatchJobSizeFilter) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 2
+	// write "UpperBound"
+	err = en.Append(0x82, 0xaa, 0x55, 0x70, 0x70, 0x65, 0x72, 0x42, 0x6f, 0x75, 0x6e, 0x64)
+	if err != nil {
+		return
+	}
+	err = en.WriteInt64(int64(z.UpperBound))
+	if err != nil {
+		err = msgp.WrapError(err, "UpperBound")
+		return
+	}
+	// write "LowerBound"
+	err = en.Append(0xaa, 0x4c, 0x6f, 0x77, 0x65, 0x72, 0x42, 0x6f, 0x75, 0x6e, 0x64)
+	if err != nil {
+		return
+	}
+	err = en.WriteInt64(int64(z.LowerBound))
+	if err != nil {
+		err = msgp.WrapError(err, "LowerBound")
+		return
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z BatchJobSizeFilter) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 2
+	// string "UpperBound"
+	o = append(o, 0x82, 0xaa, 0x55, 0x70, 0x70, 0x65, 0x72, 0x42, 0x6f, 0x75, 0x6e, 0x64)
+	o = msgp.AppendInt64(o, int64(z.UpperBound))
+	// string "LowerBound"
+	o = append(o, 0xaa, 0x4c, 0x6f, 0x77, 0x65, 0x72, 0x42, 0x6f, 0x75, 0x6e, 0x64)
+	o = msgp.AppendInt64(o, int64(z.LowerBound))
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *BatchJobSizeFilter) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadMapHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, bts, err = msgp.ReadMapKeyZC(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "UpperBound":
+			{
+				var zb0002 int64
+				zb0002, bts, err = msgp.ReadInt64Bytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "UpperBound")
+					return
+				}
+				z.UpperBound = BatchJobSize(zb0002)
+			}
+		case "LowerBound":
+			{
+				var zb0003 int64
+				zb0003, bts, err = msgp.ReadInt64Bytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "LowerBound")
+					return
+				}
+				z.LowerBound = BatchJobSize(zb0003)
+			}
+		default:
+			bts, err = msgp.Skip(bts)
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z BatchJobSizeFilter) Msgsize() (s int) {
+	s = 1 + 11 + msgp.Int64Size + 11 + msgp.Int64Size
+	return
+}
+
+// DecodeMsg implements msgp.Decodable
+func (z *BatchJobSnowball) DecodeMsg(dc *msgp.Reader) (err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, err = dc.ReadMapHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, err = dc.ReadMapKeyPtr()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "Disable":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "Disable")
+					return
+				}
+				z.Disable = nil
+			} else {
+				if z.Disable == nil {
+					z.Disable = new(bool)
+				}
+				*z.Disable, err = dc.ReadBool()
+				if err != nil {
+					err = msgp.WrapError(err, "Disable")
+					return
+				}
+			}
+		case "Batch":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "Batch")
+					return
+				}
+				z.Batch = nil
+			} else {
+				if z.Batch == nil {
+					z.Batch = new(int)
+				}
+				*z.Batch, err = dc.ReadInt()
+				if err != nil {
+					err = msgp.WrapError(err, "Batch")
+					return
+				}
+			}
+		case "InMemory":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "InMemory")
+					return
+				}
+				z.InMemory = nil
+			} else {
+				if z.InMemory == nil {
+					z.InMemory = new(bool)
+				}
+				*z.InMemory, err = dc.ReadBool()
+				if err != nil {
+					err = msgp.WrapError(err, "InMemory")
+					return
+				}
+			}
+		case "Compress":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "Compress")
+					return
+				}
+				z.Compress = nil
+			} else {
+				if z.Compress == nil {
+					z.Compress = new(bool)
+				}
+				*z.Compress, err = dc.ReadBool()
+				if err != nil {
+					err = msgp.WrapError(err, "Compress")
+					return
+				}
+			}
+		case "SmallerThan":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "SmallerThan")
+					return
+				}
+				z.SmallerThan = nil
+			} else {
+				if z.SmallerThan == nil {
+					z.SmallerThan = new(string)
+				}
+				*z.SmallerThan, err = dc.ReadString()
+				if err != nil {
+					err = msgp.WrapError(err, "SmallerThan")
+					return
+				}
+			}
+		case "SkipErrs":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "SkipErrs")
+					return
+				}
+				z.SkipErrs = nil
+			} else {
+				if z.SkipErrs == nil {
+					z.SkipErrs = new(bool)
+				}
+				*z.SkipErrs, err = dc.ReadBool()
+				if err != nil {
+					err = msgp.WrapError(err, "SkipErrs")
+					return
+				}
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z *BatchJobSnowball) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 6
+	// write "Disable"
+	err = en.Append(0x86, 0xa7, 0x44, 0x69, 0x73, 0x61, 0x62, 0x6c, 0x65)
+	if err != nil {
+		return
+	}
+	if z.Disable == nil {
+		err = en.WriteNil()
+		if err != nil {
+			return
+		}
+	} else {
+		err = en.WriteBool(*z.Disable)
+		if err != nil {
+			err = msgp.WrapError(err, "Disable")
+			return
+		}
+	}
+	// write "Batch"
+	err = en.Append(0xa5, 0x42, 0x61, 0x74, 0x63, 0x68)
+	if err != nil {
+		return
+	}
+	if z.Batch == nil {
+		err = en.WriteNil()
+		if err != nil {
+			return
+		}
+	} else {
+		err = en.WriteInt(*z.Batch)
+		if err != nil {
+			err = msgp.WrapError(err, "Batch")
+			return
+		}
+	}
+	// write "InMemory"
+	err = en.Append(0xa8, 0x49, 0x6e, 0x4d, 0x65, 0x6d, 0x6f, 0x72, 0x79)
+	if err != nil {
+		return
+	}
+	if z.InMemory == nil {
+		err = en.WriteNil()
+		if err != nil {
+			return
+		}
+	} else {
+		err = en.WriteBool(*z.InMemory)
+		if err != nil {
+			err = msgp.WrapError(err, "InMemory")
+			return
+		}
+	}
+	// write "Compress"
+	err = en.Append(0xa8, 0x43, 0x6f, 0x6d, 0x70, 0x72, 0x65, 0x73, 0x73)
+	if err != nil {
+		return
+	}
+	if z.Compress == nil {
+		err = en.WriteNil()
+		if err != nil {
+			return
+		}
+	} else {
+		err = en.WriteBool(*z.Compress)
+		if err != nil {
+			err = msgp.WrapError(err, "Compress")
+			return
+		}
+	}
+	// write "SmallerThan"
+	err = en.Append(0xab, 0x53, 0x6d, 0x61, 0x6c, 0x6c, 0x65, 0x72, 0x54, 0x68, 0x61, 0x6e)
+	if err != nil {
+		return
+	}
+	if z.SmallerThan == nil {
+		err = en.WriteNil()
+		if err != nil {
+			return
+		}
+	} else {
+		err = en.WriteString(*z.SmallerThan)
+		if err != nil {
+			err = msgp.WrapError(err, "SmallerThan")
+			return
+		}
+	}
+	// write "SkipErrs"
+	err = en.Append(0xa8, 0x53, 0x6b, 0x69, 0x70, 0x45, 0x72, 0x72, 0x73)
+	if err != nil {
+		return
+	}
+	if z.SkipErrs == nil {
+		err = en.WriteNil()
+		if err != nil {
+			return
+		}
+	} else {
+		err = en.WriteBool(*z.SkipErrs)
+		if err != nil {
+			err = msgp.WrapError(err, "SkipErrs")
+			return
+		}
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z *BatchJobSnowball) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 6
+	// string "Disable"
+	o = append(o, 0x86, 0xa7, 0x44, 0x69, 0x73, 0x61, 0x62, 0x6c, 0x65)
+	if z.Disable == nil {
+		o = msgp.AppendNil(o)
+	} else {
+		o = msgp.AppendBool(o, *z.Disable)
+	}
+	// string "Batch"
+	o = append(o, 0xa5, 0x42, 0x61, 0x74, 0x63, 0x68)
+	if z.Batch == nil {
+		o = msgp.AppendNil(o)
+	} else {
+		o = msgp.AppendInt(o, *z.Batch)
+	}
+	// string "InMemory"
+	o = append(o, 0xa8, 0x49, 0x6e, 0x4d, 0x65, 0x6d, 0x6f, 0x72, 0x79)
+	if z.InMemory == nil {
+		o = msgp.AppendNil(o)
+	} else {
+		o = msgp.AppendBool(o, *z.InMemory)
+	}
+	// string "Compress"
+	o = append(o, 0xa8, 0x43, 0x6f, 0x6d, 0x70, 0x72, 0x65, 0x73, 0x73)
+	if z.Compress == nil {
+		o = msgp.AppendNil(o)
+	} else {
+		o = msgp.AppendBool(o, *z.Compress)
+	}
+	// string "SmallerThan"
+	o = append(o, 0xab, 0x53, 0x6d, 0x61, 0x6c, 0x6c, 0x65, 0x72, 0x54, 0x68, 0x61, 0x6e)
+	if z.SmallerThan == nil {
+		o = msgp.AppendNil(o)
+	} else {
+		o = msgp.AppendString(o, *z.SmallerThan)
+	}
+	// string "SkipErrs"
+	o = append(o, 0xa8, 0x53, 0x6b, 0x69, 0x70, 0x45, 0x72, 0x72, 0x73)
+	if z.SkipErrs == nil {
+		o = msgp.AppendNil(o)
+	} else {
+		o = msgp.AppendBool(o, *z.SkipErrs)
+	}
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *BatchJobSnowball) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadMapHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, bts, err = msgp.ReadMapKeyZC(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "Disable":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.Disable = nil
+			} else {
+				if z.Disable == nil {
+					z.Disable = new(bool)
+				}
+				*z.Disable, bts, err = msgp.ReadBoolBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Disable")
+					return
+				}
+			}
+		case "Batch":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.Batch = nil
+			} else {
+				if z.Batch == nil {
+					z.Batch = new(int)
+				}
+				*z.Batch, bts, err = msgp.ReadIntBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Batch")
+					return
+				}
+			}
+		case "InMemory":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.InMemory = nil
+			} else {
+				if z.InMemory == nil {
+					z.InMemory = new(bool)
+				}
+				*z.InMemory, bts, err = msgp.ReadBoolBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "InMemory")
+					return
+				}
+			}
+		case "Compress":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.Compress = nil
+			} else {
+				if z.Compress == nil {
+					z.Compress = new(bool)
+				}
+				*z.Compress, bts, err = msgp.ReadBoolBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Compress")
+					return
+				}
+			}
+		case "SmallerThan":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.SmallerThan = nil
+			} else {
+				if z.SmallerThan == nil {
+					z.SmallerThan = new(string)
+				}
+				*z.SmallerThan, bts, err = msgp.ReadStringBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "SmallerThan")
+					return
+				}
+			}
+		case "SkipErrs":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.SkipErrs = nil
+			} else {
+				if z.SkipErrs == nil {
+					z.SkipErrs = new(bool)
+				}
+				*z.SkipErrs, bts, err = msgp.ReadBoolBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "SkipErrs")
+					return
+				}
+			}
+		default:
+			bts, err = msgp.Skip(bts)
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z *BatchJobSnowball) Msgsize() (s int) {
+	s = 1 + 8
+	if z.Disable == nil {
+		s += msgp.NilSize
+	} else {
+		s += msgp.BoolSize
+	}
+	s += 6
+	if z.Batch == nil {
+		s += msgp.NilSize
+	} else {
+		s += msgp.IntSize
+	}
+	s += 9
+	if z.InMemory == nil {
+		s += msgp.NilSize
+	} else {
+		s += msgp.BoolSize
+	}
+	s += 9
+	if z.Compress == nil {
+		s += msgp.NilSize
+	} else {
+		s += msgp.BoolSize
+	}
+	s += 12
+	if z.SmallerThan == nil {
+		s += msgp.NilSize
+	} else {
+		s += msgp.StringPrefixSize + len(*z.SmallerThan)
+	}
+	s += 9
+	if z.SkipErrs == nil {
+		s += msgp.NilSize
+	} else {
+		s += msgp.BoolSize
+	}
+	return
+}
@@ -347,3 +347,229 @@ func BenchmarkDecodeBatchJobRetry(b *testing.B) {
 		}
 	}
 }
+
+func TestMarshalUnmarshalBatchJobSizeFilter(t *testing.T) {
+	v := BatchJobSizeFilter{}
+	bts, err := v.MarshalMsg(nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	left, err := v.UnmarshalMsg(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after UnmarshalMsg(): %q", len(left), left)
+	}
+
+	left, err = msgp.Skip(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after Skip(): %q", len(left), left)
+	}
+}
+
+func BenchmarkMarshalMsgBatchJobSizeFilter(b *testing.B) {
+	v := BatchJobSizeFilter{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgBatchJobSizeFilter(b *testing.B) {
+	v := BatchJobSizeFilter{}
+	bts := make([]byte, 0, v.Msgsize())
+	bts, _ = v.MarshalMsg(bts[0:0])
+	b.SetBytes(int64(len(bts)))
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		bts, _ = v.MarshalMsg(bts[0:0])
+	}
+}
+
+func BenchmarkUnmarshalBatchJobSizeFilter(b *testing.B) {
+	v := BatchJobSizeFilter{}
+	bts, _ := v.MarshalMsg(nil)
+	b.ReportAllocs()
+	b.SetBytes(int64(len(bts)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		_, err := v.UnmarshalMsg(bts)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestEncodeDecodeBatchJobSizeFilter(t *testing.T) {
+	v := BatchJobSizeFilter{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeBatchJobSizeFilter Msgsize() is inaccurate")
+	}
+
+	vn := BatchJobSizeFilter{}
+	err := msgp.Decode(&buf, &vn)
+	if err != nil {
+		t.Error(err)
+	}
+
+	buf.Reset()
+	msgp.Encode(&buf, &v)
+	err = msgp.NewReader(&buf).Skip()
+	if err != nil {
+		t.Error(err)
+	}
+}
+
+func BenchmarkEncodeBatchJobSizeFilter(b *testing.B) {
+	v := BatchJobSizeFilter{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	en := msgp.NewWriter(msgp.Nowhere)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.EncodeMsg(en)
+	}
+	en.Flush()
+}
+
+func BenchmarkDecodeBatchJobSizeFilter(b *testing.B) {
+	v := BatchJobSizeFilter{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	rd := msgp.NewEndlessReader(buf.Bytes(), b)
+	dc := msgp.NewReader(rd)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		err := v.DecodeMsg(dc)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestMarshalUnmarshalBatchJobSnowball(t *testing.T) {
+	v := BatchJobSnowball{}
+	bts, err := v.MarshalMsg(nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	left, err := v.UnmarshalMsg(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after UnmarshalMsg(): %q", len(left), left)
+	}
+
+	left, err = msgp.Skip(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after Skip(): %q", len(left), left)
+	}
+}
+
+func BenchmarkMarshalMsgBatchJobSnowball(b *testing.B) {
+	v := BatchJobSnowball{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgBatchJobSnowball(b *testing.B) {
+	v := BatchJobSnowball{}
+	bts := make([]byte, 0, v.Msgsize())
+	bts, _ = v.MarshalMsg(bts[0:0])
+	b.SetBytes(int64(len(bts)))
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		bts, _ = v.MarshalMsg(bts[0:0])
+	}
+}
+
+func BenchmarkUnmarshalBatchJobSnowball(b *testing.B) {
+	v := BatchJobSnowball{}
+	bts, _ := v.MarshalMsg(nil)
+	b.ReportAllocs()
+	b.SetBytes(int64(len(bts)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		_, err := v.UnmarshalMsg(bts)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestEncodeDecodeBatchJobSnowball(t *testing.T) {
+	v := BatchJobSnowball{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeBatchJobSnowball Msgsize() is inaccurate")
+	}
+
+	vn := BatchJobSnowball{}
+	err := msgp.Decode(&buf, &vn)
+	if err != nil {
+		t.Error(err)
+	}
+
+	buf.Reset()
+	msgp.Encode(&buf, &v)
+	err = msgp.NewReader(&buf).Skip()
+	if err != nil {
+		t.Error(err)
+	}
+}
+
+func BenchmarkEncodeBatchJobSnowball(b *testing.B) {
+	v := BatchJobSnowball{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	en := msgp.NewWriter(msgp.Nowhere)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.EncodeMsg(en)
+	}
+	en.Flush()
+}
+
+func BenchmarkDecodeBatchJobSnowball(b *testing.B) {
+	v := BatchJobSnowball{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	rd := msgp.NewEndlessReader(buf.Bytes(), b)
+	dc := msgp.NewReader(rd)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		err := v.DecodeMsg(dc)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
@@ -0,0 +1,136 @@
+// Copyright (c) 2015-2023 MinIO, Inc.
+//
+// This file is part of MinIO Object Storage stack
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package cmd
+
+import (
+	"fmt"
+	"testing"
+)
+
+func TestBatchJobSizeInRange(t *testing.T) {
+	tests := []struct {
+		objSize    int64
+		sizeFilter BatchJobSizeFilter
+		want       bool
+	}{
+		{
+			// 1Mib < 2Mib < 10MiB -> in range
+			objSize: 2 << 20,
+			sizeFilter: BatchJobSizeFilter{
+				UpperBound: 10 << 20,
+				LowerBound: 1 << 20,
+			},
+			want: true,
+		},
+		{
+			// 2KiB < 1 MiB -> out of range from left
+			objSize: 2 << 10,
+			sizeFilter: BatchJobSizeFilter{
+				UpperBound: 10 << 20,
+				LowerBound: 1 << 20,
+			},
+			want: false,
+		},
+		{
+			// 11MiB > 10 MiB -> out of range from right
+			objSize: 11 << 20,
+			sizeFilter: BatchJobSizeFilter{
+				UpperBound: 10 << 20,
+				LowerBound: 1 << 20,
+			},
+			want: false,
+		},
+		{
+			//  2MiB < 10MiB -> in range
+			objSize: 2 << 20,
+			sizeFilter: BatchJobSizeFilter{
+				UpperBound: 10 << 20,
+			},
+			want: true,
+		},
+		{
+			//  2MiB > 1MiB -> in range
+			objSize: 2 << 20,
+			sizeFilter: BatchJobSizeFilter{
+				LowerBound: 1 << 20,
+			},
+			want: true,
+		},
+	}
+
+	for i, test := range tests {
+		t.Run(fmt.Sprintf("test-%d", i+1), func(t *testing.T) {
+			if got := test.sizeFilter.InRange(test.objSize); got != test.want {
+				t.Fatalf("Expected %v but got %v", test.want, got)
+			}
+		})
+	}
+}
+
+func TestBatchJobSizeValidate(t *testing.T) {
+	tests := []struct {
+		sizeFilter BatchJobSizeFilter
+		err        error
+	}{
+		{
+			// Unspecified size filter is a valid filter
+			sizeFilter: BatchJobSizeFilter{
+				UpperBound: 0,
+				LowerBound: 0,
+			},
+			err: nil,
+		},
+		{
+			sizeFilter: BatchJobSizeFilter{
+				UpperBound: 0,
+				LowerBound: 1 << 20,
+			},
+			err: nil,
+		},
+		{
+			sizeFilter: BatchJobSizeFilter{
+				UpperBound: 10 << 20,
+				LowerBound: 0,
+			},
+			err: nil,
+		},
+		{
+			// LowerBound > UpperBound -> empty range
+			sizeFilter: BatchJobSizeFilter{
+				UpperBound: 1 << 20,
+				LowerBound: 10 << 20,
+			},
+			err: errInvalidBatchJobSizeFilter,
+		},
+		{
+			// LowerBound == UpperBound -> empty range
+			sizeFilter: BatchJobSizeFilter{
+				UpperBound: 1 << 20,
+				LowerBound: 1 << 20,
+			},
+			err: errInvalidBatchJobSizeFilter,
+		},
+	}
+	for i, test := range tests {
+		t.Run(fmt.Sprintf("test-%d", i+1), func(t *testing.T) {
+			if err := test.sizeFilter.Validate(); err != test.err {
+				t.Fatalf("Expected %v but got %v", test.err, err)
+			}
+		})
+	}
+}
@@ -155,6 +155,7 @@ type BatchJobReplicateSource struct {
 	Endpoint string                        `yaml:"endpoint" json:"endpoint"`
 	Path     string                        `yaml:"path" json:"path"`
 	Creds    BatchJobReplicateCredentials  `yaml:"credentials" json:"credentials"`
+	Snowball BatchJobSnowball              `yaml:"snowball" json:"snowball"`
 }

 // ValidPath returns true if path is valid
@@ -469,6 +469,12 @@ func (z *BatchJobReplicateSource) DecodeMsg(dc *msgp.Reader) (err error) {
 					}
 				}
 			}
+		case "Snowball":
+			err = z.Snowball.DecodeMsg(dc)
+			if err != nil {
+				err = msgp.WrapError(err, "Snowball")
+				return
+			}
 		default:
 			err = dc.Skip()
 			if err != nil {
@@ -482,9 +488,9 @@ func (z *BatchJobReplicateSource) DecodeMsg(dc *msgp.Reader) (err error) {

 // EncodeMsg implements msgp.Encodable
 func (z *BatchJobReplicateSource) EncodeMsg(en *msgp.Writer) (err error) {
-	// map header, size 6
+	// map header, size 7
 	// write "Type"
-	err = en.Append(0x86, 0xa4, 0x54, 0x79, 0x70, 0x65)
+	err = en.Append(0x87, 0xa4, 0x54, 0x79, 0x70, 0x65)
 	if err != nil {
 		return
 	}
@@ -569,15 +575,25 @@ func (z *BatchJobReplicateSource) EncodeMsg(en *msgp.Writer) (err error) {
 		err = msgp.WrapError(err, "Creds", "SessionToken")
 		return
 	}
+	// write "Snowball"
+	err = en.Append(0xa8, 0x53, 0x6e, 0x6f, 0x77, 0x62, 0x61, 0x6c, 0x6c)
+	if err != nil {
+		return
+	}
+	err = z.Snowball.EncodeMsg(en)
+	if err != nil {
+		err = msgp.WrapError(err, "Snowball")
+		return
+	}
 	return
 }

 // MarshalMsg implements msgp.Marshaler
 func (z *BatchJobReplicateSource) MarshalMsg(b []byte) (o []byte, err error) {
 	o = msgp.Require(b, z.Msgsize())
-	// map header, size 6
+	// map header, size 7
 	// string "Type"
-	o = append(o, 0x86, 0xa4, 0x54, 0x79, 0x70, 0x65)
+	o = append(o, 0x87, 0xa4, 0x54, 0x79, 0x70, 0x65)
 	o = msgp.AppendString(o, string(z.Type))
 	// string "Bucket"
 	o = append(o, 0xa6, 0x42, 0x75, 0x63, 0x6b, 0x65, 0x74)
@@ -603,6 +619,13 @@ func (z *BatchJobReplicateSource) MarshalMsg(b []byte) (o []byte, err error) {
 	// string "SessionToken"
 	o = append(o, 0xac, 0x53, 0x65, 0x73, 0x73, 0x69, 0x6f, 0x6e, 0x54, 0x6f, 0x6b, 0x65, 0x6e)
 	o = msgp.AppendString(o, z.Creds.SessionToken)
+	// string "Snowball"
+	o = append(o, 0xa8, 0x53, 0x6e, 0x6f, 0x77, 0x62, 0x61, 0x6c, 0x6c)
+	o, err = z.Snowball.MarshalMsg(o)
+	if err != nil {
+		err = msgp.WrapError(err, "Snowball")
+		return
+	}
 	return
 }

@@ -699,6 +722,12 @@ func (z *BatchJobReplicateSource) UnmarshalMsg(bts []byte) (o []byte, err error)
 					}
 				}
 			}
+		case "Snowball":
+			bts, err = z.Snowball.UnmarshalMsg(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Snowball")
+				return
+			}
 		default:
 			bts, err = msgp.Skip(bts)
 			if err != nil {
@@ -713,7 +742,7 @@ func (z *BatchJobReplicateSource) 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 *BatchJobReplicateSource) Msgsize() (s int) {
-	s = 1 + 5 + msgp.StringPrefixSize + len(string(z.Type)) + 7 + msgp.StringPrefixSize + len(z.Bucket) + 7 + msgp.StringPrefixSize + len(z.Prefix) + 9 + msgp.StringPrefixSize + len(z.Endpoint) + 5 + msgp.StringPrefixSize + len(z.Path) + 6 + 1 + 10 + msgp.StringPrefixSize + len(z.Creds.AccessKey) + 10 + msgp.StringPrefixSize + len(z.Creds.SecretKey) + 13 + msgp.StringPrefixSize + len(z.Creds.SessionToken)
+	s = 1 + 5 + msgp.StringPrefixSize + len(string(z.Type)) + 7 + msgp.StringPrefixSize + len(z.Bucket) + 7 + msgp.StringPrefixSize + len(z.Prefix) + 9 + msgp.StringPrefixSize + len(z.Endpoint) + 5 + msgp.StringPrefixSize + len(z.Path) + 6 + 1 + 10 + msgp.StringPrefixSize + len(z.Creds.AccessKey) + 10 + msgp.StringPrefixSize + len(z.Creds.SecretKey) + 13 + msgp.StringPrefixSize + len(z.Creds.SessionToken) + 9 + z.Snowball.Msgsize()
 	return
 }

@@ -154,7 +154,6 @@ type BatchJobKeyRotateV1 struct {
 	Flags      BatchJobKeyRotateFlags      `yaml:"flags" json:"flags"`
 	Bucket     string                      `yaml:"bucket" json:"bucket"`
 	Prefix     string                      `yaml:"prefix" json:"prefix"`
-	Endpoint   string                      `yaml:"endpoint" json:"endpoint"`
 	Encryption BatchJobKeyRotateEncryption `yaml:"encryption" json:"encryption"`
 }

@@ -261,6 +260,9 @@ func (r *BatchJobKeyRotateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 	if err := ri.load(ctx, api, job); err != nil {
 		return err
 	}
+	if ri.Complete {
+		return nil
+	}

 	globalBatchJobsMetrics.save(job.ID, ri)
 	lastObject := ri.Object
@@ -356,9 +358,9 @@ func (r *BatchJobKeyRotateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 	ctx, cancel := context.WithCancel(ctx)

 	results := make(chan ObjectInfo, 100)
-	if err := api.Walk(ctx, r.Bucket, r.Prefix, results, ObjectOptions{
-		WalkMarker: lastObject,
-		WalkFilter: skip,
+	if err := api.Walk(ctx, r.Bucket, r.Prefix, results, WalkOptions{
+		Marker: lastObject,
+		Filter: skip,
 	}); err != nil {
 		cancel()
 		// Do not need to retry if we can't list objects on source.
@@ -377,14 +379,14 @@ func (r *BatchJobKeyRotateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 			defer wk.Give()
 			for attempts := 1; attempts <= retryAttempts; attempts++ {
 				attempts := attempts
-				stopFn := globalBatchJobsMetrics.trace(batchKeyRotationMetricObject, job.ID, attempts, result)
+				stopFn := globalBatchJobsMetrics.trace(batchJobMetricKeyRotation, job.ID, attempts)
 				success := true
 				if err := r.KeyRotate(ctx, api, result); err != nil {
-					stopFn(err)
+					stopFn(result, err)
 					logger.LogIf(ctx, err)
 					success = false
 				} else {
-					stopFn(nil)
+					stopFn(result, nil)
 				}
 				ri.trackCurrentBucketObject(r.Bucket, result, success)
 				ri.RetryAttempts = attempts
@@ -398,6 +400,10 @@ func (r *BatchJobKeyRotateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 					time.Sleep(delay + time.Duration(rnd.Float64()*float64(delay)))
 				}
 			}
+
+			if wait := globalBatchConfig.KeyRotationWait(); wait > 0 {
+				time.Sleep(wait)
+			}
 		}()
 	}
 	wk.Wait()
@@ -409,12 +409,6 @@ func (z *BatchJobKeyRotateV1) DecodeMsg(dc *msgp.Reader) (err error) {
 				err = msgp.WrapError(err, "Prefix")
 				return
 			}
-		case "Endpoint":
-			z.Endpoint, err = dc.ReadString()
-			if err != nil {
-				err = msgp.WrapError(err, "Endpoint")
-				return
-			}
 		case "Encryption":
 			err = z.Encryption.DecodeMsg(dc)
 			if err != nil {
@@ -434,9 +428,9 @@ func (z *BatchJobKeyRotateV1) DecodeMsg(dc *msgp.Reader) (err error) {

 // EncodeMsg implements msgp.Encodable
 func (z *BatchJobKeyRotateV1) EncodeMsg(en *msgp.Writer) (err error) {
-	// map header, size 6
+	// map header, size 5
 	// write "APIVersion"
-	err = en.Append(0x86, 0xaa, 0x41, 0x50, 0x49, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e)
+	err = en.Append(0x85, 0xaa, 0x41, 0x50, 0x49, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e)
 	if err != nil {
 		return
 	}
@@ -501,16 +495,6 @@ func (z *BatchJobKeyRotateV1) EncodeMsg(en *msgp.Writer) (err error) {
 		err = msgp.WrapError(err, "Prefix")
 		return
 	}
-	// write "Endpoint"
-	err = en.Append(0xa8, 0x45, 0x6e, 0x64, 0x70, 0x6f, 0x69, 0x6e, 0x74)
-	if err != nil {
-		return
-	}
-	err = en.WriteString(z.Endpoint)
-	if err != nil {
-		err = msgp.WrapError(err, "Endpoint")
-		return
-	}
 	// write "Encryption"
 	err = en.Append(0xaa, 0x45, 0x6e, 0x63, 0x72, 0x79, 0x70, 0x74, 0x69, 0x6f, 0x6e)
 	if err != nil {
@@ -527,9 +511,9 @@ func (z *BatchJobKeyRotateV1) EncodeMsg(en *msgp.Writer) (err error) {
 // MarshalMsg implements msgp.Marshaler
 func (z *BatchJobKeyRotateV1) MarshalMsg(b []byte) (o []byte, err error) {
 	o = msgp.Require(b, z.Msgsize())
-	// map header, size 6
+	// map header, size 5
 	// string "APIVersion"
-	o = append(o, 0x86, 0xaa, 0x41, 0x50, 0x49, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e)
+	o = append(o, 0x85, 0xaa, 0x41, 0x50, 0x49, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e)
 	o = msgp.AppendString(o, z.APIVersion)
 	// string "Flags"
 	o = append(o, 0xa5, 0x46, 0x6c, 0x61, 0x67, 0x73)
@@ -561,9 +545,6 @@ func (z *BatchJobKeyRotateV1) MarshalMsg(b []byte) (o []byte, err error) {
 	// string "Prefix"
 	o = append(o, 0xa6, 0x50, 0x72, 0x65, 0x66, 0x69, 0x78)
 	o = msgp.AppendString(o, z.Prefix)
-	// string "Endpoint"
-	o = append(o, 0xa8, 0x45, 0x6e, 0x64, 0x70, 0x6f, 0x69, 0x6e, 0x74)
-	o = msgp.AppendString(o, z.Endpoint)
 	// string "Encryption"
 	o = append(o, 0xaa, 0x45, 0x6e, 0x63, 0x72, 0x79, 0x70, 0x74, 0x69, 0x6f, 0x6e)
 	o, err = z.Encryption.MarshalMsg(o)
@@ -651,12 +632,6 @@ func (z *BatchJobKeyRotateV1) UnmarshalMsg(bts []byte) (o []byte, err error) {
 				err = msgp.WrapError(err, "Prefix")
 				return
 			}
-		case "Endpoint":
-			z.Endpoint, bts, err = msgp.ReadStringBytes(bts)
-			if err != nil {
-				err = msgp.WrapError(err, "Endpoint")
-				return
-			}
 		case "Encryption":
 			bts, err = z.Encryption.UnmarshalMsg(bts)
 			if err != nil {
@@ -677,7 +652,7 @@ func (z *BatchJobKeyRotateV1) 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 *BatchJobKeyRotateV1) Msgsize() (s int) {
-	s = 1 + 11 + msgp.StringPrefixSize + len(z.APIVersion) + 6 + 1 + 7 + z.Flags.Filter.Msgsize() + 7 + z.Flags.Notify.Msgsize() + 6 + z.Flags.Retry.Msgsize() + 7 + msgp.StringPrefixSize + len(z.Bucket) + 7 + msgp.StringPrefixSize + len(z.Prefix) + 9 + msgp.StringPrefixSize + len(z.Endpoint) + 11 + z.Encryption.Msgsize()
+	s = 1 + 11 + msgp.StringPrefixSize + len(z.APIVersion) + 6 + 1 + 7 + z.Flags.Filter.Msgsize() + 7 + z.Flags.Notify.Msgsize() + 6 + z.Flags.Retry.Msgsize() + 7 + msgp.StringPrefixSize + len(z.Bucket) + 7 + msgp.StringPrefixSize + len(z.Prefix) + 11 + z.Encryption.Msgsize()
 	return
 }

@@ -8,13 +8,14 @@ func _() {
 	// An "invalid array index" compiler error signifies that the constant values have changed.
 	// Re-run the stringer command to generate them again.
 	var x [1]struct{}
-	_ = x[batchReplicationMetricObject-0]
-	_ = x[batchKeyRotationMetricObject-1]
+	_ = x[batchJobMetricReplication-0]
+	_ = x[batchJobMetricKeyRotation-1]
+	_ = x[batchJobMetricExpire-2]
 }

-const _batchJobMetric_name = "batchReplicationMetricObjectbatchKeyRotationMetricObject"
+const _batchJobMetric_name = "ReplicationKeyRotationExpire"

-var _batchJobMetric_index = [...]uint8{0, 28, 56}
+var _batchJobMetric_index = [...]uint8{0, 11, 22, 28}

 func (i batchJobMetric) String() string {
 	if i >= batchJobMetric(len(_batchJobMetric_index)-1) {
@@ -95,7 +95,7 @@ func newStreamingBitrotWriter(disk StorageAPI, volume, filePath string, length i
 	h := algo.New()

 	bw := &streamingBitrotWriter{
-		iow:          ioutil.NewDeadlineWriter(w, diskMaxTimeout),
+		iow:          ioutil.NewDeadlineWriter(w, globalDriveConfig.GetMaxTimeout()),
 		closeWithErr: w.CloseWithError,
 		h:            h,
 		shardSize:    shardSize,
@@ -20,12 +20,14 @@ package cmd
 import (
 	"context"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"io"
 	"net/http"
 	"net/url"
 	"os"
 	"reflect"
+	"strings"
 	"time"

 	"github.com/minio/minio-go/v7/pkg/set"
@@ -82,23 +84,40 @@ func (s1 ServerSystemConfig) Diff(s2 ServerSystemConfig) error {
 				ep.Platform, s2.MinioEndpoints[i].Platform)
 		}
 	}
-	if !reflect.DeepEqual(s1.MinioEnv, s2.MinioEnv) {
-		var missing []string
-		var mismatching []string
-		for k, v := range s1.MinioEnv {
-			ev, ok := s2.MinioEnv[k]
-			if !ok {
-				missing = append(missing, k)
-			} else if v != ev {
-				mismatching = append(mismatching, k)
-			}
-		}
-		if len(mismatching) > 0 {
-			return fmt.Errorf(`Expected same MINIO_ environment variables and values across all servers: Missing environment values: %s / Mismatch environment values: %s`, missing, mismatching)
-		}
-		return fmt.Errorf(`Expected same MINIO_ environment variables and values across all servers: Missing environment values: %s`, missing)
+	if reflect.DeepEqual(s1.MinioEnv, s2.MinioEnv) {
+		return nil
 	}
-	return nil
+
+	// Report differences in environment variables.
+	var missing []string
+	var mismatching []string
+	for k, v := range s1.MinioEnv {
+		ev, ok := s2.MinioEnv[k]
+		if !ok {
+			missing = append(missing, k)
+		} else if v != ev {
+			mismatching = append(mismatching, k)
+		}
+	}
+	var extra []string
+	for k := range s2.MinioEnv {
+		_, ok := s1.MinioEnv[k]
+		if !ok {
+			extra = append(extra, k)
+		}
+	}
+	msg := "Expected same MINIO_ environment variables and values across all servers: "
+	if len(missing) > 0 {
+		msg += fmt.Sprintf(`Missing environment values: %v. `, missing)
+	}
+	if len(mismatching) > 0 {
+		msg += fmt.Sprintf(`Mismatching environment values: %v. `, mismatching)
+	}
+	if len(extra) > 0 {
+		msg += fmt.Sprintf(`Extra environment values: %v. `, extra)
+	}
+
+	return errors.New(strings.TrimSpace(msg))
 }

 var skipEnvs = map[string]struct{}{
@@ -51,6 +51,7 @@ import (
 	sse "github.com/minio/minio/internal/bucket/encryption"
 	objectlock "github.com/minio/minio/internal/bucket/object/lock"
 	"github.com/minio/minio/internal/bucket/replication"
+	"github.com/minio/minio/internal/config/cache"
 	"github.com/minio/minio/internal/config/dns"
 	"github.com/minio/minio/internal/crypto"
 	"github.com/minio/minio/internal/event"
@@ -668,6 +669,8 @@ func (api objectAPIHandlers) DeleteMultipleObjectsHandler(w http.ResponseWriter,
 			continue
 		}

+		defer globalCacheConfig.Delete(bucket, dobj.ObjectName)
+
 		if replicateDeletes && (dobj.DeleteMarkerReplicationStatus() == replication.Pending || dobj.VersionPurgeStatus() == Pending) {
 			// copy so we can re-add null ID.
 			dobj := dobj
@@ -1192,7 +1195,7 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 	})

 	var opts ObjectOptions
-	opts, err = putOpts(ctx, r, bucket, object, metadata)
+	opts, err = putOptsFromReq(ctx, r, bucket, object, metadata)
 	if err != nil {
 		writeErrorResponseHeadersOnly(w, toAPIError(ctx, err))
 		return
@@ -1343,6 +1346,22 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 					continue
 				}

+				asize, err := objInfo.GetActualSize()
+				if err != nil {
+					asize = objInfo.Size
+				}
+
+				globalCacheConfig.Set(&cache.ObjectInfo{
+					Key:          objInfo.Name,
+					Bucket:       objInfo.Bucket,
+					ETag:         getDecryptedETag(formValues, objInfo, false),
+					ModTime:      objInfo.ModTime,
+					Expires:      objInfo.Expires.UTC().Format(http.TimeFormat),
+					CacheControl: objInfo.CacheControl,
+					Metadata:     cleanReservedKeys(objInfo.UserDefined),
+					Size:         asize,
+				})
+
 				fanOutResp = append(fanOutResp, minio.PutObjectFanOutResponse{
 					Key:          objInfo.Name,
 					ETag:         getDecryptedETag(formValues, objInfo, false),
@@ -1399,10 +1418,12 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 		return
 	}

