fix: incomingBytes() to update via handleMessages() (#18891)

previous change #18880 was incomplete

.codespellrc

@@ -0,0 +1,8 @@
+
+[codespell]
+# certs_test.go - has lots of ceritificates.
+skip = go.mod,go.sum,*.txt,LICENSE,*.zip,.git,*.pdf,*.svg,.codespellrc,CREDITS,certs_test.go
+check-hidden = true
+ignore-regex = \b(newfolder/afile|filterIn|HelpES)\b
+ignore-words-list = inout,bui,to,bu,te,ot,toi,ist,parms,flate
+

.github/workflows/codespell.yml

@@ -0,0 +1,20 @@
+---
+name: Codespell
+
+on:
+  pull_request:
+    branches: [master]
+
+permissions:
+  contents: read
+
+jobs:
+  codespell:
+    name: Check for spelling errors
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+      - name: Codespell
+        uses: codespell-project/actions-codespell@v2

.github/workflows/issues.yaml

@@ -0,0 +1,18 @@
+# @format
+
+name: Issue Workflow
+
+on:
+  issues:
+    types:
+      - opened
+
+jobs:
+  add-to-project:
+    name: Add issue to project
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/add-to-project@v0.5.0
+        with:
+          project-url: https://github.com/orgs/miniohq/projects/2
+          github-token: ${{ secrets.BOT_PAT }}

.github/workflows/replication.yaml

@@ -54,3 +54,8 @@ jobs:
           sudo sysctl net.ipv6.conf.default.disable_ipv6=0
           make test-site-replication-minio
 
+      - name: Test Versioning
+        run: |
+          sudo sysctl net.ipv6.conf.all.disable_ipv6=0
+          sudo sysctl net.ipv6.conf.default.disable_ipv6=0
+          make test-versioning

.gitignore

@@ -43,3 +43,4 @@ 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

Makefile

@@ -46,7 +46,7 @@ lint-fix: getdeps ## runs golangci-lint suite of linters with automatic fixes
 check: test
 test: verifiers build build-debugging ## builds minio, runs linters, tests
 	@echo "Running unit tests"
-	@MINIO_API_REQUESTS_MAX=10000 CGO_ENABLED=0 go test -tags kqueue ./...
+	@MINIO_API_REQUESTS_MAX=10000 CGO_ENABLED=0 go test -v -tags kqueue ./...
 
 test-root-disable: install-race
 	@echo "Running minio root lockdown tests"
@@ -59,6 +59,10 @@ 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-versioning: install-race
+	@echo "Running minio versioning tests"
+	@env bash $(PWD)/docs/bucket/versioning/versioning-tests.sh
+
 test-configfile: install-race
 	@env bash $(PWD)/docs/distributed/distributed-from-config-file.sh
 
@@ -112,7 +116,6 @@ verify-healing: ## verify healing and replacing disks with minio binary
 	@echo "Verify healing build with race"
 	@GORACE=history_size=7 CGO_ENABLED=1 go build -race -tags kqueue -trimpath --ldflags "$(LDFLAGS)" -o $(PWD)/minio 1>/dev/null
 	@(env bash $(PWD)/buildscripts/verify-healing.sh)
-	@(env bash $(PWD)/buildscripts/unaligned-healing.sh)
 	@(env bash $(PWD)/buildscripts/heal-inconsistent-versions.sh)
 
 verify-healing-with-root-disks: ## verify healing root disks

buildscripts/cross-compile.sh

@@ -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() {

buildscripts/unaligned-healing.sh

@@ -1,177 +0,0 @@
-#!/bin/bash -e
-#
-
-set -E
-set -o pipefail
-set -x
-
-if [ ! -x "$PWD/minio" ]; then
-	echo "minio executable binary not found in current directory"
-	exit 1
-fi
-
-WORK_DIR="$PWD/.verify-$RANDOM"
-MINIO_CONFIG_DIR="$WORK_DIR/.minio"
-MINIO_OLD=("$PWD/minio.RELEASE.2021-11-24T23-19-33Z" --config-dir "$MINIO_CONFIG_DIR" server)
-MINIO=("$PWD/minio" --config-dir "$MINIO_CONFIG_DIR" server)
-
-function download_old_release() {
-	if [ ! -f minio.RELEASE.2021-11-24T23-19-33Z ]; then
-		curl --silent -O https://dl.minio.io/server/minio/release/linux-amd64/archive/minio.RELEASE.2021-11-24T23-19-33Z
-		chmod a+x minio.RELEASE.2021-11-24T23-19-33Z
-	fi
-}
-
-function start_minio_16drive() {
-	start_port=$1
-
-	export MINIO_ROOT_USER=minio
-	export MINIO_ROOT_PASSWORD=minio123
-	export MC_HOST_minio="http://minio:minio123@127.0.0.1:${start_port}/"
-	unset MINIO_KMS_AUTO_ENCRYPTION         # do not auto-encrypt objects
-	export _MINIO_SHARD_DISKTIME_DELTA="5s" # do not change this as its needed for tests
-	export MINIO_CI_CD=1
-
-	MC_BUILD_DIR="mc-$RANDOM"
-	if ! git clone --quiet https://github.com/minio/mc "$MC_BUILD_DIR"; then
-		echo "failed to download https://github.com/minio/mc"
-		purge "${MC_BUILD_DIR}"
-		exit 1
-	fi
-
-	(cd "${MC_BUILD_DIR}" && go build -o "$WORK_DIR/mc")
-
-	# remove mc source.
-	purge "${MC_BUILD_DIR}"
-
-	"${MINIO_OLD[@]}" --address ":$start_port" "${WORK_DIR}/xl{1...16}" >"${WORK_DIR}/server1.log" 2>&1 &
-	pid=$!
-	disown $pid
-	sleep 30
-
-	if ! ps -p ${pid} 1>&2 >/dev/null; then
-		echo "server1 log:"
-		cat "${WORK_DIR}/server1.log"
-		echo "FAILED"
-		purge "$WORK_DIR"
-		exit 1
-	fi
-
-	shred --iterations=1 --size=5241856 - 1>"${WORK_DIR}/unaligned" 2>/dev/null
-	"${WORK_DIR}/mc" mb minio/healing-shard-bucket --quiet
-	"${WORK_DIR}/mc" cp \
-		"${WORK_DIR}/unaligned" \
-		minio/healing-shard-bucket/unaligned \
-		--disable-multipart --quiet
-
-	## "unaligned" object name gets consistently distributed
-	## to disks in following distribution order
-	##
-	## NOTE: if you change the name make sure to change the
-	## distribution order present here
-	##
-	## [15, 16, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14]
-
-	## make sure to remove the "last" data shard
-	rm -rf "${WORK_DIR}/xl14/healing-shard-bucket/unaligned"
-	sleep 10
-	## Heal the shard
-	"${WORK_DIR}/mc" admin heal --quiet --recursive minio/healing-shard-bucket
-	## then remove any other data shard let's pick first disk
-	## - 1st data shard.
-	rm -rf "${WORK_DIR}/xl3/healing-shard-bucket/unaligned"
-	sleep 10
-
-	go install github.com/minio/minio/docs/debugging/s3-check-md5@latest
-	if ! s3-check-md5 \
-		-debug \
-		-access-key minio \
-		-secret-key minio123 \
-		-endpoint http://127.0.0.1:${start_port}/ 2>&1 | grep CORRUPTED; then
-		echo "server1 log:"
-		cat "${WORK_DIR}/server1.log"
-		echo "FAILED"
-		purge "$WORK_DIR"
-		exit 1
-	fi
-
-	pkill minio
-	sleep 3
-
-	"${MINIO[@]}" --address ":$start_port" "${WORK_DIR}/xl{1...16}" >"${WORK_DIR}/server1.log" 2>&1 &
-	pid=$!
-	disown $pid
-	sleep 30
-
-	if ! ps -p ${pid} 1>&2 >/dev/null; then
-		echo "server1 log:"
-		cat "${WORK_DIR}/server1.log"
-		echo "FAILED"
-		purge "$WORK_DIR"
-		exit 1
-	fi
-
-	if ! s3-check-md5 \
-		-debug \
-		-access-key minio \
-		-secret-key minio123 \
-		-endpoint http://127.0.0.1:${start_port}/ 2>&1 | grep INTACT; then
-		echo "server1 log:"
-		cat "${WORK_DIR}/server1.log"
-		echo "FAILED"
-		mkdir -p inspects
-		(
-			cd inspects
-			"${WORK_DIR}/mc" support inspect minio/healing-shard-bucket/unaligned/**
-		)
-
-		"${WORK_DIR}/mc" mb play/inspects
-		"${WORK_DIR}/mc" mirror inspects play/inspects
-
-		purge "$WORK_DIR"
-		exit 1
-	fi
-
-	"${WORK_DIR}/mc" admin heal --quiet --recursive minio/healing-shard-bucket
-
-	if ! s3-check-md5 \
-		-debug \
-		-access-key minio \
-		-secret-key minio123 \
-		-endpoint http://127.0.0.1:${start_port}/ 2>&1 | grep INTACT; then
-		echo "server1 log:"
-		cat "${WORK_DIR}/server1.log"
-		echo "FAILED"
-		mkdir -p inspects
-		(
-			cd inspects
-			"${WORK_DIR}/mc" support inspect minio/healing-shard-bucket/unaligned/**
-		)
-
-		"${WORK_DIR}/mc" mb play/inspects
-		"${WORK_DIR}/mc" mirror inspects play/inspects
-
-		purge "$WORK_DIR"
-		exit 1
-	fi
-
-	pkill minio
-	sleep 3
-}
-
-function main() {
-	download_old_release
-
-	start_port=$(shuf -i 10000-65000 -n 1)
-
-	start_minio_16drive ${start_port}
-}
-
-function purge() {
-	rm -rf "$1"
-}
-
-(main "$@")
-rv=$?
-purge "$WORK_DIR"
-exit "$rv"

cmd/admin-bucket-handlers.go

@@ -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
 	}
 
@@ -680,24 +680,17 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 			if _, ok := bucketMap[bucket]; !ok {
 				opts := MakeBucketOptions{
 					LockEnabled: config.Enabled(),
+					ForceCreate: true, // ignore if it already exists
 				}
 				err = objectAPI.MakeBucket(ctx, bucket, opts)
 				if err != nil {
-					if _, ok := err.(BucketExists); !ok {
-						rpt.SetStatus(bucket, fileName, err)
-						continue
-					}
+					rpt.SetStatus(bucket, fileName, err)
+					continue
 				}
 				v := newBucketMetadata(bucket)
 				bucketMap[bucket] = &v
 			}
 
-			// Deny object locking configuration settings on existing buckets without object lock enabled.
-			if _, _, err = globalBucketMetadataSys.GetObjectLockConfig(bucket); err != nil {
-				rpt.SetStatus(bucket, fileName, err)
-				continue
-			}
-
 			bucketMap[bucket].ObjectLockConfigXML = configData
 			bucketMap[bucket].ObjectLockConfigUpdatedAt = updatedAt
 			rpt.SetStatus(bucket, fileName, nil)
@@ -724,11 +717,11 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				continue
 			}
 			if _, ok := bucketMap[bucket]; !ok {
-				if err = objectAPI.MakeBucket(ctx, bucket, MakeBucketOptions{}); err != nil {
-					if _, ok := err.(BucketExists); !ok {
-						rpt.SetStatus(bucket, fileName, err)
-						continue
-					}
+				if err = objectAPI.MakeBucket(ctx, bucket, MakeBucketOptions{
+					ForceCreate: true, // ignore if it already exists
+				}); err != nil {
+					rpt.SetStatus(bucket, fileName, err)
+					continue
 				}
 				v := newBucketMetadata(bucket)
 				bucketMap[bucket] = &v
@@ -776,12 +769,12 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 
 		// create bucket if it does not exist yet.
 		if _, ok := bucketMap[bucket]; !ok {
-			err = objectAPI.MakeBucket(ctx, bucket, MakeBucketOptions{})
+			err = objectAPI.MakeBucket(ctx, bucket, MakeBucketOptions{
+				ForceCreate: true, // ignore if it already exists
+			})
 			if err != nil {
-				if _, ok := err.(BucketExists); !ok {
-					rpt.SetStatus(bucket, "", err)
-					continue
-				}
+				rpt.SetStatus(bucket, "", err)
+				continue
 			}
 			v := newBucketMetadata(bucket)
 			bucketMap[bucket] = &v

cmd/admin-handlers-idp-ldap.go

@@ -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)
+}

cmd/admin-handlers-users.go

@@ -614,72 +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
@@ -713,23 +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,
@@ -748,32 +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
@@ -781,19 +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)
@@ -814,7 +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
@@ -2514,3 +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{}
+}

cmd/admin-handlers.go

@@ -79,6 +79,213 @@ const (
 	mgmtForceStop   = "forceStop"
 )
 
+// ServerUpdateV2Handler - POST /minio/admin/v3/update?updateURL={updateURL}&type=2
+// ----------
+// updates all minio servers and restarts them gracefully.
+func (a adminAPIHandlers) ServerUpdateV2Handler(w http.ResponseWriter, r *http.Request) {
+	ctx := r.Context()
+
+	objectAPI, _ := validateAdminReq(ctx, w, r, policy.ServerUpdateAdminAction)
+	if objectAPI == nil {
+		return
+	}
+
+	if globalInplaceUpdateDisabled || currentReleaseTime.IsZero() {
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrMethodNotAllowed), r.URL)
+		return
+	}
+
+	vars := mux.Vars(r)
+	updateURL := vars["updateURL"]
+	dryRun := r.Form.Get("dry-run") == "true"
+
+	mode := getMinioMode()
+	if updateURL == "" {
+		updateURL = minioReleaseInfoURL
+		if runtime.GOOS == globalWindowsOSName {
+			updateURL = minioReleaseWindowsInfoURL
+		}
+	}
+
+	u, err := url.Parse(updateURL)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	content, err := downloadReleaseURL(u, updateTimeout, mode)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	sha256Sum, lrTime, releaseInfo, err := parseReleaseData(content)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	u.Path = path.Dir(u.Path) + SlashSeparator + releaseInfo
+	// Download Binary Once
+	binC, bin, err := downloadBinary(u, mode)
+	if err != nil {
+		logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	updateStatus := madmin.ServerUpdateStatusV2{DryRun: dryRun}
+	peerResults := make(map[string]madmin.ServerPeerUpdateStatus)
+
+	local := globalLocalNodeName
+	if local == "" {
+		local = "127.0.0.1"
+	}
+
+	failedClients := make(map[int]struct{})
+
+	if globalIsDistErasure {
+		// Push binary to other servers
+		for idx, nerr := range globalNotificationSys.VerifyBinary(ctx, u, sha256Sum, releaseInfo, binC) {
+			if nerr.Err != nil {
+				peerResults[nerr.Host.String()] = madmin.ServerPeerUpdateStatus{
+					Host:           nerr.Host.String(),
+					Err:            nerr.Err.Error(),
+					CurrentVersion: Version,
+				}
+				failedClients[idx] = struct{}{}
+			} else {
+				peerResults[nerr.Host.String()] = madmin.ServerPeerUpdateStatus{
+					Host:           nerr.Host.String(),
+					CurrentVersion: Version,
+					UpdatedVersion: lrTime.Format(minioReleaseTagTimeLayout),
+				}
+			}
+		}
+	}
+
+	if lrTime.Sub(currentReleaseTime) > 0 {
+		if err = verifyBinary(u, sha256Sum, releaseInfo, mode, bytes.NewReader(bin)); err != nil {
+			peerResults[local] = madmin.ServerPeerUpdateStatus{
+				Host:           local,
+				Err:            err.Error(),
+				CurrentVersion: Version,
+			}
+		} else {
+			peerResults[local] = madmin.ServerPeerUpdateStatus{
+				Host:           local,
+				CurrentVersion: Version,
+				UpdatedVersion: lrTime.Format(minioReleaseTagTimeLayout),
+			}
+		}
+	} else {
+		peerResults[local] = madmin.ServerPeerUpdateStatus{
+			Host:           local,
+			Err:            fmt.Sprintf("server is already running the latest version: %s", Version),
+			CurrentVersion: Version,
+		}
+	}
+
+	if !dryRun {
+		if globalIsDistErasure {
+			ng := WithNPeers(len(globalNotificationSys.peerClients))
+			for idx, client := range globalNotificationSys.peerClients {
+				_, ok := failedClients[idx]
+				if ok {
+					continue
+				}
+				client := client
+				ng.Go(ctx, func() error {
+					return client.CommitBinary(ctx)
+				}, idx, *client.host)
+			}
+
+			for _, nerr := range ng.Wait() {
+				if nerr.Err != nil {
+					prs, ok := peerResults[nerr.Host.String()]
+					if ok {
+						prs.Err = nerr.Err.Error()
+						peerResults[nerr.Host.String()] = prs
+					} else {
+						peerResults[nerr.Host.String()] = madmin.ServerPeerUpdateStatus{
+							Host:           nerr.Host.String(),
+							Err:            nerr.Err.Error(),
+							CurrentVersion: Version,
+							UpdatedVersion: lrTime.Format(minioReleaseTagTimeLayout),
+						}
+					}
+				}
+			}
+		}
+		prs := peerResults[local]
+		if prs.Err == "" {
+			if err = commitBinary(); err != nil {
+				prs.Err = err.Error()
+			}
+			peerResults[local] = prs
+		}
+	}
+
+	prs, ok := peerResults[local]
+	if ok {
+		prs.WaitingDrives = waitingDrivesNode()
+		peerResults[local] = prs
+	}
+
+	if globalIsDistErasure {
+		// Notify all other MinIO peers signal service.
+		ng := WithNPeers(len(globalNotificationSys.peerClients))
+		for idx, client := range globalNotificationSys.peerClients {
+			_, ok := failedClients[idx]
+			if ok {
+				continue
+			}
+			client := client
+			ng.Go(ctx, func() error {
+				prs, ok := peerResults[client.String()]
+				if ok && prs.CurrentVersion != prs.UpdatedVersion && prs.UpdatedVersion != "" {
+					return client.SignalService(serviceRestart, "", dryRun)
+				}
+				return nil
+			}, idx, *client.host)
+		}
+
+		for _, nerr := range ng.Wait() {
+			if nerr.Err != nil {
+				waitingDrives := map[string]madmin.DiskMetrics{}
+				jerr := json.Unmarshal([]byte(nerr.Err.Error()), &waitingDrives)
+				if jerr == nil {
+					prs, ok := peerResults[nerr.Host.String()]
+					if ok {
+						prs.WaitingDrives = waitingDrives
+						peerResults[nerr.Host.String()] = prs
+					}
+					continue
+				}
+			}
+		}
+	}
+
+	for _, pr := range peerResults {
+		updateStatus.Results = append(updateStatus.Results, pr)
+	}
+
+	// Marshal API response
+	jsonBytes, err := json.Marshal(updateStatus)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	writeSuccessResponseJSON(w, jsonBytes)
+
+	if !dryRun {
+		if lrTime.Sub(currentReleaseTime) > 0 {
+			globalServiceSignalCh <- serviceRestart
+		}
+	}
+}
+
 // ServerUpdateHandler - POST /minio/admin/v3/update?updateURL={updateURL}
 // ----------
 // updates all minio servers and restarts them gracefully.
@@ -90,7 +297,7 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 		return
 	}
 
-	if globalInplaceUpdateDisabled {
+	if globalInplaceUpdateDisabled || currentReleaseTime.IsZero() {
 		// if MINIO_UPDATE=off - inplace update is disabled, mostly in containers.
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrMethodNotAllowed), r.URL)
 		return
@@ -124,14 +331,7 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 		return
 	}
 
-	u.Path = path.Dir(u.Path) + SlashSeparator + releaseInfo
-	crTime, err := GetCurrentReleaseTime()
-	if err != nil {
-		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
-		return
-	}
-
-	if lrTime.Sub(crTime) <= 0 {
+	if lrTime.Sub(currentReleaseTime) <= 0 {
 		updateStatus := madmin.ServerUpdateStatus{
 			CurrentVersion: Version,
 			UpdatedVersion: Version,
@@ -148,8 +348,10 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 		return
 	}
 
+	u.Path = path.Dir(u.Path) + SlashSeparator + releaseInfo
+
 	// Download Binary Once
-	reader, err := downloadBinary(u, mode)
+	binC, bin, err := downloadBinary(u, mode)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
@@ -157,7 +359,7 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 	}
 
 	// Push binary to other servers
-	for _, nerr := range globalNotificationSys.VerifyBinary(ctx, u, sha256Sum, releaseInfo, reader) {
+	for _, nerr := range globalNotificationSys.VerifyBinary(ctx, u, sha256Sum, releaseInfo, binC) {
 		if nerr.Err != nil {
 			err := AdminError{
 				Code:       AdminUpdateApplyFailure,
@@ -171,7 +373,7 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 		}
 	}
 
-	err = verifyBinary(u, sha256Sum, releaseInfo, mode, reader)
+	err = verifyBinary(u, sha256Sum, releaseInfo, mode, bytes.NewReader(bin))
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
@@ -227,8 +429,6 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 // ServiceHandler - POST /minio/admin/v3/service?action={action}
 // ----------
 // Supports following actions:
-// - restart (restarts all the MinIO instances in a setup)
-// - stop (stops all the MinIO instances in a setup)
 // - freeze (freezes all incoming S3 API calls)
 // - unfreeze (unfreezes previously frozen S3 API calls)
 func (a adminAPIHandlers) ServiceHandler(w http.ResponseWriter, r *http.Request) {
@@ -238,7 +438,83 @@ func (a adminAPIHandlers) ServiceHandler(w http.ResponseWriter, r *http.Request)
 	action := vars["action"]
 
 	var serviceSig serviceSignal
-	switch madmin.ServiceAction(action) {
+	switch act := madmin.ServiceAction(action); act {
+	case madmin.ServiceActionFreeze:
+		serviceSig = serviceFreeze
+	case madmin.ServiceActionUnfreeze:
+		serviceSig = serviceUnFreeze
+	default:
+		process := act == madmin.ServiceActionRestart || act == madmin.ServiceActionStop
+		if process {
+			apiErr := errorCodes.ToAPIErr(ErrMalformedPOSTRequest)
+			apiErr.Description = "process actions are not supported via this API anymore, please upgrade 'mc' or madmin-go to use ServiceV2() API"
+			writeErrorResponseJSON(ctx, w, apiErr, r.URL)
+			return
+		}
+		logger.LogIf(ctx, fmt.Errorf("Unrecognized service action %s requested", action), logger.Application)
+		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrMalformedPOSTRequest), r.URL)
+		return
+	}
+
+	objectAPI, _ := validateAdminReq(ctx, w, r, policy.ServiceFreezeAdminAction)
+	if objectAPI == nil {
+		return
+	}
+
+	// Notify all other MinIO peers signal service.
+	for _, nerr := range globalNotificationSys.SignalService(serviceSig) {
+		if nerr.Err != nil {
+			logger.GetReqInfo(ctx).SetTags("peerAddress", nerr.Host.String())
+			logger.LogIf(ctx, nerr.Err)
+		}
+	}
+
+	// Reply to the client before restarting, stopping MinIO server.
+	writeSuccessResponseHeadersOnly(w)
+
+	switch serviceSig {
+	case serviceFreeze:
+		freezeServices()
+	case serviceUnFreeze:
+		unfreezeServices()
+	}
+}
+
+type servicePeerResult struct {
+	Host          string                        `json:"host"`
+	Err           string                        `json:"err,omitempty"`
+	WaitingDrives map[string]madmin.DiskMetrics `json:"waitingDrives,omitempty"`
+}
+
+type serviceResult struct {
+	Action  madmin.ServiceAction `json:"action"`
+	DryRun  bool                 `json:"dryRun"`
+	Results []servicePeerResult  `json:"results,omitempty"`
+}
+
+// ServiceV2Handler - POST /minio/admin/v3/service?action={action}&type=2
+// ----------
+// Supports following actions:
+// - restart (restarts all the MinIO instances in a setup)
+// - stop (stops all the MinIO instances in a setup)
+// - freeze (freezes all incoming S3 API calls)
+// - unfreeze (unfreezes previously frozen S3 API calls)
+//
+// This newer API now returns back status per remote peer and local regarding
+// if a "restart/stop" was successful or not. Service signal now supports
+// a dry-run that helps skip the nodes that may have hung drives. By default
+// restart/stop will ignore the servers that are hung on drives. You can use
+// 'force' param to force restart even with hung drives if needed.
+func (a adminAPIHandlers) ServiceV2Handler(w http.ResponseWriter, r *http.Request) {
+	ctx := r.Context()
+
+	vars := mux.Vars(r)
+	action := vars["action"]
+	dryRun := r.Form.Get("dry-run") == "true"
+
+	var serviceSig serviceSignal
+	act := madmin.ServiceAction(action)
+	switch act {
 	case madmin.ServiceActionRestart:
 		serviceSig = serviceRestart
 	case madmin.ServiceActionStop:
@@ -267,15 +543,55 @@ func (a adminAPIHandlers) ServiceHandler(w http.ResponseWriter, r *http.Request)
 	}
 