+	etag := getDecryptedETag(formValues, objInfo, false)
+
 	// We must not use the http.Header().Set method here because some (broken)
 	// clients expect the ETag header key to be literally "ETag" - not "Etag" (case-sensitive).
 	// Therefore, we have to set the ETag directly as map entry.
-	w.Header()[xhttp.ETag] = []string{`"` + objInfo.ETag + `"`}
+	w.Header()[xhttp.ETag] = []string{`"` + etag + `"`}

 	// Set the relevant version ID as part of the response header.
 	if objInfo.VersionID != "" && objInfo.VersionID != nullVersionID {
@@ -1413,6 +1434,22 @@ func (api objectAPIHandlers) PostPolicyBucketHandler(w http.ResponseWriter, r *h
 		w.Header().Set(xhttp.Location, obj)
 	}

+	asize, err := objInfo.GetActualSize()
+	if err != nil {
+		asize = objInfo.Size
+	}
+
+	defer globalCacheConfig.Set(&cache.ObjectInfo{
+		Key:          objInfo.Name,
+		Bucket:       objInfo.Bucket,
+		ETag:         etag,
+		ModTime:      objInfo.ModTime,
+		Expires:      objInfo.ExpiresStr(),
+		CacheControl: objInfo.CacheControl,
+		Metadata:     cleanReservedKeys(objInfo.UserDefined),
+		Size:         asize,
+	})
+
 	// Notify object created event.
 	defer sendEvent(eventArgs{
 		EventName:    event.ObjectCreatedPost,
@@ -1620,9 +1657,11 @@ func (api objectAPIHandlers) DeleteBucketHandler(w http.ResponseWriter, r *http.
 	}

 	// Return an error if the bucket does not exist
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil && !forceDelete {
-		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
-		return
+	if !forceDelete {
+		if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
+			writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
+			return
+		}
 	}

 	// Attempt to delete bucket.
@@ -25,6 +25,7 @@ type lcEventSrc uint8
 //revive:disable:var-naming Underscores is used here to indicate where common prefix ends and the enumeration name begins
 const (
 	lcEventSrc_None lcEventSrc = iota
+	lcEventSrc_Heal
 	lcEventSrc_Scanner
 	lcEventSrc_Decom
 	lcEventSrc_Rebal
@@ -22,6 +22,7 @@ import (
 	"io"
 	"net/http"
 	"strconv"
+	"time"

 	"github.com/minio/minio/internal/bucket/lifecycle"
 	xhttp "github.com/minio/minio/internal/http"
@@ -86,6 +87,41 @@ func (api objectAPIHandlers) PutBucketLifecycleHandler(w http.ResponseWriter, r
 		return
 	}

+	// Create a map of updated set of rules in request
+	updatedRules := make(map[string]lifecycle.Rule, len(bucketLifecycle.Rules))
+	for _, rule := range bucketLifecycle.Rules {
+		updatedRules[rule.ID] = rule
+	}
+
+	// Get list of rules for the bucket from disk
+	meta, err := globalBucketMetadataSys.GetConfigFromDisk(ctx, bucket)
+	if err != nil {
+		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
+		return
+	}
+	expiryRuleRemoved := false
+	if len(meta.LifecycleConfigXML) > 0 {
+		var lcCfg lifecycle.Lifecycle
+		if err := xml.Unmarshal(meta.LifecycleConfigXML, &lcCfg); err != nil {
+			writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
+			return
+		}
+		for _, rl := range lcCfg.Rules {
+			updRule, ok := updatedRules[rl.ID]
+			// original rule had expiry that is no longer in the new config,
+			// or rule is present but missing expiration flags
+			if (!rl.Expiration.IsNull() || !rl.NoncurrentVersionExpiration.IsNull()) &&
+				(!ok || (updRule.Expiration.IsNull() && updRule.NoncurrentVersionExpiration.IsNull())) {
+				expiryRuleRemoved = true
+			}
+		}
+	}
+
+	if bucketLifecycle.HasExpiry() || expiryRuleRemoved {
+		currtime := time.Now()
+		bucketLifecycle.ExpiryUpdatedAt = &currtime
+	}
+
 	configData, err := xml.Marshal(bucketLifecycle)
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
@@ -142,6 +178,8 @@ func (api objectAPIHandlers) GetBucketLifecycleHandler(w http.ResponseWriter, r
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
+	// explicitly set ExpiryUpdatedAt nil as its meant for internal consumption only
+	config.ExpiryUpdatedAt = nil

 	configData, err := xml.Marshal(config)
 	if err != nil {
@@ -42,7 +42,9 @@ import (
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/minio/internal/s3select"
 	"github.com/minio/pkg/v2/env"
+	xnet "github.com/minio/pkg/v2/net"
 	"github.com/minio/pkg/v2/workers"
+	"github.com/zeebo/xxh3"
 )

 const (
@@ -217,17 +219,24 @@ type transitionState struct {
 	numWorkers int
 	killCh     chan struct{}

-	activeTasks int32
+	activeTasks          atomic.Int64
+	missedImmediateTasks atomic.Int64

 	lastDayMu    sync.RWMutex
 	lastDayStats map[string]*lastDayTierStats
 }

 func (t *transitionState) queueTransitionTask(oi ObjectInfo, event lifecycle.Event, src lcEventSrc) {
+	task := transitionTask{objInfo: oi, event: event, src: src}
 	select {
 	case <-t.ctx.Done():
-	case t.transitionCh <- transitionTask{objInfo: oi, event: event, src: src}:
+	case t.transitionCh <- task:
 	default:
+		switch src {
+		case lcEventSrc_s3PutObject, lcEventSrc_s3CopyObject, lcEventSrc_s3CompleteMultipartUpload:
+			// Update missed immediate tasks only for incoming requests.
+			t.missedImmediateTasks.Add(1)
+		}
 	}
 }

@@ -237,7 +246,7 @@ var globalTransitionState *transitionState
 // via its Init method.
 func newTransitionState(ctx context.Context) *transitionState {
 	return &transitionState{
-		transitionCh: make(chan transitionTask, 10000),
+		transitionCh: make(chan transitionTask, 100000),
 		ctx:          ctx,
 		killCh:       make(chan struct{}),
 		lastDayStats: make(map[string]*lastDayTierStats),
@@ -262,8 +271,14 @@ func (t *transitionState) PendingTasks() int {
 }

 // ActiveTasks returns the number of active (ongoing) ILM transition tasks.
-func (t *transitionState) ActiveTasks() int {
-	return int(atomic.LoadInt32(&t.activeTasks))
+func (t *transitionState) ActiveTasks() int64 {
+	return t.activeTasks.Load()
+}
+
+// MissedImmediateTasks returns the number of tasks - deferred to scanner due
+// to tasks channel being backlogged.
+func (t *transitionState) MissedImmediateTasks() int64 {
+	return t.missedImmediateTasks.Load()
 }

 // worker waits for transition tasks
@@ -278,11 +293,13 @@ func (t *transitionState) worker(objectAPI ObjectLayer) {
 			if !ok {
 				return
 			}
-			atomic.AddInt32(&t.activeTasks, 1)
+			t.activeTasks.Add(1)
 			if err := transitionObject(t.ctx, objectAPI, task.objInfo, newLifecycleAuditEvent(task.src, task.event)); err != nil {
-				if !isErrVersionNotFound(err) && !isErrObjectNotFound(err) {
-					logger.LogIf(t.ctx, fmt.Errorf("Transition to %s failed for %s/%s version:%s with %w",
-						task.event.StorageClass, task.objInfo.Bucket, task.objInfo.Name, task.objInfo.VersionID, err))
+				if !isErrVersionNotFound(err) && !isErrObjectNotFound(err) && !xnet.IsNetworkOrHostDown(err, false) {
+					if !strings.Contains(err.Error(), "use of closed network connection") {
+						logger.LogIf(t.ctx, fmt.Errorf("Transition to %s failed for %s/%s version:%s with %w",
+							task.event.StorageClass, task.objInfo.Bucket, task.objInfo.Name, task.objInfo.VersionID, err))
+					}
 				}
 			} else {
 				ts := tierStats{
@@ -294,7 +311,7 @@ func (t *transitionState) worker(objectAPI ObjectLayer) {
 				}
 				t.addLastDayStats(task.event.StorageClass, ts)
 			}
-			atomic.AddInt32(&t.activeTasks, -1)
+			t.activeTasks.Add(-1)
 		}
 	}
 }
@@ -444,7 +461,8 @@ func genTransitionObjName(bucket string) (string, error) {
 		return "", err
 	}
 	us := u.String()
-	obj := fmt.Sprintf("%s/%s/%s/%s/%s", globalDeploymentID(), bucket, us[0:2], us[2:4], us)
+	hash := xxh3.HashString(pathJoin(globalDeploymentID(), bucket))
+	obj := fmt.Sprintf("%s/%s/%s/%s", strconv.FormatUint(hash, 16), us[0:2], us[2:4], us)
 	return obj, nil
 }

@@ -489,9 +507,13 @@ func auditTierActions(ctx context.Context, tier string, bytes int64) func(err er
 		}

 		if err == nil {
-			op.TimeToResponseNS = time.Since(startTime).Nanoseconds()
+			since := time.Since(startTime)
+			op.TimeToResponseNS = since.Nanoseconds()
+			globalTierMetrics.Observe(tier, since)
+			globalTierMetrics.logSuccess(tier)
 		} else {
 			op.Error = err.Error()
+			globalTierMetrics.logFailure(tier)
 		}

 		logger.GetReqInfo(ctx).AppendTags("tierStats", op)
@@ -862,6 +884,7 @@ func (oi ObjectInfo) ToLifecycleOpts() lifecycle.ObjectOpts {
 		UserTags:         oi.UserTags,
 		VersionID:        oi.VersionID,
 		ModTime:          oi.ModTime,
+		Size:             oi.Size,
 		IsLatest:         oi.IsLatest,
 		NumVersions:      oi.NumVersions,
 		DeleteMarker:     oi.DeleteMarker,
@@ -19,6 +19,7 @@ package cmd

 import (
 	"context"
+	"encoding/xml"
 	"errors"
 	"fmt"
 	"sync"
@@ -177,6 +178,40 @@ func (sys *BucketMetadataSys) save(ctx context.Context, meta BucketMetadata) err
 // Delete delete the bucket metadata for the specified bucket.
 // must be used by all callers instead of using Update() with nil configData.
 func (sys *BucketMetadataSys) Delete(ctx context.Context, bucket string, configFile string) (updatedAt time.Time, err error) {
+	if configFile == bucketLifecycleConfig {
+		// Get bucket config from current site
+		meta, e := globalBucketMetadataSys.GetConfigFromDisk(ctx, bucket)
+		if e != nil && !errors.Is(e, errConfigNotFound) {
+			return updatedAt, e
+		}
+		var expiryRuleRemoved bool
+		if len(meta.LifecycleConfigXML) > 0 {
+			var lcCfg lifecycle.Lifecycle
+			if err := xml.Unmarshal(meta.LifecycleConfigXML, &lcCfg); err != nil {
+				return updatedAt, err
+			}
+			// find a single expiry rule set the flag
+			for _, rl := range lcCfg.Rules {
+				if !rl.Expiration.IsNull() || !rl.NoncurrentVersionExpiration.IsNull() {
+					expiryRuleRemoved = true
+					break
+				}
+			}
+		}
+
+		// Form empty ILM details with `ExpiryUpdatedAt` field and save
+		var cfgData []byte
+		if expiryRuleRemoved {
+			var lcCfg lifecycle.Lifecycle
+			currtime := time.Now()
+			lcCfg.ExpiryUpdatedAt = &currtime
+			cfgData, err = xml.Marshal(lcCfg)
+			if err != nil {
+				return updatedAt, err
+			}
+		}
+		return sys.updateAndParse(ctx, bucket, configFile, cfgData, false)
+	}
 	return sys.updateAndParse(ctx, bucket, configFile, nil, false)
 }

@@ -267,7 +302,10 @@ func (sys *BucketMetadataSys) GetLifecycleConfig(bucket string) (*lifecycle.Life
 		}
 		return nil, time.Time{}, err
 	}
-	if meta.lifecycleConfig == nil {
+	// there could be just `ExpiryUpdatedAt` field populated as part
+	// of last delete all. Treat this situation as not lifecycle configuration
+	// available
+	if meta.lifecycleConfig == nil || len(meta.lifecycleConfig.Rules) == 0 {
 		return nil, time.Time{}, BucketLifecycleNotFound{Bucket: bucket}
 	}
 	return meta.lifecycleConfig, meta.LifecycleConfigUpdatedAt, nil
@@ -437,21 +475,8 @@ func (sys *BucketMetadataSys) Init(ctx context.Context, buckets []BucketInfo, ob
 	return nil
 }

-func (sys *BucketMetadataSys) loadBucketMetadata(ctx context.Context, bucket BucketInfo) error {
-	meta, err := loadBucketMetadata(ctx, sys.objAPI, bucket.Name)
-	if err != nil {
-		return err
-	}
-
-	sys.Lock()
-	sys.metadataMap[bucket.Name] = meta
-	sys.Unlock()
-
-	return nil
-}
-
 // concurrently load bucket metadata to speed up loading bucket metadata.
-func (sys *BucketMetadataSys) concurrentLoad(ctx context.Context, buckets []BucketInfo) {
+func (sys *BucketMetadataSys) concurrentLoad(ctx context.Context, buckets []BucketInfo, failedBuckets map[string]struct{}) {
 	g := errgroup.WithNErrs(len(buckets))
 	bucketMetas := make([]BucketMetadata, len(buckets))
 	for index := range buckets {
@@ -491,6 +516,10 @@ func (sys *BucketMetadataSys) concurrentLoad(ctx context.Context, buckets []Buck

 	for i, meta := range bucketMetas {
 		if errs[i] != nil {
+			if failedBuckets == nil {
+				failedBuckets = make(map[string]struct{})
+			}
+			failedBuckets[buckets[i].Name] = struct{}{}
 			continue
 		}
 		globalEventNotifier.set(buckets[i], meta)   // set notification targets
@@ -498,7 +527,7 @@ func (sys *BucketMetadataSys) concurrentLoad(ctx context.Context, buckets []Buck
 	}
 }

-func (sys *BucketMetadataSys) refreshBucketsMetadataLoop(ctx context.Context) {
+func (sys *BucketMetadataSys) refreshBucketsMetadataLoop(ctx context.Context, failedBuckets map[string]struct{}) {
 	const bucketMetadataRefresh = 15 * time.Minute

 	sleeper := newDynamicSleeper(2, 150*time.Millisecond, false)
@@ -513,7 +542,7 @@ func (sys *BucketMetadataSys) refreshBucketsMetadataLoop(ctx context.Context) {
 			buckets, err := sys.objAPI.ListBuckets(ctx, BucketOptions{})
 			if err != nil {
 				logger.LogIf(ctx, err)
-				continue
+				break
 			}

 			// Handle if we have some buckets in-memory those are stale.
@@ -525,38 +554,49 @@ func (sys *BucketMetadataSys) refreshBucketsMetadataLoop(ctx context.Context) {
 			}
 			sys.RemoveStaleBuckets(diskBuckets)

-			for _, bucket := range buckets {
+			for i := range buckets {
 				wait := sleeper.Timer(ctx)

-				err := sys.loadBucketMetadata(ctx, bucket)
+				meta, err := loadBucketMetadata(ctx, sys.objAPI, buckets[i].Name)
 				if err != nil {
 					logger.LogIf(ctx, err)
 					wait() // wait to proceed to next entry.
 					continue
 				}

+				sys.Lock()
+				sys.metadataMap[buckets[i].Name] = meta
+				sys.Unlock()
+
+				// Initialize the failed buckets
+				if _, ok := failedBuckets[buckets[i].Name]; ok {
+					globalEventNotifier.set(buckets[i], meta)
+					globalBucketTargetSys.set(buckets[i], meta)
+					delete(failedBuckets, buckets[i].Name)
+				}
+
 				wait() // wait to proceed to next entry.
 			}
-
-			t.Reset(bucketMetadataRefresh)
 		}
+		t.Reset(bucketMetadataRefresh)
 	}
 }

 // Loads bucket metadata for all buckets into BucketMetadataSys.
 func (sys *BucketMetadataSys) init(ctx context.Context, buckets []BucketInfo) {
 	count := 100 // load 100 bucket metadata at a time.
+	failedBuckets := make(map[string]struct{})
 	for {
 		if len(buckets) < count {
-			sys.concurrentLoad(ctx, buckets)
+			sys.concurrentLoad(ctx, buckets, failedBuckets)
 			break
 		}
-		sys.concurrentLoad(ctx, buckets[:count])
+		sys.concurrentLoad(ctx, buckets[:count], failedBuckets)
 		buckets = buckets[count:]
 	}

 	if globalIsDistErasure {
-		go sys.refreshBucketsMetadataLoop(ctx)
+		go sys.refreshBucketsMetadataLoop(ctx, failedBuckets)
 	}
 }

@@ -121,7 +121,7 @@ func getConditionValues(r *http.Request, lc string, cred auth.Credentials) map[s
 		"CurrentTime":      {currTime.Format(time.RFC3339)},
 		"EpochTime":        {strconv.FormatInt(currTime.Unix(), 10)},
 		"SecureTransport":  {strconv.FormatBool(r.TLS != nil)},
-		"SourceIp":         {handlers.GetSourceIP(r)},
+		"SourceIp":         {handlers.GetSourceIPRaw(r)},
 		"UserAgent":        {r.UserAgent()},
 		"Referer":          {r.Referer()},
 		"principaltype":    {principalType},
@@ -137,6 +137,10 @@ func getConditionValues(r *http.Request, lc string, cred auth.Credentials) map[s
 	}

 	cloneHeader := r.Header.Clone()
+	if v := cloneHeader.Get("x-amz-signature-age"); v != "" {
+		args["signatureAge"] = []string{v}
+		cloneHeader.Del("x-amz-signature-age")
+	}

 	if userTags := cloneHeader.Get(xhttp.AmzObjectTagging); userTags != "" {
 		tag, _ := tags.ParseObjectTags(userTags)
@@ -337,13 +337,13 @@ type SMA struct {
 	filledBuf bool
 }

-func newSMA(len int) *SMA {
-	if len <= 0 {
-		len = defaultWindowSize
+func newSMA(ln int) *SMA {
+	if ln <= 0 {
+		ln = defaultWindowSize
 	}
 	return &SMA{
-		buf:    make([]float64, len),
-		window: len,
+		buf:    make([]float64, ln),
+		window: ln,
 		idx:    0,
 	}
 }
@@ -23,7 +23,6 @@ import (
 	"errors"
 	"fmt"
 	"io"
-	"math"
 	"math/rand"
 	"net/http"
 	"net/url"
@@ -373,15 +372,14 @@ func checkReplicateDelete(ctx context.Context, bucket string, dobj ObjectToDelet
 			if oi.DeleteMarker && (validReplStatus || replicate) {
 				dsc.Set(newReplicateTargetDecision(tgtArn, replicate, sync))
 				continue
-			} else {
-				// can be the case that other cluster is down and duplicate `mc rm --vid`
-				// is issued - this still needs to be replicated back to the other target
-				if !oi.VersionPurgeStatus.Empty() {
-					replicate = oi.VersionPurgeStatus == Pending || oi.VersionPurgeStatus == Failed
-					dsc.Set(newReplicateTargetDecision(tgtArn, replicate, sync))
-				}
-				continue
 			}
+			// can be the case that other cluster is down and duplicate `mc rm --vid`
+			// is issued - this still needs to be replicated back to the other target
+			if !oi.VersionPurgeStatus.Empty() {
+				replicate = oi.VersionPurgeStatus == Pending || oi.VersionPurgeStatus == Failed
+				dsc.Set(newReplicateTargetDecision(tgtArn, replicate, sync))
+			}
+			continue
 		}
 		tgt := globalBucketTargetSys.GetRemoteTargetClient(bucket, tgtArn)
 		// the target online status should not be used here while deciding
@@ -639,19 +637,23 @@ func replicateDeleteToTarget(ctx context.Context, dobj DeletedObjectReplicationI
 				IsReplicationReadyForDeleteMarker: true,
 			},
 		})
+		serr := ErrorRespToObjectError(err, dobj.Bucket, dobj.ObjectName, dobj.VersionID)
 		switch {
-		case isErrMethodNotAllowed(ErrorRespToObjectError(err, dobj.Bucket, dobj.ObjectName)):
+		case isErrMethodNotAllowed(serr):
 			// delete marker already replicated
 			if dobj.VersionID == "" && rinfo.VersionPurgeStatus.Empty() {
 				rinfo.ReplicationStatus = replication.Completed
 				return
 			}
-		case isErrObjectNotFound(ErrorRespToObjectError(err, dobj.Bucket, dobj.ObjectName)):
+		case isErrObjectNotFound(serr), isErrVersionNotFound(serr):
 			// version being purged is already not found on target.
 			if !rinfo.VersionPurgeStatus.Empty() {
 				rinfo.VersionPurgeStatus = Complete
 				return
 			}
+		case isErrReadQuorum(serr), isErrWriteQuorum(serr):
+			// destination has some quorum issues, perform removeObject() anyways
+			// to complete the operation.
 		default:
 			if err != nil && minio.IsNetworkOrHostDown(err, true) && !globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
 				globalBucketTargetSys.markOffline(tgt.EndpointURL())
@@ -777,6 +779,7 @@ func putReplicationOpts(ctx context.Context, sc string, objInfo ObjectInfo) (put
 		UserMetadata:    meta,
 		ContentType:     objInfo.ContentType,
 		ContentEncoding: objInfo.ContentEncoding,
+		Expires:         objInfo.Expires,
 		StorageClass:    sc,
 		Internal: minio.AdvancedPutOptions{
 			SourceVersionID:    objInfo.VersionID,
@@ -1382,7 +1385,6 @@ func (ri ReplicateObjectInfo) replicateAll(ctx context.Context, objectAPI Object
 		rinfo.Duration = time.Since(startTime)
 	}()

-	rAction = replicateAll
 	oi, cerr := tgt.StatObject(ctx, tgt.Bucket, object, minio.StatObjectOptions{
 		VersionID: objInfo.VersionID,
 		Internal: minio.AdvancedGetOptions{
@@ -1419,16 +1421,19 @@ func (ri ReplicateObjectInfo) replicateAll(ctx context.Context, objectAPI Object
 			}
 			return
 		}
-	}
-	// if target returns error other than NoSuchKey, defer replication attempt
-	if cerr != nil {
+	} else {
+		// if target returns error other than NoSuchKey, defer replication attempt
 		if minio.IsNetworkOrHostDown(cerr, true) && !globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
 			globalBucketTargetSys.markOffline(tgt.EndpointURL())
 		}

-		errResp := minio.ToErrorResponse(cerr)
-		switch errResp.Code {
-		case "NoSuchKey", "NoSuchVersion", "SlowDownRead":
+		serr := ErrorRespToObjectError(cerr, bucket, object, objInfo.VersionID)
+		switch {
+		case isErrMethodNotAllowed(serr):
+			rAction = replicateAll
+		case isErrObjectNotFound(serr), isErrVersionNotFound(serr):
+			rAction = replicateAll
+		case isErrReadQuorum(serr), isErrWriteQuorum(serr):
 			rAction = replicateAll
 		default:
 			rinfo.Err = cerr
@@ -1462,6 +1467,24 @@ func (ri ReplicateObjectInfo) replicateAll(ctx context.Context, objectAPI Object
 				ReplicationRequest: true, // always set this to distinguish between `mc mirror` replication and serverside
 			},
 		}
+		if tagTmStr, ok := objInfo.UserDefined[ReservedMetadataPrefixLower+TaggingTimestamp]; ok {
+			ondiskTimestamp, err := time.Parse(time.RFC3339, tagTmStr)
+			if err == nil {
+				dstOpts.Internal.TaggingTimestamp = ondiskTimestamp
+			}
+		}
+		if retTmStr, ok := objInfo.UserDefined[ReservedMetadataPrefixLower+ObjectLockRetentionTimestamp]; ok {
+			ondiskTimestamp, err := time.Parse(time.RFC3339, retTmStr)
+			if err == nil {
+				dstOpts.Internal.RetentionTimestamp = ondiskTimestamp
+			}
+		}
+		if lholdTmStr, ok := objInfo.UserDefined[ReservedMetadataPrefixLower+ObjectLockLegalHoldTimestamp]; ok {
+			ondiskTimestamp, err := time.Parse(time.RFC3339, lholdTmStr)
+			if err == nil {
+				dstOpts.Internal.LegalholdTimestamp = ondiskTimestamp
+			}
+		}
 		if _, rinfo.Err = c.CopyObject(ctx, tgt.Bucket, object, tgt.Bucket, object, getCopyObjMetadata(objInfo, tgt.StorageClass), srcOpts, dstOpts); rinfo.Err != nil {
 			rinfo.ReplicationStatus = replication.Failed
 			logger.LogIf(ctx, fmt.Errorf("unable to replicate metadata for object %s/%s(%s): %s", bucket, objInfo.Name, objInfo.VersionID, rinfo.Err))
@@ -1691,12 +1714,13 @@ type ReplicationPool struct {
 	activeWorkers    int32
 	activeMRFWorkers int32

-	objLayer ObjectLayer
-	ctx      context.Context
-	priority string
-	mu       sync.RWMutex
-	mrfMU    sync.Mutex
-	resyncer *replicationResyncer
+	objLayer   ObjectLayer
+	ctx        context.Context
+	priority   string
+	maxWorkers int
+	mu         sync.RWMutex
+	mrfMU      sync.Mutex
+	resyncer   *replicationResyncer

 	// workers:
 	workers    []chan ReplicationWorkerOperation
@@ -1742,9 +1766,13 @@ const (
 func NewReplicationPool(ctx context.Context, o ObjectLayer, opts replicationPoolOpts) *ReplicationPool {
 	var workers, failedWorkers int
 	priority := "auto"
+	maxWorkers := WorkerMaxLimit
 	if opts.Priority != "" {
 		priority = opts.Priority
 	}
+	if opts.MaxWorkers > 0 {
+		maxWorkers = opts.MaxWorkers
+	}
 	switch priority {
 	case "fast":
 		workers = WorkerMaxLimit
@@ -1756,7 +1784,13 @@ func NewReplicationPool(ctx context.Context, o ObjectLayer, opts replicationPool
 		workers = WorkerAutoDefault
 		failedWorkers = MRFWorkerAutoDefault
 	}
+	if maxWorkers > 0 && workers > maxWorkers {
+		workers = maxWorkers
+	}

+	if maxWorkers > 0 && failedWorkers > maxWorkers {
+		failedWorkers = maxWorkers
+	}
 	pool := &ReplicationPool{
 		workers:         make([]chan ReplicationWorkerOperation, 0, workers),
 		lrgworkers:      make([]chan ReplicationWorkerOperation, 0, LargeWorkerCount),
@@ -1768,6 +1802,7 @@ func NewReplicationPool(ctx context.Context, o ObjectLayer, opts replicationPool
 		ctx:             ctx,
 		objLayer:        o,
 		priority:        priority,
+		maxWorkers:      maxWorkers,
 	}

 	pool.AddLargeWorkers()
@@ -1923,7 +1958,7 @@ func (p *ReplicationPool) ResizeWorkers(n, checkOld int) {
 }

 // ResizeWorkerPriority sets replication failed workers pool size
-func (p *ReplicationPool) ResizeWorkerPriority(pri string) {
+func (p *ReplicationPool) ResizeWorkerPriority(pri string, maxWorkers int) {
 	var workers, mrfWorkers int
 	p.mu.Lock()
 	switch pri {
@@ -1937,13 +1972,21 @@ func (p *ReplicationPool) ResizeWorkerPriority(pri string) {
 		workers = WorkerAutoDefault
 		mrfWorkers = MRFWorkerAutoDefault
 		if len(p.workers) < WorkerAutoDefault {
-			workers = int(math.Min(float64(len(p.workers)+1), WorkerAutoDefault))
+			workers = min(len(p.workers)+1, WorkerAutoDefault)
 		}
 		if p.mrfWorkerSize < MRFWorkerAutoDefault {
-			mrfWorkers = int(math.Min(float64(p.mrfWorkerSize+1), MRFWorkerAutoDefault))
+			mrfWorkers = min(p.mrfWorkerSize+1, MRFWorkerAutoDefault)
 		}
 	}
+	if maxWorkers > 0 && workers > maxWorkers {
+		workers = maxWorkers
+	}
+
+	if maxWorkers > 0 && mrfWorkers > maxWorkers {
+		mrfWorkers = maxWorkers
+	}
 	p.priority = pri
+	p.maxWorkers = maxWorkers
 	p.mu.Unlock()
 	p.ResizeWorkers(workers, 0)
 	p.ResizeFailedWorkers(mrfWorkers)
@@ -2017,6 +2060,7 @@ func (p *ReplicationPool) queueReplicaTask(ri ReplicateObjectInfo) {
 		globalReplicationPool.queueMRFSave(ri.ToMRFEntry())
 		p.mu.RLock()
 		prio := p.priority
+		maxWorkers := p.maxWorkers
 		p.mu.RUnlock()
 		switch prio {
 		case "fast":
@@ -2024,16 +2068,18 @@ func (p *ReplicationPool) queueReplicaTask(ri ReplicateObjectInfo) {
 		case "slow":
 			logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to keep up with incoming traffic - we recommend increasing replication priority with `mc admin config set api replication_priority=auto`"), string(replicationSubsystem))
 		default:
-			if p.ActiveWorkers() < WorkerMaxLimit {
+			maxWorkers = min(maxWorkers, WorkerMaxLimit)
+			if p.ActiveWorkers() < maxWorkers {
 				p.mu.RLock()
-				workers := int(math.Min(float64(len(p.workers)+1), WorkerMaxLimit))
+				workers := min(len(p.workers)+1, maxWorkers)
 				existing := len(p.workers)
 				p.mu.RUnlock()
 				p.ResizeWorkers(workers, existing)
 			}
-			if p.ActiveMRFWorkers() < MRFWorkerMaxLimit {
+			maxMRFWorkers := min(maxWorkers, MRFWorkerMaxLimit)
+			if p.ActiveMRFWorkers() < maxMRFWorkers {
 				p.mu.RLock()
-				workers := int(math.Min(float64(p.mrfWorkerSize+1), MRFWorkerMaxLimit))
+				workers := min(p.mrfWorkerSize+1, maxMRFWorkers)
 				p.mu.RUnlock()
 				p.ResizeFailedWorkers(workers)
 			}
@@ -2071,6 +2117,7 @@ func (p *ReplicationPool) queueReplicaDeleteTask(doi DeletedObjectReplicationInf
 		globalReplicationPool.queueMRFSave(doi.ToMRFEntry())
 		p.mu.RLock()
 		prio := p.priority
+		maxWorkers := p.maxWorkers
 		p.mu.RUnlock()
 		switch prio {
 		case "fast":
@@ -2078,9 +2125,10 @@ func (p *ReplicationPool) queueReplicaDeleteTask(doi DeletedObjectReplicationInf
 		case "slow":
 			logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to keep up with incoming deletes - we recommend increasing replication priority with `mc admin config set api replication_priority=auto`"), string(replicationSubsystem))
 		default:
-			if p.ActiveWorkers() < WorkerMaxLimit {
+			maxWorkers = min(maxWorkers, WorkerMaxLimit)
+			if p.ActiveWorkers() < maxWorkers {
 				p.mu.RLock()
-				workers := int(math.Min(float64(len(p.workers)+1), WorkerMaxLimit))
+				workers := min(len(p.workers)+1, maxWorkers)
 				existing := len(p.workers)
 				p.mu.RUnlock()
 				p.ResizeWorkers(workers, existing)
@@ -2090,13 +2138,12 @@ func (p *ReplicationPool) queueReplicaDeleteTask(doi DeletedObjectReplicationInf
 }

 type replicationPoolOpts struct {
-	Priority string
+	Priority   string
+	MaxWorkers int
 }

 func initBackgroundReplication(ctx context.Context, objectAPI ObjectLayer) {
-	globalReplicationPool = NewReplicationPool(ctx, objectAPI, replicationPoolOpts{
-		Priority: globalAPIConfig.getReplicationPriority(),
-	})
+	globalReplicationPool = NewReplicationPool(ctx, objectAPI, globalAPIConfig.getReplicationOpts())
 	globalReplicationStats = NewReplicationStats(ctx, objectAPI)
 	go globalReplicationStats.trackEWMA()
 }
@@ -2572,7 +2619,7 @@ func (s *replicationResyncer) resyncBucket(ctx context.Context, objectAPI Object

 	// Walk through all object versions - Walk() is always in ascending order needed to ensure
 	// delete marker replicated to target after object version is first created.
-	if err := objectAPI.Walk(ctx, opts.bucket, "", objInfoCh, ObjectOptions{}); err != nil {
+	if err := objectAPI.Walk(ctx, opts.bucket, "", objInfoCh, WalkOptions{}); err != nil {
 		logger.LogIf(ctx, err)
 		return
 	}
@@ -2945,7 +2992,7 @@ func getReplicationDiff(ctx context.Context, objAPI ObjectLayer, bucket string,
 	}

 	objInfoCh := make(chan ObjectInfo, 10)
-	if err := objAPI.Walk(ctx, bucket, opts.Prefix, objInfoCh, ObjectOptions{}); err != nil {
+	if err := objAPI.Walk(ctx, bucket, opts.Prefix, objInfoCh, WalkOptions{}); err != nil {
 		logger.LogIf(ctx, err)
 		return nil, err
 	}
@@ -3156,10 +3203,11 @@ func (p *ReplicationPool) persistMRF() {
 			if !ok {
 				return
 			}
+			entries[e.versionID] = e
+
 			if len(entries) >= mrfMaxEntries {
 				saveMRFToDisk()
 			}
-			entries[e.versionID] = e
 		}
 	}
 }
@@ -3183,16 +3231,22 @@ func (p *ReplicationPool) queueMRFSave(entry MRFReplicateEntry) {
 		select {
 		case p.mrfSaveCh <- entry:
 		default:
+			atomic.AddUint64(&globalReplicationStats.mrfStats.TotalDroppedCount, 1)
+			atomic.AddUint64(&globalReplicationStats.mrfStats.TotalDroppedBytes, uint64(entry.sz))
 		}
 	}
 }

-func (p *ReplicationPool) persistToDrive(ctx context.Context, v MRFReplicateEntries, data []byte) {
+func (p *ReplicationPool) persistToDrive(ctx context.Context, v MRFReplicateEntries) {
 	newReader := func() io.ReadCloser {
 		r, w := io.Pipe()
 		go func() {
+			// Initialize MRF meta header.
+			var data [4]byte
+			binary.LittleEndian.PutUint16(data[0:2], mrfMetaFormat)
+			binary.LittleEndian.PutUint16(data[2:4], mrfMetaVersion)
 			mw := msgp.NewWriter(w)
-			n, err := mw.Write(data)
+			n, err := mw.Write(data[:])
 			if err != nil {
 				w.CloseWithError(err)
 				return
@@ -3208,7 +3262,11 @@ func (p *ReplicationPool) persistToDrive(ctx context.Context, v MRFReplicateEntr
 		return r
 	}

-	for _, localDrive := range globalLocalDrives {
+	globalLocalDrivesMu.RLock()
+	localDrives := globalLocalDrives
+	globalLocalDrivesMu.RUnlock()
+
+	for _, localDrive := range localDrives {
 		r := newReader()
 		err := localDrive.CreateFile(ctx, minioMetaBucket, pathJoin(replicationMRFDir, globalLocalNodeNameHex+".bin"), -1, r)
 		r.Close()
@@ -3230,15 +3288,10 @@ func (p *ReplicationPool) saveMRFEntries(ctx context.Context, entries map[string

 	v := MRFReplicateEntries{
 		Entries: entries,
-		Version: mrfMetaVersionV1,
+		Version: mrfMetaVersion,
 	}
-	data := make([]byte, 4, v.Msgsize()+4)

-	// Initialize the resync meta header.
-	binary.LittleEndian.PutUint16(data[0:2], mrfMetaFormat)
-	binary.LittleEndian.PutUint16(data[2:4], mrfMetaVersion)
-
-	p.persistToDrive(ctx, v, data)
+	p.persistToDrive(ctx, v)
 }

 // load mrf entries from disk
@@ -3249,8 +3302,8 @@ func (p *ReplicationPool) loadMRF() (mrfRec MRFReplicateEntries, err error) {
 		if !p.initialized() {
 			return re, nil
 		}
-		data := make([]byte, 4)
-		n, err := rc.Read(data)
+		var data [4]byte
+		n, err := rc.Read(data[:])
 		if err != nil {
 			return re, err
 		}
@@ -3275,7 +3328,12 @@ func (p *ReplicationPool) loadMRF() (mrfRec MRFReplicateEntries, err error) {

 		return re, nil
 	}
-	for _, localDrive := range globalLocalDrives {
+
+	globalLocalDrivesMu.RLock()
+	localDrives := globalLocalDrives
+	globalLocalDrivesMu.RUnlock()
+
+	for _, localDrive := range localDrives {
 		rc, err := localDrive.ReadFileStream(p.ctx, minioMetaBucket, pathJoin(replicationMRFDir, globalLocalNodeNameHex+".bin"), 0, -1)
 		if err != nil {
 			continue
@@ -3364,7 +3422,7 @@ func (p *ReplicationPool) initialized() bool {
 }