 	// Notify all other MinIO peers signal service.
-	for _, nerr := range globalNotificationSys.SignalService(serviceSig) {
-		if nerr.Err != nil {
-			logger.GetReqInfo(ctx).SetTags("peerAddress", nerr.Host.String())
-			logger.LogIf(ctx, nerr.Err)
+	srvResult := serviceResult{Action: act, Results: []servicePeerResult{}}
+
+	process := act == madmin.ServiceActionRestart || act == madmin.ServiceActionStop
+	if process {
+		localhost := globalLocalNodeName
+		if globalLocalNodeName == "" {
+			localhost = "127.0.0.1"
+		}
+		waitingDrives := waitingDrivesNode()
+		srvResult.Results = append(srvResult.Results, servicePeerResult{
+			Host:          localhost,
+			WaitingDrives: waitingDrives,
+		})
+	}
+
+	if globalIsDistErasure {
+		for _, nerr := range globalNotificationSys.SignalServiceV2(serviceSig, dryRun) {
+			if nerr.Err != nil && process {
+				waitingDrives := map[string]madmin.DiskMetrics{}
+				jerr := json.Unmarshal([]byte(nerr.Err.Error()), &waitingDrives)
+				if jerr == nil {
+					srvResult.Results = append(srvResult.Results, servicePeerResult{
+						Host:          nerr.Host.String(),
+						WaitingDrives: waitingDrives,
+					})
+					continue
+				}
+			}
+			errStr := ""
+			if nerr.Err != nil {
+				errStr = nerr.Err.Error()
+			}
+			srvResult.Results = append(srvResult.Results, servicePeerResult{
+				Host: nerr.Host.String(),
+				Err:  errStr,
+			})
 		}
 	}
 
+	srvResult.DryRun = dryRun
+
+	buf, err := json.Marshal(srvResult)
+	if err != nil {
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
 	// Reply to the client before restarting, stopping MinIO server.
-	writeSuccessResponseHeadersOnly(w)
+	writeSuccessResponseJSON(w, buf)
 
 	switch serviceSig {
 	case serviceFreeze:
@@ -283,7 +599,9 @@ func (a adminAPIHandlers) ServiceHandler(w http.ResponseWriter, r *http.Request)
 	case serviceUnFreeze:
 		unfreezeServices()
 	case serviceRestart, serviceStop:
-		globalServiceSignalCh <- serviceSig
+		if !dryRun {
+			globalServiceSignalCh <- serviceSig
+		}
 	}
 }
 
@@ -340,7 +658,7 @@ 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)
@@ -1297,6 +1615,7 @@ 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 {
@@ -1313,7 +1632,7 @@ 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 {
@@ -1366,6 +1685,7 @@ func (a adminAPIHandlers) ObjectSpeedTestHandler(w http.ResponseWriter, r *http.
 		autotune:         autotune,
 		storageClass:     storageClass,
 		bucketName:       customBucket,
+		enableSha256:     enableSha256,
 	})
 	var prevResult madmin.SpeedTestResult
 	for {
@@ -1911,7 +2231,7 @@ func getServerInfo(ctx context.Context, poolsInfoEnabled bool, r *http.Request)
 		}
 	}
 
-	log, audit := fetchLoggerInfo()
+	log, audit := fetchLoggerInfo(ctx)
 
 	// Get the notification target info
 	notifyTarget := fetchLambdaInfo()
@@ -2337,7 +2657,7 @@ func fetchHealthInfo(healthCtx context.Context, objectAPI ObjectLayer, query *ur
 	}
 
 	go func() {
-		defer close(healthInfoCh)
+		defer xioutil.SafeClose(healthInfoCh)
 
 		partialWrite(healthInfo) // Write first message with only version and deployment id populated
 		getAndWritePlatformInfo()
@@ -2660,19 +2980,26 @@ 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)})
 		}
 	}
 
@@ -2719,14 +3046,14 @@ func getClusterMetaInfo(ctx context.Context) []byte {
 	resultCh := make(chan madmin.ClusterRegistrationInfo)
 
 	go func() {
-		defer close(resultCh)
+		defer xioutil.SafeClose(resultCh)
 
 		ci := madmin.ClusterRegistrationInfo{}
 		ci.Info.NoOfServerPools = len(globalEndpoints)
-		ci.Info.NoOfServers = len(globalEndpoints.Hostnames())
+		ci.Info.NoOfServers = totalNodeCount()
 		ci.Info.MinioVersion = Version
 
-		si := objectAPI.StorageInfo(ctx)
+		si := objectAPI.StorageInfo(ctx, true)
 
 		ci.Info.NoOfDrives = len(si.Disks)
 		for _, disk := range si.Disks {
@@ -2963,7 +3290,7 @@ func (a adminAPIHandlers) InspectDataHandler(w http.ResponseWriter, r *http.Requ
 	}
 
 	// save the format.json as part of inspect by default
-	if volume != minioMetaBucket && file != formatConfigFile {
+	if !(volume == minioMetaBucket && file == formatConfigFile) {
 		err = o.GetRawData(ctx, minioMetaBucket, formatConfigFile, rawDataFn)
 	}
 	if !errors.Is(err, errFileNotFound) {
@@ -3045,7 +3372,7 @@ func createHostAnonymizerForFSMode() map[string]string {
 // anonymizeHost - Add entries related to given endpoint in the host anonymizer map
 // The health report data can contain the hostname in various forms e.g. host, host:port,
 // host:port/drivepath, full url (http://host:port/drivepath)
-// The anonymizer map will have mappings for all these varients for efficiently replacing
+// The anonymizer map will have mappings for all these variants for efficiently replacing
 // any of these strings to the anonymized versions at the time of health report generation.
 func anonymizeHost(hostAnonymizer map[string]string, endpoint Endpoint, poolNum int, srvrNum int) {
 	if len(endpoint.Host) == 0 {
@@ -3093,7 +3420,7 @@ func anonymizeHost(hostAnonymizer map[string]string, endpoint Endpoint, poolNum
 	}
 }
 
-// createHostAnonymizer - Creats a map of various strings to corresponding anonymized names
+// createHostAnonymizer - Creates a map of various strings to corresponding anonymized names
 func createHostAnonymizer() map[string]string {
 	if !globalIsDistErasure {
 		return createHostAnonymizerForFSMode()

cmd/admin-handlers_test.go

@@ -168,6 +168,7 @@ func testServiceSignalReceiver(cmd cmdType, t *testing.T) {
 func getServiceCmdRequest(cmd cmdType, cred auth.Credentials) (*http.Request, error) {
 	queryVal := url.Values{}
 	queryVal.Set("action", string(cmd.toServiceAction()))
+	queryVal.Set("type", "2")
 	resource := adminPathPrefix + adminAPIVersionPrefix + "/service?" + queryVal.Encode()
 	req, err := newTestRequest(http.MethodPost, resource, 0, nil)
 	if err != nil {
@@ -220,12 +221,18 @@ func testServicesCmdHandler(cmd cmdType, t *testing.T) {
 	rec := httptest.NewRecorder()
 	adminTestBed.router.ServeHTTP(rec, req)
 
+	resp, _ := io.ReadAll(rec.Body)
 	if rec.Code != http.StatusOK {
-		resp, _ := io.ReadAll(rec.Body)
 		t.Errorf("Expected to receive %d status code but received %d. Body (%s)",
 			http.StatusOK, rec.Code, string(resp))
 	}
 
+	result := &serviceResult{}
+	if err := json.Unmarshal(resp, result); err != nil {
+		t.Error(err)
+	}
+	_ = result
+
 	// Wait until testServiceSignalReceiver() called in a goroutine quits.
 	wg.Wait()
 }

cmd/admin-heal-ops.go

@@ -28,6 +28,7 @@ import (
 	"time"
 
 	"github.com/minio/madmin-go/v3"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 )
 
@@ -800,7 +801,7 @@ func (h *healSequence) healItems(objAPI ObjectLayer, bucketsOnly bool) error {
 func (h *healSequence) traverseAndHeal(objAPI ObjectLayer) {
 	bucketsOnly := false // Heals buckets and objects also.
 	h.traverseAndHealDoneCh <- h.healItems(objAPI, bucketsOnly)
-	close(h.traverseAndHealDoneCh)
+	xioutil.SafeClose(h.traverseAndHealDoneCh)
 }
 
 // healMinioSysMeta - heals all files under a given meta prefix, returns a function

cmd/admin-router.go

@@ -154,9 +154,16 @@ func registerAdminRouter(router *mux.Router, enableConfigOps bool) {
 	}
 
 	for _, adminVersion := range adminVersions {
-		// Restart and stop MinIO service.
+		// Restart and stop MinIO service type=2
+		adminRouter.Methods(http.MethodPost).Path(adminVersion+"/service").HandlerFunc(adminMiddleware(adminAPI.ServiceV2Handler, traceAllFlag)).Queries("action", "{action:.*}", "type", "2")
+
+		// Deprecated: Restart and stop MinIO service.
 		adminRouter.Methods(http.MethodPost).Path(adminVersion+"/service").HandlerFunc(adminMiddleware(adminAPI.ServiceHandler, traceAllFlag)).Queries("action", "{action:.*}")
-		// Update MinIO servers.
+
+		// Update all MinIO servers type=2
+		adminRouter.Methods(http.MethodPost).Path(adminVersion+"/update").HandlerFunc(adminMiddleware(adminAPI.ServerUpdateV2Handler, traceAllFlag)).Queries("updateURL", "{updateURL:.*}", "type", "2")
+
+		// Deprecated: Update MinIO servers.
 		adminRouter.Methods(http.MethodPost).Path(adminVersion+"/update").HandlerFunc(adminMiddleware(adminAPI.ServerUpdateHandler, traceAllFlag)).Queries("updateURL", "{updateURL:.*}")
 
 		// Info operations
@@ -300,6 +307,12 @@ 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))

cmd/admin-server-info.go

@@ -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 {

cmd/api-errors.go

@@ -430,6 +430,9 @@ const (
 	ErrLambdaARNInvalid
 	ErrLambdaARNNotFound
 
+	// New Codes for GetObjectAttributes and GetObjectVersionAttributes
+	ErrInvalidAttributeName
+
 	apiErrCodeEnd // This is used only for the testing code
 )
 
@@ -956,7 +959,7 @@ var errorCodes = errorCodeMap{
 	},
 	ErrReplicationBandwidthLimitError: {
 		Code:           "XMinioAdminReplicationBandwidthLimitError",
-		Description:    "Bandwidth limit for remote target must be atleast 100MBps",
+		Description:    "Bandwidth limit for remote target must be at least 100MBps",
 		HTTPStatusCode: http.StatusBadRequest,
 	},
 	ErrReplicationNoExistingObjects: {
@@ -2063,6 +2066,11 @@ var errorCodes = errorCodeMap{
 		Description:    "The specified policy is not found.",
 		HTTPStatusCode: http.StatusNotFound,
 	},
+	ErrInvalidAttributeName: {
+		Code:           "InvalidArgument",
+		Description:    "Invalid attribute name specified.",
+		HTTPStatusCode: http.StatusBadRequest,
+	},
 	// Add your error structure here.
 }
 

cmd/api-router.go

@@ -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()
@@ -227,6 +227,9 @@ func registerAPIRouter(router *mux.Router) {
 		// HeadObject
 		router.Methods(http.MethodHead).Path("/{object:.+}").HandlerFunc(
 			collectAPIStats("headobject", maxClients(gz(httpTraceAll(api.HeadObjectHandler)))))
+		// GetObjectAttribytes
+		router.Methods(http.MethodGet).Path("/{object:.+}").HandlerFunc(
+			collectAPIStats("getobjectattributes", maxClients(gz(httpTraceHdrs(api.GetObjectAttributesHandler))))).Queries("attributes", "")
 		// CopyObjectPart
 		router.Methods(http.MethodPut).Path("/{object:.+}").
 			HeadersRegexp(xhttp.AmzCopySource, ".*?(\\/|%2F).*?").

cmd/auth-handler_test.go

@@ -287,7 +287,7 @@ func mustNewSignedRequest(method string, urlStr string, contentLength int64, bod
 	req := mustNewRequest(method, urlStr, contentLength, body, t)
 	cred := globalActiveCred
 	if err := signRequestV4(req, cred.AccessKey, cred.SecretKey); err != nil {
-		t.Fatalf("Unable to inititalized new signed http request %s", err)
+		t.Fatalf("Unable to initialized new signed http request %s", err)
 	}
 	return req
 }
@@ -298,7 +298,7 @@ func mustNewSignedV2Request(method string, urlStr string, contentLength int64, b
 	req := mustNewRequest(method, urlStr, contentLength, body, t)
 	cred := globalActiveCred
 	if err := signRequestV2(req, cred.AccessKey, cred.SecretKey); err != nil {
-		t.Fatalf("Unable to inititalized new signed http request %s", err)
+		t.Fatalf("Unable to initialized new signed http request %s", err)
 	}
 	return req
 }
@@ -309,7 +309,7 @@ func mustNewPresignedV2Request(method string, urlStr string, contentLength int64
 	req := mustNewRequest(method, urlStr, contentLength, body, t)
 	cred := globalActiveCred
 	if err := preSignV2(req, cred.AccessKey, cred.SecretKey, time.Now().Add(10*time.Minute).Unix()); err != nil {
-		t.Fatalf("Unable to inititalized new signed http request %s", err)
+		t.Fatalf("Unable to initialized new signed http request %s", err)
 	}
 	return req
 }
@@ -320,7 +320,7 @@ func mustNewPresignedRequest(method string, urlStr string, contentLength int64,
 	req := mustNewRequest(method, urlStr, contentLength, body, t)
 	cred := globalActiveCred
 	if err := preSignV4(req, cred.AccessKey, cred.SecretKey, time.Now().Add(10*time.Minute).Unix()); err != nil {
-		t.Fatalf("Unable to inititalized new signed http request %s", err)
+		t.Fatalf("Unable to initialized new signed http request %s", err)
 	}
 	return req
 }
@@ -491,7 +491,7 @@ func TestValidateAdminSignature(t *testing.T) {
 	for i, testCase := range testCases {
 		req := mustNewRequest(http.MethodGet, "http://localhost:9000/", 0, nil, t)
 		if err := signRequestV4(req, testCase.AccessKey, testCase.SecretKey); err != nil {
-			t.Fatalf("Unable to inititalized new signed http request %s", err)
+			t.Fatalf("Unable to initialized new signed http request %s", err)
 		}
 		_, _, s3Error := validateAdminSignature(ctx, req, globalMinioDefaultRegion)
 		if s3Error != testCase.ErrCode {

cmd/background-newdisks-heal-ops.go

@@ -347,7 +347,6 @@ func initAutoHeal(ctx context.Context, objAPI ObjectLayer) {
 
 	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)
 	}
 }
@@ -377,7 +376,7 @@ func getLocalDisksToHeal() (disksToHeal Endpoints) {
 var newDiskHealingTimeout = newDynamicTimeout(30*time.Second, 10*time.Second)
 
 func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint) error {
-	disk, format, err := connectEndpoint(endpoint)
+	disk, format, _, err := connectEndpoint(endpoint)
 	if err != nil {
 		return fmt.Errorf("Error: %w, %s", err, endpoint)
 	}
@@ -491,6 +490,10 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	}
 
 	for _, disk := range disks {
+		if disk == nil {
+			continue
+		}
+
 		t, err := loadHealingTracker(ctx, disk)
 		if err != nil {
 			if !errors.Is(err, errFileNotFound) {

cmd/batch-expire.go

@@ -32,10 +32,12 @@ import (
 	"github.com/minio/minio-go/v7/pkg/tags"
 	"github.com/minio/minio/internal/bucket/versioning"
 	xhttp "github.com/minio/minio/internal/http"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/env"
 	"github.com/minio/pkg/v2/wildcard"
 	"github.com/minio/pkg/v2/workers"
+	"gopkg.in/yaml.v3"
 )
 
 // expire: # Expire objects that match a condition
@@ -43,7 +45,7 @@ import (
 //   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
+//     - type: object  # regular objects with zero or 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"
@@ -80,19 +82,41 @@ import (
 
 // BatchJobExpirePurge type accepts non-negative versions to be retained
 type BatchJobExpirePurge struct {
+	line, col      int
 	RetainVersions int `yaml:"retainVersions" json:"retainVersions"`
 }
 
+var _ yaml.Unmarshaler = &BatchJobExpirePurge{}
+
+// UnmarshalYAML - BatchJobExpirePurge extends unmarshal to extract line, col
+func (p *BatchJobExpirePurge) UnmarshalYAML(val *yaml.Node) error {
+	type purge BatchJobExpirePurge
+	var tmp purge
+	err := val.Decode(&tmp)
+	if err != nil {
+		return err
+	}
+
+	*p = BatchJobExpirePurge(tmp)
+	p.line, p.col = val.Line, val.Column
+	return nil
+}
+
 // 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 BatchJobYamlErr{
+			line: p.line,
+			col:  p.col,
+			msg:  "retainVersions must be >= 0",
+		}
 	}
 	return nil
 }
 
 // BatchJobExpireFilter holds all the filters currently supported for batch replication
 type BatchJobExpireFilter struct {
+	line, col     int
 	OlderThan     time.Duration       `yaml:"olderThan,omitempty" json:"olderThan"`
 	CreatedBefore *time.Time          `yaml:"createdBefore,omitempty" json:"createdBefore"`
 	Tags          []BatchJobKV        `yaml:"tags,omitempty" json:"tags"`
@@ -103,6 +127,22 @@ type BatchJobExpireFilter struct {
 	Purge         BatchJobExpirePurge `yaml:"purge" json:"purge"`
 }
 
+var _ yaml.Unmarshaler = &BatchJobExpireFilter{}
+
+// UnmarshalYAML - BatchJobExpireFilter extends unmarshal to extract line, col
+// information
+func (ef *BatchJobExpireFilter) UnmarshalYAML(value *yaml.Node) error {
+	type expFilter BatchJobExpireFilter
+	var tmp expFilter
+	err := value.Decode(&tmp)
+	if err != nil {
+		return err
+	}
+	*ef = BatchJobExpireFilter(tmp)
+	ef.line, ef.col = value.Line, value.Column
+	return err
+}
+
 // 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 {
@@ -194,10 +234,18 @@ func (ef BatchJobExpireFilter) Validate() error {
 	case BatchJobExpireObject:
 	case BatchJobExpireDeleted:
 		if len(ef.Tags) > 0 || len(ef.Metadata) > 0 {
-			return errors.New("invalid batch-expire rule filter")
+			return BatchJobYamlErr{
+				line: ef.line,
+				col:  ef.col,
+				msg:  "delete type filter can't have tags or metadata",
+			}
 		}
 	default:
-		return errors.New("invalid batch-expire type")
+		return BatchJobYamlErr{
+			line: ef.line,
+			col:  ef.col,
+			msg:  "invalid batch-expire type",
+		}
 	}
 
 	for _, tag := range ef.Tags {
@@ -218,7 +266,11 @@ func (ef BatchJobExpireFilter) Validate() error {
 		return err
 	}
 	if ef.CreatedBefore != nil && !ef.CreatedBefore.Before(time.Now()) {
-		return errors.New("CreatedBefore is in the future")
+		return BatchJobYamlErr{
+			line: ef.line,
+			col:  ef.col,
+			msg:  "CreatedBefore is in the future",
+		}
 	}
 	return nil
 }
@@ -226,6 +278,7 @@ func (ef BatchJobExpireFilter) Validate() error {
 // BatchJobExpire represents configuration parameters for a batch expiration
 // job typically supplied in yaml form
 type BatchJobExpire struct {
+	line, col       int
 	APIVersion      string                 `yaml:"apiVersion" json:"apiVersion"`
 	Bucket          string                 `yaml:"bucket" json:"bucket"`
 	Prefix          string                 `yaml:"prefix" json:"prefix"`
@@ -234,6 +287,22 @@ type BatchJobExpire struct {
 	Rules           []BatchJobExpireFilter `yaml:"rules" json:"rules"`
 }
 
+var _ yaml.Unmarshaler = &BatchJobExpire{}
+
+// UnmarshalYAML - BatchJobExpire extends default unmarshal to extract line, col information.
+func (r *BatchJobExpire) UnmarshalYAML(val *yaml.Node) error {
+	type expireJob BatchJobExpire
+	var tmp expireJob
+	err := val.Decode(&tmp)
+	if err != nil {
+		return err
+	}
+
+	*r = BatchJobExpire(tmp)
+	r.line, r.col = val.Line, val.Column
+	return nil
+}
+
 // 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 == "" {
@@ -490,7 +559,7 @@ func (r *BatchJobExpire) Start(ctx context.Context, api ObjectLayer, job BatchJo
 				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.
+				// persist in-memory state immediately before exiting due to context cancellation.
 				logger.LogIf(ctx, ri.updateAfter(ctx, api, 0, job))
 				return
 
@@ -580,7 +649,7 @@ func (r *BatchJobExpire) Start(ctx context.Context, api ObjectLayer, job BatchJo
 		case expireCh <- toDel:
 		}
 	}
-	close(expireCh)
+	xioutil.SafeClose(expireCh)
 
 	wk.Wait() // waits for all expire goroutines to complete
 
@@ -590,7 +659,7 @@ func (r *BatchJobExpire) Start(ctx context.Context, api ObjectLayer, job BatchJo
 
 	// Close the saverQuitCh - this also triggers saving in-memory state
 	// immediately one last time before we exit this method.
-	close(saverQuitCh)
+	xioutil.SafeClose(saverQuitCh)
 
 	// Notify expire jobs final status to the configured endpoint
 	buf, _ := json.Marshal(ri)

cmd/batch-expire_test.go

@@ -30,7 +30,7 @@ expire: # Expire objects that match a condition
   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
+    - type: object  # regular objects with zero or 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"

cmd/batch-handlers.go

@@ -47,12 +47,13 @@ import (
 	"github.com/minio/minio/internal/hash"
 	xhttp "github.com/minio/minio/internal/http"
 	"github.com/minio/minio/internal/ioutil"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/console"
 	"github.com/minio/pkg/v2/env"
 	"github.com/minio/pkg/v2/policy"
 	"github.com/minio/pkg/v2/workers"
-	"gopkg.in/yaml.v2"
+	"gopkg.in/yaml.v3"
 )
 
 var globalBatchConfig batch.Config
@@ -545,7 +546,7 @@ func (r BatchJobReplicateV1) writeAsArchive(ctx context.Context, objAPI ObjectLa
 	}
 
 	go func() {
-		defer close(input)
+		defer xioutil.SafeClose(input)
 
 		for _, entry := range entries {
 			gr, err := objAPI.GetObjectNInfo(ctx, r.Source.Bucket,
@@ -1038,7 +1039,7 @@ func (r *BatchJobReplicateV1) Start(ctx context.Context, api ObjectLayer, job Ba
 
 	if !*r.Source.Snowball.Disable && r.Source.Type.isMinio() && r.Target.Type.isMinio() {
 		go func() {
-			defer close(slowCh)
+			defer xioutil.SafeClose(slowCh)
 
 			// Snowball currently needs the high level minio-go Client, not the Core one
 			cl, err := miniogo.New(u.Host, &miniogo.Options{
@@ -1564,7 +1565,7 @@ func (a adminAPIHandlers) StartBatchJob(w http.ResponseWriter, r *http.Request)
 	}
 
 	job := &BatchJobRequest{}
-	if err = yaml.UnmarshalStrict(buf, job); err != nil {
+	if err = yaml.Unmarshal(buf, job); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -1809,7 +1810,7 @@ func (j *BatchJobPool) queueJob(req *BatchJobRequest) error {
 	select {
 	case <-j.ctx.Done():
 		j.once.Do(func() {
-			close(j.jobCh)
+			xioutil.SafeClose(j.jobCh)
 		})
 	case j.jobCh <- req:
 	default:

cmd/batch-job-common-types.go

@@ -18,26 +18,66 @@
 package cmd
 
 import (
-	"errors"
+	"fmt"
 	"strings"
 	"time"
 
 	"github.com/dustin/go-humanize"
 	"github.com/minio/pkg/v2/wildcard"
+	"gopkg.in/yaml.v3"
 )
 
 //go:generate msgp -file $GOFILE
+//msgp:ignore BatchJobYamlErr
+
+// BatchJobYamlErr can be used to return yaml validation errors with line,
+// column information guiding user to fix syntax errors
+type BatchJobYamlErr struct {
+	line, col int
+	msg       string
+}
+
+// message returns the error message excluding line, col information.
+// Intended to be used in unit tests.
+func (b BatchJobYamlErr) message() string {
+	return b.msg
+}
+
+// Error implements Error interface
+func (b BatchJobYamlErr) Error() string {
+	return fmt.Sprintf("%s\n Hint: error near line: %d, col: %d", b.msg, b.line, b.col)
+}
 
 // BatchJobKV is a key-value data type which supports wildcard matching
 type BatchJobKV struct {
-	Key   string `yaml:"key" json:"key"`
-	Value string `yaml:"value" json:"value"`
+	line, col int
+	Key       string `yaml:"key" json:"key"`
+	Value     string `yaml:"value" json:"value"`
+}
+
+var _ yaml.Unmarshaler = &BatchJobKV{}
+
+// UnmarshalYAML - BatchJobKV extends default unmarshal to extract line, col information.
+func (kv *BatchJobKV) UnmarshalYAML(val *yaml.Node) error {
+	type jobKV BatchJobKV
+	var tmp jobKV
+	err := val.Decode(&tmp)
+	if err != nil {
+		return err
+	}
+	*kv = BatchJobKV(tmp)
+	kv.line, kv.col = val.Line, val.Column
+	return nil
 }
 
 // Validate returns an error if key is empty
 func (kv BatchJobKV) Validate() error {
 	if kv.Key == "" {
-		return errInvalidArgument
+		return BatchJobYamlErr{
+			line: kv.line,
+			col:  kv.col,
+			msg:  "key can't be empty",
+		}
 	}
 	return nil
 }
@@ -61,24 +101,66 @@ func (kv BatchJobKV) Match(ikv BatchJobKV) bool {
 // BatchJobNotification stores notification endpoint and token information.
 // Used by batch jobs to notify of their status.
 type BatchJobNotification struct {
-	Endpoint string `yaml:"endpoint" json:"endpoint"`
-	Token    string `yaml:"token" json:"token"`
+	line, col int
+	Endpoint  string `yaml:"endpoint" json:"endpoint"`
+	Token     string `yaml:"token" json:"token"`
+}
+
+var _ yaml.Unmarshaler = &BatchJobNotification{}
+
+// UnmarshalYAML - BatchJobNotification extends unmarshal to extract line, column information
+func (b *BatchJobNotification) UnmarshalYAML(val *yaml.Node) error {
+	type notification BatchJobNotification
+	var tmp notification
+	err := val.Decode(&tmp)
+	if err != nil {
+		return err
+	}
+
+	*b = BatchJobNotification(tmp)
+	b.line, b.col = val.Line, val.Column
+	return nil
 }
 
 // BatchJobRetry stores retry configuration used in the event of failures.
 type BatchJobRetry struct {
-	Attempts int           `yaml:"attempts" json:"attempts"` // number of retry attempts
-	Delay    time.Duration `yaml:"delay" json:"delay"`       // delay between each retries
+	line, col int
+	Attempts  int           `yaml:"attempts" json:"attempts"` // number of retry attempts
+	Delay     time.Duration `yaml:"delay" json:"delay"`       // delay between each retries
+}
+
+var _ yaml.Unmarshaler = &BatchJobRetry{}
+
+// UnmarshalYAML - BatchJobRetry extends unmarshal to extract line, column information
+func (r *BatchJobRetry) UnmarshalYAML(val *yaml.Node) error {
+	type retry BatchJobRetry
+	var tmp retry
+	err := val.Decode(&tmp)
+	if err != nil {
+		return err
+	}
+
+	*r = BatchJobRetry(tmp)
+	r.line, r.col = val.Line, val.Column
+	return nil
 }
 
 // Validate validates input replicate retries.
 func (r BatchJobRetry) Validate() error {
 	if r.Attempts < 0 {
-		return errInvalidArgument
+		return BatchJobYamlErr{
+			line: r.line,
+			col:  r.col,
+			msg:  "Invalid arguments specified",
+		}
 	}
 
 	if r.Delay < 0 {
-		return errInvalidArgument
+		return BatchJobYamlErr{
+			line: r.line,
+			col:  r.col,
+			msg:  "Invalid arguments specified",
+		}
 	}
 
 	return nil
@@ -96,6 +178,7 @@ func (r BatchJobRetry) Validate() error {
 
 // BatchJobSnowball describes the snowball feature when replicating objects from a local source to a remote target
 type BatchJobSnowball struct {
+	line, col   int
 	Disable     *bool   `yaml:"disable" json:"disable"`
 	Batch       *int    `yaml:"batch" json:"batch"`
 	InMemory    *bool   `yaml:"inmemory" json:"inmemory"`
@@ -104,21 +187,60 @@ type BatchJobSnowball struct {
 	SkipErrs    *bool   `yaml:"skipErrs" json:"skipErrs"`
 }
 
+var _ yaml.Unmarshaler = &BatchJobSnowball{}
+
+// UnmarshalYAML - BatchJobSnowball extends unmarshal to extract line, column information
+func (b *BatchJobSnowball) UnmarshalYAML(val *yaml.Node) error {
+	type snowball BatchJobSnowball
+	var tmp snowball
+	err := val.Decode(&tmp)
+	if err != nil {
+		return err
+	}
+
+	*b = BatchJobSnowball(tmp)
+	b.line, b.col = val.Line, val.Column
+	return nil
+}
+
 // 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")
+		return BatchJobYamlErr{
+			line: b.line,
+			col:  b.col,
+			msg:  "batch number should be non positive zero",
+		}
 	}
 	_, err := humanize.ParseBytes(*b.SmallerThan)
-	return err
+	return BatchJobYamlErr{
+		line: b.line,
+		col:  b.col,
+		msg:  err.Error(),
+	}
 }
 
 // BatchJobSizeFilter supports size based filters - LesserThan and GreaterThan
 type BatchJobSizeFilter struct {
+	line, col  int
 	UpperBound BatchJobSize `yaml:"lessThan" json:"lessThan"`
 	LowerBound BatchJobSize `yaml:"greaterThan" json:"greaterThan"`
 }
 
+// UnmarshalYAML - BatchJobSizeFilter extends unmarshal to extract line, column information
+func (sf *BatchJobSizeFilter) UnmarshalYAML(val *yaml.Node) error {
+	type sizeFilter BatchJobSizeFilter
+	var tmp sizeFilter
+	err := val.Decode(&tmp)
+	if err != nil {
+		return err
+	}
+
+	*sf = BatchJobSizeFilter(tmp)
+	sf.line, sf.col = val.Line, val.Column
+	return nil
+}
+
 // 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
@@ -134,12 +256,14 @@ func (sf BatchJobSizeFilter) InRange(sz int64) bool {
 	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 BatchJobYamlErr{
+			line: sf.line,
+			col:  sf.col,
+			msg:  "invalid batch-job size filter",
+		}
 	}
 	return nil
 }

cmd/batch-job-common-types_test.go

@@ -83,6 +83,10 @@ func TestBatchJobSizeInRange(t *testing.T) {
 }
 
 func TestBatchJobSizeValidate(t *testing.T) {
+	errInvalidBatchJobSizeFilter := BatchJobYamlErr{
+		msg: "invalid batch-job size filter",
+	}
+
 	tests := []struct {
 		sizeFilter BatchJobSizeFilter
 		err        error
@@ -128,8 +132,16 @@ func TestBatchJobSizeValidate(t *testing.T) {
 	}
 	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)
+			err := test.sizeFilter.Validate()
+			if err != nil {
+				gotErr := err.(BatchJobYamlErr)
+				testErr := test.err.(BatchJobYamlErr)
+				if gotErr.message() != testErr.message() {
+					t.Fatalf("Expected %v but got %v", test.err, err)
+				}
+			}
+			if err == nil && test.err != nil {
+				t.Fatalf("Expected %v but got nil", test.err)
 			}
 		})
 	}

cmd/bitrot-streaming.go

@@ -134,7 +134,7 @@ func (b *streamingBitrotReader) Close() error {
 		return nil
 	}
 	if closer, ok := b.rc.(io.Closer); ok {
-		// drain the body for connection re-use at network layer.
+		// drain the body for connection reuse at network layer.
 		xhttp.DrainBody(struct {
 			io.Reader
 			io.Closer

cmd/bootstrap-peer-server.go

@@ -19,71 +19,47 @@ package cmd
 
 import (
 	"context"
-	"encoding/json"
 	"errors"
 	"fmt"
-	"io"
-	"net/http"
-	"net/url"
-	"os"
+	"math/rand"
 	"reflect"
 	"strings"
+	"sync"
 	"time"
 
 	"github.com/minio/minio-go/v7/pkg/set"
-	xhttp "github.com/minio/minio/internal/http"
+	"github.com/minio/minio/internal/grid"
 	"github.com/minio/minio/internal/logger"
-	"github.com/minio/minio/internal/rest"
-	"github.com/minio/mux"
 	"github.com/minio/pkg/v2/env"
 )
 
-const (
-	bootstrapRESTVersion       = "v1"
-	bootstrapRESTVersionPrefix = SlashSeparator + bootstrapRESTVersion
-	bootstrapRESTPrefix        = minioReservedBucketPath + "/bootstrap"
-	bootstrapRESTPath          = bootstrapRESTPrefix + bootstrapRESTVersionPrefix
-)
-
-const (
-	bootstrapRESTMethodHealth = "/health"
-	bootstrapRESTMethodVerify = "/verify"
-)
-
 // To abstract a node over network.
 type bootstrapRESTServer struct{}
 
+//go:generate msgp -file=$GOFILE
+
 // ServerSystemConfig - captures information about server configuration.
 type ServerSystemConfig struct {
-	MinioEndpoints EndpointServerPools
-	MinioEnv       map[string]string
+	NEndpoints int
+	CmdLines   []string
+	MinioEnv   map[string]string
 }
 
 // Diff - returns error on first difference found in two configs.
-func (s1 ServerSystemConfig) Diff(s2 ServerSystemConfig) error {
-	if s1.MinioEndpoints.NEndpoints() != s2.MinioEndpoints.NEndpoints() {
-		return fmt.Errorf("Expected number of endpoints %d, seen %d", s1.MinioEndpoints.NEndpoints(),
-			s2.MinioEndpoints.NEndpoints())
+func (s1 *ServerSystemConfig) Diff(s2 *ServerSystemConfig) error {
+	ns1 := s1.NEndpoints
+	ns2 := s2.NEndpoints
+	if ns1 != ns2 {
+		return fmt.Errorf("Expected number of endpoints %d, seen %d", ns1, ns2)
 	}
 
-	for i, ep := range s1.MinioEndpoints {
-		if ep.CmdLine != s2.MinioEndpoints[i].CmdLine {
-			return fmt.Errorf("Expected command line argument %s, seen %s", ep.CmdLine,
-				s2.MinioEndpoints[i].CmdLine)
-		}
-		if ep.SetCount != s2.MinioEndpoints[i].SetCount {
-			return fmt.Errorf("Expected set count %d, seen %d", ep.SetCount,
-				s2.MinioEndpoints[i].SetCount)
-		}
-		if ep.DrivesPerSet != s2.MinioEndpoints[i].DrivesPerSet {
-			return fmt.Errorf("Expected drives pet set %d, seen %d", ep.DrivesPerSet,
-				s2.MinioEndpoints[i].DrivesPerSet)
-		}
-		if ep.Platform != s2.MinioEndpoints[i].Platform {
-			return fmt.Errorf("Expected platform '%s', found to be on '%s'",
-				ep.Platform, s2.MinioEndpoints[i].Platform)
+	for i, cmdLine := range s1.CmdLines {
+		if cmdLine != s2.CmdLines[i] {
+			return fmt.Errorf("Expected command line argument %s, seen %s", cmdLine,
+				s2.CmdLines[i])
 		}
 	}
+
 	if reflect.DeepEqual(s1.MinioEnv, s2.MinioEnv) {
 		return nil
 	}
@@ -131,7 +107,7 @@ var skipEnvs = map[string]struct{}{
 	"MINIO_SECRET_KEY":    {},
 }
 
-func getServerSystemCfg() ServerSystemConfig {
+func getServerSystemCfg() *ServerSystemConfig {
 	envs := env.List("MINIO_")
 	envValues := make(map[string]string, len(envs))
 	for _, envK := range envs {
@@ -144,131 +120,113 @@ func getServerSystemCfg() ServerSystemConfig {
 		}
 		envValues[envK] = logger.HashString(env.Get(envK, ""))
 	}
-	return ServerSystemConfig{
-		MinioEndpoints: globalEndpoints,
-		MinioEnv:       envValues,
+	scfg := &ServerSystemConfig{NEndpoints: globalEndpoints.NEndpoints(), MinioEnv: envValues}
+	var cmdLines []string
+	for _, ep := range globalEndpoints {
+		cmdLines = append(cmdLines, ep.CmdLine)
 	}
+	scfg.CmdLines = cmdLines
+	return scfg
 }
 
-func (b *bootstrapRESTServer) writeErrorResponse(w http.ResponseWriter, err error) {
-	w.WriteHeader(http.StatusForbidden)
-	w.Write([]byte(err.Error()))
+func (s *bootstrapRESTServer) VerifyHandler(params *grid.MSS) (*ServerSystemConfig, *grid.RemoteErr) {
+	return getServerSystemCfg(), nil
 }
 
-// HealthHandler returns success if request is valid
-func (b *bootstrapRESTServer) HealthHandler(w http.ResponseWriter, r *http.Request) {}
-
-func (b *bootstrapRESTServer) VerifyHandler(w http.ResponseWriter, r *http.Request) {
-	ctx := newContext(r, w, "VerifyHandler")
-
-	if err := storageServerRequestValidate(r); err != nil {
-		b.writeErrorResponse(w, err)
-		return
-	}
-
-	cfg := getServerSystemCfg()
-	logger.LogIf(ctx, json.NewEncoder(w).Encode(&cfg))
-}
+var serverVerifyHandler = grid.NewSingleHandler[*grid.MSS, *ServerSystemConfig](grid.HandlerServerVerify, grid.NewMSS, func() *ServerSystemConfig { return &ServerSystemConfig{} })
 
 // registerBootstrapRESTHandlers - register bootstrap rest router.
-func registerBootstrapRESTHandlers(router *mux.Router) {
-	h := func(f http.HandlerFunc) http.HandlerFunc {
-		return collectInternodeStats(httpTraceHdrs(f))
-	}
-
+func registerBootstrapRESTHandlers(gm *grid.Manager) {
 	server := &bootstrapRESTServer{}
-	subrouter := router.PathPrefix(bootstrapRESTPrefix).Subrouter()
-
-	subrouter.Methods(http.MethodPost).Path(bootstrapRESTVersionPrefix + bootstrapRESTMethodHealth).HandlerFunc(
-		h(server.HealthHandler))
-
-	subrouter.Methods(http.MethodPost).Path(bootstrapRESTVersionPrefix + bootstrapRESTMethodVerify).HandlerFunc(
-		h(server.VerifyHandler))
+	logger.FatalIf(serverVerifyHandler.Register(gm, server.VerifyHandler), "unable to register handler")
 }
 
 // client to talk to bootstrap NEndpoints.
 type bootstrapRESTClient struct {
-	endpoint   Endpoint
-	restClient *rest.Client
+	gridConn *grid.Connection
 }
 
-// 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 *bootstrapRESTClient) 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)
+// Verify function verifies the server config.
+func (client *bootstrapRESTClient) Verify(ctx context.Context, srcCfg *ServerSystemConfig) (err error) {
+	if newObjectLayerFn() != nil {
+		return nil
 	}
 
-	respBody, err = client.restClient.Call(ctx, method, values, body, length)
-	if err == nil {
-		return respBody, nil
+	recvCfg, err := serverVerifyHandler.Call(ctx, client.gridConn, grid.NewMSSWith(map[string]string{}))
+	if err != nil {
+		return err
 	}
 
-	return nil, err
+	return srcCfg.Diff(recvCfg)
 }
 
 // Stringer provides a canonicalized representation of node.
 func (client *bootstrapRESTClient) String() string {
-	return client.endpoint.String()
+	return client.gridConn.String()
 }
 
-// Verify - fetches system server config.
-func (client *bootstrapRESTClient) Verify(ctx context.Context, srcCfg ServerSystemConfig) (err error) {
-	if newObjectLayerFn() != nil {
-		return nil
-	}
-	respBody, err := client.callWithContext(ctx, bootstrapRESTMethodVerify, nil, nil, -1)
-	if err != nil {
-		return
-	}
-	defer xhttp.DrainBody(respBody)
-	recvCfg := ServerSystemConfig{}
-	if err = json.NewDecoder(respBody).Decode(&recvCfg); err != nil {
-		return err
-	}
-	return srcCfg.Diff(recvCfg)
-}
-
-func verifyServerSystemConfig(ctx context.Context, endpointServerPools EndpointServerPools) error {
+func verifyServerSystemConfig(ctx context.Context, endpointServerPools EndpointServerPools, gm *grid.Manager) error {
 	srcCfg := getServerSystemCfg()
-	clnts := newBootstrapRESTClients(endpointServerPools)
+	clnts := newBootstrapRESTClients(endpointServerPools, gm)
 	var onlineServers int
 	var offlineEndpoints []error
 	var incorrectConfigs []error
 	var retries int
+	var mu sync.Mutex
 	for onlineServers < len(clnts)/2 {
+		var wg sync.WaitGroup
+		wg.Add(len(clnts))
+		onlineServers = 0
 		for _, clnt := range clnts {
-			if err := clnt.Verify(ctx, srcCfg); err != nil {
-				bootstrapTraceMsg(fmt.Sprintf("clnt.Verify: %v, endpoint: %v", err, clnt.endpoint))
-				if !isNetworkError(err) {
-					logger.LogOnceIf(ctx, fmt.Errorf("%s has incorrect configuration: %w", clnt.String(), err), clnt.String())
-					incorrectConfigs = append(incorrectConfigs, fmt.Errorf("%s has incorrect configuration: %w", clnt.String(), err))
-				} else {
-					offlineEndpoints = append(offlineEndpoints, fmt.Errorf("%s is unreachable: %w", clnt.String(), err))
+			go func(clnt *bootstrapRESTClient) {
+				defer wg.Done()
+
+				if clnt.gridConn.State() != grid.StateConnected {
+					mu.Lock()
+					offlineEndpoints = append(offlineEndpoints, fmt.Errorf("%s is unreachable: %w", clnt, grid.ErrDisconnected))
+					mu.Unlock()
+					return
 				}
-				continue
-			}
-			onlineServers++
+
+				ctx, cancel := context.WithTimeout(ctx, 2*time.Second)
+				defer cancel()
+
+				err := clnt.Verify(ctx, srcCfg)
+				mu.Lock()
+				if err != nil {
+					bootstrapTraceMsg(fmt.Sprintf("clnt.Verify: %v, endpoint: %s", err, clnt))
+					if !isNetworkError(err) {
+						logger.LogOnceIf(context.Background(), fmt.Errorf("%s has incorrect configuration: %w", clnt, err), "incorrect_"+clnt.String())
+						incorrectConfigs = append(incorrectConfigs, fmt.Errorf("%s has incorrect configuration: %w", clnt, err))
+					} else {
+						offlineEndpoints = append(offlineEndpoints, fmt.Errorf("%s is unreachable: %w", clnt, err))
+					}
+				} else {
+					onlineServers++
+				}
+				mu.Unlock()
+			}(clnt)
 		}
+		wg.Wait()
+
 		select {
 		case <-ctx.Done():
 			return ctx.Err()
 		default:
-			// Sleep for a while - so that we don't go into
-			// 100% CPU when half the endpoints are offline.
-			time.Sleep(100 * time.Millisecond)
+			// Sleep and stagger to avoid blocked CPU and thundering
+			// herd upon start up sequence.
+			time.Sleep(25*time.Millisecond + time.Duration(rand.Int63n(int64(100*time.Millisecond))))
 			retries++
 			// after 20 retries start logging that servers are not reachable yet
 			if retries >= 20 {
-				logger.Info(fmt.Sprintf("Waiting for atleast %d remote servers with valid configuration to be online", len(clnts)/2))
+				logger.Info(fmt.Sprintf("Waiting for at least %d remote servers with valid configuration to be online", len(clnts)/2))
 				if len(offlineEndpoints) > 0 {
 					logger.Info(fmt.Sprintf("Following servers are currently offline or unreachable %s", offlineEndpoints))
 				}
 				if len(incorrectConfigs) > 0 {
 					logger.Info(fmt.Sprintf("Following servers have mismatching configuration %s", incorrectConfigs))
 				}
-				retries = 0 // reset to log again after 5 retries.
+				retries = 0 // reset to log again after 20 retries.
 			}
 			offlineEndpoints = nil
 			incorrectConfigs = nil
@@ -277,39 +235,20 @@ func verifyServerSystemConfig(ctx context.Context, endpointServerPools EndpointS
 	return nil
 }
 