 // getMRF returns MRF entries for this node.
-func (p *ReplicationPool) getMRF(ctx context.Context, bucket string) (ch chan madmin.ReplicationMRF, err error) {
+func (p *ReplicationPool) getMRF(ctx context.Context, bucket string) (ch <-chan madmin.ReplicationMRF, err error) {
 	mrfRec, err := p.loadMRF()
 	if err != nil {
 		return nil, err
@@ -3374,7 +3432,7 @@ func (p *ReplicationPool) getMRF(ctx context.Context, bucket string) (ch chan ma
 	go func() {
 		defer close(mrfCh)
 		for vID, e := range mrfRec.Entries {
-			if e.Bucket != bucket && bucket != "" {
+			if bucket != "" && e.Bucket != bucket {
 				continue
 			}
 			select {
@@ -157,15 +157,11 @@ func performCallhome(ctx context.Context) {
 }

 const (
-	healthURL    = "https://subnet.min.io/api/health/upload"
-	healthURLDev = "http://localhost:9000/api/health/upload"
+	subnetHealthPath = "/api/health/upload"
 )

 func sendHealthInfo(ctx context.Context, healthInfo madmin.HealthInfo) error {
-	url := healthURL
-	if globalIsCICD {
-		url = healthURLDev
-	}
+	url := globalSubnetConfig.BaseURL + subnetHealthPath

 	filename := fmt.Sprintf("health_%s.json.gz", UTCNow().Format("20060102150405"))
 	url += "?filename=" + filename
@@ -46,10 +46,10 @@ import (
 	"github.com/inconshreveable/mousetrap"
 	dns2 "github.com/miekg/dns"
 	"github.com/minio/cli"
+	consoleapi "github.com/minio/console/api"
+	"github.com/minio/console/api/operations"
 	consoleoauth2 "github.com/minio/console/pkg/auth/idp/oauth2"
 	consoleCerts "github.com/minio/console/pkg/certs"
-	"github.com/minio/console/restapi"
-	"github.com/minio/console/restapi/operations"
 	"github.com/minio/kes-go"
 	"github.com/minio/madmin-go/v3"
 	"github.com/minio/minio-go/v7"
@@ -178,6 +178,23 @@ func minioConfigToConsoleFeatures() {
 	os.Setenv("CONSOLE_MINIO_REGION", globalSite.Region)
 	os.Setenv("CONSOLE_CERT_PASSWD", env.Get("MINIO_CERT_PASSWD", ""))

+	// This section sets Browser (console) stored config
+	if valueSCP := globalBrowserConfig.GetCSPolicy(); valueSCP != "" {
+		os.Setenv("CONSOLE_SECURE_CONTENT_SECURITY_POLICY", valueSCP)
+	}
+
+	if hstsSeconds := globalBrowserConfig.GetHSTSSeconds(); hstsSeconds > 0 {
+		isubdom := globalBrowserConfig.IsHSTSIncludeSubdomains()
+		isprel := globalBrowserConfig.IsHSTSPreload()
+		os.Setenv("CONSOLE_SECURE_STS_SECONDS", strconv.Itoa(hstsSeconds))
+		os.Setenv("CONSOLE_SECURE_STS_INCLUDE_SUB_DOMAINS", isubdom)
+		os.Setenv("CONSOLE_SECURE_STS_PRELOAD", isprel)
+	}
+
+	if valueRefer := globalBrowserConfig.GetReferPolicy(); valueRefer != "" {
+		os.Setenv("CONSOLE_SECURE_REFERRER_POLICY", valueRefer)
+	}
+
 	globalSubnetConfig.ApplyEnv()
 }

@@ -207,7 +224,7 @@ func buildOpenIDConsoleConfig() consoleoauth2.OpenIDPCfg {
 	return m
 }

-func initConsoleServer() (*restapi.Server, error) {
+func initConsoleServer() (*consoleapi.Server, error) {
 	// unset all console_ environment variables.
 	for _, cenv := range env.List(consolePrefix) {
 		os.Unsetenv(cenv)
@@ -225,9 +242,9 @@ func initConsoleServer() (*restapi.Server, error) {
 	}

 	// set certs before other console initialization
-	restapi.GlobalRootCAs, restapi.GlobalPublicCerts, restapi.GlobalTLSCertsManager = globalRootCAs, globalPublicCerts, globalTLSCerts
+	consoleapi.GlobalRootCAs, consoleapi.GlobalPublicCerts, consoleapi.GlobalTLSCertsManager = globalRootCAs, globalPublicCerts, globalTLSCerts

-	swaggerSpec, err := loads.Embedded(restapi.SwaggerJSON, restapi.FlatSwaggerJSON)
+	swaggerSpec, err := loads.Embedded(consoleapi.SwaggerJSON, consoleapi.FlatSwaggerJSON)
 	if err != nil {
 		return nil, err
 	}
@@ -238,18 +255,18 @@ func initConsoleServer() (*restapi.Server, error) {
 		// Disable console logging if server debug log is not enabled
 		noLog := func(string, ...interface{}) {}

-		restapi.LogInfo = noLog
-		restapi.LogError = noLog
+		consoleapi.LogInfo = noLog
+		consoleapi.LogError = noLog
 		api.Logger = noLog
 	}

 	// Pass in console application config. This needs to happen before the
 	// ConfigureAPI() call.
-	restapi.GlobalMinIOConfig = restapi.MinIOConfig{
+	consoleapi.GlobalMinIOConfig = consoleapi.MinIOConfig{
 		OpenIDProviders: buildOpenIDConsoleConfig(),
 	}

-	server := restapi.NewServer(api)
+	server := consoleapi.NewServer(api)
 	// register all APIs
 	server.ConfigureAPI()

@@ -257,16 +274,16 @@ func initConsoleServer() (*restapi.Server, error) {

 	server.Host = globalMinioConsoleHost
 	server.Port = consolePort
-	restapi.Port = globalMinioConsolePort
-	restapi.Hostname = globalMinioConsoleHost
+	consoleapi.Port = globalMinioConsolePort
+	consoleapi.Hostname = globalMinioConsoleHost

 	if globalIsTLS {
 		// If TLS certificates are provided enforce the HTTPS.
 		server.EnabledListeners = []string{"https"}
 		server.TLSPort = consolePort
 		// Need to store tls-port, tls-host un config variables so secure.middleware can read from there
-		restapi.TLSPort = globalMinioConsolePort
-		restapi.Hostname = globalMinioConsoleHost
+		consoleapi.TLSPort = globalMinioConsolePort
+		consoleapi.Hostname = globalMinioConsoleHost
 	}

 	return server, nil
@@ -310,84 +327,121 @@ func checkUpdate(mode string) {
 	logger.Info(prepareUpdateMessage("Run `mc admin update`", lrTime.Sub(crTime)))
 }

-func newConfigDirFromCtx(ctx *cli.Context, option string, getDefaultDir func() string) (*ConfigDir, bool) {
-	var dir string
-	var dirSet bool
-
-	switch {
-	case ctx.IsSet(option):
-		dir = ctx.String(option)
-		dirSet = true
-	case ctx.GlobalIsSet(option):
-		dir = ctx.GlobalString(option)
-		dirSet = true
-		// cli package does not expose parent's option option.  Below code is workaround.
-		if dir == "" || dir == getDefaultDir() {
-			dirSet = false // Unset to false since GlobalIsSet() true is a false positive.
-			if ctx.Parent().GlobalIsSet(option) {
-				dir = ctx.Parent().GlobalString(option)
-				dirSet = true
-			}
-		}
-	default:
-		// Neither local nor global option is provided.  In this case, try to use
-		// default directory.
+func newConfigDir(dir string, dirSet bool, getDefaultDir func() string) (*ConfigDir, error) {
+	if dir == "" {
 		dir = getDefaultDir()
-		if dir == "" {
-			logger.FatalIf(errInvalidArgument, "%s option must be provided", option)
-		}
 	}

 	if dir == "" {
-		logger.FatalIf(errors.New("empty directory"), "%s directory cannot be empty", option)
+		if !dirSet {
+			return nil, fmt.Errorf("missing option must be provided")
+		}
+		return nil, fmt.Errorf("provided option cannot be empty")
 	}

 	// Disallow relative paths, figure out absolute paths.
 	dirAbs, err := filepath.Abs(dir)
-	logger.FatalIf(err, "Unable to fetch absolute path for %s=%s", option, dir)
+	if err != nil {
+		return nil, err
+	}
+	err = mkdirAllIgnorePerm(dirAbs)
+	if err != nil {
+		return nil, fmt.Errorf("unable to create the directory `%s`: %w", dirAbs, err)
+	}

-	logger.FatalIf(mkdirAllIgnorePerm(dirAbs), "Unable to create directory specified %s=%s", option, dir)
-
-	return &ConfigDir{path: dirAbs}, dirSet
+	return &ConfigDir{path: dirAbs}, nil
 }

-func handleCommonCmdArgs(ctx *cli.Context) {
+func buildServerCtxt(ctx *cli.Context, ctxt *serverCtxt) (err error) {
 	// Get "json" flag from command line argument and
-	// enable json and quite modes if json flag is turned on.
-	globalCLIContext.JSON = ctx.IsSet("json") || ctx.GlobalIsSet("json")
-	if globalCLIContext.JSON {
-		logger.EnableJSON()
-	}
-
+	ctxt.JSON = ctx.IsSet("json") || ctx.GlobalIsSet("json")
 	// Get quiet flag from command line argument.
-	globalCLIContext.Quiet = ctx.IsSet("quiet") || ctx.GlobalIsSet("quiet")
-	if globalCLIContext.Quiet {
-		logger.EnableQuiet()
-	}
-
+	ctxt.Quiet = ctx.IsSet("quiet") || ctx.GlobalIsSet("quiet")
 	// Get anonymous flag from command line argument.
-	globalCLIContext.Anonymous = ctx.IsSet("anonymous") || ctx.GlobalIsSet("anonymous")
-	if globalCLIContext.Anonymous {
-		logger.EnableAnonymous()
-	}
-
+	ctxt.Anonymous = ctx.IsSet("anonymous") || ctx.GlobalIsSet("anonymous")
 	// Fetch address option
-	addr := ctx.GlobalString("address")
-	if addr == "" || addr == ":"+GlobalMinioDefaultPort {
-		addr = ctx.String("address")
+	ctxt.Addr = ctx.GlobalString("address")
+	if ctxt.Addr == "" || ctxt.Addr == ":"+GlobalMinioDefaultPort {
+		ctxt.Addr = ctx.String("address")
 	}

 	// Fetch console address option
-	consoleAddr := ctx.GlobalString("console-address")
-	if consoleAddr == "" {
-		consoleAddr = ctx.String("console-address")
+	ctxt.ConsoleAddr = ctx.GlobalString("console-address")
+	if ctxt.ConsoleAddr == "" {
+		ctxt.ConsoleAddr = ctx.String("console-address")
 	}

+	// Check "no-compat" flag from command line argument.
+	ctxt.StrictS3Compat = !(ctx.IsSet("no-compat") || ctx.GlobalIsSet("no-compat"))
+	ctxt.PreAllocate = ctx.IsSet("pre-allocate") || ctx.GlobalIsSet("pre-allocate")
+
+	switch {
+	case ctx.IsSet("config-dir"):
+		ctxt.ConfigDir = ctx.String("config-dir")
+		ctxt.configDirSet = true
+	case ctx.GlobalIsSet("config-dir"):
+		ctxt.ConfigDir = ctx.GlobalString("config-dir")
+		ctxt.configDirSet = true
+	}
+
+	switch {
+	case ctx.IsSet("certs-dir"):
+		ctxt.CertsDir = ctx.String("certs-dir")
+		ctxt.certsDirSet = true
+	case ctx.GlobalIsSet("certs-dir"):
+		ctxt.CertsDir = ctx.GlobalString("certs-dir")
+		ctxt.certsDirSet = true
+	}
+
+	ctxt.FTP = ctx.StringSlice("ftp")
+	ctxt.SFTP = ctx.StringSlice("sftp")
+
+	ctxt.Interface = ctx.String("interface")
+	ctxt.UserTimeout = ctx.Duration("conn-user-timeout")
+	ctxt.ConnReadDeadline = ctx.Duration("conn-read-deadline")
+	ctxt.ConnWriteDeadline = ctx.Duration("conn-write-deadline")
+
+	ctxt.ShutdownTimeout = ctx.Duration("shutdown-timeout")
+	ctxt.IdleTimeout = ctx.Duration("idle-timeout")
+	ctxt.ReadHeaderTimeout = ctx.Duration("read-header-timeout")
+
+	if conf := ctx.String("config"); len(conf) > 0 {
+		err = mergeServerCtxtFromConfigFile(conf, ctxt)
+	} else {
+		err = mergeDisksLayoutFromArgs(serverCmdArgs(ctx), ctxt)
+	}
+
+	return err
+}
+
+func handleCommonArgs(ctxt serverCtxt) {
+	if ctxt.JSON {
+		logger.EnableJSON()
+	}
+	if ctxt.Quiet {
+		logger.EnableQuiet()
+	}
+	if ctxt.Anonymous {
+		logger.EnableAnonymous()
+	}
+
+	consoleAddr := ctxt.ConsoleAddr
+	addr := ctxt.Addr
+	configDir := ctxt.ConfigDir
+	configSet := ctxt.configDirSet
+	certsDir := ctxt.CertsDir
+	certsSet := ctxt.certsDirSet
+
 	if consoleAddr == "" {
 		p, err := xnet.GetFreePort()
 		if err != nil {
 			logger.FatalIf(err, "Unable to get free port for Console UI on the host")
 		}
+		// hold the port
+		l, err := net.Listen("TCP", fmt.Sprintf(":%s", p.String()))
+		if err == nil {
+			defer l.Close()
+		}
 		consoleAddr = net.JoinHostPort("", p.String())
 	}

@@ -417,16 +471,12 @@ func handleCommonCmdArgs(ctx *cli.Context) {

 	globalMinioAddr = addr

-	// Check "no-compat" flag from command line argument.
-	globalCLIContext.StrictS3Compat = true
-	if ctx.IsSet("no-compat") || ctx.GlobalIsSet("no-compat") {
-		globalCLIContext.StrictS3Compat = false
-	}
-
 	// Set all config, certs and CAs directories.
-	var configSet, certsSet bool
-	globalConfigDir, configSet = newConfigDirFromCtx(ctx, "config-dir", defaultConfigDir.Get)
-	globalCertsDir, certsSet = newConfigDirFromCtx(ctx, "certs-dir", defaultCertsDir.Get)
+	var err error
+	globalConfigDir, err = newConfigDir(configDir, configSet, defaultConfigDir.Get)
+	logger.FatalIf(err, "Unable to initialize the (deprecated) config directory")
+	globalCertsDir, err = newConfigDir(certsDir, certsSet, defaultCertsDir.Get)
+	logger.FatalIf(err, "Unable to initialize the certs directory")

 	// Remove this code when we deprecate and remove config-dir.
 	// This code is to make sure we inherit from the config-dir
@@ -438,9 +488,11 @@ func handleCommonCmdArgs(ctx *cli.Context) {
 	globalCertsCADir = &ConfigDir{path: filepath.Join(globalCertsDir.Get(), certsCADir)}

 	logger.FatalIf(mkdirAllIgnorePerm(globalCertsCADir.Get()), "Unable to create certs CA directory at %s", globalCertsCADir.Get())
+}

-	// Check if we have configured a custom DNS cache TTL.
+func runDNSCache(ctx *cli.Context) {
 	dnsTTL := ctx.Duration("dns-cache-ttl")
+	// Check if we have configured a custom DNS cache TTL.
 	if dnsTTL <= 0 {
 		dnsTTL = 10 * time.Minute
 	}
@@ -621,7 +673,7 @@ func loadEnvVarsFromFiles() {
 	}
 }

-func handleCommonEnvVars() {
+func serverHandleEnvVars() {
 	var err error
 	globalBrowserEnabled, err = config.ParseBool(env.Get(config.EnvBrowser, config.EnableOn))
 	if err != nil {
@@ -749,6 +801,10 @@ func handleCommonEnvVars() {
 		}
 	}

+	globalDisableFreezeOnBoot = env.Get("_MINIO_DISABLE_API_FREEZE_ON_BOOT", "") == "true" || serverDebugLog
+}
+
+func loadRootCredentials() {
 	// At this point, either both environment variables
 	// are defined or both are not defined.
 	// Check both cases and authenticate them if correctly defined
@@ -763,6 +819,9 @@ func handleCommonEnvVars() {
 		user = env.Get(config.EnvAccessKey, "")
 		password = env.Get(config.EnvSecretKey, "")
 		hasCredentials = true
+	} else if globalServerCtxt.RootUser != "" && globalServerCtxt.RootPwd != "" {
+		user, password = globalServerCtxt.RootUser, globalServerCtxt.RootPwd
+		hasCredentials = true
 	}
 	if hasCredentials {
 		cred, err := auth.CreateCredentials(user, password)
@@ -782,8 +841,6 @@ func handleCommonEnvVars() {
 	} else {
 		globalActiveCred = auth.DefaultCredentials
 	}
-
-	globalDisableFreezeOnBoot = env.Get("_MINIO_DISABLE_API_FREEZE_ON_BOOT", "") == "true" || serverDebugLog
 }

 // Initialize KMS global variable after valiadating and loading the configuration.
@@ -24,12 +24,17 @@ import (
 	"strings"
 	"sync"

+	"github.com/minio/minio/internal/config/browser"
+
 	"github.com/minio/madmin-go/v3"
 	"github.com/minio/minio/internal/config"
 	"github.com/minio/minio/internal/config/api"
+	"github.com/minio/minio/internal/config/batch"
+	"github.com/minio/minio/internal/config/cache"
 	"github.com/minio/minio/internal/config/callhome"
 	"github.com/minio/minio/internal/config/compress"
 	"github.com/minio/minio/internal/config/dns"
+	"github.com/minio/minio/internal/config/drive"
 	"github.com/minio/minio/internal/config/etcd"
 	"github.com/minio/minio/internal/config/heal"
 	xldap "github.com/minio/minio/internal/config/identity/ldap"
@@ -68,6 +73,10 @@ func initHelp() {
 		config.ScannerSubSys:        scanner.DefaultKVS,
 		config.SubnetSubSys:         subnet.DefaultKVS,
 		config.CallhomeSubSys:       callhome.DefaultKVS,
+		config.DriveSubSys:          drive.DefaultKVS,
+		config.CacheSubSys:          cache.DefaultKVS,
+		config.BatchSubSys:          batch.DefaultKVS,
+		config.BrowserSubSys:        browser.DefaultKVS,
 	}
 	for k, v := range notify.DefaultNotificationKVS {
 		kvs[k] = v
@@ -95,6 +104,10 @@ func initHelp() {
 			Description: "enable callhome to MinIO SUBNET",
 			Optional:    true,
 		},
+		config.HelpKV{
+			Key:         config.DriveSubSys,
+			Description: "enable drive specific settings",
+		},
 		config.HelpKV{
 			Key:         config.SiteSubSys,
 			Description: "label the server and its location",
@@ -107,6 +120,10 @@ func initHelp() {
 			Key:         config.ScannerSubSys,
 			Description: "manage namespace scanning for usage calculation, lifecycle, healing and more",
 		},
+		config.HelpKV{
+			Key:         config.BatchSubSys,
+			Description: "manage batch job workers and wait times",
+		},
 		config.HelpKV{
 			Key:         config.CompressionSubSys,
 			Description: "enable server side compression of objects",
@@ -206,6 +223,17 @@ func initHelp() {
 			Key:         config.EtcdSubSys,
 			Description: "persist IAM assets externally to etcd",
 		},
+		config.HelpKV{
+			Key:         config.CacheSubSys,
+			Type:        "string",
+			Description: "enable cache plugin on MinIO for GET/HEAD requests",
+			Optional:    true,
+		},
+		config.HelpKV{
+			Key:         config.BrowserSubSys,
+			Description: "manage Browser HTTP specific features, such as Security headers, etc.",
+			Optional:    true,
+		},
 	}

 	if globalIsErasure {
@@ -227,6 +255,7 @@ func initHelp() {
 		config.EtcdSubSys:           etcd.Help,
 		config.CompressionSubSys:    compress.Help,
 		config.HealSubSys:           heal.Help,
+		config.BatchSubSys:          batch.Help,
 		config.ScannerSubSys:        scanner.Help,
 		config.IdentityOpenIDSubSys: openid.Help,
 		config.IdentityLDAPSubSys:   xldap.Help,
@@ -250,6 +279,9 @@ func initHelp() {
 		config.LambdaWebhookSubSys:  lambda.HelpWebhook,
 		config.SubnetSubSys:         subnet.HelpSubnet,
 		config.CallhomeSubSys:       callhome.HelpCallhome,
+		config.DriveSubSys:          drive.HelpDrive,
+		config.CacheSubSys:          cache.Help,
+		config.BrowserSubSys:        browser.Help,
 	}

 	config.RegisterHelpSubSys(helpMap)
@@ -285,6 +317,10 @@ func validateSubSysConfig(ctx context.Context, s config.Config, subSys string, o
 		if _, err := api.LookupConfig(s[config.APISubSys][config.Default]); err != nil {
 			return err
 		}
+	case config.BatchSubSys:
+		if _, err := batch.LookupConfig(s[config.BatchSubSys][config.Default]); err != nil {
+			return err
+		}
 	case config.StorageClassSubSys:
 		if objAPI == nil {
 			return errServerNotInitialized
@@ -357,6 +393,14 @@ func validateSubSysConfig(ctx context.Context, s config.Config, subSys string, o
 		if cfg.Enabled() && !globalSubnetConfig.Registered() {
 			return errors.New("Deployment is not registered with SUBNET. Please register the deployment via 'mc license register ALIAS'")
 		}
+	case config.DriveSubSys:
+		if _, err := drive.LookupConfig(s[config.DriveSubSys][config.Default]); err != nil {
+			return err
+		}
+	case config.CacheSubSys:
+		if _, err := cache.LookupConfig(s[config.CacheSubSys][config.Default], globalRemoteTargetTransport); err != nil {
+			return err
+		}
 	case config.PolicyOPASubSys:
 		// In case legacy OPA config is being set, we treat it as if the
 		// AuthZPlugin is being set.
@@ -372,6 +416,10 @@ func validateSubSysConfig(ctx context.Context, s config.Config, subSys string, o
 				return err
 			}
 		}
+	case config.BrowserSubSys:
+		if _, err := browser.LookupConfig(s[config.BrowserSubSys][config.Default]); err != nil {
+			return err
+		}
 	default:
 		if config.LoggerSubSystems.Contains(subSys) {
 			if err := logger.ValidateSubSysConfig(ctx, s, subSys); err != nil {
@@ -540,6 +588,12 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 			return fmt.Errorf("Unable to apply heal config: %w", err)
 		}
 		globalHealConfig.Update(healCfg)
+	case config.BatchSubSys:
+		batchCfg, err := batch.LookupConfig(s[config.BatchSubSys][config.Default])
+		if err != nil {
+			return fmt.Errorf("Unable to apply batch config: %w", err)
+		}
+		globalBatchConfig.Update(batchCfg)
 	case config.ScannerSubSys:
 		scannerCfg, err := scanner.LookupConfig(s[config.ScannerSubSys][config.Default])
 		if err != nil {
@@ -618,7 +672,7 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 		if err != nil {
 			logger.LogIf(ctx, fmt.Errorf("Unable to parse subnet configuration: %w", err))
 		} else {
-			globalSubnetConfig.Update(subnetConfig)
+			globalSubnetConfig.Update(subnetConfig, globalIsCICD)
 			globalSubnetConfig.ApplyEnv() // update environment settings for Console UI
 		}
 	case config.CallhomeSubSys:
@@ -632,6 +686,28 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 				initCallhome(ctx, objAPI)
 			}
 		}
+	case config.DriveSubSys:
+		if driveConfig, err := drive.LookupConfig(s[config.DriveSubSys][config.Default]); err != nil {
+			logger.LogIf(ctx, fmt.Errorf("Unable to load drive config: %w", err))
+		} else {
+			err := globalDriveConfig.Update(driveConfig)
+			if err != nil {
+				logger.LogIf(ctx, fmt.Errorf("Unable to update drive config: %v", err))
+			}
+		}
+	case config.CacheSubSys:
+		cacheCfg, err := cache.LookupConfig(s[config.CacheSubSys][config.Default], globalRemoteTargetTransport)
+		if err != nil {
+			logger.LogIf(ctx, fmt.Errorf("Unable to load cache config: %w", err))
+		} else {
+			globalCacheConfig.Update(cacheCfg)
+		}
+	case config.BrowserSubSys:
+		browserCfg, err := browser.LookupConfig(s[config.BrowserSubSys][config.Default])
+		if err != nil {
+			return fmt.Errorf("Unable to apply browser config: %w", err)
+		}
+		globalBrowserConfig.Update(browserCfg)
 	}
 	globalServerConfigMu.Lock()
 	defer globalServerConfigMu.Unlock()
@@ -121,13 +121,13 @@ func readBackgroundHealInfo(ctx context.Context, objAPI ObjectLayer) backgroundH
 	buf, err := readConfig(ctx, objAPI, backgroundHealInfoPath)
 	if err != nil {
 		if !errors.Is(err, errConfigNotFound) {
-			logger.LogIf(ctx, err)
+			logger.LogOnceIf(ctx, err, backgroundHealInfoPath)
 		}
 		return backgroundHealInfo{}
 	}
 	var info backgroundHealInfo
 	if err = json.Unmarshal(buf, &info); err != nil {
-		logger.LogIf(ctx, err)
+		logger.LogOnceIf(ctx, err, backgroundHealInfoPath)
 	}
 	return info
 }
@@ -203,7 +203,7 @@ func runDataScanner(ctx context.Context, objAPI ObjectLayer) {
 			results := make(chan DataUsageInfo, 1)
 			go storeDataUsageInBackend(ctx, objAPI, results)
 			err := objAPI.NSScanner(ctx, results, uint32(cycleInfo.current), scanMode)
-			logger.LogIf(ctx, err)
+			logger.LogOnceIf(ctx, err, "ns-scanner")
 			res := map[string]string{"cycle": strconv.FormatUint(cycleInfo.current, 10)}
 			if err != nil {
 				res["error"] = err.Error()
@@ -223,7 +223,7 @@ func runDataScanner(ctx context.Context, objAPI ObjectLayer) {
 				binary.LittleEndian.PutUint64(tmp, cycleInfo.next)
 				tmp, _ = cycleInfo.MarshalMsg(tmp)
 				err = saveConfig(ctx, objAPI, dataUsageBloomNamePath, tmp)
-				logger.LogIf(ctx, err)
+				logger.LogOnceIf(ctx, err, dataUsageBloomNamePath)
 			}
 		}
 	}
@@ -477,7 +477,7 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 			item.heal.enabled = item.heal.enabled && f.healObjectSelect > 0

 			sz, err := f.getSize(item)
-			if err != nil {
+			if err != nil && err != errIgnoreFileContrib {
 				wait() // wait to proceed to next entry.
 				if err != errSkipFile && f.dataUsageScannerDebug {
 					console.Debugf(scannerLogPrefix+" getSize \"%v/%v\" returned err: %v\n", bucket, item.objectPath(), err)
@@ -495,8 +495,10 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 			// object.
 			delete(abandonedChildren, pathJoin(item.bucket, item.objectPath()))

-			into.addSizes(sz)
-			into.Objects++
+			if err != errIgnoreFileContrib {
+				into.addSizes(sz)
+				into.Objects++
+			}

 			wait() // wait to proceed to next entry.

@@ -579,12 +581,12 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 						foundAny = true
 						break
 					}
-					if next := f.updateCache.searchParent(parent); next == nil {
+					next := f.updateCache.searchParent(parent)
+					if next == nil {
 						foundAny = true
 						break
-					} else {
-						parent = *next
 					}
+					parent = *next
 				}
 				if !foundAny {
 					// Add non-compacted empty entry.
@@ -727,7 +729,7 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 							versionID: "",
 						}, madmin.HealItemObject)
 						if !isErrObjectNotFound(err) && !isErrVersionNotFound(err) {
-							logger.LogIf(ctx, err)
+							logger.LogOnceIf(ctx, err, entry.name)
 						}
 						foundObjs = foundObjs || err == nil
 						return
@@ -744,7 +746,7 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 						}, madmin.HealItemObject)
 						stopFn(int(ver.Size))
 						if !isErrObjectNotFound(err) && !isErrVersionNotFound(err) {
-							logger.LogIf(ctx, err)
+							logger.LogOnceIf(ctx, err, fiv.Name)
 						}
 						if err == nil {
 							successVersions++
@@ -917,16 +919,17 @@ func (i *scannerItem) applyHealing(ctx context.Context, o ObjectLayer, oi Object
 	return 0
 }

-func (i *scannerItem) applyLifecycle(ctx context.Context, o ObjectLayer, oi ObjectInfo) (applied bool, size int64) {
+func (i *scannerItem) applyLifecycle(ctx context.Context, o ObjectLayer, oi ObjectInfo) (action lifecycle.Action, size int64) {
 	size, err := oi.GetActualSize()
 	if i.debug {
 		logger.LogIf(ctx, err)
 	}
 	if i.lifeCycle == nil {
-		return false, size
+		return action, size
 	}

 	versionID := oi.VersionID
+	vcfg, _ := globalBucketVersioningSys.Get(i.bucket)
 	rCfg, _ := globalBucketObjectLockSys.Get(i.bucket)
 	replcfg, _ := getReplicationConfig(ctx, i.bucket)
 	lcEvt := evalActionFromLifecycle(ctx, *i.lifeCycle, rCfg, replcfg, oi)
@@ -938,7 +941,7 @@ func (i *scannerItem) applyLifecycle(ctx context.Context, o ObjectLayer, oi Obje
 		}
 	}
 	defer func() {
-		if applied {
+		if lcEvt.Action != lifecycle.NoneAction {
 			numVersions := uint64(1)
 			if lcEvt.Action == lifecycle.DeleteAllVersionsAction {
 				numVersions = uint64(oi.NumVersions)
@@ -948,14 +951,21 @@ func (i *scannerItem) applyLifecycle(ctx context.Context, o ObjectLayer, oi Obje
 	}()

 	switch lcEvt.Action {
-	case lifecycle.DeleteAction, lifecycle.DeleteVersionAction, lifecycle.DeleteRestoredAction, lifecycle.DeleteRestoredVersionAction, lifecycle.DeleteAllVersionsAction:
-		return applyLifecycleAction(lcEvt, lcEventSrc_Scanner, oi), 0
-	case lifecycle.TransitionAction, lifecycle.TransitionVersionAction:
-		return applyLifecycleAction(lcEvt, lcEventSrc_Scanner, oi), size
-	default:
-		// No action.
-		return false, size
+	// This version doesn't contribute towards sizeS only when it is permanently deleted.
+	// This can happen when,
+	// - ExpireObjectAllVersions flag is enabled
+	// - NoncurrentVersionExpiration is applicable
+	case lifecycle.DeleteVersionAction, lifecycle.DeleteAllVersionsAction:
+		size = 0
+	case lifecycle.DeleteAction:
+		// On a non-versioned bucket, DeleteObject removes the only version permanently.
+		if !vcfg.PrefixEnabled(oi.Name) {
+			size = 0
+		}
 	}
+
+	applyLifecycleAction(lcEvt, lcEventSrc_Scanner, oi)
+	return lcEvt.Action, size
 }

 // applyTierObjSweep removes remote object pending deletion and the free-version
@@ -1097,15 +1107,22 @@ func (i *scannerItem) applyVersionActions(ctx context.Context, o ObjectLayer, fi
 // The resulting size on disk will always be returned.
 // The metadata will be compared to consensus on the object layer before any changes are applied.
 // If no metadata is supplied, -1 is returned if no action is taken.
-func (i *scannerItem) applyActions(ctx context.Context, o ObjectLayer, oi ObjectInfo, sizeS *sizeSummary) int64 {
+func (i *scannerItem) applyActions(ctx context.Context, o ObjectLayer, oi ObjectInfo, sizeS *sizeSummary) (objDeleted bool, size int64) {
 	done := globalScannerMetrics.time(scannerMetricILM)
-	applied, size := i.applyLifecycle(ctx, o, oi)
+	var action lifecycle.Action
+	action, size = i.applyLifecycle(ctx, o, oi)
 	done()

+	// Note: objDeleted is true if and only if action ==
+	// lifecycle.DeleteAllVersionsAction
+	if action == lifecycle.DeleteAllVersionsAction {
+		return true, 0
+	}
+
 	// For instance, an applied lifecycle means we remove/transitioned an object
 	// from the current deployment, which means we don't have to call healing
 	// routine even if we are asked to do via heal flag.
-	if !applied {
+	if action == lifecycle.NoneAction {
 		if i.heal.enabled {
 			done := globalScannerMetrics.time(scannerMetricHealCheck)
 			size = i.applyHealing(ctx, o, oi)
@@ -1116,7 +1133,7 @@ func (i *scannerItem) applyActions(ctx context.Context, o ObjectLayer, oi Object
 				err := o.CheckAbandonedParts(ctx, i.bucket, i.objectPath(), madmin.HealOpts{Remove: healDeleteDangling})
 				done()
 				if err != nil {
-					logger.LogIf(ctx, fmt.Errorf("unable to check object %s/%s for abandoned data: %w", i.bucket, i.objectPath(), err))
+					logger.LogOnceIf(ctx, fmt.Errorf("unable to check object %s/%s for abandoned data: %w", i.bucket, i.objectPath(), err), i.objectPath())
 				}
 			}
 		}
@@ -1126,7 +1143,7 @@ func (i *scannerItem) applyActions(ctx context.Context, o ObjectLayer, oi Object
 		i.healReplication(ctx, o, oi.Clone(), sizeS)
 		done()
 	}
-	return size
+	return false, size
 }

 func evalActionFromLifecycle(ctx context.Context, lc lifecycle.Lifecycle, lr lock.Retention, rcfg *replication.Config, obj ObjectInfo) lifecycle.Event {
@@ -1182,7 +1199,7 @@ func applyExpiryOnTransitionedObject(ctx context.Context, objLayer ObjectLayer,
 		if isErrObjectNotFound(err) || isErrVersionNotFound(err) {
 			return false
 		}
-		logger.LogIf(ctx, err)
+		logger.LogOnceIf(ctx, err, obj.Name)
 		return false
 	}
 	// Notification already sent in *expireTransitionedObject*, just return 'true' here.
@@ -56,7 +56,7 @@ func storeDataUsageInBackend(ctx context.Context, objAPI ObjectLayer, dui <-chan
 			attempts = 1
 		}
 		if err = saveConfig(ctx, objAPI, dataUsageObjNamePath, dataUsageJSON); err != nil {
-			logger.LogIf(ctx, err)
+			logger.LogOnceIf(ctx, err, dataUsageObjNamePath)
 		}
 		attempts++
 	}
@@ -18,12 +18,15 @@
 package cmd

 import (
+	"errors"
 	"fmt"
+	"net/url"
 	"runtime"
 	"sort"
 	"strconv"
 	"strings"

+	"github.com/cespare/xxhash/v2"
 	"github.com/minio/minio-go/v7/pkg/set"
 	"github.com/minio/minio/internal/config"
 	"github.com/minio/pkg/v2/ellipses"
@@ -335,13 +338,128 @@ const (

 var globalCustomErasureDriveCount = false

-// CreateServerEndpoints - validates and creates new endpoints from input args, supports
-// both ellipses and without ellipses transparently.
-func createServerEndpoints(serverAddr string, args ...string) (
-	endpointServerPools EndpointServerPools, setupType SetupType, err error,
-) {
+type node struct {
+	nodeName string
+	disks    []string
+}
+
+type endpointsList []node
+
+func (el *endpointsList) add(arg string) error {
+	u, err := url.Parse(arg)
+	if err != nil {
+		return err
+	}
+	found := false
+	list := *el
+	for i := range list {
+		if list[i].nodeName == u.Host {
+			list[i].disks = append(list[i].disks, u.String())
+			found = true
+			break
+		}
+	}
+	if !found {
+		list = append(list, node{nodeName: u.Host, disks: []string{u.String()}})
+	}
+	*el = list
+	return nil
+}
+
+// buildDisksLayoutFromConfFile supports with and without ellipses transparently.
+func buildDisksLayoutFromConfFile(pools [][]string) (layout disksLayout, err error) {
+	if len(pools) == 0 {
+		return layout, errInvalidArgument
+	}
+
+	for _, list := range pools {
+		var endpointsList endpointsList
+
+		for _, arg := range list {
+			switch {
+			case ellipses.HasList(arg):
+				patterns, err := ellipses.FindListPatterns(arg)
+				if err != nil {
+					return layout, err
+				}
+				for _, exp := range patterns.Expand() {
+					for _, ep := range exp {
+						if err := endpointsList.add(ep); err != nil {
+							return layout, err
+						}
+					}
+				}
+			case ellipses.HasEllipses(arg):
+				patterns, err := ellipses.FindEllipsesPatterns(arg)
+				if err != nil {
+					return layout, err
+				}
+				for _, exp := range patterns.Expand() {
+					if err := endpointsList.add(strings.Join(exp, "")); err != nil {
+						return layout, err
+					}
+				}
+			default:
+				if err := endpointsList.add(arg); err != nil {
+					return layout, err
+				}
+			}
+		}
+
+		var stopping bool
+		var singleNode bool
+		var eps []string
+
+		for i := 0; ; i++ {
+			for _, node := range endpointsList {
+				if node.nodeName == "" {
+					singleNode = true
+				}
+
+				if len(node.disks) <= i {
+					stopping = true
+					continue
+				}
+				if stopping {
+					return layout, errors.New("number of disks per node does not match")
+				}
+				eps = append(eps, node.disks[i])
+			}
+			if stopping {
+				break
+			}
+		}
+
+		for _, node := range endpointsList {
+			if node.nodeName != "" && singleNode {
+				return layout, errors.New("all arguments must but either single node or distributed")
+			}
+		}
+
+		setArgs, err := GetAllSets(eps...)
+		if err != nil {
+			return layout, err
+		}
+
+		h := xxhash.New()
+		for _, s := range setArgs {
+			for _, d := range s {
+				h.WriteString(d)
+			}
+		}
+
+		layout.pools = append(layout.pools, poolDisksLayout{
+			cmdline: fmt.Sprintf("hash:%x", h.Sum(nil)),
+			layout:  setArgs,
+		})
+	}
+	return
+}
+
+// mergeDisksLayoutFromArgs supports with and without ellipses transparently.
+func mergeDisksLayoutFromArgs(args []string, ctxt *serverCtxt) (err error) {
 	if len(args) == 0 {
-		return nil, -1, errInvalidArgument
+		return errInvalidArgument
 	}

 	ok := true
@@ -349,44 +467,42 @@ func createServerEndpoints(serverAddr string, args ...string) (
 		ok = ok && !ellipses.HasEllipses(arg)
 	}