-func newBootstrapRESTClients(endpointServerPools EndpointServerPools) []*bootstrapRESTClient {
-	seenHosts := set.NewStringSet()
+func newBootstrapRESTClients(endpointServerPools EndpointServerPools, gm *grid.Manager) []*bootstrapRESTClient {
+	seenClient := set.NewStringSet()
 	var clnts []*bootstrapRESTClient
 	for _, ep := range endpointServerPools {
 		for _, endpoint := range ep.Endpoints {
-			if seenHosts.Contains(endpoint.Host) {
+			if endpoint.IsLocal {
 				continue
 			}
-			seenHosts.Add(endpoint.Host)
-
-			// Only proceed for remote endpoints.
-			if !endpoint.IsLocal {
-				cl := newBootstrapRESTClient(endpoint)
-				if serverDebugLog {
-					cl.restClient.TraceOutput = os.Stdout
-				}
-				clnts = append(clnts, cl)
+			if seenClient.Contains(endpoint.Host) {
+				continue
 			}
+			seenClient.Add(endpoint.Host)
+			clnts = append(clnts, &bootstrapRESTClient{gm.Connection(endpoint.GridHost())})
 		}
 	}
 	return clnts
 }
-
-// Returns a new bootstrap client.
-func newBootstrapRESTClient(endpoint Endpoint) *bootstrapRESTClient {
-	serverURL := &url.URL{
-		Scheme: endpoint.Scheme,
-		Host:   endpoint.Host,
-		Path:   bootstrapRESTPath,
-	}
-
-	restClient := rest.NewClient(serverURL, globalInternodeTransport, newCachedAuthToken())
-	restClient.HealthCheckFn = nil
-
-	return &bootstrapRESTClient{endpoint: endpoint, restClient: restClient}
-}

cmd/bootstrap-peer-server_gen.go

@@ -0,0 +1,270 @@
+package cmd
+
+// Code generated by github.com/tinylib/msgp DO NOT EDIT.
+
+import (
+	"github.com/tinylib/msgp/msgp"
+)
+
+// DecodeMsg implements msgp.Decodable
+func (z *ServerSystemConfig) 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 "NEndpoints":
+			z.NEndpoints, err = dc.ReadInt()
+			if err != nil {
+				err = msgp.WrapError(err, "NEndpoints")
+				return
+			}
+		case "CmdLines":
+			var zb0002 uint32
+			zb0002, err = dc.ReadArrayHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "CmdLines")
+				return
+			}
+			if cap(z.CmdLines) >= int(zb0002) {
+				z.CmdLines = (z.CmdLines)[:zb0002]
+			} else {
+				z.CmdLines = make([]string, zb0002)
+			}
+			for za0001 := range z.CmdLines {
+				z.CmdLines[za0001], err = dc.ReadString()
+				if err != nil {
+					err = msgp.WrapError(err, "CmdLines", za0001)
+					return
+				}
+			}
+		case "MinioEnv":
+			var zb0003 uint32
+			zb0003, err = dc.ReadMapHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "MinioEnv")
+				return
+			}
+			if z.MinioEnv == nil {
+				z.MinioEnv = make(map[string]string, zb0003)
+			} else if len(z.MinioEnv) > 0 {
+				for key := range z.MinioEnv {
+					delete(z.MinioEnv, key)
+				}
+			}
+			for zb0003 > 0 {
+				zb0003--
+				var za0002 string
+				var za0003 string
+				za0002, err = dc.ReadString()
+				if err != nil {
+					err = msgp.WrapError(err, "MinioEnv")
+					return
+				}
+				za0003, err = dc.ReadString()
+				if err != nil {
+					err = msgp.WrapError(err, "MinioEnv", za0002)
+					return
+				}
+				z.MinioEnv[za0002] = za0003
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z *ServerSystemConfig) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 3
+	// write "NEndpoints"
+	err = en.Append(0x83, 0xaa, 0x4e, 0x45, 0x6e, 0x64, 0x70, 0x6f, 0x69, 0x6e, 0x74, 0x73)
+	if err != nil {
+		return
+	}
+	err = en.WriteInt(z.NEndpoints)
+	if err != nil {
+		err = msgp.WrapError(err, "NEndpoints")
+		return
+	}
+	// write "CmdLines"
+	err = en.Append(0xa8, 0x43, 0x6d, 0x64, 0x4c, 0x69, 0x6e, 0x65, 0x73)
+	if err != nil {
+		return
+	}
+	err = en.WriteArrayHeader(uint32(len(z.CmdLines)))
+	if err != nil {
+		err = msgp.WrapError(err, "CmdLines")
+		return
+	}
+	for za0001 := range z.CmdLines {
+		err = en.WriteString(z.CmdLines[za0001])
+		if err != nil {
+			err = msgp.WrapError(err, "CmdLines", za0001)
+			return
+		}
+	}
+	// write "MinioEnv"
+	err = en.Append(0xa8, 0x4d, 0x69, 0x6e, 0x69, 0x6f, 0x45, 0x6e, 0x76)
+	if err != nil {
+		return
+	}
+	err = en.WriteMapHeader(uint32(len(z.MinioEnv)))
+	if err != nil {
+		err = msgp.WrapError(err, "MinioEnv")
+		return
+	}
+	for za0002, za0003 := range z.MinioEnv {
+		err = en.WriteString(za0002)
+		if err != nil {
+			err = msgp.WrapError(err, "MinioEnv")
+			return
+		}
+		err = en.WriteString(za0003)
+		if err != nil {
+			err = msgp.WrapError(err, "MinioEnv", za0002)
+			return
+		}
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z *ServerSystemConfig) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 3
+	// string "NEndpoints"
+	o = append(o, 0x83, 0xaa, 0x4e, 0x45, 0x6e, 0x64, 0x70, 0x6f, 0x69, 0x6e, 0x74, 0x73)
+	o = msgp.AppendInt(o, z.NEndpoints)
+	// string "CmdLines"
+	o = append(o, 0xa8, 0x43, 0x6d, 0x64, 0x4c, 0x69, 0x6e, 0x65, 0x73)
+	o = msgp.AppendArrayHeader(o, uint32(len(z.CmdLines)))
+	for za0001 := range z.CmdLines {
+		o = msgp.AppendString(o, z.CmdLines[za0001])
+	}
+	// string "MinioEnv"
+	o = append(o, 0xa8, 0x4d, 0x69, 0x6e, 0x69, 0x6f, 0x45, 0x6e, 0x76)
+	o = msgp.AppendMapHeader(o, uint32(len(z.MinioEnv)))
+	for za0002, za0003 := range z.MinioEnv {
+		o = msgp.AppendString(o, za0002)
+		o = msgp.AppendString(o, za0003)
+	}
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *ServerSystemConfig) 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 "NEndpoints":
+			z.NEndpoints, bts, err = msgp.ReadIntBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "NEndpoints")
+				return
+			}
+		case "CmdLines":
+			var zb0002 uint32
+			zb0002, bts, err = msgp.ReadArrayHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "CmdLines")
+				return
+			}
+			if cap(z.CmdLines) >= int(zb0002) {
+				z.CmdLines = (z.CmdLines)[:zb0002]
+			} else {
+				z.CmdLines = make([]string, zb0002)
+			}
+			for za0001 := range z.CmdLines {
+				z.CmdLines[za0001], bts, err = msgp.ReadStringBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "CmdLines", za0001)
+					return
+				}
+			}
+		case "MinioEnv":
+			var zb0003 uint32
+			zb0003, bts, err = msgp.ReadMapHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "MinioEnv")
+				return
+			}
+			if z.MinioEnv == nil {
+				z.MinioEnv = make(map[string]string, zb0003)
+			} else if len(z.MinioEnv) > 0 {
+				for key := range z.MinioEnv {
+					delete(z.MinioEnv, key)
+				}
+			}
+			for zb0003 > 0 {
+				var za0002 string
+				var za0003 string
+				zb0003--
+				za0002, bts, err = msgp.ReadStringBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "MinioEnv")
+					return
+				}
+				za0003, bts, err = msgp.ReadStringBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "MinioEnv", za0002)
+					return
+				}
+				z.MinioEnv[za0002] = za0003
+			}
+		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 *ServerSystemConfig) Msgsize() (s int) {
+	s = 1 + 11 + msgp.IntSize + 9 + msgp.ArrayHeaderSize
+	for za0001 := range z.CmdLines {
+		s += msgp.StringPrefixSize + len(z.CmdLines[za0001])
+	}
+	s += 9 + msgp.MapHeaderSize
+	if z.MinioEnv != nil {
+		for za0002, za0003 := range z.MinioEnv {
+			_ = za0003
+			s += msgp.StringPrefixSize + len(za0002) + msgp.StringPrefixSize + len(za0003)
+		}
+	}
+	return
+}

cmd/bootstrap-peer-server_gen_test.go

@@ -0,0 +1,123 @@
+package cmd
+
+// Code generated by github.com/tinylib/msgp DO NOT EDIT.
+
+import (
+	"bytes"
+	"testing"
+
+	"github.com/tinylib/msgp/msgp"
+)
+
+func TestMarshalUnmarshalServerSystemConfig(t *testing.T) {
+	v := ServerSystemConfig{}
+	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 BenchmarkMarshalMsgServerSystemConfig(b *testing.B) {
+	v := ServerSystemConfig{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgServerSystemConfig(b *testing.B) {
+	v := ServerSystemConfig{}
+	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 BenchmarkUnmarshalServerSystemConfig(b *testing.B) {
+	v := ServerSystemConfig{}
+	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 TestEncodeDecodeServerSystemConfig(t *testing.T) {
+	v := ServerSystemConfig{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeServerSystemConfig Msgsize() is inaccurate")
+	}
+
+	vn := ServerSystemConfig{}
+	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 BenchmarkEncodeServerSystemConfig(b *testing.B) {
+	v := ServerSystemConfig{}
+	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 BenchmarkDecodeServerSystemConfig(b *testing.B) {
+	v := ServerSystemConfig{}
+	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)
+		}
+	}
+}

cmd/bucket-handlers.go

@@ -424,7 +424,7 @@ func (api objectAPIHandlers) DeleteMultipleObjectsHandler(w http.ResponseWriter,
 		return
 	}
 
-	// Content-Md5 is requied should be set
+	// Content-Md5 is required should be set
 	// http://docs.aws.amazon.com/AmazonS3/latest/API/multiobjectdeleteapi.html
 	if _, ok := r.Header[xhttp.ContentMD5]; !ok {
 		writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrMissingContentMD5), r.URL)

cmd/bucket-handlers_test.go

@@ -252,7 +252,7 @@ func testHeadBucketHandler(obj ObjectLayer, instanceType, bucketName string, api
 		},
 		// Test case - 3.
 		// Testing for signature mismatch error.
-		// setting invalid acess and secret key.
+		// setting invalid access and secret key.
 		{
 			bucketName:         bucketName,
 			accessKey:          "abcd",
@@ -415,7 +415,7 @@ func testListMultipartUploadsHandler(obj ObjectLayer, instanceType, bucketName s
 			shouldPass:         false,
 		},
 		// Test case - 6.
-		// Setting a negative value to max-uploads paramater, should result in http.StatusBadRequest.
+		// Setting a negative value to max-uploads parameter, should result in http.StatusBadRequest.
 		{
 			bucket:             bucketName,
 			prefix:             "",

cmd/bucket-lifecycle.go

@@ -39,6 +39,7 @@ import (
 	"github.com/minio/minio/internal/bucket/lifecycle"
 	"github.com/minio/minio/internal/event"
 	xhttp "github.com/minio/minio/internal/http"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/minio/internal/s3select"
 	"github.com/minio/pkg/v2/env"
@@ -118,8 +119,8 @@ func (es *expiryState) PendingTasks() int {
 // close closes work channels exactly once.
 func (es *expiryState) close() {
 	es.once.Do(func() {
-		close(es.byDaysCh)
-		close(es.byNewerNoncurrentCh)
+		xioutil.SafeClose(es.byDaysCh)
+		xioutil.SafeClose(es.byNewerNoncurrentCh)
 	})
 }
 
@@ -507,9 +508,13 @@ 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)

cmd/bucket-listobjects-handlers.go

@@ -111,7 +111,7 @@ func (api objectAPIHandlers) listObjectVersionsHandler(w http.ResponseWriter, r
 
 	listObjectVersions := objectAPI.ListObjectVersions
 
-	// Inititate a list object versions operation based on the input params.
+	// Initiate a list object versions operation based on the input params.
 	// On success would return back ListObjectsInfo object to be
 	// marshaled into S3 compatible XML header.
 	listObjectVersionsInfo, err := listObjectVersions(ctx, bucket, prefix, marker, versionIDMarker, delimiter, maxkeys)
@@ -201,10 +201,10 @@ func (api objectAPIHandlers) listObjectsV2Handler(ctx context.Context, w http.Re
 	)
 
 	if r.Header.Get(xMinIOExtract) == "true" && strings.Contains(prefix, archivePattern) {
-		// Inititate a list objects operation inside a zip file based in the input params
+		// Initiate a list objects operation inside a zip file based in the input params
 		listObjectsV2Info, err = listObjectsV2InArchive(ctx, objectAPI, bucket, prefix, token, delimiter, maxKeys, fetchOwner, startAfter)
 	} else {
-		// Inititate a list objects operation based on the input params.
+		// Initiate a list objects operation based on the input params.
 		// On success would return back ListObjectsInfo object to be
 		// marshaled into S3 compatible XML header.
 		listObjectsV2Info, err = objectAPI.ListObjectsV2(ctx, bucket, prefix, token, delimiter, maxKeys, fetchOwner, startAfter)
@@ -304,7 +304,7 @@ func (api objectAPIHandlers) ListObjectsV1Handler(w http.ResponseWriter, r *http
 
 	listObjects := objectAPI.ListObjects
 
-	// Inititate a list objects operation based on the input params.
+	// Initiate a list objects operation based on the input params.
 	// On success would return back ListObjectsInfo object to be
 	// marshaled into S3 compatible XML header.
 	listObjectsInfo, err := listObjects(ctx, bucket, prefix, marker, delimiter, maxKeys)

cmd/bucket-policy-handlers_test.go

@@ -317,7 +317,7 @@ func testPutBucketPolicyHandler(obj ObjectLayer, instanceType, bucketName string
 		if err != nil {
 			t.Fatalf("Test %d: %s: Failed to create HTTP request for PutBucketPolicyHandler: <ERROR> %v", i+1, instanceType, err)
 		}
-		// Since `apiRouter` satisfies `http.Handler` it has a ServeHTTP to execute the logic ofthe handler.
+		// Since `apiRouter` satisfies `http.Handler` it has a ServeHTTP to execute the logic of the handler.
 		// Call the ServeHTTP to execute the handler.
 		apiRouter.ServeHTTP(recV4, reqV4)
 		if recV4.Code != testCase.expectedRespStatus {
@@ -331,7 +331,7 @@ func testPutBucketPolicyHandler(obj ObjectLayer, instanceType, bucketName string
 		if err != nil {
 			t.Fatalf("Test %d: %s: Failed to create HTTP request for PutBucketPolicyHandler: <ERROR> %v", i+1, instanceType, err)
 		}
-		// Since `apiRouter` satisfies `http.Handler` it has a ServeHTTP to execute the logic ofthe handler.
+		// Since `apiRouter` satisfies `http.Handler` it has a ServeHTTP to execute the logic of the handler.
 		// Call the ServeHTTP to execute the handler.
 		apiRouter.ServeHTTP(recV2, reqV2)
 		if recV2.Code != testCase.expectedRespStatus {
@@ -407,7 +407,7 @@ func testGetBucketPolicyHandler(obj ObjectLayer, instanceType, bucketName string
 		if err != nil {
 			t.Fatalf("Test %d: Failed to create HTTP request for PutBucketPolicyHandler: <ERROR> %v", i+1, err)
 		}
-		// Since `apiRouter` satisfies `http.Handler` it has a ServeHTTP to execute the logic ofthe handler.
+		// Since `apiRouter` satisfies `http.Handler` it has a ServeHTTP to execute the logic of the handler.
 		// Call the ServeHTTP to execute the handler.
 		apiRouter.ServeHTTP(recV4, reqV4)
 		if recV4.Code != testPolicy.expectedRespStatus {
@@ -421,7 +421,7 @@ func testGetBucketPolicyHandler(obj ObjectLayer, instanceType, bucketName string
 		if err != nil {
 			t.Fatalf("Test %d: Failed to create HTTP request for PutBucketPolicyHandler: <ERROR> %v", i+1, err)
 		}
-		// Since `apiRouter` satisfies `http.Handler` it has a ServeHTTP to execute the logic ofthe handler.
+		// Since `apiRouter` satisfies `http.Handler` it has a ServeHTTP to execute the logic of the handler.
 		// Call the ServeHTTP to execute the handler.
 		apiRouter.ServeHTTP(recV2, reqV2)
 		if recV2.Code != testPolicy.expectedRespStatus {

cmd/bucket-policy.go

@@ -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},

cmd/bucket-replication-utils.go

@@ -219,7 +219,7 @@ type ReplicateDecision struct {
 	targetsMap map[string]replicateTargetDecision
 }
 
-// ReplicateAny returns true if atleast one target qualifies for replication
+// ReplicateAny returns true if at least one target qualifies for replication
 func (d ReplicateDecision) ReplicateAny() bool {
 	for _, t := range d.targetsMap {
 		if t.Replicate {
@@ -229,7 +229,7 @@ func (d ReplicateDecision) ReplicateAny() bool {
 	return false
 }
 
-// Synchronous returns true if atleast one target qualifies for synchronous replication
+// Synchronous returns true if at least one target qualifies for synchronous replication
 func (d ReplicateDecision) Synchronous() bool {
 	for _, t := range d.targetsMap {
 		if t.Synchronous {
@@ -336,7 +336,7 @@ type ReplicationState struct {
 	DeleteMarker              bool                   // represents DeleteMarker replication state
 	ReplicationTimeStamp      time.Time              // timestamp when last replication activity happened
 	ReplicationStatusInternal string                 // stringified representation of all replication activity
-	// VersionPurgeStatusInternal is internally in the format "arn1=PENDING;arn2=COMMPLETED;"
+	// VersionPurgeStatusInternal is internally in the format "arn1=PENDING;arn2=COMPLETED;"
 	VersionPurgeStatusInternal string                            // stringified representation of all version purge statuses
 	ReplicateDecisionStr       string                            // stringified representation of replication decision for each target
 	Targets                    map[string]replication.StatusType // map of ARN->replication status for ongoing replication activity

cmd/bucket-replication.go

@@ -23,7 +23,6 @@ import (
 	"errors"
 	"fmt"
 	"io"
-	"math"
 	"math/rand"
 	"net/http"
 	"net/url"
@@ -48,6 +47,7 @@ import (
 	"github.com/minio/minio/internal/event"
 	"github.com/minio/minio/internal/hash"
 	xhttp "github.com/minio/minio/internal/http"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/tinylib/msgp/msgp"
 	"github.com/zeebo/xxh3"
@@ -496,7 +496,7 @@ func replicateDelete(ctx context.Context, dobj DeletedObjectReplicationInfo, obj
 		}
 		tgtClnt := globalBucketTargetSys.GetRemoteTargetClient(bucket, tgtEntry.Arn)
 		if tgtClnt == nil {
-			// Skip stale targets if any and log them to be missing atleast once.
+			// Skip stale targets if any and log them to be missing at least once.
 			logger.LogOnceIf(ctx, fmt.Errorf("failed to get target for bucket:%s arn:%s", bucket, tgtEntry.Arn), tgtEntry.Arn)
 			sendEvent(eventArgs{
 				EventName:  event.ObjectReplicationNotTracked,
@@ -684,7 +684,7 @@ func replicateDeleteToTarget(ctx context.Context, dobj DeletedObjectReplicationI
 		} else {
 			rinfo.VersionPurgeStatus = Failed
 		}
-		logger.LogIf(ctx, fmt.Errorf("Unable to replicate delete marker to %s/%s(%s): %s", tgt.Bucket, dobj.ObjectName, versionID, rmErr))
+		logger.LogIf(ctx, fmt.Errorf("unable to replicate delete marker to %s: %s/%s(%s): %w", tgt.EndpointURL(), tgt.Bucket, dobj.ObjectName, versionID, rmErr))
 		if rmErr != nil && minio.IsNetworkOrHostDown(rmErr, true) && !globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
 			globalBucketTargetSys.markOffline(tgt.EndpointURL())
 		}
@@ -1205,7 +1205,7 @@ func (ri ReplicateObjectInfo) replicateObject(ctx context.Context, objectAPI Obj
 	}
 
 	if tgt.Bucket == "" {
-		logger.LogIf(ctx, fmt.Errorf("unable to replicate object %s(%s), bucket is empty", objInfo.Name, objInfo.VersionID))
+		logger.LogIf(ctx, fmt.Errorf("unable to replicate object %s(%s), bucket is empty for target %s", objInfo.Name, objInfo.VersionID, tgt.EndpointURL()))
 		sendEvent(eventArgs{
 			EventName:  event.ObjectReplicationNotTracked,
 			BucketName: bucket,
@@ -1231,7 +1231,7 @@ func (ri ReplicateObjectInfo) replicateObject(ctx context.Context, objectAPI Obj
 
 	putOpts, err := putReplicationOpts(ctx, tgt.StorageClass, objInfo)
 	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("failed to get target for replication bucket:%s err:%w", bucket, err))
+		logger.LogIf(ctx, fmt.Errorf("failure setting options for replication bucket:%s err:%w", bucket, err))
 		sendEvent(eventArgs{
 			EventName:  event.ObjectReplicationNotTracked,
 			BucketName: bucket,
@@ -1266,14 +1266,14 @@ func (ri ReplicateObjectInfo) replicateObject(ctx context.Context, objectAPI Obj
 			r, objInfo, putOpts); rinfo.Err != nil {
 			if minio.ToErrorResponse(rinfo.Err).Code != "PreconditionFailed" {
 				rinfo.ReplicationStatus = replication.Failed
-				logger.LogIf(ctx, fmt.Errorf("unable to replicate for object %s/%s(%s): %s", bucket, objInfo.Name, objInfo.VersionID, rinfo.Err))
+				logger.LogIf(ctx, fmt.Errorf("unable to replicate for object %s/%s(%s): %s (target: %s)", bucket, objInfo.Name, objInfo.VersionID, rinfo.Err, tgt.EndpointURL()))
 			}
 		}
 	} else {
 		if _, rinfo.Err = c.PutObject(ctx, tgt.Bucket, object, r, size, "", "", putOpts); rinfo.Err != nil {
 			if minio.ToErrorResponse(rinfo.Err).Code != "PreconditionFailed" {
 				rinfo.ReplicationStatus = replication.Failed
-				logger.LogIf(ctx, fmt.Errorf("unable to replicate for object %s/%s(%s): %s", bucket, objInfo.Name, objInfo.VersionID, rinfo.Err))
+				logger.LogIf(ctx, fmt.Errorf("unable to replicate for object %s/%s(%s): %s (target: %s)", bucket, objInfo.Name, objInfo.VersionID, rinfo.Err, tgt.EndpointURL()))
 			}
 		}
 	}
@@ -1337,7 +1337,7 @@ func (ri ReplicateObjectInfo) replicateAll(ctx context.Context, objectAPI Object
 				UserAgent:  "Internal: [Replication]",
 				Host:       globalLocalNodeName,
 			})
-			logger.LogIf(ctx, fmt.Errorf("unable to update replicate metadata for %s/%s(%s): %w", bucket, object, objInfo.VersionID, err))
+			logger.LogIf(ctx, fmt.Errorf("unable to replicate to target %s for %s/%s(%s): %w", tgt.EndpointURL(), bucket, object, objInfo.VersionID, err))
 		}
 		return
 	}
@@ -1368,7 +1368,7 @@ func (ri ReplicateObjectInfo) replicateAll(ctx context.Context, objectAPI Object
 	}
 
 	if tgt.Bucket == "" {
-		logger.LogIf(ctx, fmt.Errorf("unable to replicate object %s(%s), bucket is empty", objInfo.Name, objInfo.VersionID))
+		logger.LogIf(ctx, fmt.Errorf("unable to replicate object %s(%s) to %s, target bucket is missing", objInfo.Name, objInfo.VersionID, tgt.EndpointURL()))
 		sendEvent(eventArgs{
 			EventName:  event.ObjectReplicationNotTracked,
 			BucketName: bucket,
@@ -1398,7 +1398,7 @@ func (ri ReplicateObjectInfo) replicateAll(ctx context.Context, objectAPI Object
 		if rAction == replicateNone {
 			if ri.OpType == replication.ExistingObjectReplicationType &&
 				objInfo.ModTime.Unix() > oi.LastModified.Unix() && objInfo.VersionID == nullVersionID {
-				logger.LogIf(ctx, fmt.Errorf("unable to replicate %s/%s (null). Newer version exists on target", bucket, object))
+				logger.LogIf(ctx, fmt.Errorf("unable to replicate %s/%s (null). Newer version exists on target %s", bucket, object, tgt.EndpointURL()))
 				sendEvent(eventArgs{
 					EventName:  event.ObjectReplicationNotTracked,
 					BucketName: bucket,
@@ -1488,13 +1488,13 @@ func (ri ReplicateObjectInfo) replicateAll(ctx context.Context, objectAPI Object
 		}
 		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))
+			logger.LogIf(ctx, fmt.Errorf("unable to replicate metadata for object %s/%s(%s) to target %s: %w", bucket, objInfo.Name, objInfo.VersionID, tgt.EndpointURL(), rinfo.Err))
 		}
 	} else {
 		var putOpts minio.PutObjectOptions
 		putOpts, err = putReplicationOpts(ctx, tgt.StorageClass, objInfo)
 		if err != nil {
-			logger.LogIf(ctx, fmt.Errorf("failed to get target for replication bucket:%s err:%w", bucket, err))
+			logger.LogIf(ctx, fmt.Errorf("failed to set replicate options for object %s/%s(%s) (target %s) err:%w", bucket, objInfo.Name, objInfo.VersionID, tgt.EndpointURL(), err))
 			sendEvent(eventArgs{
 				EventName:  event.ObjectReplicationNotTracked,
 				BucketName: bucket,
@@ -1528,7 +1528,7 @@ func (ri ReplicateObjectInfo) replicateAll(ctx context.Context, objectAPI Object
 				r, objInfo, putOpts); rinfo.Err != nil {
 				if minio.ToErrorResponse(rinfo.Err).Code != "PreconditionFailed" {
 					rinfo.ReplicationStatus = replication.Failed
-					logger.LogIf(ctx, fmt.Errorf("unable to replicate for object %s/%s(%s): %s", bucket, objInfo.Name, objInfo.VersionID, rinfo.Err))
+					logger.LogIf(ctx, fmt.Errorf("unable to replicate for object %s/%s(%s) to target %s: %w", bucket, objInfo.Name, objInfo.VersionID, tgt.EndpointURL(), rinfo.Err))
 				} else {
 					rinfo.ReplicationStatus = replication.Completed
 				}
@@ -1537,7 +1537,7 @@ func (ri ReplicateObjectInfo) replicateAll(ctx context.Context, objectAPI Object
 			if _, rinfo.Err = c.PutObject(ctx, tgt.Bucket, object, r, size, "", "", putOpts); rinfo.Err != nil {
 				if minio.ToErrorResponse(rinfo.Err).Code != "PreconditionFailed" {
 					rinfo.ReplicationStatus = replication.Failed
-					logger.LogIf(ctx, fmt.Errorf("unable to replicate for object %s/%s(%s): %s", bucket, objInfo.Name, objInfo.VersionID, rinfo.Err))
+					logger.LogIf(ctx, fmt.Errorf("unable to replicate for object %s/%s(%s) to target %s: %w", bucket, objInfo.Name, objInfo.VersionID, tgt.EndpointURL(), rinfo.Err))
 				} else {
 					rinfo.ReplicationStatus = replication.Completed
 				}
@@ -1895,7 +1895,7 @@ func (p *ReplicationPool) AddLargeWorkers() {
 	go func() {
 		<-p.ctx.Done()
 		for i := 0; i < LargeWorkerCount; i++ {
-			close(p.lrgworkers[i])
+			xioutil.SafeClose(p.lrgworkers[i])
 		}
 	}()
 }
@@ -1954,7 +1954,7 @@ func (p *ReplicationPool) ResizeWorkers(n, checkOld int) {
 	for len(p.workers) > n {
 		worker := p.workers[len(p.workers)-1]
 		p.workers = p.workers[:len(p.workers)-1]
-		close(worker)
+		xioutil.SafeClose(worker)
 	}
 }
 
@@ -1973,10 +1973,10 @@ func (p *ReplicationPool) ResizeWorkerPriority(pri string, maxWorkers int) {
 		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 {
@@ -2069,18 +2069,18 @@ 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:
-			maxWorkers = int(math.Min(float64(maxWorkers), WorkerMaxLimit))
+			maxWorkers = min(maxWorkers, WorkerMaxLimit)
 			if p.ActiveWorkers() < maxWorkers {
 				p.mu.RLock()
-				workers := int(math.Min(float64(len(p.workers)+1), float64(maxWorkers)))
+				workers := min(len(p.workers)+1, maxWorkers)
 				existing := len(p.workers)
 				p.mu.RUnlock()
 				p.ResizeWorkers(workers, existing)
 			}
-			maxMRFWorkers := int(math.Min(float64(maxWorkers), MRFWorkerMaxLimit))
+			maxMRFWorkers := min(maxWorkers, MRFWorkerMaxLimit)
 			if p.ActiveMRFWorkers() < maxMRFWorkers {
 				p.mu.RLock()
-				workers := int(math.Min(float64(p.mrfWorkerSize+1), float64(maxMRFWorkers)))
+				workers := min(p.mrfWorkerSize+1, maxMRFWorkers)
 				p.mu.RUnlock()
 				p.ResizeFailedWorkers(workers)
 			}
@@ -2126,10 +2126,10 @@ 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:
-			maxWorkers = int(math.Min(float64(maxWorkers), WorkerMaxLimit))
+			maxWorkers = min(maxWorkers, WorkerMaxLimit)
 			if p.ActiveWorkers() < maxWorkers {
 				p.mu.RLock()
-				workers := int(math.Min(float64(len(p.workers)+1), float64(maxWorkers)))
+				workers := min(len(p.workers)+1, maxWorkers)
 				existing := len(p.workers)
 				p.mu.RUnlock()
 				p.ResizeWorkers(workers, existing)
@@ -2256,7 +2256,7 @@ func proxyHeadToRepTarget(ctx context.Context, bucket, object string, rs *HTTPRa
 		if rs != nil {
 			h, err := rs.ToHeader()
 			if err != nil {
-				logger.LogIf(ctx, fmt.Errorf("Invalid range header for %s/%s(%s) - %w", bucket, object, opts.VersionID, err))
+				logger.LogIf(ctx, fmt.Errorf("invalid range header for %s/%s(%s) - %w", bucket, object, opts.VersionID, err))
 				continue
 			}
 			gopts.Set(xhttp.Range, h)
@@ -2349,6 +2349,127 @@ func scheduleReplication(ctx context.Context, oi ObjectInfo, o ObjectLayer, dsc
 	}
 }
 
+// proxyTaggingToRepTarget proxies tagging requests to remote targets for
+// active-active replicated setups
+func proxyTaggingToRepTarget(ctx context.Context, bucket, object string, tags *tags.Tags, opts ObjectOptions, proxyTargets *madmin.BucketTargets) (proxy proxyResult) {
+	// this option is set when active-active replication is in place between site A -> B,
+	// and request hits site B that does not have the object yet.
+	if opts.ProxyRequest || (opts.ProxyHeaderSet && !opts.ProxyRequest) { // true only when site B sets MinIOSourceProxyRequest header
+		return proxy
+	}
+	var wg sync.WaitGroup
+	errs := make([]error, len(proxyTargets.Targets))
+	for idx, t := range proxyTargets.Targets {
+		tgt := globalBucketTargetSys.GetRemoteTargetClient(bucket, t.Arn)
+		if tgt == nil || globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
+			continue
+		}
+		// if proxying explicitly disabled on remote target
+		if tgt.disableProxy {
+			continue
+		}
+		idx := idx
+		wg.Add(1)
+		go func(idx int, tgt *TargetClient) {
+			defer wg.Done()
+			var err error
+			if tags != nil {
+				popts := minio.PutObjectTaggingOptions{
+					VersionID: opts.VersionID,
+					Internal: minio.AdvancedObjectTaggingOptions{
+						ReplicationProxyRequest: "true",
+					},
+				}
+				err = tgt.PutObjectTagging(ctx, tgt.Bucket, object, tags, popts)
+			} else {
+				dopts := minio.RemoveObjectTaggingOptions{
+					VersionID: opts.VersionID,
+					Internal: minio.AdvancedObjectTaggingOptions{
+						ReplicationProxyRequest: "true",
+					},
+				}
+				err = tgt.RemoveObjectTagging(ctx, tgt.Bucket, object, dopts)
+			}
+			if err != nil {
+				errs[idx] = err
+			}
+		}(idx, tgt)
+	}
+	wg.Wait()
+
+	var (
+		terr        error
+		taggedCount int
+	)
+	for _, err := range errs {
+		if err == nil {
+			taggedCount++
+			continue
+		}
+		errCode := minio.ToErrorResponse(err).Code
+		if errCode != "NoSuchKey" && errCode != "NoSuchVersion" {
+			terr = err
+		}
+	}
+	// don't return error if at least one target was tagged successfully
+	if taggedCount == 0 && terr != nil {
+		proxy.Err = terr
+	}
+	return proxy
+}
+
+// proxyGetTaggingToRepTarget proxies get tagging requests to remote targets for
+// active-active replicated setups
+func proxyGetTaggingToRepTarget(ctx context.Context, bucket, object string, opts ObjectOptions, proxyTargets *madmin.BucketTargets) (tgs *tags.Tags, proxy proxyResult) {
+	// this option is set when active-active replication is in place between site A -> B,
+	// and request hits site B that does not have the object yet.
+	if opts.ProxyRequest || (opts.ProxyHeaderSet && !opts.ProxyRequest) { // true only when site B sets MinIOSourceProxyRequest header
+		return nil, proxy
+	}
+	var wg sync.WaitGroup
+	errs := make([]error, len(proxyTargets.Targets))
+	tagSlc := make([]map[string]string, len(proxyTargets.Targets))
+	for idx, t := range proxyTargets.Targets {
+		tgt := globalBucketTargetSys.GetRemoteTargetClient(bucket, t.Arn)
+		if tgt == nil || globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
+			continue
+		}
+		// if proxying explicitly disabled on remote target
+		if tgt.disableProxy {
+			continue
+		}
+		idx := idx
+		wg.Add(1)
+		go func(idx int, tgt *TargetClient) {
+			defer wg.Done()
+			var err error
+			gopts := minio.GetObjectTaggingOptions{
+				VersionID: opts.VersionID,
+				Internal: minio.AdvancedObjectTaggingOptions{
+					ReplicationProxyRequest: "true",
+				},
+			}
+			tgs, err = tgt.GetObjectTagging(ctx, tgt.Bucket, object, gopts)
+			if err != nil {
+				errs[idx] = err
+			} else {
+				tagSlc[idx] = tgs.ToMap()
+			}
+		}(idx, tgt)
+	}
+	wg.Wait()
+	for idx, err := range errs {
+		errCode := minio.ToErrorResponse(err).Code
+		if err != nil && errCode != "NoSuchKey" && errCode != "NoSuchVersion" {
+			return nil, proxyResult{Err: err}
+		}
+		if err == nil {
+			tgs, _ = tags.MapToObjectTags(tagSlc[idx])
+		}
+	}
+	return tgs, proxy
+}
+
 func scheduleReplicationDelete(ctx context.Context, dv DeletedObjectReplicationInfo, o ObjectLayer) {
 	globalReplicationPool.queueReplicaDeleteTask(dv)
 	for arn := range dv.ReplicationState.Targets {
@@ -2499,7 +2620,7 @@ func (s *replicationResyncer) PersistToDisk(ctx context.Context, objectAPI Objec
 				}
 				if updt {
 					if err := saveResyncStatus(ctx, bucket, brs, objectAPI); err != nil {
-						logger.LogIf(ctx, fmt.Errorf("Could not save resync metadata to drive for %s - %w", bucket, err))
+						logger.LogIf(ctx, fmt.Errorf("could not save resync metadata to drive for %s - %w", bucket, err))
 					} else {
 						lastResyncStatusSave[bucket] = brs.LastUpdate
 					}
@@ -2635,7 +2756,7 @@ func (s *replicationResyncer) resyncBucket(ctx context.Context, objectAPI Object
 	}
 	workers := make([]chan ReplicateObjectInfo, resyncParallelRoutines)
 	resultCh := make(chan TargetReplicationResyncStatus, 1)
-	defer close(resultCh)
+	defer xioutil.SafeClose(resultCh)
 	go func() {
 		for r := range resultCh {
 			s.incStats(r, opts)
@@ -2747,7 +2868,7 @@ func (s *replicationResyncer) resyncBucket(ctx context.Context, objectAPI Object
 		}
 	}
 	for i := 0; i < resyncParallelRoutines; i++ {
-		close(workers[i])
+		xioutil.SafeClose(workers[i])
 	}
 	wg.Wait()
 	resyncStatus = ResyncCompleted
@@ -2877,7 +2998,7 @@ func (p *ReplicationPool) startResyncRoutine(ctx context.Context, buckets []Buck
 		}
 		duration := time.Duration(r.Float64() * float64(time.Minute))
 		if duration < time.Second {
-			// Make sure to sleep atleast a second to avoid high CPU ticks.
+			// Make sure to sleep at least a second to avoid high CPU ticks.
 			duration = time.Second
 		}
 		time.Sleep(duration)
@@ -3003,7 +3124,7 @@ func getReplicationDiff(ctx context.Context, objAPI ObjectLayer, bucket string,
 	}
 	diffCh := make(chan madmin.DiffInfo, 4000)
 	go func() {
-		defer close(diffCh)
+		defer xioutil.SafeClose(diffCh)
 		for obj := range objInfoCh {
 			if contextCanceled(ctx) {
 				// Just consume input...
@@ -3196,7 +3317,7 @@ func (p *ReplicationPool) persistMRF() {
 			mTimer.Reset(mrfSaveInterval)
 		case <-p.ctx.Done():
 			p.mrfStopCh <- struct{}{}
-			close(p.mrfSaveCh)
+			xioutil.SafeClose(p.mrfSaveCh)
 			// We try to save if possible, but we don't care beyond that.
 			saveMRFToDisk()
 			return
@@ -3431,7 +3552,7 @@ func (p *ReplicationPool) getMRF(ctx context.Context, bucket string) (ch <-chan
 
 	mrfCh := make(chan madmin.ReplicationMRF, 100)
 	go func() {
-		defer close(mrfCh)
+		defer xioutil.SafeClose(mrfCh)
 		for vID, e := range mrfRec.Entries {
 			if bucket != "" && e.Bucket != bucket {
 				continue

cmd/bucket-targets.go

@@ -19,6 +19,7 @@ package cmd
 
 import (
 	"context"
+	"errors"
 	"fmt"
 	"net/url"
 	"sync"
@@ -351,6 +352,27 @@ func (sys *BucketTargetSys) SetTarget(ctx context.Context, bucket string, tgt *m
 			return BucketRemoteTargetNotVersioned{Bucket: tgt.TargetBucket}
 		}
 	}
+
+	// Check if target is a MinIO server and alive
+	hcCtx, cancel := context.WithTimeout(ctx, 3*time.Second)
+	scheme := "http"
+	if tgt.Secure {
+		scheme = "https"
+	}
+	result := <-sys.hcClient.Alive(hcCtx, madmin.AliveOpts{}, madmin.ServerProperties{
+		Endpoint: tgt.Endpoint,
+		Scheme:   scheme,
+	})
+
+	cancel()
+	if result.Error != nil {
+		return RemoteTargetConnectionErr{Bucket: tgt.TargetBucket, Err: result.Error, AccessKey: tgt.Credentials.AccessKey}
+	}
+	if !result.Online {
+		err := errors.New("Health check timed out after 3 seconds")
+		return RemoteTargetConnectionErr{Err: err}
+	}
+
 	sys.Lock()
 	defer sys.Unlock()
 

cmd/callhome.go

@@ -60,7 +60,7 @@ func initCallhome(ctx context.Context, objAPI ObjectLayer) {
 			// sleep for some time and try again.
 			duration := time.Duration(r.Float64() * float64(globalCallhomeConfig.FrequencyDur()))
 			if duration < time.Second {
-				// Make sure to sleep atleast a second to avoid high CPU ticks.
+				// Make sure to sleep at least a second to avoid high CPU ticks.
 				duration = time.Second
 			}
 			time.Sleep(duration)

cmd/common-main.go

@@ -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"
@@ -69,7 +69,7 @@ import (
 // serverDebugLog will enable debug printing
 var serverDebugLog = env.Get("_MINIO_SERVER_DEBUG", config.EnableOff) == config.EnableOn
 
-var shardDiskTimeDelta time.Duration
+var currentReleaseTime time.Time
 
 func init() {
 	if runtime.GOOS == "windows" {
@@ -107,16 +107,12 @@ func init() {
 	gob.Register(madmin.XFSErrorConfigs{})
 	gob.Register(map[string]interface{}{})
 
-	var err error
-	shardDiskTimeDelta, err = time.ParseDuration(env.Get("_MINIO_SHARD_DISKTIME_DELTA", "1m"))
-	if err != nil {
-		shardDiskTimeDelta = 1 * time.Minute
-	}
-
 	// All minio-go and madmin-go API operations shall be performed only once,
 	// another way to look at this is we are turning off retries.
 	minio.MaxRetry = 1
 	madmin.MaxRetry = 1
+
+	currentReleaseTime, _ = GetCurrentReleaseTime()
 }
 
 const consolePrefix = "CONSOLE_"
@@ -178,6 +174,23 @@ 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 +220,7 @@ 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 +238,9 @@ 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 +251,18 @@ 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 +270,16 @@ 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
@@ -284,9 +297,7 @@ func checkUpdate(mode string) {
 		return
 	}
 
-	// Its OK to ignore any errors during doUpdate() here.
-	crTime, err := GetCurrentReleaseTime()
-	if err != nil {
+	if currentReleaseTime.IsZero() {
 		return
 	}
 
@@ -297,8 +308,8 @@ func checkUpdate(mode string) {
 
 	var older time.Duration
 	var downloadURL string
-	if lrTime.After(crTime) {
-		older = lrTime.Sub(crTime)
+	if lrTime.After(currentReleaseTime) {
+		older = lrTime.Sub(currentReleaseTime)
 		downloadURL = getDownloadURL(releaseTimeToReleaseTag(lrTime))
 	}
 
@@ -307,7 +318,7 @@ func checkUpdate(mode string) {
 		return
 	}
 
-	logger.Info(prepareUpdateMessage("Run `mc admin update`", lrTime.Sub(crTime)))
+	logger.Info(prepareUpdateMessage("Run `mc admin update ALIAS`", lrTime.Sub(currentReleaseTime)))
 }
 
 func newConfigDir(dir string, dirSet bool, getDefaultDir func() string) (*ConfigDir, error) {
@@ -355,10 +366,7 @@ func buildServerCtxt(ctx *cli.Context, ctxt *serverCtxt) (err error) {
 	}
 
 	// Check "no-compat" flag from command line argument.
-	ctxt.StrictS3Compat = true
-	if ctx.IsSet("no-compat") || ctx.GlobalIsSet("no-compat") {
-		ctxt.StrictS3Compat = false
-	}
+	ctxt.StrictS3Compat = !(ctx.IsSet("no-compat") || ctx.GlobalIsSet("no-compat"))
 
 	switch {
 	case ctx.IsSet("config-dir"):
@@ -832,7 +840,7 @@ func loadRootCredentials() {
 // It depends on KMS env variables and global cli flags.
 func handleKMSConfig() {
 	if env.IsSet(kms.EnvKMSSecretKey) && env.IsSet(kms.EnvKESEndpoint) {
-		logger.Fatal(errors.New("ambigious KMS configuration"), fmt.Sprintf("The environment contains %q as well as %q", kms.EnvKMSSecretKey, kms.EnvKESEndpoint))
+		logger.Fatal(errors.New("ambiguous KMS configuration"), fmt.Sprintf("The environment contains %q as well as %q", kms.EnvKMSSecretKey, kms.EnvKESEndpoint))
 	}
 
 	if env.IsSet(kms.EnvKMSSecretKey) {
@@ -845,10 +853,10 @@ func handleKMSConfig() {
 	if env.IsSet(kms.EnvKESEndpoint) {
 		if env.IsSet(kms.EnvKESAPIKey) {
 			if env.IsSet(kms.EnvKESClientKey) {
-				logger.Fatal(errors.New("ambigious KMS configuration"), fmt.Sprintf("The environment contains %q as well as %q", kms.EnvKESAPIKey, kms.EnvKESClientKey))
+				logger.Fatal(errors.New("ambiguous KMS configuration"), fmt.Sprintf("The environment contains %q as well as %q", kms.EnvKESAPIKey, kms.EnvKESClientKey))
 			}
 			if env.IsSet(kms.EnvKESClientCert) {
-				logger.Fatal(errors.New("ambigious KMS configuration"), fmt.Sprintf("The environment contains %q as well as %q", kms.EnvKESAPIKey, kms.EnvKESClientCert))
+				logger.Fatal(errors.New("ambiguous KMS configuration"), fmt.Sprintf("The environment contains %q as well as %q", kms.EnvKESAPIKey, kms.EnvKESClientCert))
 			}
 		}
 		if !env.IsSet(kms.EnvKESKeyName) {

cmd/config-current.go

@@ -24,6 +24,8 @@ 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"
@@ -74,6 +76,7 @@ func initHelp() {
 		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
@@ -223,7 +226,12 @@ func initHelp() {
 		config.HelpKV{
 			Key:         config.CacheSubSys,
 			Type:        "string",
-			Description: "enable various cache optimizations on MinIO for reads",
+			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,
 		},
 	}
@@ -273,6 +281,7 @@ func initHelp() {
 		config.CallhomeSubSys:       callhome.HelpCallhome,
 		config.DriveSubSys:          drive.HelpDrive,
 		config.CacheSubSys:          cache.Help,
+		config.BrowserSubSys:        browser.Help,
 	}
 
 	config.RegisterHelpSubSys(helpMap)
@@ -407,6 +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 {
@@ -486,7 +499,7 @@ func lookupConfigs(s config.Config, objAPI ObjectLayer) {
 
 		if len(globalDomainNames) != 0 && !globalDomainIPs.IsEmpty() && globalEtcdClient != nil {
 			if globalDNSConfig != nil {
-				// if global DNS is already configured, indicate with a warning, incase
+				// if global DNS is already configured, indicate with a warning, in case
 				// users are confused.
 				logger.LogIf(ctx, fmt.Errorf("DNS store is already configured with %s, etcd is not used for DNS store", globalDNSConfig))
 			} else {
@@ -587,6 +600,7 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 			return fmt.Errorf("Unable to apply scanner config: %w", err)
 		}
 		// update dynamic scanner values.
+		scannerIdleMode.Store(scannerCfg.IdleMode)
 		scannerCycle.Store(scannerCfg.Cycle)
 		logger.LogIf(ctx, scannerSleeper.Update(scannerCfg.Delay, scannerCfg.MaxWait))
 	case config.LoggerWebhookSubSys:
@@ -689,6 +703,12 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 		} 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()

cmd/data-scanner.go

@@ -40,6 +40,7 @@ import (
 	"github.com/minio/minio/internal/color"
 	"github.com/minio/minio/internal/config/heal"
 	"github.com/minio/minio/internal/event"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/console"
 	uatomic "go.uber.org/atomic"
@@ -65,8 +66,9 @@ var (
 	globalHealConfig heal.Config
 
 	// Sleeper values are updated when config is loaded.
-	scannerSleeper = newDynamicSleeper(2, time.Second, true) // Keep defaults same as config defaults
-	scannerCycle   = uatomic.NewDuration(dataScannerStartDelay)
+	scannerSleeper  = newDynamicSleeper(2, time.Second, true) // Keep defaults same as config defaults
+	scannerCycle    = uatomic.NewDuration(dataScannerStartDelay)
+	scannerIdleMode = uatomic.NewInt32(0) // default is throttled when idle
 )
 
 // initDataScanner will start the scanner in the background.
@@ -78,7 +80,7 @@ func initDataScanner(ctx context.Context, objAPI ObjectLayer) {
 			runDataScanner(ctx, objAPI)
 			duration := time.Duration(r.Float64() * float64(scannerCycle.Load()))
 			if duration < time.Second {
-				// Make sure to sleep atleast a second to avoid high CPU ticks.
+				// Make sure to sleep at least a second to avoid high CPU ticks.
 				duration = time.Second
 			}
 			time.Sleep(duration)
@@ -246,6 +248,8 @@ type folderScanner struct {
 	healObjectSelect      uint32 // Do a heal check on an object once every n cycles. Must divide into healFolderInclude
 	scanMode              madmin.HealScanMode
 
+	weSleep func() bool
+
 	disks       []StorageAPI
 	disksQuorum int
 
@@ -299,7 +303,7 @@ type folderScanner struct {
 // The returned cache will always be valid, but may not be updated from the existing.
 // Before each operation sleepDuration is called which can be used to temporarily halt the scanner.
 // If the supplied context is canceled the function will return at the first chance.
-func scanDataFolder(ctx context.Context, disks []StorageAPI, basePath string, cache dataUsageCache, getSize getSizeFn, scanMode madmin.HealScanMode) (dataUsageCache, error) {
+func scanDataFolder(ctx context.Context, disks []StorageAPI, basePath string, cache dataUsageCache, getSize getSizeFn, scanMode madmin.HealScanMode, weSleep func() bool) (dataUsageCache, error) {
 	switch cache.Info.Name {
 	case "", dataUsageRoot:
 		return cache, errors.New("internal error: root scan attempted")
@@ -316,6 +320,7 @@ func scanDataFolder(ctx context.Context, disks []StorageAPI, basePath string, ca
 		dataUsageScannerDebug: false,
 		healObjectSelect:      0,
 		scanMode:              scanMode,
+		weSleep:               weSleep,
 		updates:               cache.Info.updates,
 		updateCurrentPath:     updatePath,
 		disks:                 disks,
@@ -372,6 +377,8 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 	done := ctx.Done()
 	scannerLogPrefix := color.Green("folder-scanner:")
 
+	noWait := func() {}
+
 	thisHash := hashPath(folder.name)
 	// Store initial compaction state.
 	wasCompacted := into.Compacted
@@ -401,8 +408,10 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 		if !f.oldCache.Info.replication.Empty() && f.oldCache.Info.replication.Config.HasActiveRules(prefix, true) {
 			replicationCfg = f.oldCache.Info.replication
 		}
-		// Check if we can skip it due to bloom filter...
-		scannerSleeper.Sleep(ctx, dataScannerSleepPerFolder)
+
+		if f.weSleep() {
+			scannerSleeper.Sleep(ctx, dataScannerSleepPerFolder)
+		}
 
 		var existingFolders, newFolders []cachedFolder
 		var foundObjects bool
@@ -453,8 +462,11 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 				return nil
 			}
 
-			// Dynamic time delay.
-			wait := scannerSleeper.Timer(ctx)
+			wait := noWait
+			if f.weSleep() {
+				// Dynamic time delay.
+				wait = scannerSleeper.Timer(ctx)
+			}
 
 			// Get file size, ignore errors.
 			item := scannerItem{
@@ -699,13 +711,16 @@ func (f *folderScanner) scanFolder(ctx context.Context, folder cachedFolder, int
 				partial: func(entries metaCacheEntries, errs []error) {
 					entry, ok := entries.resolve(&resolver)
 					if !ok {
-						// check if we can get one entry atleast
+						// check if we can get one entry at least
 						// proceed to heal nonetheless, since
 						// this object might be dangling.
 						entry, _ = entries.firstFound()
 					}
-					// wait timer per object.
-					wait := scannerSleeper.Timer(ctx)
+					wait := noWait
+					if f.weSleep() {
+						// wait timer per object.
+						wait = scannerSleeper.Timer(ctx)
+					}
 					defer wait()
 					f.updateCurrentPath(entry.name)
 					stopFn := globalScannerMetrics.log(scannerMetricHealAbandonedObject, f.root, entry.name)
@@ -1457,7 +1472,7 @@ func (d *dynamicSleeper) Sleep(ctx context.Context, base time.Duration) {
 }
 