+	var setArgs [][]string
+
 	// None of the args have ellipses use the old style.
 	if ok {
-		setArgs, err := GetAllSets(args...)
+		setArgs, err = GetAllSets(args...)
 		if err != nil {
-			return nil, -1, err
+			return err
 		}
-		endpointList, newSetupType, err := CreateEndpoints(serverAddr, setArgs...)
-		if err != nil {
-			return nil, -1, err
+		ctxt.Layout = disksLayout{
+			legacy: true,
+			pools:  []poolDisksLayout{{layout: setArgs}},
 		}
-		for i := range endpointList {
-			endpointList[i].SetPoolIndex(0)
-		}
-		endpointServerPools = append(endpointServerPools, PoolEndpoints{
-			Legacy:       true,
-			SetCount:     len(setArgs),
-			DrivesPerSet: len(setArgs[0]),
-			Endpoints:    endpointList,
-			CmdLine:      strings.Join(args, " "),
-			Platform:     fmt.Sprintf("OS: %s | Arch: %s", runtime.GOOS, runtime.GOARCH),
-		})
-		setupType = newSetupType
-		return endpointServerPools, setupType, nil
+		return
 	}

-	var poolArgs [][][]string
 	for _, arg := range args {
 		if !ellipses.HasEllipses(arg) && len(args) > 1 {
 			// TODO: support SNSD deployments to be decommissioned in future
-			return nil, -1, fmt.Errorf("all args must have ellipses for pool expansion (%w) args: %s", errInvalidArgument, args)
+			return fmt.Errorf("all args must have ellipses for pool expansion (%w) args: %s", errInvalidArgument, args)
 		}
-
-		setArgs, err := GetAllSets(arg)
+		setArgs, err = GetAllSets(arg)
 		if err != nil {
-			return nil, -1, err
+			return err
 		}
+		ctxt.Layout.pools = append(ctxt.Layout.pools, poolDisksLayout{cmdline: arg, layout: setArgs})
+	}
+	return
+}

-		poolArgs = append(poolArgs, setArgs)
+// CreateServerEndpoints - validates and creates new endpoints from input args, supports
+// both ellipses and without ellipses transparently.
+func createServerEndpoints(serverAddr string, poolArgs []poolDisksLayout, legacy bool) (
+	endpointServerPools EndpointServerPools, setupType SetupType, err error,
+) {
+	if len(poolArgs) == 0 {
+		return nil, -1, errInvalidArgument
 	}

 	poolEndpoints, setupType, err := CreatePoolEndpoints(serverAddr, poolArgs...)
@@ -396,11 +512,12 @@ func createServerEndpoints(serverAddr string, args ...string) (

 	for i, endpointList := range poolEndpoints {
 		if err = endpointServerPools.Add(PoolEndpoints{
-			SetCount:     len(poolArgs[i]),
-			DrivesPerSet: len(poolArgs[i][0]),
+			Legacy:       legacy,
+			SetCount:     len(poolArgs[i].layout),
+			DrivesPerSet: len(poolArgs[i].layout[0]),
 			Endpoints:    endpointList,
 			Platform:     fmt.Sprintf("OS: %s | Arch: %s", runtime.GOOS, runtime.GOARCH),
-			CmdLine:      args[i],
+			CmdLine:      poolArgs[i].cmdline,
 		}); err != nil {
 			return nil, -1, err
 		}
@@ -54,15 +54,20 @@ func TestCreateServerEndpoints(t *testing.T) {
 		{":9001", []string{"http://localhost:9001/export{01...64}"}, true},
 	}

-	for _, testCase := range testCases {
+	for i, testCase := range testCases {
 		testCase := testCase
 		t.Run("", func(t *testing.T) {
-			_, _, err := createServerEndpoints(testCase.serverAddr, testCase.args...)
+			srvCtxt := serverCtxt{}
+			err := mergeDisksLayoutFromArgs(testCase.args, &srvCtxt)
 			if err != nil && testCase.success {
-				t.Errorf("Expected success but failed instead %s", err)
+				t.Fatalf("Test %d: unexpected error: %v", i+1, err)
+			}
+			_, _, err = createServerEndpoints(testCase.serverAddr, srvCtxt.Layout.pools, srvCtxt.Layout.legacy)
+			if err != nil && testCase.success {
+				t.Errorf("Test %d: Expected success but failed instead %s", i+1, err)
 			}
 			if err == nil && !testCase.success {
-				t.Errorf("Expected failure but passed instead")
+				t.Errorf("Test %d: Expected failure but passed instead", i+1)
 			}
 		})
 	}
@@ -96,6 +96,11 @@ func (endpoint Endpoint) HTTPS() bool {
 	return endpoint.Scheme == "https"
 }

+// GridHost returns the host to be used for grid connections.
+func (endpoint Endpoint) GridHost() string {
+	return fmt.Sprintf("%s://%s", endpoint.Scheme, endpoint.Host)
+}
+
 // UpdateIsLocal - resolves the host and updates if it is local or not.
 func (endpoint *Endpoint) UpdateIsLocal() (err error) {
 	if !endpoint.IsLocal {
@@ -798,11 +803,23 @@ func (p PoolEndpointList) UpdateIsLocal() error {
 	return nil
 }

+func isEmptyLayout(poolsLayout ...poolDisksLayout) bool {
+	return len(poolsLayout) == 0 || len(poolsLayout[0].layout) == 0 || len(poolsLayout[0].layout[0]) == 0 || len(poolsLayout[0].layout[0][0]) == 0
+}
+
+func isSingleDriveLayout(poolsLayout ...poolDisksLayout) bool {
+	return len(poolsLayout) == 1 && len(poolsLayout[0].layout) == 1 && len(poolsLayout[0].layout[0]) == 1
+}
+
 // CreatePoolEndpoints creates a list of endpoints per pool, resolves their relevant hostnames and
 // discovers those are local or remote.
-func CreatePoolEndpoints(serverAddr string, poolArgs ...[][]string) ([]Endpoints, SetupType, error) {
+func CreatePoolEndpoints(serverAddr string, poolsLayout ...poolDisksLayout) ([]Endpoints, SetupType, error) {
 	var setupType SetupType

+	if isEmptyLayout(poolsLayout...) {
+		return nil, setupType, config.ErrInvalidErasureEndpoints(nil).Msg("invalid number of endpoints")
+	}
+
 	// Check whether serverAddr is valid for this host.
 	if err := CheckLocalServerAddr(serverAddr); err != nil {
 		return nil, setupType, err
@@ -810,11 +827,11 @@ func CreatePoolEndpoints(serverAddr string, poolArgs ...[][]string) ([]Endpoints

 	_, serverAddrPort := mustSplitHostPort(serverAddr)

-	poolEndpoints := make(PoolEndpointList, len(poolArgs))
+	poolEndpoints := make(PoolEndpointList, len(poolsLayout))

 	// For single arg, return single drive EC setup.
-	if len(poolArgs) == 1 && len(poolArgs[0]) == 1 && len(poolArgs[0][0]) == 1 && len(poolArgs[0][0][0]) == 1 {
-		endpoint, err := NewEndpoint(poolArgs[0][0][0])
+	if isSingleDriveLayout(poolsLayout...) {
+		endpoint, err := NewEndpoint(poolsLayout[0].layout[0][0])
 		if err != nil {
 			return nil, setupType, err
 		}
@@ -842,11 +859,11 @@ func CreatePoolEndpoints(serverAddr string, poolArgs ...[][]string) ([]Endpoints
 		return poolEndpoints, setupType, nil
 	}

-	for poolIdx, args := range poolArgs {
+	for poolIdx, pool := range poolsLayout {
 		var endpoints Endpoints
-		for setIdx, iargs := range args {
+		for setIdx, setLayout := range pool.layout {
 			// Convert args to endpoints
-			eps, err := NewEndpoints(iargs...)
+			eps, err := NewEndpoints(setLayout...)
 			if err != nil {
 				return nil, setupType, config.ErrInvalidErasureEndpoints(nil).Msg(err.Error())
 			}
@@ -998,6 +1015,8 @@ func CreatePoolEndpoints(serverAddr string, poolArgs ...[][]string) ([]Endpoints
 		poolEndpoints[i] = endpoints
 	}

+	// TODO: ensure that each pool has at least two nodes in a distributed setup
+
 	publicIPs := env.Get(config.EnvPublicIPs, "")
 	if len(publicIPs) == 0 {
 		updateDomainIPs(uniqueArgs)
@@ -1016,199 +1035,6 @@ func CreatePoolEndpoints(serverAddr string, poolArgs ...[][]string) ([]Endpoints
 	return poolEndpoints, setupType, nil
 }

-// CreateEndpoints - validates and creates new endpoints for given args.
-func CreateEndpoints(serverAddr string, args ...[]string) (Endpoints, SetupType, error) {
-	var endpoints Endpoints
-	var setupType SetupType
-	var err error
-
-	// Check whether serverAddr is valid for this host.
-	if err = CheckLocalServerAddr(serverAddr); err != nil {
-		return endpoints, setupType, err
-	}
-
-	_, serverAddrPort := mustSplitHostPort(serverAddr)
-
-	// For single arg, return single drive setup.
-	if len(args) == 1 && len(args[0]) == 1 {
-		var endpoint Endpoint
-		endpoint, err = NewEndpoint(args[0][0])
-		if err != nil {
-			return endpoints, setupType, err
-		}
-		if err := endpoint.UpdateIsLocal(); err != nil {
-			return endpoints, setupType, err
-		}
-		if endpoint.Type() != PathEndpointType {
-			return endpoints, setupType, config.ErrInvalidEndpoint(nil).Msg("use path style endpoint for single node setup")
-		}
-
-		endpoint.SetPoolIndex(0)
-		endpoint.SetSetIndex(0)
-		endpoint.SetDiskIndex(0)
-
-		endpoints = append(endpoints, endpoint)
-		setupType = ErasureSDSetupType
-
-		// Check for cross device mounts if any.
-		if err = checkCrossDeviceMounts(endpoints); err != nil {
-			return endpoints, setupType, config.ErrInvalidEndpoint(nil).Msg(err.Error())
-		}
-
-		return endpoints, setupType, nil
-	}
-
-	for setIdx, iargs := range args {
-		// Convert args to endpoints
-		eps, err := NewEndpoints(iargs...)
-		if err != nil {
-			return endpoints, setupType, config.ErrInvalidErasureEndpoints(nil).Msg(err.Error())
-		}
-
-		// Check for cross device mounts if any.
-		if err = checkCrossDeviceMounts(eps); err != nil {
-			return endpoints, setupType, config.ErrInvalidErasureEndpoints(nil).Msg(err.Error())
-		}
-
-		for diskIdx := range eps {
-			eps[diskIdx].SetSetIndex(setIdx)
-			eps[diskIdx].SetDiskIndex(diskIdx)
-		}
-
-		endpoints = append(endpoints, eps...)
-	}
-
-	if len(endpoints) == 0 {
-		return endpoints, setupType, config.ErrInvalidErasureEndpoints(nil).Msg("invalid number of endpoints")
-	}
-
-	// Return Erasure setup when all endpoints are path style.
-	if endpoints[0].Type() == PathEndpointType {
-		setupType = ErasureSetupType
-		return endpoints, setupType, nil
-	}
-
-	if err = endpoints.UpdateIsLocal(); err != nil {
-		return endpoints, setupType, config.ErrInvalidErasureEndpoints(nil).Msg(err.Error())
-	}
-
-	// Here all endpoints are URL style.
-	endpointPathSet := set.NewStringSet()
-	localEndpointCount := 0
-	localServerHostSet := set.NewStringSet()
-	localPortSet := set.NewStringSet()
-
-	for _, endpoint := range endpoints {
-		endpointPathSet.Add(endpoint.Path)
-		if endpoint.IsLocal {
-			localServerHostSet.Add(endpoint.Hostname())
-
-			var port string
-			_, port, err = net.SplitHostPort(endpoint.Host)
-			if err != nil {
-				port = serverAddrPort
-			}
-			localPortSet.Add(port)
-
-			localEndpointCount++
-		}
-	}
-
-	orchestrated := IsKubernetes() || IsDocker()
-	reverseProxy := (env.Get("_MINIO_REVERSE_PROXY", "") != "") && ((env.Get("MINIO_CI_CD", "") != "") || (env.Get("CI", "") != ""))
-	// If not orchestrated
-	if !orchestrated &&
-		// and not setup in reverse proxy
-		!reverseProxy {
-		// Check whether same path is not used in endpoints of a host on different port.
-		// Only verify this on baremetal setups, DNS is not available in orchestrated
-		// environments so we can't do much here.
-		pathIPMap := make(map[string]set.StringSet)
-		hostIPCache := make(map[string]set.StringSet)
-		for _, endpoint := range endpoints {
-			host := endpoint.Hostname()
-			hostIPSet, ok := hostIPCache[host]
-			if !ok {
-				hostIPSet, err = getHostIP(host)
-				if err != nil {
-					return endpoints, setupType, config.ErrInvalidErasureEndpoints(nil).Msg(fmt.Sprintf("host '%s' cannot resolve: %s", host, err))
-				}
-				hostIPCache[host] = hostIPSet
-			}
-			if IPSet, ok := pathIPMap[endpoint.Path]; ok {
-				if !IPSet.Intersection(hostIPSet).IsEmpty() {
-					return endpoints, setupType,
-						config.ErrInvalidErasureEndpoints(nil).Msg(fmt.Sprintf("same path '%s' can not be served by different port on same address", endpoint.Path))
-				}
-				pathIPMap[endpoint.Path] = IPSet.Union(hostIPSet)
-			} else {
-				pathIPMap[endpoint.Path] = hostIPSet
-			}
-		}
-	}
-
-	// Check whether same path is used for more than 1 local endpoints.
-	{
-		localPathSet := set.CreateStringSet()
-		for _, endpoint := range endpoints {
-			if !endpoint.IsLocal {
-				continue
-			}
-			if localPathSet.Contains(endpoint.Path) {
-				return endpoints, setupType,
-					config.ErrInvalidErasureEndpoints(nil).Msg(fmt.Sprintf("path '%s' cannot be served by different address on same server", endpoint.Path))
-			}
-			localPathSet.Add(endpoint.Path)
-		}
-	}
-
-	// Add missing port in all endpoints.
-	for i := range endpoints {
-		_, port, err := net.SplitHostPort(endpoints[i].Host)
-		if err != nil {
-			endpoints[i].Host = net.JoinHostPort(endpoints[i].Host, serverAddrPort)
-		} else if endpoints[i].IsLocal && serverAddrPort != port {
-			// If endpoint is local, but port is different than serverAddrPort, then make it as remote.
-			endpoints[i].IsLocal = false
-		}
-	}
-
-	// All endpoints are pointing to local host
-	if len(endpoints) == localEndpointCount {
-		// If all endpoints have same port number, Just treat it as local erasure setup
-		// using URL style endpoints.
-		if len(localPortSet) == 1 {
-			if len(localServerHostSet) > 1 {
-				return endpoints, setupType,
-					config.ErrInvalidErasureEndpoints(nil).Msg("all local endpoints should not have different hostnames/ips")
-			}
-			return endpoints, ErasureSetupType, nil
-		}
-
-		// Even though all endpoints are local, but those endpoints use different ports.
-		// This means it is DistErasure setup.
-	}
-
-	uniqueArgs := set.NewStringSet()
-	for _, endpoint := range endpoints {
-		uniqueArgs.Add(endpoint.Host)
-	}
-
-	// Error out if we have less than 2 unique servers.
-	if len(uniqueArgs.ToSlice()) < 2 && setupType == DistErasureSetupType {
-		err := fmt.Errorf("Unsupported number of endpoints (%s), minimum number of servers cannot be less than 2 in distributed setup", endpoints)
-		return endpoints, setupType, err
-	}
-
-	publicIPs := env.Get(config.EnvPublicIPs, "")
-	if len(publicIPs) == 0 {
-		updateDomainIPs(uniqueArgs)
-	}
-
-	setupType = DistErasureSetupType
-	return endpoints, setupType, nil
-}
-
 // GetLocalPeer - returns local peer value, returns globalMinioAddr
 // for FS and Erasure mode. In case of distributed server return
 // the first element from the set of peers which indicate that
@@ -225,24 +225,24 @@ func TestCreateEndpoints(t *testing.T) {

 	testCases := []struct {
 		serverAddr         string
-		args               [][]string
+		args               []string
 		expectedServerAddr string
 		expectedEndpoints  Endpoints
 		expectedSetupType  SetupType
 		expectedErr        error
 	}{
-		{"localhost", [][]string{}, "", Endpoints{}, -1, fmt.Errorf("address localhost: missing port in address")},
+		{"localhost", []string{}, "", Endpoints{}, -1, fmt.Errorf("address localhost: missing port in address")},

 		// Erasure Single Drive
-		{"localhost:9000", [][]string{{"http://localhost/d1"}}, "", Endpoints{}, -1, fmt.Errorf("use path style endpoint for FS setup")},
-		{":443", [][]string{{"/d1"}}, ":443", Endpoints{Endpoint{URL: &url.URL{Path: mustAbs("/d1")}, IsLocal: true}}, ErasureSDSetupType, nil},
-		{"localhost:10000", [][]string{{"/d1"}}, "localhost:10000", Endpoints{Endpoint{URL: &url.URL{Path: mustAbs("/d1")}, IsLocal: true}}, ErasureSDSetupType, nil},
-		{"localhost:9000", [][]string{{"https://127.0.0.1:9000/d1", "https://localhost:9001/d1", "https://example.com/d1", "https://example.com/d2"}}, "", Endpoints{}, -1, fmt.Errorf("path '/d1' can not be served by different port on same address")},
+		{"localhost:9000", []string{"http://localhost/d1"}, "", Endpoints{}, -1, fmt.Errorf("use path style endpoint for SD setup")},
+		{":443", []string{"/d1"}, ":443", Endpoints{Endpoint{URL: &url.URL{Path: mustAbs("/d1")}, IsLocal: true}}, ErasureSDSetupType, nil},
+		{"localhost:10000", []string{"/d1"}, "localhost:10000", Endpoints{Endpoint{URL: &url.URL{Path: mustAbs("/d1")}, IsLocal: true}}, ErasureSDSetupType, nil},
+		{"localhost:9000", []string{"https://127.0.0.1:9000/d1", "https://localhost:9001/d1", "https://example.com/d1", "https://example.com/d2"}, "", Endpoints{}, -1, fmt.Errorf("path '/d1' can not be served by different port on same address")},

 		// Erasure Setup with PathEndpointType
 		{
 			":1234",
-			[][]string{{"/d1", "/d2", "/d3", "/d4"}},
+			[]string{"/d1", "/d2", "/d3", "/d4"},
 			":1234",
 			Endpoints{
 				Endpoint{URL: &url.URL{Path: mustAbs("/d1")}, IsLocal: true},
@@ -253,49 +253,49 @@ func TestCreateEndpoints(t *testing.T) {
 			ErasureSetupType, nil,
 		},
 		// DistErasure Setup with URLEndpointType
-		{":9000", [][]string{{"http://localhost/d1", "http://localhost/d2", "http://localhost/d3", "http://localhost/d4"}}, ":9000", Endpoints{
+		{":9000", []string{"http://localhost/d1", "http://localhost/d2", "http://localhost/d3", "http://localhost/d4"}, ":9000", Endpoints{
 			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost:9000", Path: "/d1"}, IsLocal: true},
 			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost:9000", Path: "/d2"}, IsLocal: true},
 			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost:9000", Path: "/d3"}, IsLocal: true},
 			Endpoint{URL: &url.URL{Scheme: "http", Host: "localhost:9000", Path: "/d4"}, IsLocal: true},
-		}, ErasureSetupType, nil},
+		}, DistErasureSetupType, nil},
 		// DistErasure Setup with URLEndpointType having mixed naming to local host.
-		{"127.0.0.1:10000", [][]string{{"http://localhost/d1", "http://localhost/d2", "http://127.0.0.1/d3", "http://127.0.0.1/d4"}}, "", Endpoints{}, -1, fmt.Errorf("all local endpoints should not have different hostnames/ips")},
+		{"127.0.0.1:10000", []string{"http://localhost/d1", "http://localhost/d2", "http://127.0.0.1/d3", "http://127.0.0.1/d4"}, "", Endpoints{}, -1, fmt.Errorf("all local endpoints should not have different hostnames/ips")},

-		{":9001", [][]string{{"http://10.0.0.1:9000/export", "http://10.0.0.2:9000/export", "http://" + nonLoopBackIP + ":9001/export", "http://10.0.0.2:9001/export"}}, "", Endpoints{}, -1, fmt.Errorf("path '/export' can not be served by different port on same address")},
+		{":9001", []string{"http://10.0.0.1:9000/export", "http://10.0.0.2:9000/export", "http://" + nonLoopBackIP + ":9001/export", "http://10.0.0.2:9001/export"}, "", Endpoints{}, -1, fmt.Errorf("path '/export' can not be served by different port on same address")},

-		{":9000", [][]string{{"http://127.0.0.1:9000/export", "http://" + nonLoopBackIP + ":9000/export", "http://10.0.0.1:9000/export", "http://10.0.0.2:9000/export"}}, "", Endpoints{}, -1, fmt.Errorf("path '/export' cannot be served by different address on same server")},
+		{":9000", []string{"http://127.0.0.1:9000/export", "http://" + nonLoopBackIP + ":9000/export", "http://10.0.0.1:9000/export", "http://10.0.0.2:9000/export"}, "", Endpoints{}, -1, fmt.Errorf("path '/export' cannot be served by different address on same server")},

 		// DistErasure type
-		{"127.0.0.1:10000", [][]string{{case1Endpoint1, case1Endpoint2, "http://example.org/d3", "http://example.com/d4"}}, "127.0.0.1:10000", Endpoints{
+		{"127.0.0.1:10000", []string{case1Endpoint1, case1Endpoint2, "http://example.org/d3", "http://example.com/d4"}, "127.0.0.1:10000", Endpoints{
 			Endpoint{URL: case1URLs[0], IsLocal: case1LocalFlags[0]},
 			Endpoint{URL: case1URLs[1], IsLocal: case1LocalFlags[1]},
 			Endpoint{URL: case1URLs[2], IsLocal: case1LocalFlags[2]},
 			Endpoint{URL: case1URLs[3], IsLocal: case1LocalFlags[3]},
 		}, DistErasureSetupType, nil},

-		{"127.0.0.1:10000", [][]string{{case2Endpoint1, case2Endpoint2, "http://example.org/d3", "http://example.com/d4"}}, "127.0.0.1:10000", Endpoints{
+		{"127.0.0.1:10000", []string{case2Endpoint1, case2Endpoint2, "http://example.org/d3", "http://example.com/d4"}, "127.0.0.1:10000", Endpoints{
 			Endpoint{URL: case2URLs[0], IsLocal: case2LocalFlags[0]},
 			Endpoint{URL: case2URLs[1], IsLocal: case2LocalFlags[1]},
 			Endpoint{URL: case2URLs[2], IsLocal: case2LocalFlags[2]},
 			Endpoint{URL: case2URLs[3], IsLocal: case2LocalFlags[3]},
 		}, DistErasureSetupType, nil},

-		{":80", [][]string{{case3Endpoint1, "http://example.org:9000/d2", "http://example.com/d3", "http://example.net/d4"}}, ":80", Endpoints{
+		{":80", []string{case3Endpoint1, "http://example.org:9000/d2", "http://example.com/d3", "http://example.net/d4"}, ":80", Endpoints{
 			Endpoint{URL: case3URLs[0], IsLocal: case3LocalFlags[0]},
 			Endpoint{URL: case3URLs[1], IsLocal: case3LocalFlags[1]},
 			Endpoint{URL: case3URLs[2], IsLocal: case3LocalFlags[2]},
 			Endpoint{URL: case3URLs[3], IsLocal: case3LocalFlags[3]},
 		}, DistErasureSetupType, nil},

-		{":9000", [][]string{{case4Endpoint1, "http://example.org/d2", "http://example.com/d3", "http://example.net/d4"}}, ":9000", Endpoints{
+		{":9000", []string{case4Endpoint1, "http://example.org/d2", "http://example.com/d3", "http://example.net/d4"}, ":9000", Endpoints{
 			Endpoint{URL: case4URLs[0], IsLocal: case4LocalFlags[0]},
 			Endpoint{URL: case4URLs[1], IsLocal: case4LocalFlags[1]},
 			Endpoint{URL: case4URLs[2], IsLocal: case4LocalFlags[2]},
 			Endpoint{URL: case4URLs[3], IsLocal: case4LocalFlags[3]},
 		}, DistErasureSetupType, nil},

-		{":9000", [][]string{{case5Endpoint1, case5Endpoint2, case5Endpoint3, case5Endpoint4}}, ":9000", Endpoints{
+		{":9000", []string{case5Endpoint1, case5Endpoint2, case5Endpoint3, case5Endpoint4}, ":9000", Endpoints{
 			Endpoint{URL: case5URLs[0], IsLocal: case5LocalFlags[0]},
 			Endpoint{URL: case5URLs[1], IsLocal: case5LocalFlags[1]},
 			Endpoint{URL: case5URLs[2], IsLocal: case5LocalFlags[2]},
@@ -303,7 +303,7 @@ func TestCreateEndpoints(t *testing.T) {
 		}, DistErasureSetupType, nil},

 		// DistErasure Setup using only local host.
-		{":9003", [][]string{{"http://localhost:9000/d1", "http://localhost:9001/d2", "http://127.0.0.1:9002/d3", case6Endpoint}}, ":9003", Endpoints{
+		{":9003", []string{"http://localhost:9000/d1", "http://localhost:9001/d2", "http://127.0.0.1:9002/d3", case6Endpoint}, ":9003", Endpoints{
 			Endpoint{URL: case6URLs[0], IsLocal: case6LocalFlags[0]},
 			Endpoint{URL: case6URLs[1], IsLocal: case6LocalFlags[1]},
 			Endpoint{URL: case6URLs[2], IsLocal: case6LocalFlags[2]},
@@ -311,24 +311,32 @@ func TestCreateEndpoints(t *testing.T) {
 		}, DistErasureSetupType, nil},
 	}

-	for _, testCase := range testCases {
+	for i, testCase := range testCases {
+		i := i
 		testCase := testCase
 		t.Run("", func(t *testing.T) {
-			endpoints, setupType, err := CreateEndpoints(testCase.serverAddr, testCase.args...)
+			var srvCtxt serverCtxt
+			err := mergeDisksLayoutFromArgs(testCase.args, &srvCtxt)
+			if err != nil && testCase.expectedErr == nil {
+				t.Errorf("Test %d: unexpected error: %v", i+1, err)
+			}
+			pools, setupType, err := CreatePoolEndpoints(testCase.serverAddr, srvCtxt.Layout.pools...)
 			if err == nil && testCase.expectedErr != nil {
-				t.Errorf("error: expected = %v, got = <nil>", testCase.expectedErr)
+				t.Errorf("Test %d: expected = %v, got = <nil>", i+1, testCase.expectedErr)
 			}
 			if err == nil {
 				if setupType != testCase.expectedSetupType {
-					t.Errorf("setupType: expected = %v, got = %v", testCase.expectedSetupType, setupType)
+					t.Errorf("Test %d: setupType: expected = %v, got = %v", i+1, testCase.expectedSetupType, setupType)
 				}
+				endpoints := pools[0]
 				if len(endpoints) != len(testCase.expectedEndpoints) {
-					t.Errorf("endpoints: expected = %d, got = %d", len(testCase.expectedEndpoints),
+					t.Errorf("Test %d: endpoints: expected = %d, got = %d", i+1, len(testCase.expectedEndpoints),
 						len(endpoints))
 				} else {
 					for i, endpoint := range endpoints {
 						if testCase.expectedEndpoints[i].String() != endpoint.String() {
-							t.Errorf("endpoints: expected = %s, got = %s",
+							t.Errorf("Test %d: endpoints: expected = %s, got = %s",
+								i+1,
 								testCase.expectedEndpoints[i],
 								endpoint)
 						}
@@ -336,7 +344,7 @@ func TestCreateEndpoints(t *testing.T) {
 				}
 			}
 			if err != nil && testCase.expectedErr == nil {
-				t.Errorf("error: expected = <nil>, got = %v, testCase: %v", err, testCase)
+				t.Errorf("Test %d: error: expected = <nil>, got = %v, testCase: %v", i+1, err, testCase)
 			}
 		})
 	}
@@ -190,7 +190,7 @@ func benchmarkErasureEncode(data, parity, dataDown, parityDown int, size int64,
 			}
 			disk.Delete(context.Background(), "testbucket", "object", DeleteOptions{
 				Recursive: false,
-				Force:     false,
+				Immediate: false,
 			})
 			writers[i] = newBitrotWriter(disk, "testbucket", "object", erasure.ShardFileSize(size), DefaultBitrotAlgorithm, erasure.ShardSize())
 		}
@@ -286,6 +286,7 @@ func disksWithAllParts(ctx context.Context, onlineDisks []StorageAPI, partsMetad

 	availableDisks := make([]StorageAPI, len(onlineDisks))
 	dataErrs := make([]error, len(onlineDisks))
+
 	inconsistent := 0
 	for i, meta := range partsMetadata {
 		if !meta.IsValid() {
@@ -246,7 +246,7 @@ func TestListOnlineDisks(t *testing.T) {
 				t.Fatalf("Failed to putObject %v", err)
 			}

-			partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+			partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 			fi, err := getLatestFileInfo(ctx, partsMetadata, z.serverPools[0].sets[0].defaultParityCount, errs)
 			if err != nil {
 				t.Fatalf("Failed to getLatestFileInfo %v", err)
@@ -273,7 +273,7 @@ func TestListOnlineDisks(t *testing.T) {
 					tamperedIndex = index
 					dErr := erasureDisks[index].Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), DeleteOptions{
 						Recursive: false,
-						Force:     false,
+						Immediate: false,
 					})
 					if dErr != nil {
 						t.Fatalf("Failed to delete %s - %v", filepath.Join(object, "part.1"), dErr)
@@ -424,7 +424,7 @@ func TestListOnlineDisksSmallObjects(t *testing.T) {
 				t.Fatalf("Failed to putObject %v", err)
 			}

-			partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", true)
+			partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", true, true)
 			fi, err := getLatestFileInfo(ctx, partsMetadata, z.serverPools[0].sets[0].defaultParityCount, errs)
 			if err != nil {
 				t.Fatalf("Failed to getLatestFileInfo %v", err)
@@ -455,7 +455,7 @@ func TestListOnlineDisksSmallObjects(t *testing.T) {
 					tamperedIndex = index
 					dErr := erasureDisks[index].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
 						Recursive: false,
-						Force:     false,
+						Immediate: false,
 					})
 					if dErr != nil {
 						t.Fatalf("Failed to delete %s - %v", pathJoin(object, xlStorageFormatFile), dErr)
@@ -534,7 +534,7 @@ func TestDisksWithAllParts(t *testing.T) {
 		t.Fatalf("Failed to putObject %v", err)
 	}

-	_, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+	_, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 	readQuorum := len(erasureDisks) / 2
 	if reducedErr := reduceReadQuorumErrs(ctx, errs, objectOpIgnoredErrs, readQuorum); reducedErr != nil {
 		t.Fatalf("Failed to read xl meta data %v", reducedErr)
@@ -542,7 +542,7 @@ func TestDisksWithAllParts(t *testing.T) {

 	// Test 1: Test that all disks are returned without any failures with
 	// unmodified meta data
-	partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+	partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 	if err != nil {
 		t.Fatalf("Failed to read xl meta data %v", err)
 	}
@@ -23,6 +23,7 @@ import (
 	"errors"
 	"fmt"
 	"io"
+	"strconv"
 	"strings"
 	"sync"
 	"time"
@@ -67,11 +68,11 @@ func (fi FileInfo) DataShardFixed() bool {
 	return fi.Metadata[reservedMetadataPrefixLowerDataShardFix] == "true"
 }

-func (er erasureObjects) listAndHeal(bucket, prefix string, healEntry func(string, metaCacheEntry) error) error {
+func (er erasureObjects) listAndHeal(bucket, prefix string, scanMode madmin.HealScanMode, healEntry func(string, metaCacheEntry, madmin.HealScanMode) error) error {
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()

-	disks, _ := er.getOnlineDisksWithHealing()
+	disks, _ := er.getOnlineDisksWithHealing(false)
 	if len(disks) == 0 {
 		return errors.New("listAndHeal: No non-healing drives found")
 	}
@@ -100,7 +101,7 @@ func (er erasureObjects) listAndHeal(bucket, prefix string, healEntry func(strin
 		minDisks:       1,
 		reportNotFound: false,
 		agreed: func(entry metaCacheEntry) {
-			if err := healEntry(bucket, entry); err != nil {
+			if err := healEntry(bucket, entry, scanMode); err != nil {
 				cancel()
 			}
 		},
@@ -112,7 +113,7 @@ func (er erasureObjects) listAndHeal(bucket, prefix string, healEntry func(strin
 				entry, _ = entries.firstFound()
 			}

-			if err := healEntry(bucket, *entry); err != nil {
+			if err := healEntry(bucket, *entry, scanMode); err != nil {
 				cancel()
 				return
 			}
@@ -369,10 +370,8 @@ func (fi *FileInfo) SetHealing() {
 // Healing returns true if object is being healed (i.e fi is being passed down
 // from healObject)
 func (fi FileInfo) Healing() bool {
-	if _, ok := fi.Metadata[xMinIOHealing]; ok {
-		return true
-	}
-	return false
+	_, ok := fi.Metadata[xMinIOHealing]
+	return ok
 }

 // Heals an object by re-writing corrupt/missing erasure blocks.
@@ -409,7 +408,7 @@ func (er *erasureObjects) healObject(ctx context.Context, bucket string, object
 	}

 	// Re-read when we have lock...
-	partsMetadata, errs := readAllFileInfo(ctx, storageDisks, bucket, object, versionID, true)
+	partsMetadata, errs := readAllFileInfo(ctx, storageDisks, bucket, object, versionID, true, true)
 	if isAllNotFound(errs) {
 		err := errFileNotFound
 		if versionID != "" {
@@ -759,7 +758,8 @@ func (er *erasureObjects) healObject(ctx context.Context, bucket string, object

 		// Attempt a rename now from healed data to final location.
 		partsMetadata[i].SetHealing()
-		if _, err = disk.RenameData(ctx, minioMetaTmpBucket, tmpID, partsMetadata[i], bucket, object); err != nil {
+
+		if _, err = disk.RenameData(ctx, minioMetaTmpBucket, tmpID, partsMetadata[i], bucket, object, RenameOptions{}); err != nil {
 			return result, err
 		}

@@ -865,7 +865,7 @@ func (er *erasureObjects) healObjectDir(ctx context.Context, bucket, object stri
 					defer wg.Done()
 					_ = disk.Delete(ctx, bucket, object, DeleteOptions{
 						Recursive: false,
-						Force:     false,
+						Immediate: false,
 					})
 				}(index, disk)
 			}
@@ -1162,7 +1162,7 @@ func (er erasureObjects) HealObject(ctx context.Context, bucket, object, version

 	// Perform quick read without lock.
 	// This allows to quickly check if all is ok or all are missing.
-	_, errs := readAllFileInfo(healCtx, storageDisks, bucket, object, versionID, false)
+	_, errs := readAllFileInfo(healCtx, storageDisks, bucket, object, versionID, false, false)
 	if isAllNotFound(errs) {
 		err := errFileNotFound
 		if versionID != "" {
@@ -1201,6 +1201,10 @@ func healTrace(funcName healingMetric, startTime time.Time, bucket, object strin
 			"recreate": fmt.Sprint(opts.Recreate),
 			"mode":     fmt.Sprint(opts.ScanMode),
 		}
+		if result != nil {
+			tr.Custom["version-id"] = result.VersionID
+			tr.Custom["disks"] = strconv.Itoa(result.DiskCount)
+		}
 	}
 	if err != nil {
 		tr.Error = err.Error()
@@ -266,7 +266,7 @@ func TestHealing(t *testing.T) {
 	}

 	disk := er.getDisks()[0]
-	fileInfoPreHeal, err := disk.ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPreHeal, err := disk.ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -289,7 +289,7 @@ func TestHealing(t *testing.T) {
 		t.Fatal(err)
 	}

-	fileInfoPostHeal, err := disk.ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPostHeal, err := disk.ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -318,7 +318,7 @@ func TestHealing(t *testing.T) {
 		t.Fatal(err)
 	}

-	fileInfoPostHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPostHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -426,11 +426,11 @@ func TestHealingVersioned(t *testing.T) {
 	}

 	disk := er.getDisks()[0]
-	fileInfoPreHeal1, err := disk.ReadVersion(context.Background(), bucket, object, oi1.VersionID, false)
+	fileInfoPreHeal1, err := disk.ReadVersion(context.Background(), bucket, object, oi1.VersionID, ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
-	fileInfoPreHeal2, err := disk.ReadVersion(context.Background(), bucket, object, oi2.VersionID, false)
+	fileInfoPreHeal2, err := disk.ReadVersion(context.Background(), bucket, object, oi2.VersionID, ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -453,11 +453,11 @@ func TestHealingVersioned(t *testing.T) {
 		t.Fatal(err)
 	}

-	fileInfoPostHeal1, err := disk.ReadVersion(context.Background(), bucket, object, oi1.VersionID, false)
+	fileInfoPostHeal1, err := disk.ReadVersion(context.Background(), bucket, object, oi1.VersionID, ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
-	fileInfoPostHeal2, err := disk.ReadVersion(context.Background(), bucket, object, oi2.VersionID, false)
+	fileInfoPostHeal2, err := disk.ReadVersion(context.Background(), bucket, object, oi2.VersionID, ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -489,7 +489,7 @@ func TestHealingVersioned(t *testing.T) {
 		t.Fatal(err)
 	}

-	fileInfoPostHeal1, err = disk.ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPostHeal1, err = disk.ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -499,7 +499,7 @@ func TestHealingVersioned(t *testing.T) {
 		t.Fatal("HealObject failed")
 	}

-	fileInfoPostHeal2, err = disk.ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPostHeal2, err = disk.ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -637,7 +637,7 @@ func TestHealingDanglingObject(t *testing.T) {
 	// Restore...
 	setDisks(orgDisks[:4]...)