 // Update the current settings and cycle all waiting.
-// Parameters are the same as in the contructor.
+// Parameters are the same as in the constructor.
 func (d *dynamicSleeper) Update(factor float64, maxWait time.Duration) error {
 	d.mu.Lock()
 	defer d.mu.Unlock()
@@ -1465,7 +1480,7 @@ func (d *dynamicSleeper) Update(factor float64, maxWait time.Duration) error {
 		return nil
 	}
 	// Update values and cycle waiting.
-	close(d.cycle)
+	xioutil.SafeClose(d.cycle)
 	d.factor = factor
 	d.maxSleep = maxWait
 	d.cycle = make(chan struct{})

cmd/data-usage-cache.go

@@ -1,4 +1,4 @@
-// Copyright (c) 2015-2021 MinIO, Inc.
+// Copyright (c) 2015-2023 MinIO, Inc.
 //
 // This file is part of MinIO Object Storage stack
 //
@@ -32,6 +32,7 @@ import (
 	"time"
 
 	"github.com/cespare/xxhash/v2"
+	"github.com/dustin/go-humanize"
 	"github.com/klauspost/compress/zstd"
 	"github.com/minio/madmin-go/v3"
 	"github.com/minio/minio/internal/bucket/lifecycle"
@@ -46,6 +47,10 @@ import (
 // dataUsageHash is the hash type used.
 type dataUsageHash string
 
+// sizeHistogramV1 is size histogram V1, which has fewer intervals esp. between
+// 1024B and 1MiB.
+type sizeHistogramV1 [dataUsageBucketLenV1]uint64
+
 // sizeHistogram is a size histogram.
 type sizeHistogram [dataUsageBucketLen]uint64
 
@@ -204,8 +209,8 @@ func (r *replicationAllStats) clone() *replicationAllStats {
 	return &dst
 }
 
-//msgp:encode ignore dataUsageEntryV2 dataUsageEntryV3 dataUsageEntryV4 dataUsageEntryV5 dataUsageEntryV6
-//msgp:marshal ignore dataUsageEntryV2 dataUsageEntryV3 dataUsageEntryV4 dataUsageEntryV5 dataUsageEntryV6
+//msgp:encode ignore dataUsageEntryV2 dataUsageEntryV3 dataUsageEntryV4 dataUsageEntryV5 dataUsageEntryV6 dataUsageEntryV7
+//msgp:marshal ignore dataUsageEntryV2 dataUsageEntryV3 dataUsageEntryV4 dataUsageEntryV5 dataUsageEntryV6 dataUsageEntryV7
 
 //msgp:tuple dataUsageEntryV2
 type dataUsageEntryV2 struct {
@@ -263,14 +268,28 @@ type dataUsageEntryV6 struct {
 	Compacted        bool
 }
 
+type dataUsageEntryV7 struct {
+	Children dataUsageHashMap `msg:"ch"`
+	// These fields do no include any children.
+	Size             int64                `msg:"sz"`
+	Objects          uint64               `msg:"os"`
+	Versions         uint64               `msg:"vs"` // Versions that are not delete markers.
+	DeleteMarkers    uint64               `msg:"dms"`
+	ObjSizes         sizeHistogramV1      `msg:"szs"`
+	ObjVersions      versionsHistogram    `msg:"vh"`
+	ReplicationStats *replicationAllStats `msg:"rs,omitempty"`
+	AllTierStats     *allTierStats        `msg:"ats,omitempty"`
+	Compacted        bool                 `msg:"c"`
+}
+
 // dataUsageCache contains a cache of data usage entries latest version.
 type dataUsageCache struct {
 	Info  dataUsageCacheInfo
 	Cache map[string]dataUsageEntry
 }
 
-//msgp:encode ignore dataUsageCacheV2 dataUsageCacheV3 dataUsageCacheV4 dataUsageCacheV5 dataUsageCacheV6
-//msgp:marshal ignore dataUsageCacheV2 dataUsageCacheV3 dataUsageCacheV4 dataUsageCacheV5 dataUsageCacheV6
+//msgp:encode ignore dataUsageCacheV2 dataUsageCacheV3 dataUsageCacheV4 dataUsageCacheV5 dataUsageCacheV6 dataUsageCacheV7
+//msgp:marshal ignore dataUsageCacheV2 dataUsageCacheV3 dataUsageCacheV4 dataUsageCacheV5 dataUsageCacheV6 dataUsageCacheV7
 
 // dataUsageCacheV2 contains a cache of data usage entries version 2.
 type dataUsageCacheV2 struct {
@@ -302,6 +321,12 @@ type dataUsageCacheV6 struct {
 	Cache map[string]dataUsageEntryV6
 }
 
+// dataUsageCacheV7 contains a cache of data usage entries version 7.
+type dataUsageCacheV7 struct {
+	Info  dataUsageCacheInfo
+	Cache map[string]dataUsageEntryV7
+}
+
 //msgp:ignore dataUsageEntryInfo
 type dataUsageEntryInfo struct {
 	Name   string
@@ -750,11 +775,42 @@ func (h *sizeHistogram) add(size int64) {
 	}
 }
 
+// mergeV1 is used to migrate data usage cache from sizeHistogramV1 to
+// sizeHistogram
+func (h *sizeHistogram) mergeV1(v sizeHistogramV1) {
+	var oidx, nidx int
+	for oidx < len(v) {
+		intOld, intNew := ObjectsHistogramIntervalsV1[oidx], ObjectsHistogramIntervals[nidx]
+		// skip intervals that aren't common to both histograms
+		if intOld.start != intNew.start || intOld.end != intNew.end {
+			nidx++
+			continue
+		}
+		h[nidx] += v[oidx]
+		oidx++
+		nidx++
+	}
+}
+
 // toMap returns the map to a map[string]uint64.
 func (h *sizeHistogram) toMap() map[string]uint64 {
 	res := make(map[string]uint64, dataUsageBucketLen)
+	var splCount uint64
 	for i, count := range h {
-		res[ObjectsHistogramIntervals[i].name] = count
+		szInt := ObjectsHistogramIntervals[i]
+		switch {
+		case humanize.KiByte == szInt.start && szInt.end == humanize.MiByte-1:
+			// spl interval: [1024B, 1MiB)
+			res[szInt.name] = splCount
+		case humanize.KiByte <= szInt.start && szInt.end <= humanize.MiByte-1:
+			// intervals that fall within the spl interval above; they
+			// appear earlier in this array of intervals, see
+			// ObjectsHistogramIntervals
+			splCount += count
+			fallthrough
+		default:
+			res[szInt.name] = count
+		}
 	}
 	return res
 }
@@ -924,7 +980,7 @@ type objectIO interface {
 // The loader is optimistic and has no locking, but tries 5 times before giving up.
 // If the object is not found, a nil error with empty data usage cache is returned.
 func (d *dataUsageCache) load(ctx context.Context, store objectIO, name string) error {
-	// By defaut, empty data usage cache
+	// By default, empty data usage cache
 	*d = dataUsageCache{}
 
 	load := func(name string, timeout time.Duration) (bool, error) {
@@ -1027,7 +1083,8 @@ func (d *dataUsageCache) save(ctx context.Context, store objectIO, name string)
 // Bumping the cache version will drop data from previous versions
 // and write new data with the new version.
 const (
-	dataUsageCacheVerCurrent = 7
+	dataUsageCacheVerCurrent = 8
+	dataUsageCacheVerV7      = 7
 	dataUsageCacheVerV6      = 6
 	dataUsageCacheVerV5      = 5
 	dataUsageCacheVerV4      = 4
@@ -1275,6 +1332,34 @@ func (d *dataUsageCache) deserialize(r io.Reader) error {
 			}
 			d.Cache[k] = due
 		}
+		return nil
+	case dataUsageCacheVerV7:
+		// Zstd compressed.
+		dec, err := zstd.NewReader(r, zstd.WithDecoderConcurrency(2))
+		if err != nil {
+			return err
+		}
+		defer dec.Close()
+		dold := &dataUsageCacheV7{}
+		if err = dold.DecodeMsg(msgp.NewReader(dec)); err != nil {
+			return err
+		}
+		d.Info = dold.Info
+		d.Cache = make(map[string]dataUsageEntry, len(dold.Cache))
+		for k, v := range dold.Cache {
+			var szHist sizeHistogram
+			szHist.mergeV1(v.ObjSizes)
+			d.Cache[k] = dataUsageEntry{
+				Children:         v.Children,
+				Size:             v.Size,
+				Objects:          v.Objects,
+				Versions:         v.Versions,
+				ObjSizes:         szHist,
+				ReplicationStats: v.ReplicationStats,
+				Compacted:        v.Compacted,
+			}
+		}
+
 		return nil
 	case dataUsageCacheVerCurrent:
 		// Zstd compressed.

cmd/data-usage-cache_gen.go

@@ -1480,6 +1480,149 @@ func (z *dataUsageCacheV6) Msgsize() (s int) {
 	return
 }
 
+// DecodeMsg implements msgp.Decodable
+func (z *dataUsageCacheV7) 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 "Info":
+			err = z.Info.DecodeMsg(dc)
+			if err != nil {
+				err = msgp.WrapError(err, "Info")
+				return
+			}
+		case "Cache":
+			var zb0002 uint32
+			zb0002, err = dc.ReadMapHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "Cache")
+				return
+			}
+			if z.Cache == nil {
+				z.Cache = make(map[string]dataUsageEntryV7, zb0002)
+			} else if len(z.Cache) > 0 {
+				for key := range z.Cache {
+					delete(z.Cache, key)
+				}
+			}
+			for zb0002 > 0 {
+				zb0002--
+				var za0001 string
+				var za0002 dataUsageEntryV7
+				za0001, err = dc.ReadString()
+				if err != nil {
+					err = msgp.WrapError(err, "Cache")
+					return
+				}
+				err = za0002.DecodeMsg(dc)
+				if err != nil {
+					err = msgp.WrapError(err, "Cache", za0001)
+					return
+				}
+				z.Cache[za0001] = za0002
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *dataUsageCacheV7) 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 "Info":
+			bts, err = z.Info.UnmarshalMsg(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Info")
+				return
+			}
+		case "Cache":
+			var zb0002 uint32
+			zb0002, bts, err = msgp.ReadMapHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Cache")
+				return
+			}
+			if z.Cache == nil {
+				z.Cache = make(map[string]dataUsageEntryV7, zb0002)
+			} else if len(z.Cache) > 0 {
+				for key := range z.Cache {
+					delete(z.Cache, key)
+				}
+			}
+			for zb0002 > 0 {
+				var za0001 string
+				var za0002 dataUsageEntryV7
+				zb0002--
+				za0001, bts, err = msgp.ReadStringBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Cache")
+					return
+				}
+				bts, err = za0002.UnmarshalMsg(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Cache", za0001)
+					return
+				}
+				z.Cache[za0001] = za0002
+			}
+		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 *dataUsageCacheV7) Msgsize() (s int) {
+	s = 1 + 5 + z.Info.Msgsize() + 6 + msgp.MapHeaderSize
+	if z.Cache != nil {
+		for za0001, za0002 := range z.Cache {
+			_ = za0002
+			s += msgp.StringPrefixSize + len(za0001) + za0002.Msgsize()
+		}
+	}
+	return
+}
+
 // DecodeMsg implements msgp.Decodable
 func (z *dataUsageEntry) DecodeMsg(dc *msgp.Reader) (err error) {
 	var field []byte
@@ -2684,6 +2827,297 @@ func (z *dataUsageEntryV6) Msgsize() (s int) {
 	return
 }
 
+// DecodeMsg implements msgp.Decodable
+func (z *dataUsageEntryV7) 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 "ch":
+			err = z.Children.DecodeMsg(dc)
+			if err != nil {
+				err = msgp.WrapError(err, "Children")
+				return
+			}
+		case "sz":
+			z.Size, err = dc.ReadInt64()
+			if err != nil {
+				err = msgp.WrapError(err, "Size")
+				return
+			}
+		case "os":
+			z.Objects, err = dc.ReadUint64()
+			if err != nil {
+				err = msgp.WrapError(err, "Objects")
+				return
+			}
+		case "vs":
+			z.Versions, err = dc.ReadUint64()
+			if err != nil {
+				err = msgp.WrapError(err, "Versions")
+				return
+			}
+		case "dms":
+			z.DeleteMarkers, err = dc.ReadUint64()
+			if err != nil {
+				err = msgp.WrapError(err, "DeleteMarkers")
+				return
+			}
+		case "szs":
+			var zb0002 uint32
+			zb0002, err = dc.ReadArrayHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "ObjSizes")
+				return
+			}
+			if zb0002 != uint32(dataUsageBucketLenV1) {
+				err = msgp.ArrayError{Wanted: uint32(dataUsageBucketLenV1), Got: zb0002}
+				return
+			}
+			for za0001 := range z.ObjSizes {
+				z.ObjSizes[za0001], err = dc.ReadUint64()
+				if err != nil {
+					err = msgp.WrapError(err, "ObjSizes", za0001)
+					return
+				}
+			}
+		case "vh":
+			var zb0003 uint32
+			zb0003, err = dc.ReadArrayHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "ObjVersions")
+				return
+			}
+			if zb0003 != uint32(dataUsageVersionLen) {
+				err = msgp.ArrayError{Wanted: uint32(dataUsageVersionLen), Got: zb0003}
+				return
+			}
+			for za0002 := range z.ObjVersions {
+				z.ObjVersions[za0002], err = dc.ReadUint64()
+				if err != nil {
+					err = msgp.WrapError(err, "ObjVersions", za0002)
+					return
+				}
+			}
+		case "rs":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "ReplicationStats")
+					return
+				}
+				z.ReplicationStats = nil
+			} else {
+				if z.ReplicationStats == nil {
+					z.ReplicationStats = new(replicationAllStats)
+				}
+				err = z.ReplicationStats.DecodeMsg(dc)
+				if err != nil {
+					err = msgp.WrapError(err, "ReplicationStats")
+					return
+				}
+			}
+		case "ats":
+			if dc.IsNil() {
+				err = dc.ReadNil()
+				if err != nil {
+					err = msgp.WrapError(err, "AllTierStats")
+					return
+				}
+				z.AllTierStats = nil
+			} else {
+				if z.AllTierStats == nil {
+					z.AllTierStats = new(allTierStats)
+				}
+				err = z.AllTierStats.DecodeMsg(dc)
+				if err != nil {
+					err = msgp.WrapError(err, "AllTierStats")
+					return
+				}
+			}
+		case "c":
+			z.Compacted, err = dc.ReadBool()
+			if err != nil {
+				err = msgp.WrapError(err, "Compacted")
+				return
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *dataUsageEntryV7) 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 "ch":
+			bts, err = z.Children.UnmarshalMsg(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Children")
+				return
+			}
+		case "sz":
+			z.Size, bts, err = msgp.ReadInt64Bytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Size")
+				return
+			}
+		case "os":
+			z.Objects, bts, err = msgp.ReadUint64Bytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Objects")
+				return
+			}
+		case "vs":
+			z.Versions, bts, err = msgp.ReadUint64Bytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Versions")
+				return
+			}
+		case "dms":
+			z.DeleteMarkers, bts, err = msgp.ReadUint64Bytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "DeleteMarkers")
+				return
+			}
+		case "szs":
+			var zb0002 uint32
+			zb0002, bts, err = msgp.ReadArrayHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "ObjSizes")
+				return
+			}
+			if zb0002 != uint32(dataUsageBucketLenV1) {
+				err = msgp.ArrayError{Wanted: uint32(dataUsageBucketLenV1), Got: zb0002}
+				return
+			}
+			for za0001 := range z.ObjSizes {
+				z.ObjSizes[za0001], bts, err = msgp.ReadUint64Bytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "ObjSizes", za0001)
+					return
+				}
+			}
+		case "vh":
+			var zb0003 uint32
+			zb0003, bts, err = msgp.ReadArrayHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "ObjVersions")
+				return
+			}
+			if zb0003 != uint32(dataUsageVersionLen) {
+				err = msgp.ArrayError{Wanted: uint32(dataUsageVersionLen), Got: zb0003}
+				return
+			}
+			for za0002 := range z.ObjVersions {
+				z.ObjVersions[za0002], bts, err = msgp.ReadUint64Bytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "ObjVersions", za0002)
+					return
+				}
+			}
+		case "rs":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.ReplicationStats = nil
+			} else {
+				if z.ReplicationStats == nil {
+					z.ReplicationStats = new(replicationAllStats)
+				}
+				bts, err = z.ReplicationStats.UnmarshalMsg(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "ReplicationStats")
+					return
+				}
+			}
+		case "ats":
+			if msgp.IsNil(bts) {
+				bts, err = msgp.ReadNilBytes(bts)
+				if err != nil {
+					return
+				}
+				z.AllTierStats = nil
+			} else {
+				if z.AllTierStats == nil {
+					z.AllTierStats = new(allTierStats)
+				}
+				bts, err = z.AllTierStats.UnmarshalMsg(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "AllTierStats")
+					return
+				}
+			}
+		case "c":
+			z.Compacted, bts, err = msgp.ReadBoolBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Compacted")
+				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 *dataUsageEntryV7) Msgsize() (s int) {
+	s = 1 + 3 + z.Children.Msgsize() + 3 + msgp.Int64Size + 3 + msgp.Uint64Size + 3 + msgp.Uint64Size + 4 + msgp.Uint64Size + 4 + msgp.ArrayHeaderSize + (dataUsageBucketLenV1 * (msgp.Uint64Size)) + 3 + msgp.ArrayHeaderSize + (dataUsageVersionLen * (msgp.Uint64Size)) + 3
+	if z.ReplicationStats == nil {
+		s += msgp.NilSize
+	} else {
+		s += z.ReplicationStats.Msgsize()
+	}
+	s += 4
+	if z.AllTierStats == nil {
+		s += msgp.NilSize
+	} else {
+		s += z.AllTierStats.Msgsize()
+	}
+	s += 2 + msgp.BoolSize
+	return
+}
+
 // DecodeMsg implements msgp.Decodable
 func (z *dataUsageHash) DecodeMsg(dc *msgp.Reader) (err error) {
 	{
@@ -3705,6 +4139,84 @@ func (z *sizeHistogram) Msgsize() (s int) {
 	return
 }
 
+// DecodeMsg implements msgp.Decodable
+func (z *sizeHistogramV1) DecodeMsg(dc *msgp.Reader) (err error) {
+	var zb0001 uint32
+	zb0001, err = dc.ReadArrayHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	if zb0001 != uint32(dataUsageBucketLenV1) {
+		err = msgp.ArrayError{Wanted: uint32(dataUsageBucketLenV1), Got: zb0001}
+		return
+	}
+	for za0001 := range z {
+		z[za0001], err = dc.ReadUint64()
+		if err != nil {
+			err = msgp.WrapError(err, za0001)
+			return
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z *sizeHistogramV1) EncodeMsg(en *msgp.Writer) (err error) {
+	err = en.WriteArrayHeader(uint32(dataUsageBucketLenV1))
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for za0001 := range z {
+		err = en.WriteUint64(z[za0001])
+		if err != nil {
+			err = msgp.WrapError(err, za0001)
+			return
+		}
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z *sizeHistogramV1) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	o = msgp.AppendArrayHeader(o, uint32(dataUsageBucketLenV1))
+	for za0001 := range z {
+		o = msgp.AppendUint64(o, z[za0001])
+	}
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *sizeHistogramV1) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadArrayHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	if zb0001 != uint32(dataUsageBucketLenV1) {
+		err = msgp.ArrayError{Wanted: uint32(dataUsageBucketLenV1), Got: zb0001}
+		return
+	}
+	for za0001 := range z {
+		z[za0001], bts, err = msgp.ReadUint64Bytes(bts)
+		if err != nil {
+			err = msgp.WrapError(err, za0001)
+			return
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z *sizeHistogramV1) Msgsize() (s int) {
+	s = msgp.ArrayHeaderSize + (dataUsageBucketLenV1 * (msgp.Uint64Size))
+	return
+}
+
 // DecodeMsg implements msgp.Decodable
 func (z *tierStats) DecodeMsg(dc *msgp.Reader) (err error) {
 	var field []byte

cmd/data-usage-cache_gen_test.go

@@ -1084,6 +1084,119 @@ func BenchmarkDecodesizeHistogram(b *testing.B) {
 	}
 }
 
+func TestMarshalUnmarshalsizeHistogramV1(t *testing.T) {
+	v := sizeHistogramV1{}
+	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 BenchmarkMarshalMsgsizeHistogramV1(b *testing.B) {
+	v := sizeHistogramV1{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgsizeHistogramV1(b *testing.B) {
+	v := sizeHistogramV1{}
+	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 BenchmarkUnmarshalsizeHistogramV1(b *testing.B) {
+	v := sizeHistogramV1{}
+	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 TestEncodeDecodesizeHistogramV1(t *testing.T) {
+	v := sizeHistogramV1{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodesizeHistogramV1 Msgsize() is inaccurate")
+	}
+
+	vn := sizeHistogramV1{}
+	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 BenchmarkEncodesizeHistogramV1(b *testing.B) {
+	v := sizeHistogramV1{}
+	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 BenchmarkDecodesizeHistogramV1(b *testing.B) {
+	v := sizeHistogramV1{}
+	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 TestMarshalUnmarshaltierStats(t *testing.T) {
 	v := tierStats{}
 	bts, err := v.MarshalMsg(nil)

cmd/data-usage-cache_test.go

@@ -0,0 +1,98 @@
+// 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"
+
+	"github.com/dustin/go-humanize"
+)
+
+func TestSizeHistogramToMap(t *testing.T) {
+	tests := []struct {
+		sizes []int64
+		want  map[string]uint64
+	}{
+		{
+			sizes: []int64{100, 1000, 72_000, 100_000},
+			want: map[string]uint64{
+				"LESS_THAN_1024_B":         2,
+				"BETWEEN_64_KB_AND_256_KB": 2,
+				"BETWEEN_1024B_AND_1_MB":   2,
+			},
+		},
+		{
+			sizes: []int64{100, 1000, 2000, 100_000, 13 * humanize.MiByte},
+			want: map[string]uint64{
+				"LESS_THAN_1024_B":         2,
+				"BETWEEN_1024_B_AND_64_KB": 1,
+				"BETWEEN_64_KB_AND_256_KB": 1,
+				"BETWEEN_1024B_AND_1_MB":   2,
+				"BETWEEN_10_MB_AND_64_MB":  1,
+			},
+		},
+	}
+	for i, test := range tests {
+		t.Run(fmt.Sprintf("Test-%d", i), func(t *testing.T) {
+			var h sizeHistogram
+			for _, sz := range test.sizes {
+				h.add(sz)
+			}
+			got := h.toMap()
+			exp := test.want
+			// what is in exp is in got
+			for k := range exp {
+				if exp[k] != got[k] {
+					t.Fatalf("interval %s: Expected %d values but got %d values\n", k, exp[k], got[k])
+				}
+			}
+			// what is absent in exp is absent in got too
+			for k := range got {
+				if _, ok := exp[k]; !ok && got[k] > 0 {
+					t.Fatalf("Unexpected interval: %s has value %d\n", k, got[k])
+				}
+			}
+		})
+	}
+}
+
+func TestMigrateSizeHistogramFromV1(t *testing.T) {
+	tests := []struct {
+		v    sizeHistogramV1
+		want sizeHistogram
+	}{
+		{
+			v:    sizeHistogramV1{0: 10, 1: 20, 2: 3},
+			want: sizeHistogram{0: 10, 5: 20, 6: 3},
+		},
+		{
+			v:    sizeHistogramV1{0: 10, 1: 20, 2: 3, 3: 4, 4: 5, 5: 6, 6: 7},
+			want: sizeHistogram{0: 10, 5: 20, 6: 3, 7: 4, 8: 5, 9: 6, 10: 7},
+		},
+	}
+	for i, test := range tests {
+		t.Run(fmt.Sprintf("test-%d", i), func(t *testing.T) {
+			var got sizeHistogram
+			got.mergeV1(test.v)
+			if got != test.want {
+				t.Fatalf("Expected %v but got %v", test.want, got)
+			}
+		})
+	}
+}

cmd/data-usage.go

@@ -41,7 +41,7 @@ const (
 	dataUsageCacheName = ".usage-cache.bin"
 )
 
-// storeDataUsageInBackend will store all objects sent on the gui channel until closed.
+// storeDataUsageInBackend will store all objects sent on the dui channel until closed.
 func storeDataUsageInBackend(ctx context.Context, objAPI ObjectLayer, dui <-chan DataUsageInfo) {
 	attempts := 1
 	for dataUsageInfo := range dui {

cmd/data-usage_test.go

@@ -62,7 +62,9 @@ func TestDataUsageUpdate(t *testing.T) {
 		return
 	}
 
-	got, err := scanDataFolder(context.Background(), nil, base, dataUsageCache{Info: dataUsageCacheInfo{Name: bucket}}, getSize, 0)
+	weSleep := func() bool { return false }
+
+	got, err := scanDataFolder(context.Background(), nil, base, dataUsageCache{Info: dataUsageCacheInfo{Name: bucket}}, getSize, 0, weSleep)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -80,7 +82,7 @@ func TestDataUsageUpdate(t *testing.T) {
 			size:    1322310,
 			flatten: true,
 			objs:    8,
-			oSizes:  sizeHistogram{0: 2, 1: 6},
+			oSizes:  sizeHistogram{0: 2, 1: 3, 2: 2, 4: 1},
 		},
 		{
 			path:   "/",
@@ -92,7 +94,7 @@ func TestDataUsageUpdate(t *testing.T) {
 			path:   "/dir1",
 			size:   1302010,
 			objs:   5,
-			oSizes: sizeHistogram{0: 1, 1: 4},
+			oSizes: sizeHistogram{0: 1, 1: 1, 2: 2, 4: 1},
 		},
 		{
 			path:  "/dir1/dira",
@@ -165,7 +167,6 @@ func TestDataUsageUpdate(t *testing.T) {
 			size: 200,
 		},
 	}
-
 	createUsageTestFiles(t, base, bucket, files)
 	err = os.RemoveAll(filepath.Join(base, bucket, "dir1/dira/dirasub/dcfile"))
 	if err != nil {
@@ -173,7 +174,7 @@ func TestDataUsageUpdate(t *testing.T) {
 	}
 	// Changed dir must be picked up in this many cycles.
 	for i := 0; i < dataUsageUpdateDirCycles; i++ {
-		got, err = scanDataFolder(context.Background(), nil, base, got, getSize, 0)
+		got, err = scanDataFolder(context.Background(), nil, base, got, getSize, 0, weSleep)
 		got.Info.NextCycle++
 		if err != nil {
 			t.Fatal(err)
@@ -192,14 +193,14 @@ func TestDataUsageUpdate(t *testing.T) {
 			size:    363515,
 			flatten: true,
 			objs:    14,
-			oSizes:  sizeHistogram{0: 7, 1: 7},
+			oSizes:  sizeHistogram{0: 7, 1: 5, 2: 2},
 		},
 		{
 			path:    "/dir1",
 			size:    342210,
 			objs:    7,
 			flatten: false,
-			oSizes:  sizeHistogram{0: 2, 1: 5},
+			oSizes:  sizeHistogram{0: 2, 1: 3, 2: 2},
 		},
 		{
 			path:   "/newfolder",
@@ -280,7 +281,10 @@ func TestDataUsageUpdatePrefix(t *testing.T) {
 		}
 		return
 	}
-	got, err := scanDataFolder(context.Background(), nil, base, dataUsageCache{Info: dataUsageCacheInfo{Name: "bucket"}}, getSize, 0)
+
+	weSleep := func() bool { return false }
+
+	got, err := scanDataFolder(context.Background(), nil, base, dataUsageCache{Info: dataUsageCacheInfo{Name: "bucket"}}, getSize, 0, weSleep)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -303,7 +307,7 @@ func TestDataUsageUpdatePrefix(t *testing.T) {
 			path:   "flat",
 			size:   1322310 + expectSize,
 			objs:   8 + expectSize,
-			oSizes: sizeHistogram{0: 2 + expectSize, 1: 6},
+			oSizes: sizeHistogram{0: 2 + expectSize, 1: 3, 2: 2, 4: 1},
 		},
 		{
 			path:   "bucket/",
@@ -316,7 +320,7 @@ func TestDataUsageUpdatePrefix(t *testing.T) {
 			path:   "bucket/dir1",
 			size:   1302010,
 			objs:   5,
-			oSizes: sizeHistogram{0: 1, 1: 4},
+			oSizes: sizeHistogram{0: 1, 1: 1, 2: 2, 4: 1},
 		},
 		{
 			// Gets compacted at this level...
@@ -414,7 +418,7 @@ func TestDataUsageUpdatePrefix(t *testing.T) {
 	}
 	// Changed dir must be picked up in this many cycles.
 	for i := 0; i < dataUsageUpdateDirCycles; i++ {
-		got, err = scanDataFolder(context.Background(), nil, base, got, getSize, 0)
+		got, err = scanDataFolder(context.Background(), nil, base, got, getSize, 0, weSleep)
 		got.Info.NextCycle++
 		if err != nil {
 			t.Fatal(err)
@@ -431,13 +435,13 @@ func TestDataUsageUpdatePrefix(t *testing.T) {
 			path:   "flat",
 			size:   363515 + expectSize,
 			objs:   14 + expectSize,
-			oSizes: sizeHistogram{0: 7 + expectSize, 1: 7},
+			oSizes: sizeHistogram{0: 7 + expectSize, 1: 5, 2: 2},
 		},
 		{
 			path:   "bucket/dir1",
 			size:   342210,
 			objs:   7,
-			oSizes: sizeHistogram{0: 2, 1: 5},
+			oSizes: sizeHistogram{0: 2, 1: 3, 2: 2},
 		},
 		{
 			path:   "bucket/",
@@ -562,7 +566,8 @@ func TestDataUsageCacheSerialize(t *testing.T) {
 		}
 		return
 	}
-	want, err := scanDataFolder(context.Background(), nil, base, dataUsageCache{Info: dataUsageCacheInfo{Name: bucket}}, getSize, 0)
+	weSleep := func() bool { return false }
+	want, err := scanDataFolder(context.Background(), nil, base, dataUsageCache{Info: dataUsageCacheInfo{Name: bucket}}, getSize, 0, weSleep)
 	if err != nil {
 		t.Fatal(err)
 	}

cmd/encryption-v1.go

@@ -209,9 +209,6 @@ func DecryptETags(ctx context.Context, k kms.KMS, objects []ObjectInfo) error {
 // uploaded by the user using multipart mechanism:
 // initiate new multipart, upload part, complete upload
 func (o *ObjectInfo) isMultipart() bool {
-	if len(o.Parts) == 0 {
-		return false
-	}
 	_, encrypted := crypto.IsEncrypted(o.UserDefined)
 	if encrypted {
 		if !crypto.IsMultiPart(o.UserDefined) {
@@ -228,7 +225,7 @@ func (o *ObjectInfo) isMultipart() bool {
 	// Further check if this object is uploaded using multipart mechanism
 	// by the user and it is not about Erasure internally splitting the
 	// object into parts in PutObject()
-	return !(o.backendType == BackendErasure && len(o.ETag) == 32)
+	return len(o.ETag) != 32
 }
 
 // ParseSSECopyCustomerRequest parses the SSE-C header fields of the provided request.
@@ -667,7 +664,7 @@ func (d *DecryptBlocksReader) buildDecrypter(partID int) error {
 	mac.Write(partIDbin[:])
 	partEncryptionKey := mac.Sum(nil)
 
-	// Limit the reader, so the decryptor doesnt receive bytes
+	// Limit the reader, so the decryptor doesn't receive bytes
 	// from the next part (different DARE stream)
 	encLenToRead := d.parts[d.partIndex].Size - d.partEncRelOffset
 	decrypter, err := newDecryptReaderWithObjectKey(io.LimitReader(d.reader, encLenToRead), partEncryptionKey, d.startSeqNum)
@@ -754,7 +751,7 @@ func (o ObjectInfo) DecryptedSize() (int64, error) {
 // However, DecryptETag does not try to decrypt the ETag if
 // it consists of a 128 bit hex value (32 hex chars) and exactly
 // one '-' followed by a 32-bit number.
-// This special case adresses randomly-generated ETags generated
+// This special case addresses randomly-generated ETags generated
 // by the MinIO server when running in non-compat mode. These
 // random ETags are not encrypt.
 //
@@ -1082,6 +1079,30 @@ func (o *ObjectInfo) metadataDecrypter() objectMetaDecryptFn {
 	}
 }
 
+// decryptChecksums will attempt to decode checksums and return it/them if set.
+// if part > 0, and we have the checksum for the part that will be returned.
+func (o *ObjectInfo) decryptPartsChecksums() {
+	data := o.Checksum
+	if len(data) == 0 {
+		return
+	}
+	if _, encrypted := crypto.IsEncrypted(o.UserDefined); encrypted {
+		decrypted, err := o.metadataDecrypter()("object-checksum", data)
+		if err != nil {
+			logger.LogIf(GlobalContext, err)
+			return
+		}
+		data = decrypted
+	}
+	cs := hash.ReadPartCheckSums(data)
+	if len(cs) == len(o.Parts) {
+		for i := range o.Parts {
+			o.Parts[i].Checksums = cs[i]
+		}
+	}
+	return
+}
+
 // metadataEncryptFn provides an encryption function for metadata.
 // Will return nil, nil if unencrypted.
 func (o *ObjectInfo) metadataEncryptFn(headers http.Header) (objectMetaEncryptFn, error) {

cmd/endpoint.go

@@ -103,7 +103,7 @@ func (endpoint Endpoint) GridHost() string {
 
 // UpdateIsLocal - resolves the host and updates if it is local or not.
 func (endpoint *Endpoint) UpdateIsLocal() (err error) {
-	if !endpoint.IsLocal {
+	if endpoint.Host != "" {
 		endpoint.IsLocal, err = isLocalHost(endpoint.Hostname(), endpoint.Port(), globalMinioPort)
 		if err != nil {
 			return err
@@ -386,6 +386,31 @@ func (l EndpointServerPools) NEndpoints() (count int) {
 	return count
 }
 
+// GridHosts will return all peers, including local.
+// in websocket grid compatible format, The local peer
+// is returned as a separate string.
+func (l EndpointServerPools) GridHosts() (gridHosts []string, gridLocal string) {
+	seenHosts := set.NewStringSet()
+	for _, ep := range l {
+		for _, endpoint := range ep.Endpoints {
+			u := endpoint.GridHost()
+			if seenHosts.Contains(u) {
+				continue
+			}
+			seenHosts.Add(u)
+
+			// Set local endpoint
+			if endpoint.IsLocal {
+				gridLocal = u
+			}
+
+			gridHosts = append(gridHosts, u)
+		}
+	}
+
+	return gridHosts, gridLocal
+}
+
 // Hostnames - returns list of unique hostnames
 func (l EndpointServerPools) Hostnames() []string {
 	foundSet := set.NewStringSet()
@@ -435,7 +460,9 @@ func (l EndpointServerPools) peers() (peers []string, local string) {
 			peer := endpoint.Host
 			if endpoint.IsLocal {
 				if _, port := mustSplitHostPort(peer); port == globalMinioPort {
-					local = peer
+					if local == "" {
+						local = peer
+					}
 				}
 			}
 
@@ -522,6 +549,16 @@ func (endpoints Endpoints) UpdateIsLocal() error {
 					continue
 				}
 
+				if endpoints[i].Host == "" {
+					resolvedList[i] = true
+					endpoints[i].IsLocal = true
+					epsResolved++
+					if !foundLocal {
+						foundLocal = true
+					}
+					continue
+				}
+
 				// Log the message to console about the host resolving
 				reqInfo := (&logger.ReqInfo{}).AppendTags(
 					"host",
@@ -705,6 +742,17 @@ func (p PoolEndpointList) UpdateIsLocal() error {
 						continue
 					}
 
+					if endpoint.Host == "" {
+						if !foundLocal {
+							foundLocal = true
+						}
+						endpoint.IsLocal = true
+						endpoints[j] = endpoint
+						epsResolved++
+						resolvedList[endpoint] = true
+						continue
+					}
+
 					// Log the message to console about the host resolving
 					reqInfo := (&logger.ReqInfo{}).AppendTags(
 						"host",
@@ -859,6 +907,8 @@ func CreatePoolEndpoints(serverAddr string, poolsLayout ...poolDisksLayout) ([]E
 		return poolEndpoints, setupType, nil
 	}
 
+	uniqueArgs := set.NewStringSet()
+
 	for poolIdx, pool := range poolsLayout {
 		var endpoints Endpoints
 		for setIdx, setLayout := range pool.layout {
@@ -889,26 +939,10 @@ func CreatePoolEndpoints(serverAddr string, poolsLayout ...poolDisksLayout) ([]E
 		poolEndpoints[poolIdx] = endpoints
 	}
 
-	for _, endpoints := range poolEndpoints {
-		// Return Erasure setup when all endpoints are path style.
-		if endpoints[0].Type() == PathEndpointType {
-			setupType = ErasureSetupType
-		}
-		if endpoints[0].Type() == URLEndpointType && setupType != DistErasureSetupType {
-			setupType = DistErasureSetupType
-		}
-	}
-
-	if setupType == ErasureSetupType {
-		return poolEndpoints, setupType, nil
-	}
-
 	if err := poolEndpoints.UpdateIsLocal(); err != nil {
 		return nil, setupType, config.ErrInvalidErasureEndpoints(nil).Msg(err.Error())
 	}
 
-	uniqueArgs := set.NewStringSet()
-
 	for i, endpoints := range poolEndpoints {
 		// Here all endpoints are URL style.
 		endpointPathSet := set.NewStringSet()
@@ -918,7 +952,7 @@ func CreatePoolEndpoints(serverAddr string, poolsLayout ...poolDisksLayout) ([]E
 
 		for _, endpoint := range endpoints {
 			endpointPathSet.Add(endpoint.Path)
-			if endpoint.IsLocal {
+			if endpoint.IsLocal && endpoint.Host != "" {
 				localServerHostSet.Add(endpoint.Hostname())
 
 				_, port, err := net.SplitHostPort(endpoint.Host)
@@ -934,9 +968,8 @@ func CreatePoolEndpoints(serverAddr string, poolsLayout ...poolDisksLayout) ([]E
 		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 {
+		// and not setup in reverse proxy
+		if !orchestrated && !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.
@@ -944,14 +977,18 @@ func CreatePoolEndpoints(serverAddr string, poolsLayout ...poolDisksLayout) ([]E
 			hostIPCache := make(map[string]set.StringSet)
 			for _, endpoint := range endpoints {
 				host := endpoint.Hostname()
-				hostIPSet, ok := hostIPCache[host]
-				if !ok {
-					var err error
-					hostIPSet, err = getHostIP(host)
-					if err != nil {
-						return nil, setupType, config.ErrInvalidErasureEndpoints(nil).Msg(fmt.Sprintf("host '%s' cannot resolve: %s", host, err))
+				var hostIPSet set.StringSet
+				if host != "" {
+					var ok bool
+					hostIPSet, ok = hostIPCache[host]
+					if !ok {
+						var err error
+						hostIPSet, err = getHostIP(host)
+						if err != nil {
+							return nil, setupType, config.ErrInvalidErasureEndpoints(nil).Msg(fmt.Sprintf("host '%s' cannot resolve: %s", host, err))
+						}
+						hostIPCache[host] = hostIPSet
 					}
-					hostIPCache[host] = hostIPSet
 				}
 				if IPSet, ok := pathIPMap[endpoint.Path]; ok {
 					if !IPSet.Intersection(hostIPSet).IsEmpty() {
@@ -982,12 +1019,14 @@ func CreatePoolEndpoints(serverAddr string, poolsLayout ...poolDisksLayout) ([]E
 
 		// 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
+			if endpoints[i].Host != "" {
+				_, 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
+				}
 			}
 		}
 
@@ -1006,29 +1045,37 @@ func CreatePoolEndpoints(serverAddr string, poolsLayout ...poolDisksLayout) ([]E
 			// This means it is DistErasure setup.
 		}
 
-		poolEndpoints[i] = endpoints
-
 		for _, endpoint := range endpoints {
-			uniqueArgs.Add(endpoint.Host)
+			if endpoint.Host != "" {
+				uniqueArgs.Add(endpoint.Host)
+			} else {
+				uniqueArgs.Add(net.JoinHostPort("localhost", serverAddrPort))
+			}
 		}
 
 		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)
 	}
 
+	erasureType := len(uniqueArgs.ToSlice()) == 1
+
 	for _, endpoints := range poolEndpoints {
 		// Return Erasure setup when all endpoints are path style.
 		if endpoints[0].Type() == PathEndpointType {
 			setupType = ErasureSetupType
+			break
 		}
-		if endpoints[0].Type() == URLEndpointType && setupType != DistErasureSetupType {
-			setupType = DistErasureSetupType
+		if endpoints[0].Type() == URLEndpointType {
+			if erasureType {
+				setupType = ErasureSetupType
+			} else {
+				setupType = DistErasureSetupType
+			}
+			break
 		}
 	}
 

cmd/endpoint_test.go

@@ -258,7 +258,7 @@ func TestCreateEndpoints(t *testing.T) {
 			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},
-		}, DistErasureSetupType, nil},
+		}, ErasureSetupType, nil},
 		// DistErasure Setup with URLEndpointType having mixed naming to local host.
 		{"127.0.0.1:10000", []string{"http://localhost/d1", "http://localhost/d2", "http://127.0.0.1/d3", "http://127.0.0.1/d4"}, "", Endpoints{}, -1, fmt.Errorf("all local endpoints should not have different hostnames/ips")},
 

cmd/erasure-common.go

@@ -42,7 +42,7 @@ func (er erasureObjects) getOnlineDisks() (newDisks []StorageAPI) {
 			if disks[i] == nil {
 				return
 			}
-			di, err := disks[i].DiskInfo(context.Background(), false)
+			di, err := disks[i].DiskInfo(context.Background(), DiskInfoOptions{})
 			if err != nil || di.Healing {
 				// - Do not consume disks which are not reachable
 				//   unformatted or simply not accessible for some reason.

cmd/erasure-decode.go

@@ -25,6 +25,7 @@ import (
 	"sync"
 	"sync/atomic"
 
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 )
 
@@ -118,7 +119,7 @@ func (p *parallelReader) Read(dst [][]byte) ([][]byte, error) {
 	}
 
 	readTriggerCh := make(chan bool, len(p.readers))
-	defer close(readTriggerCh) // close the channel upon return
+	defer xioutil.SafeClose(readTriggerCh) // close the channel upon return
 
 	for i := 0; i < p.dataBlocks; i++ {
 		// Setup read triggers for p.dataBlocks number of reads so that it reads in parallel.
@@ -158,7 +159,7 @@ func (p *parallelReader) Read(dst [][]byte) ([][]byte, error) {
 			bufIdx := p.readerToBuf[i]
 			if p.buf[bufIdx] == nil {
 				// Reading first time on this disk, hence the buffer needs to be allocated.
-				// Subsequent reads will re-use this buffer.
+				// Subsequent reads will reuse this buffer.
 				p.buf[bufIdx] = make([]byte, p.shardSize)
 			}
 			// For the last shard, the shardsize might be less than previous shard sizes.

cmd/erasure-decode_test.go

@@ -144,7 +144,7 @@ func TestErasureDecode(t *testing.T) {
 		}
 		if err == nil {
 			if content := writer.Bytes(); !bytes.Equal(content, data[test.offset:test.offset+test.length]) {
-				t.Errorf("Test %d: read retruns wrong file content.", i)
+				t.Errorf("Test %d: read returns wrong file content.", i)
 			}
 		}
 

cmd/erasure-healing-common.go

@@ -191,19 +191,6 @@ func filterOnlineDisksInplace(fi FileInfo, partsMetadata []FileInfo, onlineDisks
 	}
 }
 
-// Extracts list of disk mtimes from FileInfo slice and returns, skips
-// slice elements that have errors.
-func listObjectDiskMtimes(partsMetadata []FileInfo) (diskMTimes []time.Time) {
-	diskMTimes = bootModtimes(len(partsMetadata))
-	for index, metadata := range partsMetadata {
-		if metadata.IsValid() {
-			// Once the file is found, save the disk mtime saved on disk.
-			diskMTimes[index] = metadata.DiskMTime
-		}
-	}
-	return diskMTimes
-}
-
 // Notes:
 // There are 5 possible states a disk could be in,
 // 1. __online__             - has the latest copy of xl.meta - returned by listOnlineDisks
@@ -277,13 +264,6 @@ func disksWithAllParts(ctx context.Context, onlineDisks []StorageAPI, partsMetad
 	errs []error, latestMeta FileInfo, bucket, object string,
 	scanMode madmin.HealScanMode) ([]StorageAPI, []error, time.Time,
 ) {
-	var diskMTime time.Time
-	var shardFix bool
-	if !latestMeta.DataShardFixed() {
-		diskMTime = pickValidDiskTimeWithQuorum(partsMetadata,
-			latestMeta.Erasure.DataBlocks)
-	}
-
 	availableDisks := make([]StorageAPI, len(onlineDisks))
 	dataErrs := make([]error, len(onlineDisks))
 
@@ -351,23 +331,6 @@ func disksWithAllParts(ctx context.Context, onlineDisks []StorageAPI, partsMetad
 			}
 		}
 
-		if !diskMTime.Equal(timeSentinel) && !diskMTime.IsZero() {
-			if !partsMetadata[i].AcceptableDelta(diskMTime, shardDiskTimeDelta) {
-				// not with in acceptable delta, skip.
-				// If disk mTime mismatches it is considered outdated
-				// https://github.com/minio/minio/pull/13803
-				//
-				// This check only is active if we could find maximally
-				// occurring disk mtimes that are somewhat same across
-				// the quorum. Allowing to skip those shards which we
-				// might think are wrong.
-				shardFix = true
-				partsMetadata[i] = FileInfo{}
-				dataErrs[i] = errFileCorrupt
-				continue
-			}
-		}
-
 		// Always check data, if we got it.
 		if (len(meta.Data) > 0 || meta.Size == 0) && len(meta.Parts) > 0 {
 			checksumInfo := meta.Erasure.GetChecksumInfo(meta.Parts[0].Number)
@@ -410,10 +373,5 @@ func disksWithAllParts(ctx context.Context, onlineDisks []StorageAPI, partsMetad
 		}
 	}
 
-	if shardFix {
-		// Only when shard is fixed return an appropriate disk mtime value.
-		return availableDisks, dataErrs, diskMTime
-	} // else return timeSentinel for disk mtime
-
 	return availableDisks, dataErrs, timeSentinel
 }

cmd/erasure-healing-common_test.go

@@ -31,7 +31,7 @@ import (
 
 // Returns the latest updated FileInfo files and error in case of failure.
 func getLatestFileInfo(ctx context.Context, partsMetadata []FileInfo, defaultParityCount int, errs []error) (FileInfo, error) {
-	// There should be atleast half correct entries, if not return failure
+	// There should be at least half correct entries, if not return failure
 	expectedRQuorum := len(partsMetadata) / 2
 	if defaultParityCount == 0 {
 		// if parity count is '0', we expected all entries to be present.
@@ -57,7 +57,7 @@ func getLatestFileInfo(ctx context.Context, partsMetadata []FileInfo, defaultPar
 		return FileInfo{}, errErasureReadQuorum
 	}
 
-	// Interate through all the modTimes and count the FileInfo(s) with latest time.
+	// Iterate through all the modTimes and count the FileInfo(s) with latest time.
 	for index, t := range modTimes {
 		if partsMetadata[index].IsValid() && t.Equal(modTime) {
 			latestFileInfo = partsMetadata[index]

cmd/erasure-healing.go

@@ -33,8 +33,6 @@ import (
 	"github.com/minio/pkg/v2/sync/errgroup"
 )
 
-const reservedMetadataPrefixLowerDataShardFix = ReservedMetadataPrefixLower + "data-shard-fix"
-
 //go:generate stringer -type=healingMetric -trimprefix=healingMetric $GOFILE
 
 type healingMetric uint8
@@ -45,29 +43,6 @@ const (
 	healingMetricCheckAbandonedParts
 )
 
-// AcceptableDelta returns 'true' if the fi.DiskMTime is under
-// acceptable delta of "delta" duration with maxTime.
-//
-// This code is primarily used for heuristic detection of
-// incorrect shards, as per https://github.com/minio/minio/pull/13803
-//
-// This check only is active if we could find maximally
-// occurring disk mtimes that are somewhat same across
-// the quorum. Allowing to skip those shards which we
-// might think are wrong.
-func (fi FileInfo) AcceptableDelta(maxTime time.Time, delta time.Duration) bool {
-	diff := maxTime.Sub(fi.DiskMTime)
-	if diff < 0 {
-		diff = -diff
-	}
-	return diff < delta
-}
-
-// DataShardFixed - data shard fixed?
-func (fi FileInfo) DataShardFixed() bool {
-	return fi.Metadata[reservedMetadataPrefixLowerDataShardFix] == "true"
-}
-
 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()
@@ -108,7 +83,7 @@ func (er erasureObjects) listAndHeal(bucket, prefix string, scanMode madmin.Heal
 		partial: func(entries metaCacheEntries, _ []error) {
 			entry, ok := entries.resolve(&resolver)
 			if !ok {
-				// check if we can get one entry atleast
+				// check if we can get one entry at least
 				// proceed to heal nonetheless.
 				entry, _ = entries.firstFound()
 			}
@@ -128,159 +103,6 @@ func (er erasureObjects) listAndHeal(bucket, prefix string, scanMode madmin.Heal
 	return nil
 }
 