-	fileInfoPreHeal, err := disks[0].ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPreHeal, err := disks[0].ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -647,14 +647,14 @@ func TestHealingDanglingObject(t *testing.T) {
 	}

 	if err = objLayer.HealObjects(ctx, bucket, "", madmin.HealOpts{Remove: true},
-		func(bucket, object, vid string) error {
-			_, err := objLayer.HealObject(ctx, bucket, object, vid, madmin.HealOpts{Remove: true})
+		func(bucket, object, vid string, scanMode madmin.HealScanMode) error {
+			_, err := objLayer.HealObject(ctx, bucket, object, vid, madmin.HealOpts{ScanMode: scanMode, Remove: true})
 			return err
 		}); err != nil {
 		t.Fatal(err)
 	}

-	fileInfoPostHeal, err := disks[0].ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPostHeal, err := disks[0].ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -684,7 +684,7 @@ func TestHealingDanglingObject(t *testing.T) {

 	setDisks(orgDisks[:4]...)
 	disk := getDisk(0)
-	fileInfoPreHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPreHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -694,15 +694,15 @@ func TestHealingDanglingObject(t *testing.T) {
 	}

 	if err = objLayer.HealObjects(ctx, bucket, "", madmin.HealOpts{Remove: true},
-		func(bucket, object, vid string) error {
-			_, err := objLayer.HealObject(ctx, bucket, object, vid, madmin.HealOpts{Remove: true})
+		func(bucket, object, vid string, scanMode madmin.HealScanMode) error {
+			_, err := objLayer.HealObject(ctx, bucket, object, vid, madmin.HealOpts{ScanMode: scanMode, Remove: true})
 			return err
 		}); err != nil {
 		t.Fatal(err)
 	}

 	disk = getDisk(0)
-	fileInfoPostHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPostHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -733,7 +733,7 @@ func TestHealingDanglingObject(t *testing.T) {
 	setDisks(orgDisks[:4]...)

 	disk = getDisk(0)
-	fileInfoPreHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPreHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -743,15 +743,15 @@ func TestHealingDanglingObject(t *testing.T) {
 	}

 	if err = objLayer.HealObjects(ctx, bucket, "", madmin.HealOpts{Remove: true},
-		func(bucket, object, vid string) error {
-			_, err := objLayer.HealObject(ctx, bucket, object, vid, madmin.HealOpts{Remove: true})
+		func(bucket, object, vid string, scanMode madmin.HealScanMode) error {
+			_, err := objLayer.HealObject(ctx, bucket, object, vid, madmin.HealOpts{ScanMode: scanMode, Remove: true})
 			return err
 		}); err != nil {
 		t.Fatal(err)
 	}

 	disk = getDisk(0)
-	fileInfoPostHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", false)
+	fileInfoPostHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", ReadOptions{ReadData: false, Healing: true})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -836,7 +836,7 @@ func TestHealCorrectQuorum(t *testing.T) {
 		er := set.sets[0]
 		erasureDisks := er.getDisks()

-		fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+		fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 		nfi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 		if errors.Is(err, errFileNotFound) {
 			continue
@@ -848,7 +848,7 @@ func TestHealCorrectQuorum(t *testing.T) {
 		for i := 0; i < nfi.Erasure.ParityBlocks; i++ {
 			erasureDisks[i].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
 				Recursive: false,
-				Force:     false,
+				Immediate: false,
 			})
 		}

@@ -858,12 +858,12 @@ func TestHealCorrectQuorum(t *testing.T) {
 			t.Fatal(err)
 		}

-		fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+		fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 		if countErrs(errs, nil) != len(fileInfos) {
 			t.Fatal("Expected all xl.meta healed, but partial heal detected")
 		}

-		fileInfos, errs = readAllFileInfo(ctx, erasureDisks, minioMetaBucket, cfgFile, "", false)
+		fileInfos, errs = readAllFileInfo(ctx, erasureDisks, minioMetaBucket, cfgFile, "", false, true)
 		nfi, err = getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 		if errors.Is(err, errFileNotFound) {
 			continue
@@ -875,7 +875,7 @@ func TestHealCorrectQuorum(t *testing.T) {
 		for i := 0; i < nfi.Erasure.ParityBlocks; i++ {
 			erasureDisks[i].Delete(context.Background(), minioMetaBucket, pathJoin(cfgFile, xlStorageFormatFile), DeleteOptions{
 				Recursive: false,
-				Force:     false,
+				Immediate: false,
 			})
 		}

@@ -885,7 +885,7 @@ func TestHealCorrectQuorum(t *testing.T) {
 			t.Fatal(err)
 		}

-		fileInfos, errs = readAllFileInfo(ctx, erasureDisks, minioMetaBucket, cfgFile, "", false)
+		fileInfos, errs = readAllFileInfo(ctx, erasureDisks, minioMetaBucket, cfgFile, "", false, true)
 		if countErrs(errs, nil) != len(fileInfos) {
 			t.Fatal("Expected all xl.meta healed, but partial heal detected")
 		}
@@ -959,7 +959,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	firstDisk := erasureDisks[0]
 	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
 		Recursive: false,
-		Force:     false,
+		Immediate: false,
 	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
@@ -970,7 +970,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 		t.Fatalf("Failed to heal object - %v", err)
 	}

-	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 	fi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
@@ -982,7 +982,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {

 	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), DeleteOptions{
 		Recursive: false,
-		Force:     false,
+		Immediate: false,
 	})
 	if err != nil {
 		t.Errorf("Failure during deleting part.1 - %v", err)
@@ -998,7 +998,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 		t.Errorf("Expected nil but received %v", err)
 	}

-	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 	nfi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
@@ -1012,7 +1012,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {

 	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), DeleteOptions{
 		Recursive: false,
-		Force:     false,
+		Immediate: false,
 	})
 	if err != nil {
 		t.Errorf("Failure during deleting part.1 - %v", err)
@@ -1029,7 +1029,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 		t.Errorf("Expected nil but received %v", err)
 	}

-	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 	nfi, err = getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
@@ -1046,7 +1046,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	for i := 0; i <= nfi.Erasure.DataBlocks; i++ {
 		erasureDisks[i].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
 			Recursive: false,
-			Force:     false,
+			Immediate: false,
 		})
 	}

@@ -1133,7 +1133,7 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {
 	firstDisk := erasureDisks[0]

 	// Test 1: Remove the object backend files from the first disk.
-	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 	fi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
@@ -1141,7 +1141,7 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {

 	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
 		Recursive: false,
-		Force:     false,
+		Immediate: false,
 	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
@@ -1156,7 +1156,7 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {
 		t.Errorf("Expected xl.meta file to be present but stat failed - %v", err)
 	}

-	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 	nfi1, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
@@ -1179,7 +1179,7 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {
 		t.Errorf("Expected nil but received %v", err)
 	}

-	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 	nfi2, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
@@ -1196,7 +1196,7 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {
 	for i := 0; i <= nfi2.Erasure.DataBlocks; i++ {
 		erasureDisks[i].Delete(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), DeleteOptions{
 			Recursive: false,
-			Force:     false,
+			Immediate: false,
 		})
 	}

@@ -1277,7 +1277,7 @@ func TestHealObjectCorruptedParts(t *testing.T) {
 	firstDisk := erasureDisks[0]
 	secondDisk := erasureDisks[1]

-	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
+	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false, true)
 	fi, err := getLatestFileInfo(ctx, fileInfos, er.defaultParityCount, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
@@ -1296,7 +1296,7 @@ func TestHealObjectCorruptedParts(t *testing.T) {
 	// Test 1, remove part.1
 	err = firstDisk.Delete(context.Background(), bucket, pathJoin(object, fi.DataDir, "part.1"), DeleteOptions{
 		Recursive: false,
-		Force:     false,
+		Immediate: false,
 	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
@@ -1344,7 +1344,7 @@ func TestHealObjectCorruptedParts(t *testing.T) {

 	err = secondDisk.Delete(context.Background(), bucket, object, DeleteOptions{
 		Recursive: true,
-		Force:     false,
+		Immediate: false,
 	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
@@ -1435,7 +1435,7 @@ func TestHealObjectErasure(t *testing.T) {
 	// Delete the whole object folder
 	err = firstDisk.Delete(context.Background(), bucket, object, DeleteOptions{
 		Recursive: true,
-		Force:     false,
+		Immediate: false,
 	})
 	if err != nil {
 		t.Fatalf("Failed to delete a file - %v", err)
@@ -148,11 +148,36 @@ func hashOrder(key string, cardinality int) []int {
 	return nums
 }

+var readFileInfoIgnoredErrs = append(objectOpIgnoredErrs,
+	errFileNotFound,
+	errVolumeNotFound,
+	errFileVersionNotFound,
+	io.ErrUnexpectedEOF, // some times we would read without locks, ignore these errors
+	io.EOF,              // some times we would read without locks, ignore these errors
+)
+
+func readFileInfo(ctx context.Context, disk StorageAPI, bucket, object, versionID string, opts ReadOptions) (FileInfo, error) {
+	fi, err := disk.ReadVersion(ctx, bucket, object, versionID, opts)
+
+	if err != nil && !IsErr(err, readFileInfoIgnoredErrs...) {
+		logger.LogOnceIf(ctx, fmt.Errorf("Drive %s, path (%s/%s) returned an error (%w)",
+			disk.String(), bucket, object, err),
+			disk.String())
+	}
+
+	return fi, err
+}
+
 // Reads all `xl.meta` metadata as a FileInfo slice.
 // Returns error slice indicating the failed metadata reads.
-func readAllFileInfo(ctx context.Context, disks []StorageAPI, bucket, object, versionID string, readData bool) ([]FileInfo, []error) {
+func readAllFileInfo(ctx context.Context, disks []StorageAPI, bucket, object, versionID string, readData, healing bool) ([]FileInfo, []error) {
 	metadataArray := make([]FileInfo, len(disks))

+	opts := ReadOptions{
+		ReadData: readData,
+		Healing:  healing,
+	}
+
 	g := errgroup.WithNErrs(len(disks))
 	// Read `xl.meta` in parallel across disks.
 	for index := range disks {
@@ -161,33 +186,12 @@ func readAllFileInfo(ctx context.Context, disks []StorageAPI, bucket, object, ve
 			if disks[index] == nil {
 				return errDiskNotFound
 			}
-			metadataArray[index], err = disks[index].ReadVersion(ctx, bucket, object, versionID, readData)
+			metadataArray[index], err = readFileInfo(ctx, disks[index], bucket, object, versionID, opts)
 			return err
 		}, index)
 	}

-	ignoredErrs := []error{
-		errFileNotFound,
-		errVolumeNotFound,
-		errFileVersionNotFound,
-		io.ErrUnexpectedEOF, // some times we would read without locks, ignore these errors
-		io.EOF,              // some times we would read without locks, ignore these errors
-	}
-	ignoredErrs = append(ignoredErrs, objectOpIgnoredErrs...)
-	errs := g.Wait()
-	for index, err := range errs {
-		if err == nil {
-			continue
-		}
-		if !IsErr(err, ignoredErrs...) {
-			logger.LogOnceIf(ctx, fmt.Errorf("Drive %s, path (%s/%s) returned an error (%w)",
-				disks[index], bucket, object, err),
-				disks[index].String())
-		}
-	}
-
-	// Return all the metadata.
-	return metadataArray, errs
+	return metadataArray, g.Wait()
 }

 // shuffleDisksAndPartsMetadataByIndex this function should be always used by GetObjectNInfo()
@@ -158,6 +158,7 @@ func (fi FileInfo) ToObjectInfo(bucket, object string, versioned bool) ObjectInf
 		ContentEncoding:  fi.Metadata["content-encoding"],
 		NumVersions:      fi.NumVersions,
 		SuccessorModTime: fi.SuccessorModTime,
+		CacheControl:     fi.Metadata["cache-control"],
 	}

 	if exp, ok := fi.Metadata["expires"]; ok {
@@ -72,7 +72,7 @@ func (er erasureObjects) checkUploadIDExists(ctx context.Context, bucket, object

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

 	readQuorum, writeQuorum, err := objectQuorumFromMeta(ctx, partsMetadata, errs, er.defaultParityCount)
 	if err != nil {
@@ -118,7 +118,7 @@ func (er erasureObjects) removePartMeta(bucket, object, uploadID, dataDir string
 		g.Go(func() error {
 			_ = storageDisks[index].Delete(context.TODO(), minioMetaMultipartBucket, curpartPath+".meta", DeleteOptions{
 				Recursive: false,
-				Force:     false,
+				Immediate: false,
 			})

 			return nil
@@ -145,11 +145,11 @@ func (er erasureObjects) removeObjectPart(bucket, object, uploadID, dataDir stri
 			// the object. The presence of parts that don't belong in the object doesn't affect correctness.
 			_ = storageDisks[index].Delete(context.TODO(), minioMetaMultipartBucket, curpartPath, DeleteOptions{
 				Recursive: false,
-				Force:     false,
+				Immediate: false,
 			})
 			_ = storageDisks[index].Delete(context.TODO(), minioMetaMultipartBucket, curpartPath+".meta", DeleteOptions{
 				Recursive: false,
-				Force:     false,
+				Immediate: false,
 			})

 			return nil
@@ -185,7 +185,7 @@ func (er erasureObjects) deleteAll(ctx context.Context, bucket, prefix string) {
 			defer wg.Done()
 			disk.Delete(ctx, bucket, prefix, DeleteOptions{
 				Recursive: true,
-				Force:     false,
+				Immediate: false,
 			})
 		}(disk)
 	}
@@ -200,11 +200,11 @@ func (er erasureObjects) cleanupStaleUploadsOnDisk(ctx context.Context, disk Sto
 	readDirFn(pathJoin(diskPath, minioMetaMultipartBucket), func(shaDir string, typ os.FileMode) error {
 		readDirFn(pathJoin(diskPath, minioMetaMultipartBucket, shaDir), func(uploadIDDir string, typ os.FileMode) error {
 			uploadIDPath := pathJoin(shaDir, uploadIDDir)
-			fi, err := disk.ReadVersion(ctx, minioMetaMultipartBucket, uploadIDPath, "", false)
+			fi, err := disk.ReadVersion(ctx, minioMetaMultipartBucket, uploadIDPath, "", ReadOptions{})
 			if err != nil {
 				return nil
 			}
-			w := xioutil.NewDeadlineWorker(diskMaxTimeout)
+			w := xioutil.NewDeadlineWorker(globalDriveConfig.GetMaxTimeout())
 			return w.Run(func() error {
 				wait := deletedCleanupSleeper.Timer(ctx)
 				if now.Sub(fi.ModTime) > expiry {
@@ -218,7 +218,7 @@ func (er erasureObjects) cleanupStaleUploadsOnDisk(ctx context.Context, disk Sto
 		if err != nil {
 			return nil
 		}
-		w := xioutil.NewDeadlineWorker(diskMaxTimeout)
+		w := xioutil.NewDeadlineWorker(globalDriveConfig.GetMaxTimeout())
 		return w.Run(func() error {
 			wait := deletedCleanupSleeper.Timer(ctx)
 			if now.Sub(vi.Created) > expiry {
@@ -239,7 +239,7 @@ func (er erasureObjects) cleanupStaleUploadsOnDisk(ctx context.Context, disk Sto
 		if err != nil {
 			return nil
 		}
-		w := xioutil.NewDeadlineWorker(diskMaxTimeout)
+		w := xioutil.NewDeadlineWorker(globalDriveConfig.GetMaxTimeout())
 		return w.Run(func() error {
 			wait := deletedCleanupSleeper.Timer(ctx)
 			if now.Sub(vi.Created) > expiry {
@@ -564,7 +564,7 @@ func writeAllDisks(ctx context.Context, disks []StorageAPI, dstBucket, dstEntry
 			}
 			index := index
 			g.Go(func() error {
-				return disks[index].Delete(ctx, dstBucket, dstEntry, DeleteOptions{Force: true})
+				return disks[index].Delete(ctx, dstBucket, dstEntry, DeleteOptions{Immediate: true})
 			}, index)
 		}
 		// Ignore these errors.
@@ -662,12 +662,12 @@ func (er erasureObjects) PutObjectPart(ctx context.Context, bucket, object, uplo
 			// Account for padding and forced compression overhead and encryption.
 			buffer = make([]byte, data.ActualSize()+256+32+32, data.ActualSize()*2+512)
 		} else {
-			buffer = er.bp.Get()
-			defer er.bp.Put(buffer)
+			buffer = globalBytePoolCap.Get()
+			defer globalBytePoolCap.Put(buffer)
 		}
 	case size >= fi.Erasure.BlockSize:
-		buffer = er.bp.Get()
-		defer er.bp.Put(buffer)
+		buffer = globalBytePoolCap.Get()
+		defer globalBytePoolCap.Put(buffer)
 	case size < fi.Erasure.BlockSize:
 		// No need to allocate fully fi.Erasure.BlockSize buffer if the incoming data is smaller.
 		buffer = make([]byte, size, 2*size+int64(fi.Erasure.ParityBlocks+fi.Erasure.DataBlocks-1))
@@ -688,10 +688,10 @@ func (er erasureObjects) PutObjectPart(ctx context.Context, bucket, object, uplo
 	if data.Size() > bigFileThreshold {
 		// Add input readahead.
 		// We use 2 buffers, so we always have a full buffer of input.
-		bufA := er.bp.Get()
-		bufB := er.bp.Get()
-		defer er.bp.Put(bufA)
-		defer er.bp.Put(bufB)
+		bufA := globalBytePoolCap.Get()
+		bufB := globalBytePoolCap.Get()
+		defer globalBytePoolCap.Put(bufA)
+		defer globalBytePoolCap.Put(bufB)
 		ra, err := readahead.NewReaderBuffer(data, [][]byte{bufA[:fi.Erasure.BlockSize], bufB[:fi.Erasure.BlockSize]})
 		if err == nil {
 			toEncode = ra
@@ -1290,13 +1290,15 @@ func (er erasureObjects) CompleteMultipartUpload(ctx context.Context, bucket str
 		})
 	}

-	// Check if there is any offline disk and add it to the MRF list
-	for _, disk := range onlineDisks {
-		if disk != nil && disk.IsOnline() {
-			continue
+	if !opts.Speedtest && !versionsDisparity {
+		// Check if there is any offline disk and add it to the MRF list
+		for _, disk := range onlineDisks {
+			if disk != nil && disk.IsOnline() {
+				continue
+			}
+			er.addPartial(bucket, object, fi.VersionID)
+			break
 		}
-		er.addPartial(bucket, object, fi.VersionID)
-		break
 	}

 	for i := 0; i < len(onlineDisks); i++ {
@@ -50,7 +50,7 @@ import (
 )

 // list all errors which can be ignored in object operations.
-var objectOpIgnoredErrs = append(baseIgnoredErrs, errDiskAccessDenied, errUnformattedDisk)
+var objectOpIgnoredErrs = append(baseIgnoredErrs, errDiskAccessDenied, errUnformattedDisk, errDiskOngoingReq)

 // Object Operations

@@ -91,7 +91,7 @@ func (er erasureObjects) CopyObject(ctx context.Context, srcBucket, srcObject, d

 	// Read metadata associated with the object from all disks.
 	if srcOpts.VersionID != "" {
-		metaArr, errs = readAllFileInfo(ctx, storageDisks, srcBucket, srcObject, srcOpts.VersionID, true)
+		metaArr, errs = readAllFileInfo(ctx, storageDisks, srcBucket, srcObject, srcOpts.VersionID, true, false)
 	} else {
 		metaArr, errs = readAllXL(ctx, storageDisks, srcBucket, srcObject, true, false, true)
 	}
@@ -99,7 +99,7 @@ func (er erasureObjects) CopyObject(ctx context.Context, srcBucket, srcObject, d
 	readQuorum, writeQuorum, err := objectQuorumFromMeta(ctx, metaArr, errs, er.defaultParityCount)
 	if err != nil {
 		if errors.Is(err, errErasureReadQuorum) && !strings.HasPrefix(srcBucket, minioMetaBucket) {
-			_, derr := er.deleteIfDangling(ctx, srcBucket, srcObject, metaArr, errs, nil, srcOpts)
+			_, derr := er.deleteIfDangling(context.Background(), srcBucket, srcObject, metaArr, errs, nil, srcOpts)
 			if derr != nil {
 				err = derr
 			}
@@ -317,7 +317,6 @@ func (er erasureObjects) getObjectWithFileInfo(ctx context.Context, bucket, obje

 	// Reply back invalid range if the input offset and length fall out of range.
 	if startOffset > fi.Size || startOffset+length > fi.Size {
-		logger.LogIf(ctx, InvalidRange{startOffset, length, fi.Size}, logger.Application)
 		return InvalidRange{startOffset, length, fi.Size}
 	}

@@ -414,6 +413,7 @@ func (er erasureObjects) getObjectWithFileInfo(ctx context.Context, bucket, obje
 							queued:    time.Now(),
 							setIndex:  er.setIndex,
 							poolIndex: er.poolIndex,
+							scanMode:  scan,
 						})
 					})
 					// Healing is triggered and we have written
@@ -476,40 +476,50 @@ func auditDanglingObjectDeletion(ctx context.Context, bucket, object, versionID
 	auditLogInternal(ctx, opts)
 }

-func joinErrors(errs ...error) error {
-	s := make([]string, 0, len(errs))
-	nonNilErrs := make([]any, 0, len(errs))
-	for _, err := range errs {
-		if err == nil {
-			continue
+func joinErrs(errs []error) []string {
+	s := make([]string, len(errs))
+	for i := range s {
+		if errs[i] == nil {
+			s[i] = "<nil>"
+		} else {
+			s[i] = errs[i].Error()
 		}
-		s = append(s, "[%w]")
-		nonNilErrs = append(nonNilErrs, err)
 	}
-	// If all the errors were nil, return nil.
-	if len(nonNilErrs) == 0 {
-		return nil
-	}
-	allErrs := strings.Join(s, "\n")
-	return fmt.Errorf(allErrs, nonNilErrs...)
+	return s
 }

 func (er erasureObjects) deleteIfDangling(ctx context.Context, bucket, object string, metaArr []FileInfo, errs []error, dataErrs []error, opts ObjectOptions) (FileInfo, error) {
-	_, file, line, cok := runtime.Caller(1)
 	var err error
 	m, ok := isObjectDangling(metaArr, errs, dataErrs)
 	if ok {
 		tags := make(map[string]interface{}, 4)
 		tags["set"] = er.setIndex
 		tags["pool"] = er.poolIndex
-		tags["merrs"] = joinErrors(errs...)     // errors.Join(errs...)
-		tags["derrs"] = joinErrors(dataErrs...) // errors.Join(dataErrs...)
+		tags["merrs"] = joinErrs(errs)
+		tags["derrs"] = joinErrs(dataErrs)
 		if m.IsValid() {
 			tags["size"] = m.Size
 			tags["mtime"] = m.ModTime.Format(http.TimeFormat)
+			tags["data"] = m.Erasure.DataBlocks
 			tags["parity"] = m.Erasure.ParityBlocks
+		} else {
+			tags["invalid-meta"] = true
+			tags["data"] = er.setDriveCount - er.defaultParityCount
+			tags["parity"] = er.defaultParityCount
 		}

+		// count the number of offline disks
+		offline := 0
+		for i := 0; i < max(len(errs), len(dataErrs)); i++ {
+			if i < len(errs) && errs[i] == errDiskNotFound || i < len(dataErrs) && dataErrs[i] == errDiskNotFound {
+				offline++
+			}
+		}
+		if offline > 0 {
+			tags["offline"] = offline
+		}
+
+		_, file, line, cok := runtime.Caller(1)
 		if cok {
 			tags["caller"] = fmt.Sprintf("%s:%d", file, line)
 		}
@@ -536,7 +546,7 @@ func (er erasureObjects) deleteIfDangling(ctx context.Context, bucket, object st
 				if disks[index] == nil {
 					return errDiskNotFound
 				}
-				return disks[index].DeleteVersion(ctx, bucket, object, fi, false)
+				return disks[index].DeleteVersion(ctx, bucket, object, fi, false, DeleteOptions{})
 			}, index)
 		}

@@ -545,66 +555,111 @@ func (er erasureObjects) deleteIfDangling(ctx context.Context, bucket, object st
 	return m, err
 }

-func readAllXL(ctx context.Context, disks []StorageAPI, bucket, object string, readData, inclFreeVers, allParts bool) ([]FileInfo, []error) {
-	metadataArray := make([]*xlMetaV2, len(disks))
-	metaFileInfos := make([]FileInfo, len(metadataArray))
-	metadataShallowVersions := make([][]xlMetaV2ShallowVersion, len(disks))
-	var v2bufs [][]byte
-	if !readData {
-		v2bufs = make([][]byte, len(disks))
+var readRawFileInfoErrs = append(objectOpIgnoredErrs,
+	errFileNotFound,
+	errFileNameTooLong,
+	errVolumeNotFound,
+	errFileVersionNotFound,
+	io.ErrUnexpectedEOF, // some times we would read without locks, ignore these errors
+	io.EOF,              // some times we would read without locks, ignore these errors
+	msgp.ErrShortBytes,
+	context.DeadlineExceeded,
+	context.Canceled,
+)
+
+func readRawFileInfo(ctx context.Context, disk StorageAPI, bucket, object string, readData bool) (RawFileInfo, error) {
+	rf, err := disk.ReadXL(ctx, bucket, object, readData)
+
+	if err != nil && !IsErr(err, readRawFileInfoErrs...) {
+		logger.LogOnceIf(ctx, fmt.Errorf("Drive %s, path (%s/%s) returned an error (%w)",
+			disk.String(), bucket, object, err),
+			disk.String())
+	}
+	return rf, err
+}
+
+func fileInfoFromRaw(ri RawFileInfo, bucket, object string, readData, inclFreeVers, allParts bool) (FileInfo, error) {
+	var xl xlMetaV2
+	if err := xl.LoadOrConvert(ri.Buf); err != nil {
+		return FileInfo{}, err
 	}

+	fi, err := xl.ToFileInfo(bucket, object, "", inclFreeVers, allParts)
+	if err != nil {
+		return FileInfo{}, err
+	}
+
+	if !fi.IsValid() {
+		return FileInfo{}, errCorruptedFormat
+	}
+
+	versionID := fi.VersionID
+	if versionID == "" {
+		versionID = nullVersionID
+	}
+
+	fileInfo, err := xl.ToFileInfo(bucket, object, versionID, inclFreeVers, allParts)
+	if err != nil {
+		return FileInfo{}, err
+	}
+
+	if readData {
+		fileInfo.Data = xl.data.find(versionID)
+	}
+
+	fileInfo.DiskMTime = ri.DiskMTime
+	return fileInfo, nil
+}
+
+func readAllRawFileInfo(ctx context.Context, disks []StorageAPI, bucket, object string, readData bool) ([]RawFileInfo, []error) {
+	rawFileInfos := make([]RawFileInfo, len(disks))
 	g := errgroup.WithNErrs(len(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
 			}
-			rf, err := disks[index].ReadXL(ctx, bucket, object, readData)
+			rf, err := readRawFileInfo(ctx, disks[index], bucket, object, readData)
 			if err != nil {
 				return err
 			}
-			if !readData {
-				// Save the buffer so we can reuse it.
-				v2bufs[index] = rf.Buf
-			}
-
-			var xl xlMetaV2
-			if err = xl.LoadOrConvert(rf.Buf); err != nil {
-				return err
-			}
-			metadataArray[index] = &xl
-			metaFileInfos[index] = FileInfo{
-				DiskMTime: rf.DiskMTime,
-			}
+			rawFileInfos[index] = rf
 			return nil
 		}, index)
 	}

-	ignoredErrs := []error{
-		errFileNotFound,
-		errFileNameTooLong,
-		errVolumeNotFound,
-		errFileVersionNotFound,
-		io.ErrUnexpectedEOF, // some times we would read without locks, ignore these errors
-		io.EOF,              // some times we would read without locks, ignore these errors
-		msgp.ErrShortBytes,
-		context.DeadlineExceeded,
-		context.Canceled,
-	}
-	ignoredErrs = append(ignoredErrs, objectOpIgnoredErrs...)
+	return rawFileInfos, g.Wait()
+}

-	errs := g.Wait()
-	for index, err := range errs {
-		if err == nil {
+func pickLatestQuorumFilesInfo(ctx context.Context, rawFileInfos []RawFileInfo, errs []error, bucket, object string, readData, inclFreeVers, allParts bool) ([]FileInfo, []error) {
+	metadataArray := make([]*xlMetaV2, len(rawFileInfos))
+	metaFileInfos := make([]FileInfo, len(rawFileInfos))
+	metadataShallowVersions := make([][]xlMetaV2ShallowVersion, len(rawFileInfos))
+	var v2bufs [][]byte
+	if !readData {
+		v2bufs = make([][]byte, len(rawFileInfos))
+	}
+
+	// Read `xl.meta` in parallel across disks.
+	for index := range rawFileInfos {
+		rf := rawFileInfos[index]
+		if rf.Buf == nil {
 			continue
 		}
-		if !IsErr(err, ignoredErrs...) {
-			logger.LogOnceIf(ctx, fmt.Errorf("Drive %s, path (%s/%s) returned an error (%w)",
-				disks[index], bucket, object, err),
-				disks[index].String())
+		if !readData {
+			// Save the buffer so we can reuse it.
+			v2bufs[index] = rf.Buf
+		}
+
+		var xl xlMetaV2
+		if err := xl.LoadOrConvert(rf.Buf); err != nil {
+			errs[index] = err
+			continue
+		}
+		metadataArray[index] = &xl
+		metaFileInfos[index] = FileInfo{
+			DiskMTime: rf.DiskMTime,
 		}
 	}

@@ -614,7 +669,7 @@ func readAllXL(ctx context.Context, disks []StorageAPI, bucket, object string, r
 		}
 	}

-	readQuorum := (len(disks) + 1) / 2
+	readQuorum := (len(rawFileInfos) + 1) / 2
 	meta := &xlMetaV2{versions: mergeXLV2Versions(readQuorum, false, 1, metadataShallowVersions...)}
 	lfi, err := meta.ToFileInfo(bucket, object, "", inclFreeVers, allParts)
 	if err != nil {
@@ -666,22 +721,207 @@ func readAllXL(ctx context.Context, disks []StorageAPI, bucket, object string, r
 	return metaFileInfos, errs
 }

-func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object string, opts ObjectOptions, readData bool) (fi FileInfo, metaArr []FileInfo, onlineDisks []StorageAPI, err error) {
-	disks := er.getDisks()
+// Checking if an object is dangling costs some IOPS; hence implementing this function
+// which decides which condition it is useful to check if an object is dangling
+//
+//	  errs: errors from reading xl.meta in all disks
+//	   err: reduced errs
+//	bucket: the object name in question
+func shouldCheckForDangling(err error, errs []error, bucket string) bool {
+	// Avoid data in .minio.sys for now
+	if bucket == minioMetaBucket {
+		return false
+	}
+	switch {
+	// Check if we have a read quorum issue
+	case errors.Is(err, errErasureReadQuorum):
+		return true
+	// Check if the object is inexistent in most disks but not all of them
+	case errors.Is(err, errFileNotFound) || errors.Is(err, errFileVersionNotFound):
+		for i := range errs {
+			if errs[i] == nil {
+				return true
+			}
+		}
+	}
+	return false
+}

-	var errs []error
+func readAllXL(ctx context.Context, disks []StorageAPI, bucket, object string, readData, inclFreeVers, allParts bool) ([]FileInfo, []error) {
+	rawFileInfos, errs := readAllRawFileInfo(ctx, disks, bucket, object, readData)
+	return pickLatestQuorumFilesInfo(ctx, rawFileInfos, errs, bucket, object, readData, inclFreeVers, allParts)
+}

-	// Read metadata associated with the object from all disks.
-	if opts.VersionID != "" {
-		metaArr, errs = readAllFileInfo(ctx, disks, bucket, object, opts.VersionID, readData)
-	} else {
-		metaArr, errs = readAllXL(ctx, disks, bucket, object, readData, opts.InclFreeVersions, true)
+func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object string, opts ObjectOptions, readData bool) (FileInfo, []FileInfo, []StorageAPI, error) {
+	var mu sync.Mutex
+
+	rawArr := make([]RawFileInfo, er.setDriveCount)
+	metaArr := make([]FileInfo, er.setDriveCount)
+	errs := make([]error, er.setDriveCount)
+	for i := range errs {
+		errs[i] = errDiskOngoingReq
+	}
+
+	done := make(chan bool, er.setDriveCount)
+	disks := er.getDisks()
+
+	mrfCheck := make(chan FileInfo)
+	defer close(mrfCheck)
+
+	ropts := ReadOptions{
+		ReadData: readData,
+		Healing:  false,
+	}
+
+	// Ask for all disks first;
+	go func() {
+		ctx, cancel := context.WithCancel(ctx)
+		defer cancel()
+
+		wg := sync.WaitGroup{}
+		for i, disk := range disks {
+			if disk == nil {
+				done <- false
+				continue
+			}
+			wg.Add(1)
+			go func(i int, disk StorageAPI) {
+				defer wg.Done()
+				var fi FileInfo
+				var err error
+				if opts.VersionID != "" {
+					// Read a specific version ID
+					fi, err = readFileInfo(ctx, disk, bucket, object, opts.VersionID, ropts)
+					mu.Lock()
+					metaArr[i], errs[i] = fi, err
+					mu.Unlock()
+				} else {
+					// Read the latest version
+					var ri RawFileInfo
+					ri, err = readRawFileInfo(ctx, disk, bucket, object, readData)
+					mu.Lock()
+					rawArr[i], errs[i] = ri, err
+					mu.Unlock()
+					if err == nil {
+						fi, err = fileInfoFromRaw(ri, bucket, object, readData, opts.InclFreeVersions, true)
+						mu.Lock()
+						metaArr[i], errs[i] = fi, err
+						mu.Unlock()
+					}
+				}
+				done <- err == nil
+			}(i, disk)
+		}
+
+		wg.Wait()
+		close(done)
+
+		fi, ok := <-mrfCheck
+		if !ok {
+			return
+		}
+		if fi.Deleted {
+			return
+		}
+		// if one of the disk is offline, return right here no need
+		// to attempt a heal on the object.
+		if countErrs(errs, errDiskNotFound) > 0 {
+			return
+		}
+		var missingBlocks int
+		for i := range errs {
+			if errors.Is(errs[i], errFileNotFound) {
+				missingBlocks++
+			}
+		}
+		// if missing metadata can be reconstructed, attempt to reconstruct.
+		// additionally do not heal delete markers inline, let them be
+		// healed upon regular heal process.
+		if missingBlocks > 0 && missingBlocks < fi.Erasure.DataBlocks {
+			globalMRFState.addPartialOp(partialOperation{
+				bucket:    fi.Volume,
+				object:    fi.Name,
+				versionID: fi.VersionID,
+				queued:    time.Now(),
+				setIndex:  er.setIndex,
+				poolIndex: er.poolIndex,
+			})
+		}
+
+		return
+	}()
+
+	validResp := 0
+	totalResp := 0
+
+	// minDisks value is only to reduce the number of calls
+	// to the disks; this value is not accurate because we do
+	// not know the storage class of the object yet
+	minDisks := 0
+	if p := globalStorageClass.GetParityForSC(""); p > -1 {
+		minDisks = er.setDriveCount - p
+	} else {
+		minDisks = er.setDriveCount - er.defaultParityCount
+	}
+
+	calcQuorum := func() (FileInfo, []FileInfo, []StorageAPI, time.Time, string, error) {
+		readQuorum, _, err := objectQuorumFromMeta(ctx, metaArr, errs, er.defaultParityCount)
+		if err != nil {
+			return FileInfo{}, nil, nil, time.Time{}, "", err
+		}
+		err = reduceReadQuorumErrs(ctx, errs, objectOpIgnoredErrs, readQuorum)
+		if err != nil {
+			return FileInfo{}, nil, nil, time.Time{}, "", err
+		}
+		onlineDisks, modTime, etag := listOnlineDisks(disks, metaArr, errs, readQuorum)
+		fi, err := pickValidFileInfo(ctx, metaArr, modTime, etag, readQuorum)
+		if err != nil {
+			return FileInfo{}, nil, nil, time.Time{}, "", err
+		}
+
+		onlineMeta := make([]FileInfo, len(metaArr))
+		copy(onlineMeta, metaArr)
+
+		return fi, onlineMeta, onlineDisks, modTime, etag, nil
+	}
+
+	var (
+		modTime     time.Time
+		etag        string
+		fi          FileInfo
+		onlineMeta  []FileInfo
+		onlineDisks []StorageAPI
+		err         error
+	)
+
+	for success := range done {
+		totalResp++
+		if success {
+			validResp++
+		}
+		if totalResp < er.setDriveCount {
+			if !opts.FastGetObjInfo {
+				continue
+			}
+			if validResp < minDisks {
+				continue
+			}
+		}
+		if opts.VersionID == "" && totalResp == er.setDriveCount {
+			// Disks cannot agree about the latest version, pass this to a more advanced code
+			metaArr, errs = pickLatestQuorumFilesInfo(ctx, rawArr, errs, bucket, object, readData, opts.InclFreeVersions, true)
+		}
+		mu.Lock()
+		fi, onlineMeta, onlineDisks, modTime, etag, err = calcQuorum()
+		mu.Unlock()
+		if err == nil && fi.InlineData() {
+			break
+		}
 	}

-	readQuorum, _, err := objectQuorumFromMeta(ctx, metaArr, errs, er.defaultParityCount)
 	if err != nil {
-		if errors.Is(err, errErasureReadQuorum) && !strings.HasPrefix(bucket, minioMetaBucket) {
-			_, derr := er.deleteIfDangling(ctx, bucket, object, metaArr, errs, nil, opts)
+		if shouldCheckForDangling(err, errs, bucket) {
+			_, derr := er.deleteIfDangling(context.Background(), bucket, object, metaArr, errs, nil, opts)
 			if derr != nil {
 				err = derr
 			}
@@ -689,25 +929,6 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 		return fi, nil, nil, toObjectErr(err, bucket, object)
 	}

-	if reducedErr := reduceReadQuorumErrs(ctx, errs, objectOpIgnoredErrs, readQuorum); reducedErr != nil {
-		if errors.Is(reducedErr, errErasureReadQuorum) && !strings.HasPrefix(bucket, minioMetaBucket) {
-			_, derr := er.deleteIfDangling(ctx, bucket, object, metaArr, errs, nil, opts)
-			if derr != nil {
-				reducedErr = derr
-			}
-		}
-		return fi, nil, nil, toObjectErr(reducedErr, bucket, object)
-	}
-
-	// List all online disks.
-	onlineDisks, modTime, etag := listOnlineDisks(disks, metaArr, errs, readQuorum)
-
-	// Pick latest valid metadata.
-	fi, err = pickValidFileInfo(ctx, metaArr, modTime, etag, readQuorum)
-	if err != nil {
-		return fi, nil, nil, err
-	}
-
 	if !fi.Deleted && len(fi.Erasure.Distribution) != len(onlineDisks) {
 		err := fmt.Errorf("unexpected file distribution (%v) from online disks (%v), looks like backend disks have been manually modified refusing to heal %s/%s(%s)",
 			fi.Erasure.Distribution, onlineDisks, bucket, object, opts.VersionID)
@@ -715,26 +936,13 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 		return fi, nil, nil, toObjectErr(err, bucket, object, opts.VersionID)
 	}

-	filterOnlineDisksInplace(fi, metaArr, onlineDisks)
-
-	// if one of the disk is offline, return right here no need
-	// to attempt a heal on the object.
-	if countErrs(errs, errDiskNotFound) > 0 {
-		return fi, metaArr, onlineDisks, nil
-	}
-
-	var missingBlocks int
-	for i, err := range errs {
-		if err != nil && errors.Is(err, errFileNotFound) {
-			missingBlocks++
-			continue
-		}
-
+	filterOnlineDisksInplace(fi, onlineMeta, onlineDisks)
+	for i := range onlineMeta {
 		// verify metadata is valid, it has similar erasure info
 		// as well as common modtime, if modtime is not possible
 		// verify if it has common "etag" atleast.
-		if metaArr[i].IsValid() && metaArr[i].Erasure.Equal(fi.Erasure) {
-			ok := metaArr[i].ModTime.Equal(modTime)
+		if onlineMeta[i].IsValid() && onlineMeta[i].Erasure.Equal(fi.Erasure) {
+			ok := onlineMeta[i].ModTime.Equal(modTime)
 			if modTime.IsZero() || modTime.Equal(timeSentinel) {
 				ok = etag != "" && etag == fi.Metadata["etag"]
 			}
@@ -743,26 +951,13 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 			}
 		} // in all other cases metadata is corrupt, do not read from it.

-		metaArr[i] = FileInfo{}
+		onlineMeta[i] = FileInfo{}
 		onlineDisks[i] = nil
-		missingBlocks++
 	}

-	// if missing metadata can be reconstructed, attempt to reconstruct.
-	// additionally do not heal delete markers inline, let them be
-	// healed upon regular heal process.
-	if !fi.Deleted && missingBlocks > 0 && missingBlocks < readQuorum {
-		globalMRFState.addPartialOp(partialOperation{
-			bucket:    bucket,
-			object:    object,
-			versionID: fi.VersionID,
-			queued:    time.Now(),
-			setIndex:  er.setIndex,
-			poolIndex: er.poolIndex,
-		})
-	}
+	mrfCheck <- fi.ShallowCopy()

-	return fi, metaArr, onlineDisks, nil
+	return fi, onlineMeta, onlineDisks, nil
 }

 // getObjectInfo - wrapper for reading object metadata and constructs ObjectInfo.
@@ -837,7 +1032,7 @@ func renameData(ctx context.Context, disks []StorageAPI, srcBucket, srcEntry str
 			if !fi.IsValid() {
 				return errFileCorrupt
 			}
-			sign, err := disks[index].RenameData(ctx, srcBucket, srcEntry, fi, dstBucket, dstEntry)
+			sign, err := disks[index].RenameData(ctx, srcBucket, srcEntry, fi, dstBucket, dstEntry, RenameOptions{})
 			if err != nil {
 				return err
 			}
@@ -852,6 +1047,23 @@ func renameData(ctx context.Context, disks []StorageAPI, srcBucket, srcEntry str
 	var versionsDisparity bool

 	err := reduceWriteQuorumErrs(ctx, errs, objectOpIgnoredErrs, writeQuorum)
+	if err != nil {
+		dg := errgroup.WithNErrs(len(disks))
+		for index, nerr := range errs {
+			if nerr != nil {
+				continue
+			}
+			index := index
+			// When we are going to return error, attempt to delete success
+			// on some of the drives, if we cannot we do not have to notify
+			// caller this dangling object will be now scheduled to be removed
+			// via active healing.
+			dg.Go(func() error {
+				return disks[index].DeleteVersion(context.Background(), dstBucket, dstEntry, metadata[index], false, DeleteOptions{UndoWrite: true})
+			}, index)
+		}
+		dg.Wait()
+	}
 	if err == nil {
 		versions := reduceCommonVersions(diskVersions, writeQuorum)
 		for index, dversions := range diskVersions {
@@ -925,8 +1137,8 @@ func (er erasureObjects) putMetacacheObject(ctx context.Context, key string, r *
 	case size == 0:
 		buffer = make([]byte, 1) // Allocate atleast a byte to reach EOF
 	case size >= fi.Erasure.BlockSize:
-		buffer = er.bp.Get()
-		defer er.bp.Put(buffer)
+		buffer = globalBytePoolCap.Get()
+		defer globalBytePoolCap.Put(buffer)
 	case size < fi.Erasure.BlockSize:
 		// No need to allocate fully blockSizeV1 buffer if the incoming data is smaller.
 		buffer = make([]byte, size, 2*size+int64(fi.Erasure.ParityBlocks+fi.Erasure.DataBlocks-1))
@@ -1016,7 +1228,7 @@ func (er erasureObjects) PutObject(ctx context.Context, bucket string, object st
 }

 // Heal up to two versions of one object when there is disparity between disks
-func healObjectVersionsDisparity(bucket string, entry metaCacheEntry) error {
+func healObjectVersionsDisparity(bucket string, entry metaCacheEntry, scanMode madmin.HealScanMode) error {
 	if entry.isDir() {
 		return nil
 	}
@@ -1040,13 +1252,13 @@ func healObjectVersionsDisparity(bucket string, entry metaCacheEntry) error {

 	fivs, err := entry.fileInfoVersions(bucket)
 	if err != nil {
-		go healObject(bucket, entry.name, "", madmin.HealDeepScan)
+		healObject(bucket, entry.name, "", madmin.HealDeepScan)
 		return err
 	}

 	if len(fivs.Versions) <= 2 {
 		for _, version := range fivs.Versions {
-			go healObject(bucket, entry.name, version.VersionID, madmin.HealNormalScan)
+			healObject(bucket, entry.name, version.VersionID, scanMode)
 		}
 	}

@@ -1176,8 +1388,8 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 	case size == 0:
 		buffer = make([]byte, 1) // Allocate atleast a byte to reach EOF
 	case size >= fi.Erasure.BlockSize || size == -1:
-		buffer = er.bp.Get()
-		defer er.bp.Put(buffer)
+		buffer = globalBytePoolCap.Get()
+		defer globalBytePoolCap.Put(buffer)
 	case size < fi.Erasure.BlockSize:
 		// No need to allocate fully blockSizeV1 buffer if the incoming data is smaller.
 		buffer = make([]byte, size, 2*size+int64(fi.Erasure.ParityBlocks+fi.Erasure.DataBlocks-1))
@@ -1236,10 +1448,10 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 	toEncode := io.Reader(data)
 	if data.Size() > bigFileThreshold {
 		// We use 2 buffers, so we always have a full buffer of input.
-		bufA := er.bp.Get()
-		bufB := er.bp.Get()
-		defer er.bp.Put(bufA)
-		defer er.bp.Put(bufB)
+		bufA := globalBytePoolCap.Get()
+		bufB := globalBytePoolCap.Get()
+		defer globalBytePoolCap.Put(bufA)
+		defer globalBytePoolCap.Put(bufB)
 		ra, err := readahead.NewReaderBuffer(data, [][]byte{bufA[:fi.Erasure.BlockSize], bufB[:fi.Erasure.BlockSize]})
 		if err == nil {
 			toEncode = ra
@@ -1347,18 +1559,21 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st

 	// For speedtest objects do not attempt to heal them.
 	if !opts.Speedtest {
-		// Whether a disk was initially or becomes offline
-		// during this upload, send it to the MRF list.
-		for i := 0; i < len(onlineDisks); i++ {
-			if onlineDisks[i] != nil && onlineDisks[i].IsOnline() {
-				continue
+		// When there is versions disparity we are healing
+		// the content implicitly for all versions, we can
+		// avoid triggering another MRF heal for offline drives.
+		if !versionsDisparity {
+			// Whether a disk was initially or becomes offline
+			// during this upload, send it to the MRF list.
+			for i := 0; i < len(onlineDisks); i++ {
+				if onlineDisks[i] != nil && onlineDisks[i].IsOnline() {
+					continue
+				}
+
+				er.addPartial(bucket, object, fi.VersionID)
+				break
 			}
-
-			er.addPartial(bucket, object, fi.VersionID)
-			break
-		}
-
-		if versionsDisparity {
+		} else {
 			globalMRFState.addPartialOp(partialOperation{
 				bucket:      bucket,
 				object:      object,
@@ -1395,7 +1610,7 @@ func (er erasureObjects) deleteObjectVersion(ctx context.Context, bucket, object
 			if disks[index] == nil {
 				return errDiskNotFound
 			}
-			return disks[index].DeleteVersion(ctx, bucket, object, fi, forceDelMarker)
+			return disks[index].DeleteVersion(ctx, bucket, object, fi, forceDelMarker, DeleteOptions{})
 		}, index)
 	}
 	// return errors if any during deletion
@@ -1508,7 +1723,7 @@ func (er erasureObjects) DeleteObjects(ctx context.Context, bucket string, objec
 				}
 				return
 			}
-			errs := disk.DeleteVersions(ctx, bucket, dedupVersions)
+			errs := disk.DeleteVersions(ctx, bucket, dedupVersions, DeleteOptions{})
 			for i, err := range errs {
 				if err == nil {
 					continue
@@ -1588,11 +1803,11 @@ func (er erasureObjects) deletePrefix(ctx context.Context, bucket, prefix string
 			// - The prefix__XLDIR__
 			defer disks[index].Delete(ctx, bucket, dirPrefix, DeleteOptions{
 				Recursive: true,
-				Force:     true,
+				Immediate: true,
 			})
 			return disks[index].Delete(ctx, bucket, prefix, DeleteOptions{
 				Recursive: true,
-				Force:     true,
+				Immediate: true,
 			})
 		}, index)
 	}
@@ -1611,6 +1826,9 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 	auditObjectErasureSet(ctx, object, &er)

 	if opts.DeletePrefix {
+		if globalCacheConfig.Enabled() {
+			return ObjectInfo{}, toObjectErr(errMethodNotAllowed, bucket, object)
+		}
 		return ObjectInfo{}, toObjectErr(er.deletePrefix(ctx, bucket, object), bucket, object)
 	}

@@ -1833,7 +2051,7 @@ func (er erasureObjects) PutObjectMetadata(ctx context.Context, bucket, object s

 	// Read metadata associated with the object from all disks.
 	if opts.VersionID != "" {
-		metaArr, errs = readAllFileInfo(ctx, disks, bucket, object, opts.VersionID, false)
+		metaArr, errs = readAllFileInfo(ctx, disks, bucket, object, opts.VersionID, false, false)
 	} else {
 		metaArr, errs = readAllXL(ctx, disks, bucket, object, false, false, true)
 	}
@@ -1841,7 +2059,7 @@ func (er erasureObjects) PutObjectMetadata(ctx context.Context, bucket, object s
 	readQuorum, _, err := objectQuorumFromMeta(ctx, metaArr, errs, er.defaultParityCount)
 	if err != nil {
 		if errors.Is(err, errErasureReadQuorum) && !strings.HasPrefix(bucket, minioMetaBucket) {
-			_, derr := er.deleteIfDangling(ctx, bucket, object, metaArr, errs, nil, opts)
+			_, derr := er.deleteIfDangling(context.Background(), bucket, object, metaArr, errs, nil, opts)
 			if derr != nil {
 				err = derr
 			}
@@ -1906,7 +2124,7 @@ func (er erasureObjects) PutObjectTags(ctx context.Context, bucket, object strin

 	// Read metadata associated with the object from all disks.
 	if opts.VersionID != "" {
-		metaArr, errs = readAllFileInfo(ctx, disks, bucket, object, opts.VersionID, false)
+		metaArr, errs = readAllFileInfo(ctx, disks, bucket, object, opts.VersionID, false, false)
 	} else {
 		metaArr, errs = readAllXL(ctx, disks, bucket, object, false, false, true)
 	}
@@ -1914,7 +2132,7 @@ func (er erasureObjects) PutObjectTags(ctx context.Context, bucket, object strin
 	readQuorum, _, err := objectQuorumFromMeta(ctx, metaArr, errs, er.defaultParityCount)
 	if err != nil {
 		if errors.Is(err, errErasureReadQuorum) && !strings.HasPrefix(bucket, minioMetaBucket) {
-			_, derr := er.deleteIfDangling(ctx, bucket, object, metaArr, errs, nil, opts)
+			_, derr := er.deleteIfDangling(context.Background(), bucket, object, metaArr, errs, nil, opts)
 			if derr != nil {
 				err = derr
 			}
@@ -2062,7 +2280,6 @@ func (er erasureObjects) TransitionObject(ctx context.Context, bucket, object st
 	rv, err = tgtClient.Put(ctx, destObj, pr, fi.Size)
 	pr.CloseWithError(err)
 	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Unable to transition %s/%s(%s) to %s tier: %w", bucket, object, opts.VersionID, opts.Transition.Tier, err))
 		return err
 	}
 	fi.TransitionStatus = lifecycle.TransitionComplete
@@ -526,7 +526,7 @@ func TestGetObjectNoQuorum(t *testing.T) {
 			if file != "xl.meta" {
 				disk.Delete(ctx, bucket, pathJoin(object, file), DeleteOptions{
 					Recursive: true,
-					Force:     false,
+					Immediate: false,
 				})
 			}
 		}
@@ -630,7 +630,7 @@ func TestHeadObjectNoQuorum(t *testing.T) {
 			if file != "xl.meta" {
 				disk.Delete(ctx, bucket, pathJoin(object, file), DeleteOptions{
 					Recursive: true,
-					Force:     false,
+					Immediate: false,
 				})
 			}
 		}
@@ -908,7 +908,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}

-	parts1, errs1 := readAllFileInfo(ctx, erasureDisks, bucket, object1, "", false)
+	parts1, errs1 := readAllFileInfo(ctx, erasureDisks, bucket, object1, "", false, false)
 	parts1SC := globalStorageClass

 	// Object for test case 2 - No StorageClass defined, MetaData in PutObject requesting RRS Class
@@ -920,7 +920,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}

-	parts2, errs2 := readAllFileInfo(ctx, erasureDisks, bucket, object2, "", false)
+	parts2, errs2 := readAllFileInfo(ctx, erasureDisks, bucket, object2, "", false, false)
 	parts2SC := globalStorageClass

 	// Object for test case 3 - No StorageClass defined, MetaData in PutObject requesting Standard Storage Class
@@ -932,7 +932,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}

-	parts3, errs3 := readAllFileInfo(ctx, erasureDisks, bucket, object3, "", false)
+	parts3, errs3 := readAllFileInfo(ctx, erasureDisks, bucket, object3, "", false, false)
 	parts3SC := globalStorageClass

 	// Object for test case 4 - Standard StorageClass defined as Parity 6, MetaData in PutObject requesting Standard Storage Class
@@ -950,7 +950,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}

-	parts4, errs4 := readAllFileInfo(ctx, erasureDisks, bucket, object4, "", false)
+	parts4, errs4 := readAllFileInfo(ctx, erasureDisks, bucket, object4, "", false, false)
 	parts4SC := storageclass.Config{
 		Standard: storageclass.StorageClass{
 			Parity: 6,
@@ -973,7 +973,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}

-	parts5, errs5 := readAllFileInfo(ctx, erasureDisks, bucket, object5, "", false)
+	parts5, errs5 := readAllFileInfo(ctx, erasureDisks, bucket, object5, "", false, false)
 	parts5SC := globalStorageClass

 	// Object for test case 6 - RRS StorageClass defined as Parity 2, MetaData in PutObject requesting Standard Storage Class
@@ -994,7 +994,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}

-	parts6, errs6 := readAllFileInfo(ctx, erasureDisks, bucket, object6, "", false)
+	parts6, errs6 := readAllFileInfo(ctx, erasureDisks, bucket, object6, "", false, false)
 	parts6SC := storageclass.Config{
 		RRS: storageclass.StorageClass{
 			Parity: 2,
@@ -1017,7 +1017,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}

-	parts7, errs7 := readAllFileInfo(ctx, erasureDisks, bucket, object7, "", false)
+	parts7, errs7 := readAllFileInfo(ctx, erasureDisks, bucket, object7, "", false, false)
 	parts7SC := storageclass.Config{
 		Standard: storageclass.StorageClass{
 			Parity: 5,
@@ -701,10 +701,12 @@ func (set *erasureObjects) listObjectsToDecommission(ctx context.Context, bi dec
 		return fmt.Errorf("no online drives found for set with endpoints %s", set.getEndpoints())
 	}

+	listQuorum := (len(disks) + 1) / 2
+
 	// How to resolve partial results.
 	resolver := metadataResolutionParams{
-		dirQuorum: len(disks) / 2, // make sure to capture all quorum ratios
-		objQuorum: len(disks) / 2, // make sure to capture all quorum ratios
+		dirQuorum: listQuorum, // make sure to capture all quorum ratios
+		objQuorum: listQuorum, // make sure to capture all quorum ratios
 		bucket:    bi.Name,
 	}

@@ -714,7 +716,7 @@ func (set *erasureObjects) listObjectsToDecommission(ctx context.Context, bi dec
 		path:           bi.Prefix,
 		recursive:      true,
 		forwardTo:      "",
-		minDisks:       len(disks) / 2, // to capture all quorum ratios
+		minDisks:       listQuorum,
 		reportNotFound: false,
 		agreed:         fn,
 		partial: func(entries metaCacheEntries, _ []error) {
@@ -767,7 +769,7 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool

 			evt := evalActionFromLifecycle(ctx, *lc, lr, rcfg, objInfo)
 			switch {
-			case evt.Action.DeleteRestored(): // if restored copy has expired,delete it synchronously
+			case evt.Action.DeleteRestored(): // if restored copy has expired, delete it synchronously
 				applyExpiryOnTransitionedObject(ctx, z, objInfo, evt, lcEventSrc_Decom)
 				return false
 			case evt.Action.Delete():
@@ -129,7 +129,7 @@ func (z *erasureServerPools) initRebalanceMeta(ctx context.Context, buckets []st
 	}

 	// Fetch disk capacity and available space.
-	si := z.StorageInfo(ctx)
+	si := z.StorageInfo(ctx, true)
 	diskStats := make([]struct {
 		AvailableSpace uint64
 		TotalSpace     uint64
@@ -624,10 +624,12 @@ func (z *erasureServerPools) rebalanceBucket(ctx context.Context, bucket string,
 		go func() {
 			defer wg.Done()

+			listQuorum := (len(disks) + 1) / 2
+
 			// How to resolve partial results.
 			resolver := metadataResolutionParams{
-				dirQuorum: len(disks) / 2, // make sure to capture all quorum ratios
-				objQuorum: len(disks) / 2, // make sure to capture all quorum ratios
+				dirQuorum: listQuorum, // make sure to capture all quorum ratios
+				objQuorum: listQuorum, // make sure to capture all quorum ratios
 				bucket:    bucket,
 			}
 			err := listPathRaw(ctx, listPathRawOptions{
@@ -635,7 +637,7 @@ func (z *erasureServerPools) rebalanceBucket(ctx context.Context, bucket string,
 				bucket:         bucket,
 				recursive:      true,
 				forwardTo:      "",
-				minDisks:       len(disks) / 2, // to capture all quorum ratios
+				minDisks:       listQuorum,
 				reportNotFound: false,
 				agreed: func(entry metaCacheEntry) {
 					workers <- struct{}{}
@@ -38,6 +38,7 @@ import (
 	"github.com/minio/minio-go/v7/pkg/s3utils"
 	"github.com/minio/minio-go/v7/pkg/set"
 	"github.com/minio/minio-go/v7/pkg/tags"
+	"github.com/minio/minio/internal/bpool"
 	"github.com/minio/minio/internal/config/storageclass"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/sync/errgroup"
@@ -82,6 +83,21 @@ func newErasureServerPools(ctx context.Context, endpointServerPools EndpointServ
 		}
 	)

+	// Maximum number of reusable buffers per node at any given point in time.
+	n := 1024 // single node single/multiple drives set this to 1024 entries
+
+	if globalIsDistErasure {
+		n = 2048
+	}
+
+	// Initialize byte pool once for all sets, bpool size is set to
+	// setCount * setDriveCount with each memory upto blockSizeV2.
+	globalBytePoolCap = bpool.NewBytePoolCap(n, blockSizeV2, blockSizeV2*2)
+
+	if globalServerCtxt.PreAllocate {
+		globalBytePoolCap.Populate()
+	}
+
 	var localDrives []StorageAPI
 	local := endpointServerPools.FirstLocal()
 	for i, ep := range endpointServerPools {
@@ -107,12 +123,6 @@ func newErasureServerPools(ctx context.Context, endpointServerPools EndpointServ
 			return nil, err
 		}

-		for _, storageDisk := range storageDisks[i] {
-			if storageDisk != nil && storageDisk.IsLocal() {
-				localDrives = append(localDrives, storageDisk)
-			}
-		}
-
 		if deploymentID == "" {
 			// all pools should have same deployment ID
 			deploymentID = formats[i].ID
@@ -139,6 +149,18 @@ func newErasureServerPools(ctx context.Context, endpointServerPools EndpointServ
 		if distributionAlgo != "" && z.distributionAlgo == "" {
 			z.distributionAlgo = distributionAlgo
 		}
+
+		for _, storageDisk := range storageDisks[i] {
+			if storageDisk != nil && storageDisk.IsLocal() {
+				localDrives = append(localDrives, storageDisk)
+			}
+		}
+	}
+
+	if !globalIsDistErasure {
+		globalLocalDrivesMu.Lock()
+		globalLocalDrives = localDrives
+		globalLocalDrivesMu.Unlock()
 	}

 	z.decommissionCancelers = make([]context.CancelFunc, len(z.serverPools))
@@ -167,10 +189,6 @@ func newErasureServerPools(ctx context.Context, endpointServerPools EndpointServ
 		break
 	}

-	globalLocalDrivesMu.Lock()
-	globalLocalDrives = localDrives
-	defer globalLocalDrivesMu.Unlock()
-
 	return z, nil
 }

@@ -595,7 +613,7 @@ func (z *erasureServerPools) BackendInfo() (b madmin.BackendInfo) {
 	return
 }

-func (z *erasureServerPools) LocalStorageInfo(ctx context.Context) StorageInfo {
+func (z *erasureServerPools) LocalStorageInfo(ctx context.Context, metrics bool) StorageInfo {
 	var storageInfo StorageInfo

 	storageInfos := make([]StorageInfo, len(z.serverPools))
@@ -603,7 +621,7 @@ func (z *erasureServerPools) LocalStorageInfo(ctx context.Context) StorageInfo {
 	for index := range z.serverPools {
 		index := index
 		g.Go(func() error {
-			storageInfos[index] = z.serverPools[index].LocalStorageInfo(ctx)
+			storageInfos[index] = z.serverPools[index].LocalStorageInfo(ctx, metrics)
 			return nil
 		}, index)
 	}
@@ -619,8 +637,8 @@ func (z *erasureServerPools) LocalStorageInfo(ctx context.Context) StorageInfo {
 	return storageInfo
 }

-func (z *erasureServerPools) StorageInfo(ctx context.Context) StorageInfo {
-	return globalNotificationSys.StorageInfo(z)
+func (z *erasureServerPools) StorageInfo(ctx context.Context, metrics bool) StorageInfo {
+	return globalNotificationSys.StorageInfo(z, metrics)
 }

 func (z *erasureServerPools) NSScanner(ctx context.Context, updates chan<- DataUsageInfo, wantCycle uint32, healScanMode madmin.HealScanMode) error {
@@ -970,7 +988,7 @@ func (z *erasureServerPools) GetObjectInfo(ctx context.Context, bucket, object s
 // PutObject - writes an object to least used erasure pool.
 func (z *erasureServerPools) PutObject(ctx context.Context, bucket string, object string, data *PutObjReader, opts ObjectOptions) (ObjectInfo, error) {
 	// Validate put object input args.
-	if err := checkPutObjectArgs(ctx, bucket, object, z); err != nil {
+	if err := checkPutObjectArgs(ctx, bucket, object); err != nil {
 		return ObjectInfo{}, err
 	}

@@ -1945,7 +1963,7 @@ func (z *erasureServerPools) HealBucket(ctx context.Context, bucket string, opts
 // to allocate a receive channel for ObjectInfo, upon any unhandled
 // error walker returns error. Optionally if context.Done() is received
 // then Walk() stops the walker.
-func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, results chan<- ObjectInfo, opts ObjectOptions) error {
+func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, results chan<- ObjectInfo, opts WalkOptions) error {
 	if err := checkListObjsArgs(ctx, bucket, prefix, "", z); err != nil {
 		// Upon error close the channel.
 		close(results)
@@ -1967,7 +1985,7 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 				go func() {
 					defer wg.Done()

-					disks, _ := set.getOnlineDisksWithHealing()
+					disks, infos, _ := set.getOnlineDisksWithHealingAndInfo(true)
 					if len(disks) == 0 {
 						cancel()
 						return
@@ -1982,21 +2000,37 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 						}
 					}

-					askDisks := getListQuorum(opts.WalkAskDisks, set.setDriveCount)
-					var fallbackDisks []StorageAPI
+					askDisks := getListQuorum(opts.AskDisks, set.setDriveCount)
+					if askDisks == -1 {
+						newDisks := getQuorumDisks(disks, infos, (len(disks)+1)/2)
+						if newDisks != nil {
+							// If we found disks signature in quorum, we proceed to list
+							// from a single drive, shuffling of the drives is subsequently.
+							disks = newDisks
+							askDisks = 1
+						} else {
+							// If we did not find suitable disks, perform strict quorum listing
+							// as no disk agrees on quorum anymore.
+							askDisks = getListQuorum("strict", set.setDriveCount)
+						}
+					}

 					// Special case: ask all disks if the drive count is 4
 					if set.setDriveCount == 4 || askDisks > len(disks) {
 						askDisks = len(disks) // use all available drives
 					}

+					var fallbackDisks []StorageAPI
 					if askDisks > 0 && len(disks) > askDisks {
+						rand.Shuffle(len(disks), func(i, j int) {
+							disks[i], disks[j] = disks[j], disks[i]
+						})
 						fallbackDisks = disks[askDisks:]
 						disks = disks[:askDisks]
 					}

 					requestedVersions := 0
-					if opts.WalkLatestOnly {
+					if opts.LatestOnly {
 						requestedVersions = 1
 					}
 					loadEntry := func(entry metaCacheEntry) {
@@ -2004,14 +2038,14 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 							return
 						}

-						if opts.WalkLatestOnly {
+						if opts.LatestOnly {
 							fi, err := entry.fileInfo(bucket)
 							if err != nil {
 								cancel()
 								return
 							}
-							if opts.WalkFilter != nil {
-								if opts.WalkFilter(fi) {
+							if opts.Filter != nil {
+								if opts.Filter(fi) {
 									if !send(fi.ToObjectInfo(bucket, fi.Name, vcfg != nil && vcfg.Versioned(fi.Name))) {
 										return
 									}
@@ -2029,11 +2063,14 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 								return
 							}

-							versionsSorter(fivs.Versions).reverse()
+							// Note: entry.fileInfoVersions returns versions sorted in reverse chronological order based on ModTime
+							if opts.VersionsSort == WalkVersionsSortAsc {
+								versionsSorter(fivs.Versions).reverse()
+							}

 							for _, version := range fivs.Versions {
-								if opts.WalkFilter != nil {
-									if opts.WalkFilter(version) {
+								if opts.Filter != nil {
+									if opts.Filter(version) {
 										if !send(version.ToObjectInfo(bucket, version.Name, vcfg != nil && vcfg.Versioned(version.Name))) {
 											return
 										}
@@ -2047,10 +2084,13 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 						}
 					}

+					// However many we ask, versions must exist on ~50%
+					listingQuorum := (askDisks + 1) / 2
+
 					// How to resolve partial results.
 					resolver := metadataResolutionParams{
-						dirQuorum:         1,
-						objQuorum:         1,
+						dirQuorum:         listingQuorum,
+						objQuorum:         listingQuorum,
 						bucket:            bucket,
 						requestedVersions: requestedVersions,
 					}
@@ -2068,19 +2108,15 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 						path:           path,
 						filterPrefix:   filterPrefix,
 						recursive:      true,
-						forwardTo:      opts.WalkMarker,
+						forwardTo:      opts.Marker,
 						minDisks:       1,
 						reportNotFound: false,
 						agreed:         loadEntry,
 						partial: func(entries metaCacheEntries, _ []error) {
 							entry, ok := entries.resolve(&resolver)
-							if !ok {
-								// check if we can get one entry atleast
-								// proceed to heal nonetheless.
-								entry, _ = entries.firstFound()
+							if ok {
+								loadEntry(*entry)
 							}
-
-							loadEntry(*entry)
 						},
 						finished: nil,
 					}
@@ -2100,10 +2136,10 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 }

 // HealObjectFn closure function heals the object.
-type HealObjectFn func(bucket, object, versionID string) error
+type HealObjectFn func(bucket, object, versionID string, scanMode madmin.HealScanMode) error

 func (z *erasureServerPools) HealObjects(ctx context.Context, bucket, prefix string, opts madmin.HealOpts, healObjectFn HealObjectFn) error {
-	healEntry := func(bucket string, entry metaCacheEntry) error {
+	healEntry := func(bucket string, entry metaCacheEntry, scanMode madmin.HealScanMode) error {
 		if entry.isDir() {
 			return nil
 		}
@@ -2126,7 +2162,7 @@ func (z *erasureServerPools) HealObjects(ctx context.Context, bucket, prefix str
 		}
 		fivs, err := entry.fileInfoVersions(bucket)
 		if err != nil {
-			return healObjectFn(bucket, entry.name, "")
+			return healObjectFn(bucket, entry.name, "", scanMode)
 		}
 		if opts.Remove && !opts.DryRun {
 			err := z.CheckAbandonedParts(ctx, bucket, entry.name, opts)
@@ -2135,7 +2171,7 @@ func (z *erasureServerPools) HealObjects(ctx context.Context, bucket, prefix str
 			}
 		}
 		for _, version := range fivs.Versions {
-			err := healObjectFn(bucket, version.Name, version.VersionID)
+			err := healObjectFn(bucket, version.Name, version.VersionID, scanMode)
 			if err != nil && !isErrObjectNotFound(err) && !isErrVersionNotFound(err) {
 				return err
 			}
@@ -2159,7 +2195,7 @@ func (z *erasureServerPools) HealObjects(ctx context.Context, bucket, prefix str
 			go func(idx int, set *erasureObjects) {
 				defer wg.Done()

-				errs[idx] = set.listAndHeal(bucket, prefix, healEntry)
+				errs[idx] = set.listAndHeal(bucket, prefix, opts.ScanMode, healEntry)
 			}(idx, set)
 		}
 		wg.Wait()
@@ -2253,11 +2289,13 @@ type HealthOptions struct {
 // additionally with any specific heuristic information which
 // was queried
 type HealthResult struct {
-	Healthy        bool
-	HealingDrives  int
-	UnhealthyPools []struct {
+	Healthy       bool
+	HealingDrives int
+	ESHealth      []struct {
 		Maintenance   bool
 		PoolID, SetID int
+		HealthyDrives int
+		HealingDrives int
 		WriteQuorum   int
 	}
 	WriteQuorum   int
@@ -2304,29 +2342,36 @@ func (z *erasureServerPools) ReadHealth(ctx context.Context) bool {
 // can be used to query scenarios if health may be lost
 // if this node is taken down by an external orchestrator.
 func (z *erasureServerPools) Health(ctx context.Context, opts HealthOptions) HealthResult {
-	erasureSetUpCount := make([][]int, len(z.serverPools))
+	reqInfo := (&logger.ReqInfo{}).AppendTags("maintenance", strconv.FormatBool(opts.Maintenance))
+
+	type setInfo struct {
+		online  int
+		healing int
+	}
+
+	var drivesHealing int
+
+	erasureSetUpCount := make([][]setInfo, len(z.serverPools))
 	for i := range z.serverPools {
-		erasureSetUpCount[i] = make([]int, len(z.serverPools[i].sets))
+		erasureSetUpCount[i] = make([]setInfo, len(z.serverPools[i].sets))
 	}

-	diskIDs := globalNotificationSys.GetLocalDiskIDs(ctx)
-	if !opts.Maintenance {
-		diskIDs = append(diskIDs, getLocalDiskIDs(z))
-	}
+	storageInfo := z.StorageInfo(ctx, false)

-	for _, localDiskIDs := range diskIDs {
-		for _, id := range localDiskIDs {
-			poolIdx, setIdx, _, err := z.getPoolAndSet(id)
-			if err != nil {
-				logger.LogIf(ctx, err)
-				continue
+	for _, disk := range storageInfo.Disks {
+		if disk.PoolIndex > -1 && disk.SetIndex > -1 {
+			if disk.State == madmin.DriveStateOk {
+				si := erasureSetUpCount[disk.PoolIndex][disk.SetIndex]
+				si.online++
+				if disk.Healing {
+					si.healing++
+					drivesHealing++
+				}
+				erasureSetUpCount[disk.PoolIndex][disk.SetIndex] = si
 			}
-			erasureSetUpCount[poolIdx][setIdx]++
 		}
 	}

-	reqInfo := (&logger.ReqInfo{}).AppendTags("maintenance", strconv.FormatBool(opts.Maintenance))
-
 	b := z.BackendInfo()
 	poolWriteQuorums := make([]int, len(b.StandardSCData))
 	for i, data := range b.StandardSCData {
@@ -2336,23 +2381,10 @@ func (z *erasureServerPools) Health(ctx context.Context, opts HealthOptions) Hea
 		}
 	}

-	var aggHealStateResult madmin.BgHealState
 	// Check if disks are healing on in-case of VMware vsphere deployments.
 	if opts.Maintenance && opts.DeploymentType == vmware {
-		// check if local disks are being healed, if they are being healed
-		// we need to tell healthy status as 'false' so that this server
-		// is not taken down for maintenance
-		var err error
-		aggHealStateResult, err = getAggregatedBackgroundHealState(ctx, nil)
-		if err != nil {
-			logger.LogIf(logger.SetReqInfo(ctx, reqInfo), fmt.Errorf("Unable to verify global heal status: %w", err))
-			return HealthResult{
-				Healthy: false,
-			}
-		}
-
-		if len(aggHealStateResult.HealDisks) > 0 {
-			logger.LogIf(logger.SetReqInfo(ctx, reqInfo), fmt.Errorf("Total drives to be healed %d", len(aggHealStateResult.HealDisks)))
+		if drivesHealing > 0 {
+			logger.LogIf(logger.SetReqInfo(ctx, reqInfo), fmt.Errorf("Total drives to be healed %d", drivesHealing))
 		}
 	}

@@ -2372,50 +2404,41 @@ func (z *erasureServerPools) Health(ctx context.Context, opts HealthOptions) Hea
 	}

 	result := HealthResult{
-		HealingDrives: len(aggHealStateResult.HealDisks),
+		Healthy:       true,
 		WriteQuorum:   maximumWriteQuorum,
 		UsingDefaults: usingDefaults, // indicates if config was not initialized and we are using defaults on this node.
 	}

 	for poolIdx := range erasureSetUpCount {
 		for setIdx := range erasureSetUpCount[poolIdx] {
-			if erasureSetUpCount[poolIdx][setIdx] < poolWriteQuorums[poolIdx] {
+			result.ESHealth = append(result.ESHealth, struct {
+				Maintenance                               bool
+				PoolID, SetID                             int
+				HealthyDrives, HealingDrives, WriteQuorum int
+			}{
+				Maintenance:   opts.Maintenance,
+				SetID:         setIdx,
+				PoolID:        poolIdx,
+				HealthyDrives: erasureSetUpCount[poolIdx][setIdx].online,
+				HealingDrives: erasureSetUpCount[poolIdx][setIdx].healing,
+				WriteQuorum:   poolWriteQuorums[poolIdx],
+			})
+
+			if erasureSetUpCount[poolIdx][setIdx].online < poolWriteQuorums[poolIdx] {
 				logger.LogIf(logger.SetReqInfo(ctx, reqInfo),
 					fmt.Errorf("Write quorum may be lost on pool: %d, set: %d, expected write quorum: %d",
 						poolIdx, setIdx, poolWriteQuorums[poolIdx]))
-				result.UnhealthyPools = append(result.UnhealthyPools, struct {
-					Maintenance                bool
-					PoolID, SetID, WriteQuorum int
-				}{
-					Maintenance: opts.Maintenance,
-					SetID:       setIdx,
-					PoolID:      poolIdx,
-					WriteQuorum: poolWriteQuorums[poolIdx],
-				})
+				result.Healthy = false
 			}
 		}
-		if len(result.UnhealthyPools) > 0 {
-			// We have unhealthy pools return error.
-			return result
-		}
 	}

-	// when maintenance is not specified we don't have
-	// to look at the healing side of the code.
-	if !opts.Maintenance {
-		return HealthResult{
-			Healthy:       true,
-			WriteQuorum:   maximumWriteQuorum,
-			UsingDefaults: usingDefaults, // indicates if config was not initialized and we are using defaults on this node.
-		}
+	if opts.Maintenance {
+		result.Healthy = result.Healthy && drivesHealing == 0
+		result.HealingDrives = drivesHealing
 	}

-	return HealthResult{
-		Healthy:       len(aggHealStateResult.HealDisks) == 0,
-		HealingDrives: len(aggHealStateResult.HealDisks),
-		WriteQuorum:   maximumWriteQuorum,
-		UsingDefaults: usingDefaults, // indicates if config was not initialized and we are using defaults on this node.
-	}
+	return result
 }

 // PutObjectMetadata - replace or add tags to an existing object
@@ -36,7 +36,6 @@ import (
 	"github.com/minio/madmin-go/v3"
 	"github.com/minio/minio-go/v7/pkg/set"
 	"github.com/minio/minio-go/v7/pkg/tags"
-	"github.com/minio/minio/internal/bpool"
 	"github.com/minio/minio/internal/dsync"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/console"
@@ -217,6 +216,10 @@ func (s *erasureSets) connectDisks() {
 				continue
 			}
 		}
+		if cdisk != nil {
+			// Close previous offline disk.
+			cdisk.Close()
+		}

 		wg.Add(1)
 		go func(endpoint Endpoint) {
@@ -258,6 +261,12 @@ func (s *erasureSets) connectDisks() {
 			disk.SetDiskLoc(s.poolIndex, setIndex, diskIndex)
 			s.erasureDisks[setIndex][diskIndex] = disk
 			s.erasureDisksMu.Unlock()
+
+			if disk.IsLocal() && globalIsDistErasure {
+				globalLocalDrivesMu.Lock()
+				globalLocalSetDrives[s.poolIndex][setIndex][diskIndex] = disk
+				globalLocalDrivesMu.Unlock()
+			}
 		}(endpoint)
 	}

@@ -361,21 +370,6 @@ func newErasureSets(ctx context.Context, endpoints PoolEndpoints, storageDisks [

 	mutex := newNSLock(globalIsDistErasure)

-	// Number of buffers, max 2GB
-	n := (2 * humanize.GiByte) / (blockSizeV2 * 2)
-
-	// Initialize byte pool once for all sets, bpool size is set to
-	// setCount * setDriveCount with each memory upto blockSizeV2.
-	bp := bpool.NewBytePoolCap(n, blockSizeV2, blockSizeV2*2)
-
-	// Initialize byte pool for all sets, bpool size is set to
-	// setCount * setDriveCount with each memory upto blockSizeV1
-	//
-	// Number of buffers, max 10GiB
-	m := (10 * humanize.GiByte) / (blockSizeV1 * 2)
-
-	bpOld := bpool.NewBytePoolCap(m, blockSizeV1, blockSizeV1*2)
-
 	for i := 0; i < setCount; i++ {
 		s.erasureDisks[i] = make([]StorageAPI, setDriveCount)
 	}
@@ -411,6 +405,18 @@ func newErasureSets(ctx context.Context, endpoints PoolEndpoints, storageDisks [
 				if disk == nil {
 					continue
 				}
+
+				if disk.IsLocal() && globalIsDistErasure {
+					globalLocalDrivesMu.RLock()
+					ldisk := globalLocalSetDrives[poolIdx][i][j]
+					if ldisk == nil {
+						globalLocalDrivesMu.RUnlock()
+						continue
+					}
+					disk = ldisk
+					globalLocalDrivesMu.RUnlock()
+				}
+
 				innerWg.Add(1)
 				go func(disk StorageAPI, i, j int) {
 					defer innerWg.Done()
@@ -451,8 +457,6 @@ func newErasureSets(ctx context.Context, endpoints PoolEndpoints, storageDisks [
 				getLockers:         s.GetLockers(i),
 				getEndpoints:       s.GetEndpoints(i),
 				nsMutex:            mutex,
-				bp:                 bp,
-				bpOld:              bpOld,
 			}
 		}(i)
 	}
@@ -607,7 +611,7 @@ func (s *erasureSets) StorageInfo(ctx context.Context) StorageInfo {
 }

 // StorageInfo - combines output of StorageInfo across all erasure coded object sets.
-func (s *erasureSets) LocalStorageInfo(ctx context.Context) StorageInfo {
+func (s *erasureSets) LocalStorageInfo(ctx context.Context, metrics bool) StorageInfo {
 	var storageInfo StorageInfo

 	storageInfos := make([]StorageInfo, len(s.sets))
@@ -616,7 +620,7 @@ func (s *erasureSets) LocalStorageInfo(ctx context.Context) StorageInfo {
 	for index := range s.sets {
 		index := index
 		g.Go(func() error {
-			storageInfos[index] = s.sets[index].LocalStorageInfo(ctx)
+			storageInfos[index] = s.sets[index].LocalStorageInfo(ctx, metrics)
 			return nil
 		}, index)
 	}
@@ -1106,7 +1110,7 @@ func (s *erasureSets) HealFormat(ctx context.Context, dryRun bool) (res madmin.H
 	formatOpID := mustGetUUID()

 	// Initialize a new set of set formats which will be written to disk.
-	newFormatSets := newHealFormatSets(refFormat, s.setCount, s.setDriveCount, formats, sErrs)
+	newFormatSets, currentDisksInfo := newHealFormatSets(refFormat, s.setCount, s.setDriveCount, formats, sErrs)

 	if !dryRun {
 		tmpNewFormats := make([]*formatErasureV3, s.setCount*s.setDriveCount)
@@ -1149,9 +1153,27 @@ func (s *erasureSets) HealFormat(ctx context.Context, dryRun bool) (res madmin.H
 				s.erasureDisks[m][n].Close()
 			}

-			if storageDisks[index] != nil {
-				storageDisks[index].SetDiskLoc(s.poolIndex, m, n)
-				s.erasureDisks[m][n] = storageDisks[index]
+			if disk := storageDisks[index]; disk != nil {
+				disk.SetDiskLoc(s.poolIndex, m, n)
+
+				if disk.IsLocal() && driveQuorum {
+					commonWrites, commonDeletes := calcCommonWritesDeletes(currentDisksInfo[m], (s.setDriveCount+1)/2)
+					xldisk, ok := disk.(*xlStorageDiskIDCheck)
+					if ok {
+						xldisk.totalWrites.Add(commonWrites)
+						xldisk.totalDeletes.Add(commonDeletes)
+						xldisk.storage.setWriteAttribute(commonWrites)
+						xldisk.storage.setDeleteAttribute(commonDeletes)
+					}
+				}
+
+				s.erasureDisks[m][n] = disk
+
+				if disk.IsLocal() && globalIsDistErasure {
+					globalLocalDrivesMu.Lock()
+					globalLocalSetDrives[s.poolIndex][m][n] = disk
+					globalLocalDrivesMu.Unlock()
+				}
 			}
 		}

@@ -18,7 +18,6 @@
 package cmd

 import (
-	"bytes"
 	"context"
 	"encoding/base64"
 	"fmt"
@@ -26,7 +25,6 @@ import (
 	"strings"

 	"github.com/klauspost/reedsolomon"
-	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 )

@@ -74,18 +72,18 @@ func writeDataBlocks(ctx context.Context, dst io.Writer, enBlocks [][]byte, data
 			// Decrement offset.
 			offset -= int64(len(block))
 			continue
-		} else {
-			// Skip until offset.
-			block = block[offset:]
-
-			// Reset the offset for next iteration to read everything
-			// from subsequent blocks.
-			offset = 0
 		}

+		// Skip until offset.
+		block = block[offset:]
+
+		// Reset the offset for next iteration to read everything
+		// from subsequent blocks.
+		offset = 0
+
 		// We have written all the blocks, write the last remaining block.
 		if write < int64(len(block)) {
-			n, err := xioutil.Copy(dst, bytes.NewReader(block[:write]))
+			n, err := dst.Write(block[:write])
 			if err != nil {
 				// The writer will be closed incase of range queries, which will emit ErrClosedPipe.
 				// The reader pipe might be closed at ListObjects io.EOF ignore it.
@@ -94,12 +92,12 @@ func writeDataBlocks(ctx context.Context, dst io.Writer, enBlocks [][]byte, data
 				}
 				return 0, err
 			}
-			totalWritten += n
+			totalWritten += int64(n)
 			break
 		}

 		// Copy the block.
-		n, err := xioutil.Copy(dst, bytes.NewReader(block))
+		n, err := dst.Write(block)
 		if err != nil {
 			// The writer will be closed incase of range queries, which will emit ErrClosedPipe.
 			// The reader pipe might be closed at ListObjects io.EOF ignore it.
@@ -110,10 +108,10 @@ func writeDataBlocks(ctx context.Context, dst io.Writer, enBlocks [][]byte, data
 		}

 		// Decrement output size.
-		write -= n
+		write -= int64(n)

 		// Increment written.
-		totalWritten += n
+		totalWritten += int64(n)
 	}

 	// Success.
@@ -29,7 +29,6 @@ import (
 	"time"

 	"github.com/minio/madmin-go/v3"
-	"github.com/minio/minio/internal/bpool"
 	"github.com/minio/minio/internal/dsync"
 	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
@@ -65,13 +64,6 @@ type erasureObjects struct {

 	// Locker mutex map.
 	nsMutex *nsLockMap
-
-	// Byte pools used for temporary i/o buffers.
-	bp *bpool.BytePoolCap
-
-	// Byte pools used for temporary i/o buffers,
-	// legacy objects.
-	bpOld *bpool.BytePoolCap
 }

 // NewNSLock - initialize a new namespace RWLocker instance.
@@ -168,7 +160,7 @@ func getOnlineOfflineDisksStats(disksInfo []madmin.Disk) (onlineDisks, offlineDi
 }

 // getDisksInfo - fetch disks info across all other storage API.
-func getDisksInfo(disks []StorageAPI, endpoints []Endpoint) (disksInfo []madmin.Disk) {
+func getDisksInfo(disks []StorageAPI, endpoints []Endpoint, metrics bool) (disksInfo []madmin.Disk) {
 	disksInfo = make([]madmin.Disk, len(disks))

 	g := errgroup.WithNErrs(len(disks))
@@ -186,7 +178,7 @@ func getDisksInfo(disks []StorageAPI, endpoints []Endpoint) (disksInfo []madmin.
 				disksInfo[index] = di
 				return nil
 			}
-			info, err := disks[index].DiskInfo(context.TODO(), true)
+			info, err := disks[index].DiskInfo(context.TODO(), metrics)
 			di.DrivePath = info.MountPath
 			di.TotalSpace = info.Total
 			di.UsedSpace = info.Used
@@ -233,8 +225,8 @@ func getDisksInfo(disks []StorageAPI, endpoints []Endpoint) (disksInfo []madmin.
 }

 // Get an aggregated storage info across all disks.
-func getStorageInfo(disks []StorageAPI, endpoints []Endpoint) StorageInfo {
-	disksInfo := getDisksInfo(disks, endpoints)
+func getStorageInfo(disks []StorageAPI, endpoints []Endpoint, metrics bool) StorageInfo {
+	disksInfo := getDisksInfo(disks, endpoints, metrics)

 	// Sort so that the first element is the smallest.
 	sort.Slice(disksInfo, func(i, j int) bool {
@@ -253,11 +245,11 @@ func getStorageInfo(disks []StorageAPI, endpoints []Endpoint) StorageInfo {
 func (er erasureObjects) StorageInfo(ctx context.Context) StorageInfo {
 	disks := er.getDisks()
 	endpoints := er.getEndpoints()
-	return getStorageInfo(disks, endpoints)
+	return getStorageInfo(disks, endpoints, true)
 }

 // LocalStorageInfo - returns underlying local storage statistics.
-func (er erasureObjects) LocalStorageInfo(ctx context.Context) StorageInfo {
+func (er erasureObjects) LocalStorageInfo(ctx context.Context, metrics bool) StorageInfo {
 	disks := er.getDisks()
 	endpoints := er.getEndpoints()

@@ -271,10 +263,15 @@ func (er erasureObjects) LocalStorageInfo(ctx context.Context) StorageInfo {
 		}
 	}

-	return getStorageInfo(localDisks, localEndpoints)
+	return getStorageInfo(localDisks, localEndpoints, metrics)
 }

-func (er erasureObjects) getOnlineDisksWithHealing() (newDisks []StorageAPI, healing bool) {
+// getOnlineDisksWithHealingAndInfo - returns online disks and overall healing status.
+// Disks are randomly ordered, but in the following groups:
+// - Non-scanning disks
+// - Non-healing disks
+// - Healing disks (if inclHealing is true)
+func (er erasureObjects) getOnlineDisksWithHealingAndInfo(inclHealing bool) (newDisks []StorageAPI, newInfos []DiskInfo, healing bool) {
 	var wg sync.WaitGroup
 	disks := er.getDisks()
 	infos := make([]DiskInfo, len(disks))
@@ -287,48 +284,63 @@ func (er erasureObjects) getOnlineDisksWithHealing() (newDisks []StorageAPI, hea

 			disk := disks[i]
 			if disk == nil {
-				infos[i].Error = "offline drive"
+				infos[i].Error = errDiskNotFound.Error()
 				return
 			}

 			di, err := disk.DiskInfo(context.Background(), false)
-			if err != nil || di.Healing {
+			infos[i] = di
+			if err != nil {
 				// - Do not consume disks which are not reachable
 				//   unformatted or simply not accessible for some reason.
-				//
-				//
-				// - Future: skip busy disks
-				if err != nil {
-					infos[i].Error = err.Error()
-				}
-				return
+				infos[i].Error = err.Error()
 			}
-
-			infos[i] = di
 		}()
 	}
 	wg.Wait()

-	var scanningDisks []StorageAPI
+	var scanningDisks, healingDisks []StorageAPI
+	var scanningInfos, healingInfos []DiskInfo
+
 	for i, info := range infos {
 		// Check if one of the drives in the set is being healed.
 		// this information is used by scanner to skip healing
 		// this erasure set while it calculates the usage.
-		if info.Healing || info.Error != "" {
-			healing = true
+		if info.Error != "" || disks[i] == nil {
 			continue
 		}
+		if info.Healing {
+			healing = true
+			if inclHealing {
+				healingDisks = append(healingDisks, disks[i])
+				healingInfos = append(healingInfos, infos[i])
+			}
+			continue
+		}
+
 		if !info.Scanning {
 			newDisks = append(newDisks, disks[i])
+			newInfos = append(newInfos, infos[i])
 		} else {
 			scanningDisks = append(scanningDisks, disks[i])
+			scanningInfos = append(scanningInfos, infos[i])
 		}
 	}

-	// Prefer new disks over disks which are currently being scanned.
+	// Prefer non-scanning disks over disks which are currently being scanned.
 	newDisks = append(newDisks, scanningDisks...)
+	newInfos = append(newInfos, scanningInfos...)

-	return newDisks, healing
+	/// Then add healing disks.
+	newDisks = append(newDisks, healingDisks...)
+	newInfos = append(newInfos, healingInfos...)
+
+	return newDisks, newInfos, healing
+}
+
+func (er erasureObjects) getOnlineDisksWithHealing(inclHealing bool) (newDisks []StorageAPI, healing bool) {
+	newDisks, _, healing = er.getOnlineDisksWithHealingAndInfo(inclHealing)
+	return
 }

 // Clean-up previously deleted objects. from .minio.sys/tmp/.trash/
@@ -342,7 +354,7 @@ func (er erasureObjects) cleanupDeletedObjects(ctx context.Context) {
 				defer wg.Done()
 				diskPath := disk.Endpoint().Path
 				readDirFn(pathJoin(diskPath, minioMetaTmpDeletedBucket), func(ddir string, typ os.FileMode) error {
-					w := xioutil.NewDeadlineWorker(diskMaxTimeout)
+					w := xioutil.NewDeadlineWorker(globalDriveConfig.GetMaxTimeout())
 					return w.Run(func() error {
 						wait := deletedCleanupSleeper.Timer(ctx)
 						removeAll(pathJoin(diskPath, minioMetaTmpDeletedBucket, ddir))
@@ -364,7 +376,7 @@ func (er erasureObjects) nsScanner(ctx context.Context, buckets []BucketInfo, wa
 	}

 	// Collect disks we can use.
-	disks, healing := er.getOnlineDisksWithHealing()
+	disks, healing := er.getOnlineDisksWithHealing(false)
 	if len(disks) == 0 {
 		logger.LogIf(ctx, errors.New("data-scanner: all drives are offline or being healed, skipping scanner cycle"))
 		return nil
@@ -21,6 +21,7 @@ import (
 	"context"
 	"fmt"
 	"net/url"
+	"runtime"
 	"strings"
 	"sync"

@@ -36,16 +37,14 @@ import (
 type EventNotifier struct {
 	sync.RWMutex
 	targetList     *event.TargetList
-	targetResCh    chan event.TargetIDResult
 	bucketRulesMap map[string]event.RulesMap
 }

 // NewEventNotifier - creates new event notification object.
-func NewEventNotifier() *EventNotifier {
+func NewEventNotifier(ctx context.Context) *EventNotifier {
 	// targetList/bucketRulesMap/bucketRemoteTargetRulesMap are populated by NotificationSys.InitBucketTargets()
 	return &EventNotifier{
-		targetList:     event.NewTargetList(),
-		targetResCh:    make(chan event.TargetIDResult),
+		targetList:     event.NewTargetList(ctx),
 		bucketRulesMap: make(map[string]event.RulesMap),
 	}
 }
@@ -90,6 +89,11 @@ func (evnot *EventNotifier) set(bucket BucketInfo, meta BucketMetadata) {
 	evnot.AddRulesMap(bucket.Name, config.ToRulesMap())
 }

+// Targets returns all the registered targets
+func (evnot *EventNotifier) Targets() []event.Target {
+	return evnot.targetList.Targets()
+}
+
 // InitBucketTargets - initializes event notification system from notification.xml of all buckets.
 func (evnot *EventNotifier) InitBucketTargets(ctx context.Context, objAPI ObjectLayer) error {
 	if objAPI == nil {
@@ -99,17 +103,7 @@ func (evnot *EventNotifier) InitBucketTargets(ctx context.Context, objAPI Object
 	if err := evnot.targetList.Add(globalNotifyTargetList.Targets()...); err != nil {
 		return err
 	}
-
-	go func() {
-		for res := range evnot.targetResCh {
-			if res.Err != nil {
-				reqInfo := &logger.ReqInfo{}
-				reqInfo.AppendTags("targetID", res.ID.String())
-				logger.LogOnceIf(logger.SetReqInfo(GlobalContext, reqInfo), res.Err, res.ID.String())
-			}
-		}
-	}()
-
+	evnot.targetList = evnot.targetList.Init(runtime.GOMAXPROCS(0)) // TODO: make this configurable (y4m4)
 	return nil
 }

@@ -128,40 +122,6 @@ func (evnot *EventNotifier) AddRulesMap(bucketName string, rulesMap event.RulesM
 	}
 }

-// RemoveRulesMap - removes rules map for bucket name.
-func (evnot *EventNotifier) RemoveRulesMap(bucketName string, rulesMap event.RulesMap) {
-	evnot.Lock()
-	defer evnot.Unlock()
-
-	evnot.bucketRulesMap[bucketName].Remove(rulesMap)
-	if len(evnot.bucketRulesMap[bucketName]) == 0 {
-		delete(evnot.bucketRulesMap, bucketName)
-	}
-}
-
-// ConfiguredTargetIDs - returns list of configured target id's
-func (evnot *EventNotifier) ConfiguredTargetIDs() []event.TargetID {
-	if evnot == nil {
-		return nil
-	}
-
-	evnot.RLock()
-	defer evnot.RUnlock()
-
-	var targetIDs []event.TargetID
-	for _, rmap := range evnot.bucketRulesMap {
-		for _, rules := range rmap {
-			for _, targetSet := range rules {
-				for id := range targetSet {
-					targetIDs = append(targetIDs, id)
-				}
-			}
-		}
-	}
-
-	return targetIDs
-}
-
 // RemoveNotification - removes all notification configuration for bucket name.
 func (evnot *EventNotifier) RemoveNotification(bucketName string) {
 	evnot.Lock()
@@ -193,7 +153,7 @@ func (evnot *EventNotifier) Send(args eventArgs) {
 	}

 	// If MINIO_API_SYNC_EVENTS is set, send events synchronously.
-	evnot.targetList.Send(args.ToEvent(true), targetIDSet, evnot.targetResCh, globalAPIConfig.isSyncEventsEnabled())
+	evnot.targetList.Send(args.ToEvent(true), targetIDSet, globalAPIConfig.isSyncEventsEnabled())
 }

 type eventArgs struct {
@@ -123,6 +123,7 @@ type formatErasureV3 struct {
 		// to pick the right set index for an object.
 		DistributionAlgo string `json:"distributionAlgo"`
 	} `json:"xl"`
+	Info DiskInfo `json:"-"`
 }

 func (f *formatErasureV3) Drives() (drives int) {
@@ -328,6 +329,11 @@ func loadFormatErasureAll(storageDisks []StorageAPI, heal bool) ([]*formatErasur
 			if err != nil {
 				return err
 			}
+			info, err := storageDisks[index].DiskInfo(context.Background(), false)
+			if err != nil {
+				return err
+			}
+			format.Info = info
 			formats[index] = format
 			if !heal {
 				// If no healing required, make the disks valid and
@@ -364,7 +370,7 @@ func saveFormatErasure(disk StorageAPI, format *formatErasureV3, healID string)
 	// Purge any existing temporary file, okay to ignore errors here.
 	defer disk.Delete(context.TODO(), minioMetaBucket, tmpFormat, DeleteOptions{
 		Recursive: false,
-		Force:     false,
+		Immediate: false,
 	})

 	// write to unique file.
@@ -388,6 +394,12 @@ func saveFormatErasure(disk StorageAPI, format *formatErasureV3, healID string)

 // loadFormatErasure - loads format.json from disk.
 func loadFormatErasure(disk StorageAPI) (format *formatErasureV3, err error) {
+	// Ensure that the grid is online.
+	if _, err := disk.DiskInfo(context.Background(), false); err != nil {
+		if errors.Is(err, errDiskNotFound) {
+			return nil, err
+		}
+	}
 	buf, err := disk.ReadAll(context.TODO(), minioMetaBucket, formatConfigFile)
 	if err != nil {
 		// 'file not found' and 'volume not found' as
@@ -818,11 +830,15 @@ func makeFormatErasureMetaVolumes(disk StorageAPI) error {
 }

 // Initialize a new set of set formats which will be written to all disks.
-func newHealFormatSets(refFormat *formatErasureV3, setCount, setDriveCount int, formats []*formatErasureV3, errs []error) [][]*formatErasureV3 {
+func newHealFormatSets(refFormat *formatErasureV3, setCount, setDriveCount int, formats []*formatErasureV3, errs []error) ([][]*formatErasureV3, [][]DiskInfo) {
 	newFormats := make([][]*formatErasureV3, setCount)
 	for i := range refFormat.Erasure.Sets {
 		newFormats[i] = make([]*formatErasureV3, setDriveCount)
 	}
+	currentDisksInfo := make([][]DiskInfo, setCount)
+	for i := range refFormat.Erasure.Sets {
+		currentDisksInfo[i] = make([]DiskInfo, setDriveCount)
+	}
 	for i := range refFormat.Erasure.Sets {
 		for j := range refFormat.Erasure.Sets[i] {
 			if errors.Is(errs[i*setDriveCount+j], errUnformattedDisk) {
@@ -835,7 +851,10 @@ func newHealFormatSets(refFormat *formatErasureV3, setCount, setDriveCount int,
 				newFormats[i][j].Erasure.Version = refFormat.Erasure.Version
 				newFormats[i][j].Erasure.DistributionAlgo = refFormat.Erasure.DistributionAlgo
 			}
+			if format := formats[i*setDriveCount+j]; format != nil && (errs[i*setDriveCount+j] == nil) {
+				currentDisksInfo[i][j] = format.Info
+			}
 		}
 	}
-	return newFormats
+	return newFormats, currentDisksInfo
 }
@@ -513,7 +513,7 @@ func TestNewFormatSets(t *testing.T) {
 	// 16th disk is unformatted.
 	errs[15] = errUnformattedDisk

-	newFormats := newHealFormatSets(quorumFormat, setCount, setDriveCount, formats, errs)
+	newFormats, _ := newHealFormatSets(quorumFormat, setCount, setDriveCount, formats, errs)
 	if newFormats == nil {
 		t.Fatal("Unexpected failure")
 	}
@@ -248,12 +248,19 @@ func (driver *ftpDriver) CheckPasswd(c *ftp.Context, username, password string)
 	defer stopFn(err)

 	if globalIAMSys.LDAPConfig.Enabled() {
-		ldapUserDN, groupDistNames, err := globalIAMSys.LDAPConfig.Bind(username, password)
-		if err != nil {
+		sa, _, err := globalIAMSys.getServiceAccount(context.Background(), username)
+		if err != nil && !errors.Is(err, errNoSuchServiceAccount) {
 			return false, err
 		}
-		ldapPolicies, _ := globalIAMSys.PolicyDBGet(ldapUserDN, false, groupDistNames...)
-		return len(ldapPolicies) > 0, nil
+		if errors.Is(err, errNoSuchServiceAccount) {
+			ldapUserDN, groupDistNames, err := globalIAMSys.LDAPConfig.Bind(username, password)
+			if err != nil {
+				return false, err
+			}
+			ldapPolicies, _ := globalIAMSys.PolicyDBGet(ldapUserDN, groupDistNames...)
+			return len(ldapPolicies) > 0, nil
+		}
+		return subtle.ConstantTimeCompare([]byte(sa.Credentials.SecretKey), []byte(password)) == 1, nil
 	}

 	ui, ok := globalIAMSys.GetUser(context.Background(), username)
@@ -269,58 +276,70 @@ func (driver *ftpDriver) getMinIOClient(ctx *ftp.Context) (*minio.Client, error)
 		return nil, errNoSuchUser
 	}
 	if !ok && globalIAMSys.LDAPConfig.Enabled() {
-		targetUser, targetGroups, err := globalIAMSys.LDAPConfig.LookupUserDN(ctx.Sess.LoginUser())
-		if err != nil {
-			return nil, err
-		}
-		ldapPolicies, _ := globalIAMSys.PolicyDBGet(targetUser, false, targetGroups...)
-		if len(ldapPolicies) == 0 {
-			return nil, errAuthentication
-		}
-		expiryDur, err := globalIAMSys.LDAPConfig.GetExpiryDuration("")
-		if err != nil {
-			return nil, err
-		}
-		claims := make(map[string]interface{})
-		claims[expClaim] = UTCNow().Add(expiryDur).Unix()
-		claims[ldapUser] = targetUser
-		claims[ldapUserN] = ctx.Sess.LoginUser()
-
-		cred, err := auth.GetNewCredentialsWithMetadata(claims, globalActiveCred.SecretKey)
-		if err != nil {
+		sa, _, err := globalIAMSys.getServiceAccount(context.Background(), ctx.Sess.LoginUser())
+		if err != nil && !errors.Is(err, errNoSuchServiceAccount) {
 			return nil, err
 		}

-		// Set the parent of the temporary access key, this is useful
-		// in obtaining service accounts by this cred.
-		cred.ParentUser = targetUser
+		var mcreds *credentials.Credentials
+		if errors.Is(err, errNoSuchServiceAccount) {
+			targetUser, targetGroups, err := globalIAMSys.LDAPConfig.LookupUserDN(ctx.Sess.LoginUser())
+			if err != nil {
+				return nil, err
+			}
+			ldapPolicies, _ := globalIAMSys.PolicyDBGet(targetUser, targetGroups...)
+			if len(ldapPolicies) == 0 {
+				return nil, errAuthentication
+			}
+			expiryDur, err := globalIAMSys.LDAPConfig.GetExpiryDuration("")
+			if err != nil {
+				return nil, err
+			}
+			claims := make(map[string]interface{})
+			claims[expClaim] = UTCNow().Add(expiryDur).Unix()
+			claims[ldapUser] = targetUser
+			claims[ldapUserN] = ctx.Sess.LoginUser()

-		// Set this value to LDAP groups, LDAP user can be part
-		// of large number of groups
-		cred.Groups = targetGroups
+			cred, err := auth.GetNewCredentialsWithMetadata(claims, globalActiveCred.SecretKey)
+			if err != nil {
+				return nil, err
+			}

-		// Set the newly generated credentials, policyName is empty on purpose
-		// LDAP policies are applied automatically using their ldapUser, ldapGroups
-		// mapping.
-		updatedAt, err := globalIAMSys.SetTempUser(context.Background(), cred.AccessKey, cred, "")
-		if err != nil {
-			return nil, err
+			// Set the parent of the temporary access key, this is useful
+			// in obtaining service accounts by this cred.
+			cred.ParentUser = targetUser
+
+			// Set this value to LDAP groups, LDAP user can be part
+			// of large number of groups
+			cred.Groups = targetGroups
+
+			// Set the newly generated credentials, policyName is empty on purpose
+			// LDAP policies are applied automatically using their ldapUser, ldapGroups
+			// mapping.
+			updatedAt, err := globalIAMSys.SetTempUser(context.Background(), cred.AccessKey, cred, "")
+			if err != nil {
+				return nil, err
+			}
+
+			// Call hook for site replication.
+			logger.LogIf(context.Background(), globalSiteReplicationSys.IAMChangeHook(context.Background(), madmin.SRIAMItem{
+				Type: madmin.SRIAMItemSTSAcc,
+				STSCredential: &madmin.SRSTSCredential{
+					AccessKey:    cred.AccessKey,
+					SecretKey:    cred.SecretKey,
+					SessionToken: cred.SessionToken,
+					ParentUser:   cred.ParentUser,
+				},
+				UpdatedAt: updatedAt,
+			}))
+
+			mcreds = credentials.NewStaticV4(cred.AccessKey, cred.SecretKey, cred.SessionToken)
+		} else {
+			mcreds = credentials.NewStaticV4(sa.Credentials.AccessKey, sa.Credentials.SecretKey, "")
 		}

-		// Call hook for site replication.
-		logger.LogIf(context.Background(), globalSiteReplicationSys.IAMChangeHook(context.Background(), madmin.SRIAMItem{
-			Type: madmin.SRIAMItemSTSAcc,
-			STSCredential: &madmin.SRSTSCredential{
-				AccessKey:    cred.AccessKey,
-				SecretKey:    cred.SecretKey,
-				SessionToken: cred.SessionToken,
-				ParentUser:   cred.ParentUser,
-			},
-			UpdatedAt: updatedAt,
-		}))
-
 		return minio.New(driver.endpoint, &minio.Options{
-			Creds:     credentials.NewStaticV4(cred.AccessKey, cred.SecretKey, cred.SessionToken),
+			Creds:     mcreds,
 			Secure:    globalIsTLS,
 			Transport: globalRemoteFTPClientTransport,
 		})
@@ -18,20 +18,13 @@
 package cmd

 import (
-	"context"
-	"crypto/subtle"
 	"fmt"
 	"net"
-	"os"
 	"strconv"
 	"strings"
-	"time"

-	"github.com/minio/cli"
 	"github.com/minio/minio/internal/logger"
-	"github.com/pkg/sftp"
 	ftp "goftp.io/server/v2"
-	"golang.org/x/crypto/ssh"
 )

 var globalRemoteFTPClientTransport = NewRemoteTargetHTTPTransport(true)()
@@ -75,199 +68,7 @@ func (log *minioLogger) PrintResponse(sessionID string, code int, message string
 	}
 }

-func acceptSFTPConnection(conn net.Conn, config *ssh.ServerConfig) {
-	// For SSH handshake keep a 2 minute deadline, as OpensSSH default.
-	conn.SetDeadline(time.Now().Add(2 * time.Minute))
-
-	// Before use, a handshake must be performed on the incoming net.Conn.
-	sconn, chans, reqs, err := ssh.NewServerConn(conn, config)
-	if err != nil {
-		return
-	}
-
-	// Once we are done with SSH handshake, remove deadline.
-	conn.SetDeadline(time.Time{})
-
-	// The incoming Request channel must be serviced.
-	go ssh.DiscardRequests(reqs)
-
-	// Service the incoming Channel channel.
-	for newChannel := range chans {
-		// Channels have a type, depending on the application level
-		// protocol intended. In the case of an SFTP session, this is "subsystem"
-		// with a payload string of "<length=4>sftp"
-		if newChannel.ChannelType() != "session" {
-			newChannel.Reject(ssh.UnknownChannelType, "unknown channel type")
-			continue
-		}
-
-		channel, requests, err := newChannel.Accept()
-		if err != nil {
-			logger.LogOnceIf(context.Background(), fmt.Errorf("unable to accept the connection requests channel: %w", err), "accept-channel-sftp")
-			continue
-		}
-
-		// Sessions have out-of-band requests such as "shell",
-		// "pty-req" and "env".  Here we handle only the
-		// "subsystem" request.
-		go func(in <-chan *ssh.Request) {
-			for req := range in {
-				ok := false
-				if req.Type == "subsystem" {
-					if len(req.Payload) > 4 && string(req.Payload[4:]) == "sftp" {
-						ok = true
-						go handleSFTPConnection(channel, sconn)
-					}
-				}
-
-				if req.WantReply {
-					// We only reply to SSH packets that have `sftp` payload, all other
-					// packets are rejected
-					req.Reply(ok, nil)
-				}
-			}
-		}(requests)
-	}
-}
-
-func handleSFTPConnection(channel ssh.Channel, sconn *ssh.ServerConn) {
-	// Create the server instance for the channel using the handler we created above.
-	server := sftp.NewRequestServer(channel, NewSFTPDriver(sconn.Permissions), sftp.WithRSAllocator())
-	defer server.Close()
-
-	server.Serve()
-}
-
-func startSFTPServer(c *cli.Context) {
-	args := c.StringSlice("sftp")
-
-	var (
-		port          int
-		publicIP      string
-		sshPrivateKey string
-	)
-
-	var err error
-	for _, arg := range args {
-		tokens := strings.SplitN(arg, "=", 2)
-		if len(tokens) != 2 {
-			logger.Fatal(fmt.Errorf("invalid arguments passed to --sftp=%s", arg), "unable to start SFTP server")
-		}
-		switch tokens[0] {
-		case "address":
-			host, portStr, err := net.SplitHostPort(tokens[1])
-			if err != nil {
-				logger.Fatal(fmt.Errorf("invalid arguments passed to --sftp=%s (%v)", arg, err), "unable to start SFTP server")
-			}
-			port, err = strconv.Atoi(portStr)
-			if err != nil {
-				logger.Fatal(fmt.Errorf("invalid arguments passed to --sftp=%s (%v)", arg, err), "unable to start SFTP server")
-			}
-			if port < 1 || port > 65535 {
-				logger.Fatal(fmt.Errorf("invalid arguments passed to --sftp=%s, (port number must be between 1 to 65535)", arg), "unable to start SFTP server")
-			}
-			publicIP = host
-		case "ssh-private-key":
-			sshPrivateKey = tokens[1]
-		}
-	}
-
-	if port == 0 {
-		port = 8022 // Default SFTP port, since no port was given.
-	}
-
-	if sshPrivateKey == "" {
-		logger.Fatal(fmt.Errorf("invalid arguments passed, private key file is mandatory for --sftp='ssh-private-key=path/to/id_ecdsa'"), "unable to start SFTP server")
-	}
-
-	privateBytes, err := os.ReadFile(sshPrivateKey)
-	if err != nil {
-		logger.Fatal(fmt.Errorf("invalid arguments passed, private key file is not accessible: %v", err), "unable to start SFTP server")
-	}
-
-	private, err := ssh.ParsePrivateKey(privateBytes)
-	if err != nil {
-		logger.Fatal(fmt.Errorf("invalid arguments passed, private key file is not parseable: %v", err), "unable to start SFTP server")
-	}
-
-	// An SSH server is represented by a ServerConfig, which holds
-	// certificate details and handles authentication of ServerConns.
-	config := &ssh.ServerConfig{
-		PasswordCallback: func(c ssh.ConnMetadata, pass []byte) (*ssh.Permissions, error) {
-			if globalIAMSys.LDAPConfig.Enabled() {
-				targetUser, targetGroups, err := globalIAMSys.LDAPConfig.Bind(c.User(), string(pass))
-				if err != nil {
-					return nil, err
-				}
-				ldapPolicies, _ := globalIAMSys.PolicyDBGet(targetUser, false, targetGroups...)
-				if len(ldapPolicies) == 0 {
-					return nil, errAuthentication
-				}
-				return &ssh.Permissions{
-					CriticalOptions: map[string]string{
-						ldapUser:  targetUser,
-						ldapUserN: c.User(),
-					},
-					Extensions: make(map[string]string),
-				}, nil
-			}
-
-			ui, ok := globalIAMSys.GetUser(context.Background(), c.User())
-			if !ok {
-				return nil, errNoSuchUser
-			}
-
-			if subtle.ConstantTimeCompare([]byte(ui.Credentials.SecretKey), pass) == 1 {
-				return &ssh.Permissions{
-					CriticalOptions: map[string]string{
-						"accessKey": c.User(),
-					},
-					Extensions: make(map[string]string),
-				}, nil
-			}
-			return nil, errAuthentication
-		},
-	}
-
-	config.AddHostKey(private)
-
-	// Once a ServerConfig has been configured, connections can be accepted.
-	listener, err := net.Listen("tcp", net.JoinHostPort(publicIP, strconv.Itoa(port)))
-	if err != nil {
-		logger.Fatal(err, "unable to start listening on --sftp='port=%d'", port)
-	}
-
-	logger.Info(fmt.Sprintf("MinIO SFTP Server listening on %s", net.JoinHostPort(publicIP, strconv.Itoa(port))))
-
-	var tempDelay time.Duration // how long to sleep on accept failure
-
-	for {
-		conn, err := listener.Accept()
-		if err != nil {
-			// From https://github.com/golang/go/blob/4aa1efed4853ea067d665a952eee77c52faac774/src/net/http/server.go#L3046
-			if ne, ok := err.(net.Error); ok && ne.Temporary() {
-				if tempDelay == 0 {
-					tempDelay = 5 * time.Millisecond
-				} else {
-					tempDelay *= 2
-				}
-				if max := 1 * time.Second; tempDelay > max {
-					tempDelay = max
-				}
-				logger.LogOnceIf(context.Background(), fmt.Errorf("error while accepting connections: %w, retrying in %s", err, tempDelay), "accept-limit-sftp")
-				time.Sleep(tempDelay)
-				continue
-			}
-			logger.Fatal(err, "unrecoverable error while accepting new connections")
-		}
-
-		go acceptSFTPConnection(conn, config)
-	}
-}
-
-func startFTPServer(c *cli.Context) {
-	args := c.StringSlice("ftp")
-
+func startFTPServer(args []string) {
 	var (
 		port          int
 		publicIP      string
@@ -31,6 +31,7 @@ import (
 	"github.com/dustin/go-humanize"
 	"github.com/minio/minio-go/v7/pkg/s3utils"
 	"github.com/minio/minio-go/v7/pkg/set"
+	"github.com/minio/minio/internal/grid"
 	xnet "github.com/minio/pkg/v2/net"

 	"github.com/minio/minio/internal/amztime"
@@ -152,7 +153,7 @@ func setBrowserRedirectMiddleware(h http.Handler) http.Handler {
 	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		read := r.Method == http.MethodGet || r.Method == http.MethodHead
 		// Re-direction is handled specifically for browser requests.
-		if guessIsBrowserReq(r) && read && globalBrowserRedirect {
+		if !guessIsHealthCheckReq(r) && guessIsBrowserReq(r) && read && globalBrowserRedirect {
 			// Fetch the redirect location if any.
 			if u := getRedirectLocation(r); u != nil {
 				// Employ a temporary re-direct.
@@ -240,6 +241,10 @@ func guessIsRPCReq(req *http.Request) bool {
 	if req == nil {
 		return false
 	}
+	if req.Method == http.MethodGet && req.URL != nil && req.URL.Path == grid.RoutePath {
+		return true
+	}
+
 	return req.Method == http.MethodPost &&
 		strings.HasPrefix(req.URL.Path, minioReservedBucketPath+SlashSeparator)
 }
@@ -25,6 +25,7 @@ import (
 	"testing"

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

@@ -54,6 +55,14 @@ func TestGuessIsRPC(t *testing.T) {
 	if guessIsRPCReq(r) {
 		t.Fatal("Test shouldn't report as net/rpc for a non net/rpc request.")
 	}
+	r = &http.Request{
+		Proto:  "HTTP/1.1",
+		Method: http.MethodGet,
+		URL:    &url.URL{Path: grid.RoutePath},
+	}
+	if !guessIsRPCReq(r) {
+		t.Fatal("Grid RPC path not detected")
+	}
 }

 var isHTTPHeaderSizeTooLargeTests = []struct {
@@ -97,7 +97,7 @@ func getLocalBackgroundHealStatus(ctx context.Context, o ObjectLayer) (madmin.Bg
 		return status, true
 	}

-	si := o.LocalStorageInfo(ctx)
+	si := o.LocalStorageInfo(ctx, true)

 	indexed := make(map[string][]madmin.Disk)
 	for _, disk := range si.Disks {
@@ -195,6 +195,7 @@ func (er *erasureObjects) healErasureSet(ctx context.Context, buckets []string,
 		}
 		tracker.setObject("")
 		tracker.setBucket(bucket)
+
 		// Heal current bucket again in case if it is failed
 		// in the beginning of erasure set healing
 		if _, err := er.HealBucket(ctx, bucket, madmin.HealOpts{
@@ -204,12 +205,21 @@ func (er *erasureObjects) healErasureSet(ctx context.Context, buckets []string,
 			continue
 		}

+		vc, _ := globalBucketVersioningSys.Get(bucket)
+
+		// Check if the current bucket has a configured lifecycle policy
+		lc, _ := globalLifecycleSys.Get(bucket)
+
+		// Check if bucket is object locked.
+		lr, _ := globalBucketObjectLockSys.Get(bucket)
+		rcfg, _ := getReplicationConfig(ctx, bucket)
+
 		if serverDebugLog {
 			console.Debugf(color.Green("healDrive:")+" healing bucket %s content on %s erasure set\n",
 				bucket, humanize.Ordinal(er.setIndex+1))
 		}

-		disks, _ := er.getOnlineDisksWithHealing()
+		disks, _ := er.getOnlineDisksWithHealing(false)
 		if len(disks) == 0 {
 			logger.LogIf(ctx, fmt.Errorf("no online disks found to heal the bucket `%s`", bucket))
 			continue
@@ -244,6 +254,26 @@ func (er *erasureObjects) healErasureSet(ctx context.Context, buckets []string,
 			}
 		}

+		filterLifecycle := func(bucket, object string, fi FileInfo) bool {
+			if lc == nil {
+				return false
+			}
+			versioned := vc != nil && vc.Versioned(object)
+			objInfo := fi.ToObjectInfo(bucket, object, versioned)
+
+			evt := evalActionFromLifecycle(ctx, *lc, lr, rcfg, objInfo)
+			switch {
+			case evt.Action.DeleteRestored(): // if restored copy has expired,delete it synchronously
+				applyExpiryOnTransitionedObject(ctx, newObjectLayerFn(), objInfo, evt, lcEventSrc_Heal)
+				return false
+			case evt.Action.Delete():
+				globalExpiryState.enqueueByDays(objInfo, evt, lcEventSrc_Heal)
+				return true
+			default:
+				return false
+			}
+		}
+
 		// Collect updates to tracker from concurrent healEntry calls
 		results := make(chan healEntryResult, 1000)
 		go func() {
@@ -326,10 +356,17 @@ func (er *erasureObjects) healErasureSet(ctx context.Context, buckets []string,

 			var versionNotFound int
 			for _, version := range fivs.Versions {
-				// Ignore a version with a modtime newer than healing start
+				// Ignore a version with a modtime newer than healing start time.
 				if version.ModTime.After(tracker.Started) {
 					continue
 				}
+
+				// Apply lifecycle rules on the objects that are expired.
+				if filterLifecycle(bucket, version.Name, version) {
+					versionNotFound++
+					continue
+				}
+
 				if _, err := er.HealObject(ctx, bucket, encodedEntryName,
 					version.VersionID, madmin.HealOpts{
 						ScanMode: scanMode,
@@ -25,22 +25,26 @@ import (
 	"sync"
 	"time"

-	"github.com/minio/console/restapi"
+	consoleapi "github.com/minio/console/api"
 	"github.com/minio/dnscache"
 	"github.com/minio/madmin-go/v3"
 	"github.com/minio/minio-go/v7"
 	"github.com/minio/minio-go/v7/pkg/set"
+	"github.com/minio/minio/internal/bpool"
 	"github.com/minio/minio/internal/bucket/bandwidth"
 	"github.com/minio/minio/internal/config"
+	"github.com/minio/minio/internal/config/browser"
 	"github.com/minio/minio/internal/handlers"
 	"github.com/minio/minio/internal/kms"
 	"go.uber.org/atomic"

 	"github.com/dustin/go-humanize"
 	"github.com/minio/minio/internal/auth"
+	"github.com/minio/minio/internal/config/cache"
 	"github.com/minio/minio/internal/config/callhome"
 	"github.com/minio/minio/internal/config/compress"
 	"github.com/minio/minio/internal/config/dns"
+	"github.com/minio/minio/internal/config/drive"
 	idplugin "github.com/minio/minio/internal/config/identity/plugin"
 	polplugin "github.com/minio/minio/internal/config/policy/plugin"
 	"github.com/minio/minio/internal/config/storageclass"
@@ -125,13 +129,47 @@ const (
 	tlsClientSessionCacheSize = 100
 )

-var globalCLIContext = struct {
-	JSON, Quiet    bool
-	Anonymous      bool
-	StrictS3Compat bool
-}{}
+type poolDisksLayout struct {
+	cmdline string
+	layout  [][]string
+}
+
+type disksLayout struct {
+	legacy bool
+	pools  []poolDisksLayout
+}
+
+type serverCtxt struct {
+	JSON, Quiet               bool
+	Anonymous                 bool
+	StrictS3Compat            bool
+	PreAllocate               bool
+	Addr, ConsoleAddr         string
+	ConfigDir, CertsDir       string
+	configDirSet, certsDirSet bool
+	Interface                 string
+
+	RootUser, RootPwd string
+
+	FTP  []string
+	SFTP []string
+
+	UserTimeout       time.Duration
+	ConnReadDeadline  time.Duration
+	ConnWriteDeadline time.Duration
+
+	ShutdownTimeout   time.Duration
+	IdleTimeout       time.Duration
+	ReadHeaderTimeout time.Duration
+
+	// The layout of disks as interpreted
+	Layout disksLayout
+}

 var (
+	// Global user opts context
+	globalServerCtxt serverCtxt
+
 	// Indicates if the running minio server is distributed setup.
 	globalIsDistErasure = false

@@ -154,6 +192,9 @@ var (
 	// Disable redirect, default is enabled.
 	globalBrowserRedirect bool

+	// globalBrowserConfig Browser user configurable settings
+	globalBrowserConfig browser.Config
+
 	// This flag is set to 'true' when MINIO_UPDATE env is set to 'off'. Default is false.
 	globalInplaceUpdateDisabled = false

@@ -190,6 +231,7 @@ var (
 	globalBucketMonitor     *bandwidth.Monitor
 	globalPolicySys         *PolicySys
 	globalIAMSys            *IAMSys
+	globalBytePoolCap       *bpool.BytePoolCap

 	globalLifecycleSys       *LifecycleSys
 	globalBucketSSEConfigSys *BucketSSEConfigSys
@@ -242,6 +284,12 @@ var (
 	// The global callhome config
 	globalCallhomeConfig callhome.Config

+	// The global drive config
+	globalDriveConfig drive.Config
+
+	// The global cache config
+	globalCacheConfig cache.Config
+
 	// Global server's network statistics
 	globalConnStats = newConnStats()

@@ -351,7 +399,7 @@ var (

 	globalTierJournal *TierJournal

-	globalConsoleSrv *restapi.Server
+	globalConsoleSrv *consoleapi.Server

 	// handles service freeze or un-freeze S3 API calls.
 	globalServiceFreeze atomic.Value
@@ -0,0 +1,83 @@
+// Copyright (c) 2015-2023 MinIO, Inc.
+//
+// This file is part of MinIO Object Storage stack
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package cmd
+
+import (
+	"context"
+	"crypto/tls"
+	"sync/atomic"
+
+	"github.com/minio/minio-go/v7/pkg/set"
+	"github.com/minio/minio/internal/fips"
+	"github.com/minio/minio/internal/grid"
+	xhttp "github.com/minio/minio/internal/http"
+	"github.com/minio/minio/internal/rest"
+)
+
+// globalGrid is the global grid manager.
+var globalGrid atomic.Pointer[grid.Manager]
+
+// globalGridStart is a channel that will block startup of grid connections until closed.
+var globalGridStart = make(chan struct{})
+
+func initGlobalGrid(ctx context.Context, eps EndpointServerPools) error {
+	seenHosts := set.NewStringSet()
+	var hosts []string
+	var local string
+	for _, ep := range eps {
+		for _, endpoint := range ep.Endpoints {
+			u := endpoint.GridHost()
+			if seenHosts.Contains(u) {
+				continue
+			}
+			seenHosts.Add(u)
+
+			// Set local endpoint
+			if endpoint.IsLocal {
+				local = u
+			}
+			hosts = append(hosts, u)
+		}
+	}
+	lookupHost := globalDNSCache.LookupHost
+	if IsKubernetes() || IsDocker() {
+		lookupHost = nil
+	}
+	g, err := grid.NewManager(ctx, grid.ManagerOptions{
+		Dialer:       grid.ContextDialer(xhttp.DialContextWithLookupHost(lookupHost, xhttp.NewInternodeDialContext(rest.DefaultTimeout, globalTCPOptions))),
+		Local:        local,
+		Hosts:        hosts,
+		AddAuth:      newCachedAuthToken(),
+		AuthRequest:  storageServerRequestValidate,
+		BlockConnect: globalGridStart,
+		TLSConfig: &tls.Config{
+			RootCAs:          globalRootCAs,
+			CipherSuites:     fips.TLSCiphers(),
+			CurvePreferences: fips.TLSCurveIDs(),
+		},
+		// Record incoming and outgoing bytes.
+		Incoming: globalConnStats.incInternodeInputBytes,
+		Outgoing: globalConnStats.incInternodeOutputBytes,
+		TraceTo:  globalTrace,
+	})
+	if err != nil {
+		return err
+	}
+	globalGrid.Store(g)
+	return nil
+}
@@ -42,9 +42,10 @@ type apiConfig struct {
 	listQuorum       string
 	corsAllowOrigins []string
 	// total drives per erasure set across pools.
-	totalDriveCount     int
-	replicationPriority string
-	transitionWorkers   int
+	totalDriveCount       int
+	replicationPriority   string
+	replicationMaxWorkers int
+	transitionWorkers     int

 	staleUploadsExpiry          time.Duration
 	staleUploadsCleanupInterval time.Duration
@@ -152,10 +153,11 @@ func (t *apiConfig) init(cfg api.Config, setDriveCounts []int) {
 	}
 	t.listQuorum = listQuorum
 	if globalReplicationPool != nil &&
-		cfg.ReplicationPriority != t.replicationPriority {
-		globalReplicationPool.ResizeWorkerPriority(cfg.ReplicationPriority)
+		(cfg.ReplicationPriority != t.replicationPriority || cfg.ReplicationMaxWorkers != t.replicationMaxWorkers) {
+		globalReplicationPool.ResizeWorkerPriority(cfg.ReplicationPriority, cfg.ReplicationMaxWorkers)
 	}
 	t.replicationPriority = cfg.ReplicationPriority
+	t.replicationMaxWorkers = cfg.ReplicationMaxWorkers
 	if globalTransitionState != nil && cfg.TransitionWorkers != t.transitionWorkers {
 		globalTransitionState.UpdateWorkers(cfg.TransitionWorkers)
 	}
@@ -334,15 +336,21 @@ func maxClients(f http.HandlerFunc) http.HandlerFunc {
 	}
 }

-func (t *apiConfig) getReplicationPriority() string {
+func (t *apiConfig) getReplicationOpts() replicationPoolOpts {
 	t.mu.RLock()
 	defer t.mu.RUnlock()

 	if t.replicationPriority == "" {
-		return "auto"
+		return replicationPoolOpts{
+			Priority:   "auto",
+			MaxWorkers: WorkerMaxLimit,
+		}
 	}

-	return t.replicationPriority
+	return replicationPoolOpts{
+		Priority:   t.replicationPriority,
+		MaxWorkers: t.replicationMaxWorkers,
+	}
 }

 func (t *apiConfig) getTransitionWorkers() int {
@@ -111,21 +111,20 @@ var userMetadataKeyPrefixes = []string{
 	"x-minio-meta-",
 }

-// extractMetadata extracts metadata from HTTP header and HTTP queryString.
-func extractMetadata(ctx context.Context, r *http.Request) (metadata map[string]string, err error) {
-	query := r.Form
-	header := r.Header
-	metadata = make(map[string]string)
-	// Extract all query values.
-	err = extractMetadataFromMime(ctx, textproto.MIMEHeader(query), metadata)
-	if err != nil {
-		return nil, err
-	}
+// extractMetadataFromReq extracts metadata from HTTP header and HTTP queryString.
+func extractMetadataFromReq(ctx context.Context, r *http.Request) (metadata map[string]string, err error) {
+	return extractMetadata(ctx, textproto.MIMEHeader(r.Form), textproto.MIMEHeader(r.Header))
+}

-	// Extract all header values.
-	err = extractMetadataFromMime(ctx, textproto.MIMEHeader(header), metadata)
-	if err != nil {
-		return nil, err
+func extractMetadata(ctx context.Context, mimesHeader ...textproto.MIMEHeader) (metadata map[string]string, err error) {
+	metadata = make(map[string]string)
+
+	for _, hdr := range mimesHeader {
+		// Extract all query values.
+		err = extractMetadataFromMime(ctx, hdr, metadata)
+		if err != nil {
+			return nil, err
+		}
 	}

 	// Set content-type to default value if it is not set.
@@ -389,12 +388,6 @@ func errorResponseHandler(w http.ResponseWriter, r *http.Request) {
 			Description:    desc,
 			HTTPStatusCode: http.StatusUpgradeRequired,
 		}, r.URL)
-	case strings.HasPrefix(r.URL.Path, lockRESTPrefix):
-		writeErrorResponseString(r.Context(), w, APIError{
-			Code:           "XMinioLockVersionMismatch",
-			Description:    desc,
-			HTTPStatusCode: http.StatusUpgradeRequired,
-		}, r.URL)
 	case strings.HasPrefix(r.URL.Path, adminPathPrefix):
 		var desc string
 		version := extractAPIVersion(r)
@@ -90,21 +90,15 @@ func ClusterReadCheckHandler(w http.ResponseWriter, r *http.Request) {
 	writeResponse(w, http.StatusOK, nil, mimeNone)
 }

-// ReadinessCheckHandler Checks if the process is up. Always returns success.
+// ReadinessCheckHandler checks whether MinIO is up and ready to serve requests.
+// It also checks whether the KMS is available and whether etcd is reachable,
+// if configured.
 func ReadinessCheckHandler(w http.ResponseWriter, r *http.Request) {
-	LivenessCheckHandler(w, r)
-}
-
-// LivenessCheckHandler - Checks if the process is up. Always returns success.
-func LivenessCheckHandler(w http.ResponseWriter, r *http.Request) {
-	objLayer := newObjectLayerFn()
-	if objLayer == nil {
-		// Service not initialized yet
-		w.Header().Set(xhttp.MinIOServerStatus, unavailable)
+	if objLayer := newObjectLayerFn(); objLayer == nil {
+		w.Header().Set(xhttp.MinIOServerStatus, unavailable) // Service not initialized yet
 	}
-
-	peerCall := r.Header.Get(xhttp.MinIOPeerCall) != ""
-	if peerCall {
+	if r.Header.Get(xhttp.MinIOPeerCall) != "" {
+		writeResponse(w, http.StatusOK, nil, mimeNone)
 		return
 	}

@@ -120,7 +114,7 @@ func LivenessCheckHandler(w http.ResponseWriter, r *http.Request) {
 	}

 	// Verify if KMS is reachable if its configured
-	if GlobalKMS != nil && !peerCall {
+	if GlobalKMS != nil {
 		ctx, cancel := context.WithTimeout(r.Context(), time.Minute)
 		defer cancel()

@@ -153,6 +147,30 @@ func LivenessCheckHandler(w http.ResponseWriter, r *http.Request) {
 			return
 		}
 	}
-
+	writeResponse(w, http.StatusOK, nil, mimeNone)
+}
+
+// LivenessCheckHandler checks whether MinIO is up. It differs from the
+// readiness handler since a failing liveness check causes pod restarts
+// in K8S enviromnents. Therefore, it does not contact external systems.
+func LivenessCheckHandler(w http.ResponseWriter, r *http.Request) {
+	if objLayer := newObjectLayerFn(); objLayer == nil {
+		w.Header().Set(xhttp.MinIOServerStatus, unavailable) // Service not initialized yet
+	}
+	if r.Header.Get(xhttp.MinIOPeerCall) != "" {
+		writeResponse(w, http.StatusOK, nil, mimeNone)
+		return
+	}
+
+	if int(globalHTTPStats.loadRequestsInQueue()) > globalAPIConfig.getRequestsPoolCapacity() {
+		apiErr := getAPIError(ErrBusy)
+		switch r.Method {
+		case http.MethodHead:
+			writeResponse(w, apiErr.HTTPStatusCode, nil, mimeNone)
+		case http.MethodGet:
+			writeErrorResponse(r.Context(), w, apiErr, r.URL)
+		}
+		return
+	}
 	writeResponse(w, http.StatusOK, nil, mimeNone)
 }
@@ -55,6 +55,7 @@ func getOpName(name string) (op string) {
 	op = strings.Replace(op, "(*peerRESTServer)", "peer", 1)
 	op = strings.Replace(op, "(*lockRESTServer)", "lockR", 1)
 	op = strings.Replace(op, "(*stsAPIHandlers)", "sts", 1)
+	op = strings.Replace(op, "(*peerS3Server)", "s3", 1)
 	op = strings.Replace(op, "ClusterCheckHandler", "health.Cluster", 1)
 	op = strings.Replace(op, "ClusterReadCheckHandler", "health.ClusterRead", 1)
 	op = strings.Replace(op, "LivenessCheckHandler", "health.Liveness", 1)
@@ -580,7 +580,7 @@ func listIAMConfigItems(ctx context.Context, objAPI ObjectLayer, pathPrefix stri
 		// Allocate new results channel to receive ObjectInfo.
 		objInfoCh := make(chan ObjectInfo)

-		if err := objAPI.Walk(ctx, minioMetaBucket, pathPrefix, objInfoCh, ObjectOptions{}); err != nil {
+		if err := objAPI.Walk(ctx, minioMetaBucket, pathPrefix, objInfoCh, WalkOptions{}); err != nil {
 			select {
 			case ch <- itemOrErr{Err: err}:
 			case <-ctx.Done():
@@ -488,7 +488,12 @@ func (store *IAMStoreSys) PurgeExpiredSTS(ctx context.Context) error {

 // LoadIAMCache reads all IAM items and populates a new iamCache object and
 // replaces the in-memory cache object.
-func (store *IAMStoreSys) LoadIAMCache(ctx context.Context) error {
+func (store *IAMStoreSys) LoadIAMCache(ctx context.Context, firstTime bool) error {
+	bootstrapTraceMsg := func(s string) {
+		if firstTime {
+			bootstrapTraceMsg(s)
+		}
+	}
 	bootstrapTraceMsg("loading IAM data")

 	newCache := newIamCache()
@@ -655,7 +660,7 @@ func (store *IAMStoreSys) GroupNotificationHandler(ctx context.Context, group st

 // PolicyDBGet - fetches policies associated with the given user or group, and
 // additional groups if provided.
-func (store *IAMStoreSys) PolicyDBGet(name string, isGroup bool, groups ...string) ([]string, error) {
+func (store *IAMStoreSys) PolicyDBGet(name string, groups ...string) ([]string, error) {
 	if name == "" {
 		return nil, errInvalidArgument
 	}
@@ -663,19 +668,17 @@ func (store *IAMStoreSys) PolicyDBGet(name string, isGroup bool, groups ...strin
 	cache := store.rlock()
 	defer store.runlock()

-	policies, _, err := cache.policyDBGet(store, name, isGroup)
+	policies, _, err := cache.policyDBGet(store, name, false)
 	if err != nil {
 		return nil, err
 	}

-	if !isGroup {
-		for _, group := range groups {
-			ps, _, err := cache.policyDBGet(store, group, true)
-			if err != nil {
-				return nil, err
-			}
-			policies = append(policies, ps...)
+	for _, group := range groups {
+		ps, _, err := cache.policyDBGet(store, group, true)
+		if err != nil {
+			return nil, err
 		}
+		policies = append(policies, ps...)
 	}

 	return policies, nil
@@ -1219,6 +1222,9 @@ func (store *IAMStoreSys) GetPolicy(name string) (policy.Policy, error) {
 		}
 		toMerge = append(toMerge, v.Policy)
 	}
+	if len(toMerge) == 0 {
+		return policy.Policy{}, errNoSuchPolicy
+	}
 	return policy.MergePolicies(toMerge...), nil
 }

@@ -2313,6 +2319,27 @@ func (store *IAMStoreSys) ListServiceAccounts(ctx context.Context, accessKey str
 	return serviceAccounts, nil
 }

+// ListSTSAccounts - lists only STS accounts from the cache.
+func (store *IAMStoreSys) ListSTSAccounts(ctx context.Context, accessKey string) ([]auth.Credentials, error) {
+	cache := store.rlock()
+	defer store.runlock()
+
+	var stsAccounts []auth.Credentials
+	for _, u := range cache.iamSTSAccountsMap {
+		v := u.Credentials
+		if accessKey != "" && v.ParentUser == accessKey {
+			if v.IsTemp() {
+				// Hide secret key & session key here
+				v.SecretKey = ""
+				v.SessionToken = ""
+				stsAccounts = append(stsAccounts, v)
+			}
+		}
+	}
+
+	return stsAccounts, nil
+}
+
 // AddUser - adds/updates long term user account to storage.
 func (store *IAMStoreSys) AddUser(ctx context.Context, accessKey string, ureq madmin.AddOrUpdateUserReq) (updatedAt time.Time, err error) {
 	cache := store.lock()
@@ -189,7 +189,7 @@ func (sys *IAMSys) Initialized() bool {
 // Load - loads all credentials, policies and policy mappings.
 func (sys *IAMSys) Load(ctx context.Context, firstTime bool) error {
 	loadStartTime := time.Now()
-	err := sys.store.LoadIAMCache(ctx)
+	err := sys.store.LoadIAMCache(ctx, firstTime)
 	if err != nil {
 		atomic.AddUint64(&sys.TotalRefreshFailures, 1)
 		return err
@@ -1041,7 +1041,7 @@ func (sys *IAMSys) UpdateServiceAccount(ctx context.Context, accessKey string, o
 	return updatedAt, nil
 }

-// ListServiceAccounts - lists all services accounts associated to a specific user
+// ListServiceAccounts - lists all service accounts associated to a specific user
 func (sys *IAMSys) ListServiceAccounts(ctx context.Context, accessKey string) ([]auth.Credentials, error) {
 	if !sys.Initialized() {
 		return nil, errServerNotInitialized
@@ -1055,7 +1055,7 @@ func (sys *IAMSys) ListServiceAccounts(ctx context.Context, accessKey string) ([
 	}
 }

-// ListTempAccounts - lists all services accounts associated to a specific user
+// ListTempAccounts - lists all temporary service accounts associated to a specific user
 func (sys *IAMSys) ListTempAccounts(ctx context.Context, accessKey string) ([]UserIdentity, error) {
 	if !sys.Initialized() {
 		return nil, errServerNotInitialized
@@ -1069,6 +1069,20 @@ func (sys *IAMSys) ListTempAccounts(ctx context.Context, accessKey string) ([]Us
 	}
 }

+// ListSTSAccounts - lists all STS accounts associated to a specific user
+func (sys *IAMSys) ListSTSAccounts(ctx context.Context, accessKey string) ([]auth.Credentials, error) {
+	if !sys.Initialized() {
+		return nil, errServerNotInitialized
+	}
+
+	select {
+	case <-sys.configLoaded:
+		return sys.store.ListSTSAccounts(ctx, accessKey)
+	case <-ctx.Done():
+		return nil, ctx.Err()
+	}
+}
+
 // GetServiceAccount - wrapper method to get information about a service account
 func (sys *IAMSys) GetServiceAccount(ctx context.Context, accessKey string) (auth.Credentials, *policy.Policy, error) {
 	sa, embeddedPolicy, err := sys.getServiceAccount(ctx, accessKey)
@@ -1663,18 +1677,25 @@ func (sys *IAMSys) PolicyDBUpdateLDAP(ctx context.Context, isAttach bool,
 			return
 		}
 		if dn == "" {
-			err = errNoSuchUser
-			return
+			// Still attempt to detach if provided user is a DN.
+			if !isAttach && sys.LDAPConfig.IsLDAPUserDN(r.User) {
+				dn = r.User
+			} else {
+				err = errNoSuchUser
+				return
+			}
 		}
 		isGroup = false
 	} else {
-		var exists bool
-		if exists, err = sys.LDAPConfig.DoesGroupDNExist(r.Group); err != nil {
-			logger.LogIf(ctx, err)
-			return
-		} else if !exists {
-			err = errNoSuchGroup
-			return
+		if isAttach {
+			var exists bool
+			if exists, err = sys.LDAPConfig.DoesGroupDNExist(r.Group); err != nil {
+				logger.LogIf(ctx, err)
+				return
+			} else if !exists {
+				err = errNoSuchGroup
+				return
+			}
 		}
 		dn = r.Group
 		isGroup = true
@@ -1713,12 +1734,12 @@ func (sys *IAMSys) PolicyDBUpdateLDAP(ctx context.Context, isAttach bool,

 // PolicyDBGet - gets policy set on a user or group. If a list of groups is
 // given, policies associated with them are included as well.
-func (sys *IAMSys) PolicyDBGet(name string, isGroup bool, groups ...string) ([]string, error) {
+func (sys *IAMSys) PolicyDBGet(name string, groups ...string) ([]string, error) {
 	if !sys.Initialized() {
 		return nil, errServerNotInitialized
 	}

-	return sys.store.PolicyDBGet(name, isGroup, groups...)
+	return sys.store.PolicyDBGet(name, groups...)
 }

 const sessionPolicyNameExtracted = policy.SessionPolicyName + "-extracted"
@@ -1767,7 +1788,7 @@ func (sys *IAMSys) IsAllowedServiceAccount(args policy.Args, parentUser string)

 	default:
 		// Check policy for parent user of service account.
-		svcPolicies, err = sys.PolicyDBGet(parentUser, false, args.Groups...)
+		svcPolicies, err = sys.PolicyDBGet(parentUser, args.Groups...)
 		if err != nil {
 			logger.LogIf(GlobalContext, err)
 			return false
@@ -1875,7 +1896,7 @@ func (sys *IAMSys) IsAllowedSTS(args policy.Args, parentUser string) bool {
 	default:
 		// Otherwise, inherit parent user's policy
 		var err error
-		policies, err = sys.store.PolicyDBGet(parentUser, false, args.Groups...)
+		policies, err = sys.store.PolicyDBGet(parentUser, args.Groups...)
 		if err != nil {
 			logger.LogIf(GlobalContext, fmt.Errorf("error fetching policies on %s: %v", parentUser, err))
 			return false
@@ -2012,7 +2033,7 @@ func (sys *IAMSys) IsAllowed(args policy.Args) bool {
 	}

 	// Continue with the assumption of a regular user
-	policies, err := sys.PolicyDBGet(args.AccountName, false, args.Groups...)
+	policies, err := sys.PolicyDBGet(args.AccountName, args.Groups...)
 	if err != nil {
 		return false
 	}
@@ -9,20 +9,21 @@ func _() {
 	// Re-run the stringer command to generate them again.
 	var x [1]struct{}
 	_ = x[lcEventSrc_None-0]
-	_ = x[lcEventSrc_Scanner-1]
-	_ = x[lcEventSrc_Decom-2]
-	_ = x[lcEventSrc_Rebal-3]
-	_ = x[lcEventSrc_s3HeadObject-4]
-	_ = x[lcEventSrc_s3GetObject-5]
-	_ = x[lcEventSrc_s3ListObjects-6]
-	_ = x[lcEventSrc_s3PutObject-7]
-	_ = x[lcEventSrc_s3CopyObject-8]
-	_ = x[lcEventSrc_s3CompleteMultipartUpload-9]
+	_ = x[lcEventSrc_Heal-1]
+	_ = x[lcEventSrc_Scanner-2]
+	_ = x[lcEventSrc_Decom-3]
+	_ = x[lcEventSrc_Rebal-4]
+	_ = x[lcEventSrc_s3HeadObject-5]
+	_ = x[lcEventSrc_s3GetObject-6]
+	_ = x[lcEventSrc_s3ListObjects-7]
+	_ = x[lcEventSrc_s3PutObject-8]
+	_ = x[lcEventSrc_s3CopyObject-9]
+	_ = x[lcEventSrc_s3CompleteMultipartUpload-10]
 }

-const _lcEventSrc_name = "NoneScannerDecomRebals3HeadObjects3GetObjects3ListObjectss3PutObjects3CopyObjects3CompleteMultipartUpload"
+const _lcEventSrc_name = "NoneHealScannerDecomRebals3HeadObjects3GetObjects3ListObjectss3PutObjects3CopyObjects3CompleteMultipartUpload"

-var _lcEventSrc_index = [...]uint8{0, 4, 11, 16, 21, 33, 44, 57, 68, 80, 105}
+var _lcEventSrc_index = [...]uint8{0, 4, 8, 15, 20, 25, 37, 48, 61, 72, 84, 109}

 func (i lcEventSrc) String() string {
 	if i >= lcEventSrc(len(_lcEventSrc_index)-1) {
@@ -29,8 +29,7 @@ import (

 const (
 	licUpdateCycle = 24 * time.Hour * 30
-	licRenewURL    = "https://subnet.min.io/api/cluster/renew-license"
-	licRenewURLDev = "http://localhost:9000/api/cluster/renew-license"
+	licRenewPath   = "/api/cluster/renew-license"
 )

 // initlicenseUpdateJob start the periodic license update job in the background.
@@ -82,10 +81,7 @@ func licenceUpdaterLoop(ctx context.Context, objAPI ObjectLayer) {
 func performLicenseUpdate(ctx context.Context, objectAPI ObjectLayer) {
 	// the subnet license renewal api renews the license only
 	// if required e.g. when it is expiring soon
-	url := licRenewURL
-	if globalIsCICD {
-		url = licRenewURLDev
-	}
+	url := globalSubnetConfig.BaseURL + licRenewPath

 	resp, err := globalSubnetConfig.Post(url, nil)
 	if err != nil {
@@ -93,7 +89,7 @@ func performLicenseUpdate(ctx context.Context, objectAPI ObjectLayer) {
 		return
 	}

-	r := gjson.Parse(resp).Get("license")
+	r := gjson.Parse(resp).Get("license_v2")
 	if r.Index == 0 {
 		logger.LogIf(ctx, fmt.Errorf("license not found in response from %s", url))
 		return
@@ -159,7 +159,7 @@ func (l *localLocker) removeEntry(name string, args dsync.LockArgs, lri *[]lockR
 }

 func (l *localLocker) RLock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
-	if len(args.Resources) > 1 {
+	if len(args.Resources) != 1 {
 		return false, fmt.Errorf("internal error: localLocker.RLock called with more than one resource")
 	}

@@ -18,121 +18,84 @@
 package cmd

 import (
-	"bytes"
 	"context"
-	"io"
-	"net/url"
+	"errors"

 	"github.com/minio/minio/internal/dsync"
-	xhttp "github.com/minio/minio/internal/http"
-	"github.com/minio/minio/internal/rest"
+	"github.com/minio/minio/internal/grid"
 )

 // lockRESTClient is authenticable lock REST client
 type lockRESTClient struct {
-	restClient *rest.Client
-	u          *url.URL
-}
-
-func toLockError(err error) error {
-	if err == nil {
-		return nil
-	}
-
-	switch err.Error() {
-	case errLockConflict.Error():
-		return errLockConflict
-	case errLockNotFound.Error():
-		return errLockNotFound
-	}
-	return err
-}
-
-// String stringer *dsync.NetLocker* interface compatible method.
-func (client *lockRESTClient) String() string {
-	return client.u.String()
-}
-
-// Wrapper to restClient.Call to handle network errors, in case of network error the connection is marked disconnected
-// permanently. The only way to restore the connection is at the xl-sets layer by xlsets.monitorAndConnectEndpoints()
-// after verifying format.json
-func (client *lockRESTClient) callWithContext(ctx context.Context, method string, values url.Values, body io.Reader, length int64) (respBody io.ReadCloser, err error) {
-	if values == nil {
-		values = make(url.Values)
-	}
-
-	respBody, err = client.restClient.Call(ctx, method, values, body, length)
-	if err == nil {
-		return respBody, nil
-	}
-
-	return nil, toLockError(err)
+	connection *grid.Connection
 }

 // IsOnline - returns whether REST client failed to connect or not.
-func (client *lockRESTClient) IsOnline() bool {
-	return client.restClient.IsOnline()
+func (c *lockRESTClient) IsOnline() bool {
+	return c.connection.State() == grid.StateConnected
 }

 // Not a local locker
-func (client *lockRESTClient) IsLocal() bool {
+func (c *lockRESTClient) IsLocal() bool {
 	return false
 }

 // Close - marks the client as closed.
-func (client *lockRESTClient) Close() error {
-	client.restClient.Close()
+func (c *lockRESTClient) Close() error {
 	return nil
 }

-// restCall makes a call to the lock REST server.
-func (client *lockRESTClient) restCall(ctx context.Context, call string, args dsync.LockArgs) (reply bool, err error) {
-	argsBytes, err := args.MarshalMsg(metaDataPoolGet()[:0])
+// String - returns the remote host of the connection.
+func (c *lockRESTClient) String() string {
+	return c.connection.Remote
+}
+
+func (c *lockRESTClient) call(ctx context.Context, h *grid.SingleHandler[*dsync.LockArgs, *dsync.LockResp], args *dsync.LockArgs) (ok bool, err error) {
+	r, err := h.Call(ctx, c.connection, args)
 	if err != nil {
 		return false, err
 	}
-	defer metaDataPoolPut(argsBytes)
-	body := bytes.NewReader(argsBytes)
-	respBody, err := client.callWithContext(ctx, call, nil, body, body.Size())
-	defer xhttp.DrainBody(respBody)
-	switch err {
-	case nil:
-		return true, nil
-	case errLockConflict, errLockNotFound:
-		return false, nil
+	defer h.PutResponse(r)
+	ok = r.Code == dsync.RespOK
+	switch r.Code {
+	case dsync.RespLockConflict, dsync.RespLockNotFound, dsync.RespOK:
+	// no error
+	case dsync.RespLockNotInitialized:
+		err = errLockNotInitialized
 	default:
-		return false, err
+		err = errors.New(r.Err)
 	}
+	return ok, err
 }

 // RLock calls read lock REST API.
-func (client *lockRESTClient) RLock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
-	return client.restCall(ctx, lockRESTMethodRLock, args)
+func (c *lockRESTClient) RLock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
+	return c.call(ctx, lockRPCRLock, &args)
 }

 // Lock calls lock REST API.
-func (client *lockRESTClient) Lock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
-	return client.restCall(ctx, lockRESTMethodLock, args)
+func (c *lockRESTClient) Lock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
+	return c.call(ctx, lockRPCLock, &args)
 }

 // RUnlock calls read unlock REST API.
-func (client *lockRESTClient) RUnlock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
-	return client.restCall(ctx, lockRESTMethodRUnlock, args)
+func (c *lockRESTClient) RUnlock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
+	return c.call(ctx, lockRPCRUnlock, &args)
 }

-// RUnlock calls read unlock REST API.
-func (client *lockRESTClient) Refresh(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
-	return client.restCall(ctx, lockRESTMethodRefresh, args)
+// Refresh calls Refresh REST API.
+func (c *lockRESTClient) Refresh(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
+	return c.call(ctx, lockRPCRefresh, &args)
 }

 // Unlock calls write unlock RPC.
-func (client *lockRESTClient) Unlock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
-	return client.restCall(ctx, lockRESTMethodUnlock, args)
+func (c *lockRESTClient) Unlock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
+	return c.call(ctx, lockRPCUnlock, &args)
 }

 // ForceUnlock calls force unlock handler to forcibly unlock an active lock.
-func (client *lockRESTClient) ForceUnlock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
-	return client.restCall(ctx, lockRESTMethodForceUnlock, args)
+func (c *lockRESTClient) ForceUnlock(ctx context.Context, args dsync.LockArgs) (reply bool, err error) {
+	return c.call(ctx, lockRPCForceUnlock, &args)
 }

 func newLockAPI(endpoint Endpoint) dsync.NetLocker {
@@ -143,27 +106,6 @@ func newLockAPI(endpoint Endpoint) dsync.NetLocker {
 }

 // Returns a lock rest client.
-func newlockRESTClient(endpoint Endpoint) *lockRESTClient {
-	serverURL := &url.URL{
-		Scheme: endpoint.Scheme,
-		Host:   endpoint.Host,
-		Path:   pathJoin(lockRESTPrefix, lockRESTVersion),
-	}
-
-	restClient := rest.NewClient(serverURL, globalInternodeTransport, newCachedAuthToken())
-	// Use a separate client to avoid recursive calls.
-	healthClient := rest.NewClient(serverURL, globalInternodeTransport, newCachedAuthToken())
-	healthClient.NoMetrics = true
-	restClient.HealthCheckFn = func() bool {
-		ctx, cancel := context.WithTimeout(context.Background(), restClient.HealthCheckTimeout)
-		defer cancel()
-		respBody, err := healthClient.Call(ctx, lockRESTMethodHealth, nil, nil, -1)
-		xhttp.DrainBody(respBody)
-		return !isNetworkError(err)
-	}
-
-	return &lockRESTClient{u: &url.URL{
-		Scheme: endpoint.Scheme,
-		Host:   endpoint.Host,
-	}, restClient: restClient}
+func newlockRESTClient(ep Endpoint) *lockRESTClient {
+	return &lockRESTClient{globalGrid.Load().Connection(ep.GridHost())}
 }
--- a/Show More
+++ b/Show More