-// HealBucket heals a bucket if it doesn't exist on one of the disks, additionally
-// also heals the missing entries for bucket metadata files
-// `policy.json, notification.xml, listeners.json`.
-func (er erasureObjects) HealBucket(ctx context.Context, bucket string, opts madmin.HealOpts) (
-	result madmin.HealResultItem, err error,
-) {
-	storageDisks := er.getDisks()
-	storageEndpoints := er.getEndpoints()
-
-	// Heal bucket.
-	return er.healBucket(ctx, storageDisks, storageEndpoints, bucket, opts)
-}
-
-// Heal bucket - create buckets on disks where it does not exist.
-func (er erasureObjects) healBucket(ctx context.Context, storageDisks []StorageAPI, storageEndpoints []Endpoint, bucket string, opts madmin.HealOpts) (res madmin.HealResultItem, err error) {
-	// get write quorum for an object
-	writeQuorum := len(storageDisks) - er.defaultParityCount
-	if writeQuorum == er.defaultParityCount {
-		writeQuorum++
-	}
-
-	if globalTrace.NumSubscribers(madmin.TraceHealing) > 0 {
-		startTime := time.Now()
-		defer func() {
-			healTrace(healingMetricBucket, startTime, bucket, "", &opts, err, &res)
-		}()
-	}
-
-	// Initialize sync waitgroup.
-	g := errgroup.WithNErrs(len(storageDisks))
-
-	// Disk states slices
-	beforeState := make([]string, len(storageDisks))
-	afterState := make([]string, len(storageDisks))
-
-	// Make a volume entry on all underlying storage disks.
-	for index := range storageDisks {
-		index := index
-		g.Go(func() error {
-			if storageDisks[index] == nil {
-				beforeState[index] = madmin.DriveStateOffline
-				afterState[index] = madmin.DriveStateOffline
-				return errDiskNotFound
-			}
-
-			beforeState[index] = madmin.DriveStateOk
-			afterState[index] = madmin.DriveStateOk
-
-			if bucket == minioReservedBucket {
-				return nil
-			}
-
-			if _, serr := storageDisks[index].StatVol(ctx, bucket); serr != nil {
-				if serr == errDiskNotFound {
-					beforeState[index] = madmin.DriveStateOffline
-					afterState[index] = madmin.DriveStateOffline
-					return serr
-				}
-				if serr != errVolumeNotFound {
-					beforeState[index] = madmin.DriveStateCorrupt
-					afterState[index] = madmin.DriveStateCorrupt
-					return serr
-				}
-
-				beforeState[index] = madmin.DriveStateMissing
-				afterState[index] = madmin.DriveStateMissing
-
-				// mutate only if not a dry-run
-				if opts.DryRun {
-					return nil
-				}
-
-				return serr
-			}
-			return nil
-		}, index)
-	}
-
-	errs := g.Wait()
-
-	// Initialize heal result info
-	res = madmin.HealResultItem{
-		Type:         madmin.HealItemBucket,
-		Bucket:       bucket,
-		DiskCount:    len(storageDisks),
-		ParityBlocks: er.defaultParityCount,
-		DataBlocks:   len(storageDisks) - er.defaultParityCount,
-	}
-
-	for i := range beforeState {
-		res.Before.Drives = append(res.Before.Drives, madmin.HealDriveInfo{
-			UUID:     "",
-			Endpoint: storageEndpoints[i].String(),
-			State:    beforeState[i],
-		})
-	}
-
-	reducedErr := reduceReadQuorumErrs(ctx, errs, bucketOpIgnoredErrs, res.DataBlocks)
-	if errors.Is(reducedErr, errVolumeNotFound) && !opts.Recreate {
-		for i := range beforeState {
-			res.After.Drives = append(res.After.Drives, madmin.HealDriveInfo{
-				UUID:     "",
-				Endpoint: storageEndpoints[i].String(),
-				State:    madmin.DriveStateOk,
-			})
-		}
-		return res, nil
-	}
-
-	// Initialize sync waitgroup.
-	g = errgroup.WithNErrs(len(storageDisks))
-
-	// Make a volume entry on all underlying storage disks.
-	for index := range storageDisks {
-		index := index
-		g.Go(func() error {
-			if beforeState[index] == madmin.DriveStateMissing {
-				makeErr := storageDisks[index].MakeVol(ctx, bucket)
-				if makeErr == nil {
-					afterState[index] = madmin.DriveStateOk
-				}
-				return makeErr
-			}
-			return errs[index]
-		}, index)
-	}
-
-	errs = g.Wait()
-
-	reducedErr = reduceWriteQuorumErrs(ctx, errs, bucketOpIgnoredErrs, writeQuorum)
-	if reducedErr != nil {
-		// If we have exactly half the drives not available,
-		// we should still allow HealBucket to not return error.
-		// this is necessary for starting the server.
-		readQuorum := res.DataBlocks
-		switch reduceReadQuorumErrs(ctx, errs, nil, readQuorum) {
-		case nil:
-		case errDiskNotFound:
-		default:
-			return res, reducedErr
-		}
-	}
-
-	for i := range afterState {
-		res.After.Drives = append(res.After.Drives, madmin.HealDriveInfo{
-			UUID:     "",
-			Endpoint: storageEndpoints[i].String(),
-			State:    afterState[i],
-		})
-	}
-	return res, nil
-}
-
 // listAllBuckets lists all buckets from all disks. It also
 // returns the occurrence of each buckets in all disks
 func listAllBuckets(ctx context.Context, storageDisks []StorageAPI, healBuckets map[string]VolInfo, readQuorum int) error {
@@ -370,10 +192,8 @@ 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.
@@ -466,7 +286,7 @@ func (er *erasureObjects) healObject(ctx context.Context, bucket string, object
 	// used here for reconstruction. This is done to ensure that
 	// we do not skip drives that have inconsistent metadata to be
 	// skipped from purging when they are stale.
-	availableDisks, dataErrs, diskMTime := disksWithAllParts(ctx, onlineDisks, partsMetadata,
+	availableDisks, dataErrs, _ := disksWithAllParts(ctx, onlineDisks, partsMetadata,
 		errs, latestMeta, bucket, object, scanMode)
 
 	var erasure Erasure
@@ -489,7 +309,7 @@ func (er *erasureObjects) healObject(ctx context.Context, bucket string, object
 			erasure.ShardFileSize(latestMeta.Parts[0].ActualSize) < smallFileThreshold)
 	}
 
-	result.ObjectSize, err = latestMeta.GetActualSize()
+	result.ObjectSize, err = latestMeta.ToObjectInfo(bucket, object, true).GetActualSize()
 	if err != nil {
 		return result, err
 	}
@@ -760,7 +580,8 @@ 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
 		}
 
@@ -768,7 +589,10 @@ func (er *erasureObjects) healObject(ctx context.Context, bucket string, object
 		// - Remove any remaining parts from outdated disks from before transition.
 		if recreate || partsMetadata[i].IsRemote() {
 			rmDataDir := partsMetadata[i].DataDir
-			disk.DeleteVol(ctx, pathJoin(bucket, encodeDirObject(object), rmDataDir), true)
+			disk.Delete(ctx, bucket, pathJoin(encodeDirObject(object), rmDataDir), DeleteOptions{
+				Immediate: true,
+				Recursive: true,
+			})
 		}
 
 		for i, v := range result.Before.Drives {
@@ -778,20 +602,6 @@ func (er *erasureObjects) healObject(ctx context.Context, bucket string, object
 		}
 	}
 
-	if !diskMTime.Equal(timeSentinel) && !diskMTime.IsZero() {
-		// Update metadata to indicate special fix.
-		_, err = er.PutObjectMetadata(ctx, bucket, object, ObjectOptions{
-			NoLock: true,
-			UserDefined: map[string]string{
-				reservedMetadataPrefixLowerDataShardFix: "true",
-				// another reserved metadata to capture original disk-mtime
-				// captured for this version of the object, to be used
-				// possibly in future to heal other versions if possible.
-				ReservedMetadataPrefixLower + "disk-mtime": diskMTime.String(),
-			},
-		})
-	}
-
 	return result, nil
 }
 
@@ -1020,6 +830,22 @@ func isAllNotFound(errs []error) bool {
 	return len(errs) > 0
 }
 
+// isAllBucketsNotFound will return true if all the errors are either errFileNotFound
+// or errFileCorrupt
+// A 0 length slice will always return false.
+func isAllBucketsNotFound(errs []error) bool {
+	if len(errs) == 0 {
+		return false
+	}
+	notFoundCount := 0
+	for _, err := range errs {
+		if err != nil && errors.Is(err, errVolumeNotFound) {
+			notFoundCount++
+		}
+	}
+	return len(errs) == notFoundCount
+}
+
 // ObjectDir is considered dangling/corrupted if any only
 // if total disks - a combination of corrupted and missing
 // files is lesser than N/2+1 number of disks.
@@ -1045,7 +871,7 @@ func isObjectDirDangling(errs []error) (ok bool) {
 	return found < notFound && found > 0
 }
 
-// Object is considered dangling/corrupted if any only
+// Object is considered dangling/corrupted if and only
 // if total disks - a combination of corrupted and missing
 // files is lesser than number of data blocks.
 func isObjectDangling(metaArr []FileInfo, errs []error, dataErrs []error) (validMeta FileInfo, ok bool) {
@@ -1054,21 +880,25 @@ func isObjectDangling(metaArr []FileInfo, errs []error, dataErrs []error) (valid
 	// or when xl.meta is not readable in read quorum disks.
 	danglingErrsCount := func(cerrs []error) (int, int, int) {
 		var (
-			notFoundCount     int
-			corruptedCount    int
-			diskNotFoundCount int
+			notFoundCount      int
+			corruptedCount     int
+			driveNotFoundCount int
 		)
 		for _, readErr := range cerrs {
+			if readErr == nil {
+				continue
+			}
 			switch {
 			case errors.Is(readErr, errFileNotFound) || errors.Is(readErr, errFileVersionNotFound):
 				notFoundCount++
 			case errors.Is(readErr, errFileCorrupt):
 				corruptedCount++
-			case errors.Is(readErr, errDiskNotFound):
-				diskNotFoundCount++
+			default:
+				// All other errors are non-actionable
+				driveNotFoundCount++
 			}
 		}
-		return notFoundCount, corruptedCount, diskNotFoundCount
+		return notFoundCount, corruptedCount, driveNotFoundCount
 	}
 
 	ndataErrs := make([]error, len(dataErrs))

cmd/erasure-healing_test.go

@@ -362,7 +362,7 @@ func TestHealing(t *testing.T) {
 		t.Fatal(err)
 	}
 	// This would create the bucket.
-	_, err = er.HealBucket(ctx, bucket, madmin.HealOpts{
+	_, err = obj.HealBucket(ctx, bucket, madmin.HealOpts{
 		DryRun: false,
 		Remove: false,
 	})
@@ -543,7 +543,7 @@ func TestHealingVersioned(t *testing.T) {
 		t.Fatal(err)
 	}
 	// This would create the bucket.
-	_, err = er.HealBucket(ctx, bucket, madmin.HealOpts{
+	_, err = obj.HealBucket(ctx, bucket, madmin.HealOpts{
 		DryRun: false,
 		Remove: false,
 	})
@@ -1067,7 +1067,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	for i := 0; i < (nfi.Erasure.DataBlocks + nfi.Erasure.ParityBlocks); i++ {
 		stats, _ := erasureDisks[i].StatInfoFile(context.Background(), bucket, pathJoin(object, xlStorageFormatFile), false)
 		if len(stats) != 0 {
-			t.Errorf("Expected xl.meta file to be not present, but succeeeded")
+			t.Errorf("Expected xl.meta file to be not present, but succeeded")
 		}
 	}
 }

cmd/erasure-metadata-utils.go

@@ -64,7 +64,7 @@ func reduceErrs(errs []error, ignoredErrs []error) (maxCount int, maxErr error)
 		if IsErrIgnored(err, ignoredErrs...) {
 			continue
 		}
-		// Errors due to context cancelation may be wrapped - group them by context.Canceled.
+		// Errors due to context cancellation may be wrapped - group them by context.Canceled.
 		if errors.Is(err, context.Canceled) {
 			errorCounts[context.Canceled]++
 			continue

cmd/erasure-metadata-utils_test.go

@@ -94,7 +94,7 @@ func TestReduceErrs(t *testing.T) {
 			},
 			nil, nil,
 		},
-		// Checks if wrapped context cancelation errors are grouped as one.
+		// Checks if wrapped context cancellation errors are grouped as one.
 		{canceledErrs, nil, context.Canceled},
 	}
 	// Validates list of all the testcases for returning valid errors.

cmd/erasure-metadata.go

@@ -22,18 +22,15 @@ import (
 	"encoding/hex"
 	"fmt"
 	"sort"
-	"strconv"
 	"strings"
 	"time"
 
 	"github.com/minio/minio/internal/amztime"
 	"github.com/minio/minio/internal/bucket/replication"
-	"github.com/minio/minio/internal/crypto"
 	"github.com/minio/minio/internal/hash/sha256"
 	xhttp "github.com/minio/minio/internal/http"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/sync/errgroup"
-	"github.com/minio/sio"
 )
 
 // Object was stored with additional erasure codes due to degraded system at upload time
@@ -88,52 +85,6 @@ func (fi FileInfo) IsValid() bool {
 		correctIndexes)
 }
 
-func (fi FileInfo) checkMultipart() (int64, bool) {
-	if len(fi.Parts) == 0 {
-		return 0, false
-	}
-	if !crypto.IsMultiPart(fi.Metadata) {
-		return 0, false
-	}
-	var size int64
-	for _, part := range fi.Parts {
-		psize, err := sio.DecryptedSize(uint64(part.Size))
-		if err != nil {
-			return 0, false
-		}
-		size += int64(psize)
-	}
-
-	return size, len(extractETag(fi.Metadata)) != 32
-}
-
-// GetActualSize - returns the actual size of the stored object
-func (fi FileInfo) GetActualSize() (int64, error) {
-	if _, ok := fi.Metadata[ReservedMetadataPrefix+"compression"]; ok {
-		sizeStr, ok := fi.Metadata[ReservedMetadataPrefix+"actual-size"]
-		if !ok {
-			return -1, errInvalidDecompressedSize
-		}
-		size, err := strconv.ParseInt(sizeStr, 10, 64)
-		if err != nil {
-			return -1, errInvalidDecompressedSize
-		}
-		return size, nil
-	}
-	if _, ok := crypto.IsEncrypted(fi.Metadata); ok {
-		size, ok := fi.checkMultipart()
-		if !ok {
-			size, err := sio.DecryptedSize(uint64(fi.Size))
-			if err != nil {
-				err = errObjectTampered // assign correct error type
-			}
-			return int64(size), err
-		}
-		return size, nil
-	}
-	return fi.Size, nil
-}
-
 // ToObjectInfo - Converts metadata to object info.
 func (fi FileInfo) ToObjectInfo(bucket, object string, versioned bool) ObjectInfo {
 	object = decodeDirObject(object)
@@ -166,7 +117,6 @@ func (fi FileInfo) ToObjectInfo(bucket, object string, versioned bool) ObjectInf
 			objInfo.Expires = t.UTC()
 		}
 	}
-	objInfo.backendType = BackendErasure
 
 	// Extract etag from metadata.
 	objInfo.ETag = extractETag(fi.Metadata)
@@ -424,17 +374,6 @@ func findFileInfoInQuorum(ctx context.Context, metaArr []FileInfo, modTime time.
 	return FileInfo{}, errErasureReadQuorum
 }
 
-func pickValidDiskTimeWithQuorum(metaArr []FileInfo, quorum int) time.Time {
-	diskMTimes := listObjectDiskMtimes(metaArr)
-
-	diskMTime, diskMaxima := commonTimeAndOccurence(diskMTimes, shardDiskTimeDelta)
-	if diskMaxima >= quorum {
-		return diskMTime
-	}
-
-	return timeSentinel
-}
-
 // pickValidFileInfo - picks one valid FileInfo content and returns from a
 // slice of FileInfo.
 func pickValidFileInfo(ctx context.Context, metaArr []FileInfo, modTime time.Time, etag string, quorum int) (FileInfo, error) {
@@ -533,7 +472,7 @@ func listObjectParities(partsMetadata []FileInfo, errs []error) (parities []int)
 // readQuorum is the min required disks to read data.
 // writeQuorum is the min required disks to write data.
 func objectQuorumFromMeta(ctx context.Context, partsMetaData []FileInfo, errs []error, defaultParityCount int) (objectReadQuorum, objectWriteQuorum int, err error) {
-	// There should be atleast half correct entries, if not return failure
+	// There should be at least half correct entries, if not return failure
 	expectedRQuorum := len(partsMetaData) / 2
 	if defaultParityCount == 0 {
 		// if parity count is '0', we expected all entries to be present.

cmd/erasure-multipart.go

@@ -103,7 +103,7 @@ func (er erasureObjects) checkUploadIDExists(ctx context.Context, bucket, object
 	return fi, partsMetadata, err
 }
 
-// Removes part.meta given by partName belonging to a mulitpart upload from minioMetaBucket
+// Removes part.meta given by partName belonging to a multipart upload from minioMetaBucket
 func (er erasureObjects) removePartMeta(bucket, object, uploadID, dataDir string, partNumber int) {
 	uploadIDPath := er.getUploadIDDir(bucket, object, uploadID)
 	curpartPath := pathJoin(uploadIDPath, dataDir, fmt.Sprintf("part.%d", partNumber))
@@ -127,7 +127,7 @@ func (er erasureObjects) removePartMeta(bucket, object, uploadID, dataDir string
 	g.Wait()
 }
 
-// Removes part given by partName belonging to a mulitpart upload from minioMetaBucket
+// Removes part given by partName belonging to a multipart upload from minioMetaBucket
 func (er erasureObjects) removeObjectPart(bucket, object, uploadID, dataDir string, partNumber int) {
 	uploadIDPath := er.getUploadIDDir(bucket, object, uploadID)
 	curpartPath := pathJoin(uploadIDPath, dataDir, fmt.Sprintf("part.%d", partNumber))
@@ -656,18 +656,18 @@ func (er erasureObjects) PutObjectPart(ctx context.Context, bucket, object, uplo
 	var buffer []byte
 	switch size := data.Size(); {
 	case size == 0:
-		buffer = make([]byte, 1) // Allocate atleast a byte to reach EOF
+		buffer = make([]byte, 1) // Allocate at least a byte to reach EOF
 	case size == -1:
 		if size := data.ActualSize(); size > 0 && size < fi.Erasure.BlockSize {
 			// 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

cmd/erasure-object.go

@@ -227,30 +227,6 @@ func (er erasureObjects) GetObjectNInfo(ctx context.Context, bucket, object stri
 		return nil, toObjectErr(err, bucket, object)
 	}
 
-	if !fi.DataShardFixed() {
-		diskMTime := pickValidDiskTimeWithQuorum(metaArr, fi.Erasure.DataBlocks)
-		if !diskMTime.Equal(timeSentinel) && !diskMTime.IsZero() {
-			for index := range onlineDisks {
-				if onlineDisks[index] == OfflineDisk {
-					continue
-				}
-				if !metaArr[index].IsValid() {
-					continue
-				}
-				if !metaArr[index].AcceptableDelta(diskMTime, shardDiskTimeDelta) {
-					// If disk mTime mismatches it is considered outdated
-					// https://github.com/minio/minio/pull/13803
-					//
-					// This check only is active if we could find maximally
-					// occurring disk mtimes that are somewhat same across
-					// the quorum. Allowing to skip those shards which we
-					// might think are wrong.
-					onlineDisks[index] = OfflineDisk
-				}
-			}
-		}
-	}
-
 	objInfo := fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended)
 	if objInfo.DeleteMarker {
 		if opts.VersionID == "" {
@@ -476,40 +452,50 @@ 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) && errors.Is(errs[i], errDiskNotFound) || i < len(dataErrs) && errors.Is(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)
 		}
@@ -536,11 +522,26 @@ 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)
 		}
 
-		g.Wait()
+		rmDisks := make(map[string]string, len(disks))
+		for index, err := range g.Wait() {
+			var errStr, diskName string
+			if err != nil {
+				errStr = err.Error()
+			} else {
+				errStr = "<nil>"
+			}
+			if disks[index] != nil {
+				diskName = disks[index].String()
+			} else {
+				diskName = fmt.Sprintf("disk-%d", index)
+			}
+			rmDisks[diskName] = errStr
+		}
+		tags["cleanupResult"] = rmDisks
 	}
 	return m, err
 }
@@ -743,8 +744,6 @@ func readAllXL(ctx context.Context, disks []StorageAPI, bucket, object string, r
 }
 
 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)
@@ -755,14 +754,16 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 	done := make(chan bool, er.setDriveCount)
 	disks := er.getDisks()
 
-	mrfCheck := make(chan FileInfo)
-	defer close(mrfCheck)
-
 	ropts := ReadOptions{
 		ReadData: readData,
 		Healing:  false,
 	}
 
+	mrfCheck := make(chan FileInfo)
+	defer xioutil.SafeClose(mrfCheck)
+
+	var rw sync.Mutex
+
 	// Ask for all disks first;
 	go func() {
 		ctx, cancel := context.WithCancel(ctx)
@@ -774,56 +775,65 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 				done <- false
 				continue
 			}
+			if !disk.IsOnline() {
+				done <- false
+				continue
+			}
 			wg.Add(1)
 			go func(i int, disk StorageAPI) {
 				defer wg.Done()
-				var fi FileInfo
-				var err error
+
+				var (
+					fi  FileInfo
+					rfi RawFileInfo
+					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()
+					rfi, err = readRawFileInfo(ctx, disk, bucket, object, readData)
 					if err == nil {
-						fi, err = fileInfoFromRaw(ri, bucket, object, readData, opts.InclFreeVersions, true)
-						mu.Lock()
-						metaArr[i], errs[i] = fi, err
-						mu.Unlock()
+						fi, err = fileInfoFromRaw(rfi, bucket, object, readData, opts.InclFreeVersions, true)
 					}
 				}
+
+				rw.Lock()
+				rawArr[i] = rfi
+				metaArr[i], errs[i] = fi, err
+				rw.Unlock()
+
 				done <- err == nil
 			}(i, disk)
 		}
 
 		wg.Wait()
-		close(done)
+		xioutil.SafeClose(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.
@@ -854,13 +864,12 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 		minDisks = er.setDriveCount - er.defaultParityCount
 	}
 
-	calcQuorum := func() (FileInfo, []FileInfo, []StorageAPI, time.Time, string, error) {
+	calcQuorum := func(metaArr []FileInfo, errs []error) (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 {
+		if err := reduceReadQuorumErrs(ctx, errs, objectOpIgnoredErrs, readQuorum); err != nil {
 			return FileInfo{}, nil, nil, time.Time{}, "", err
 		}
 		onlineDisks, modTime, etag := listOnlineDisks(disks, metaArr, errs, readQuorum)
@@ -870,7 +879,11 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 		}
 
 		onlineMeta := make([]FileInfo, len(metaArr))
-		copy(onlineMeta, metaArr)
+		for i, disk := range onlineDisks {
+			if disk != nil {
+				onlineMeta[i] = metaArr[i]
+			}
+		}
 
 		return fi, onlineMeta, onlineDisks, modTime, etag, nil
 	}
@@ -897,20 +910,24 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 				continue
 			}
 		}
+
+		rw.Lock()
 		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)
+			fi, onlineMeta, onlineDisks, modTime, etag, err = calcQuorum(pickLatestQuorumFilesInfo(ctx,
+				rawArr, errs, bucket, object, readData, opts.InclFreeVersions, true))
+		} else {
+			fi, onlineMeta, onlineDisks, modTime, etag, err = calcQuorum(metaArr, errs)
 		}
-		mu.Lock()
-		fi, onlineMeta, onlineDisks, modTime, etag, err = calcQuorum()
-		mu.Unlock()
+		rw.Unlock()
 		if err == nil && fi.InlineData() {
 			break
 		}
 	}
 
 	if err != nil {
-		if shouldCheckForDangling(err, errs, bucket) {
+		// We can only look for dangling if we received all the responses, if we did
+		// not we simply ignore it, since we can't tell for sure if its dangling object.
+		if totalResp == er.setDriveCount && shouldCheckForDangling(err, errs, bucket) {
 			_, derr := er.deleteIfDangling(context.Background(), bucket, object, metaArr, errs, nil, opts)
 			if derr != nil {
 				err = derr
@@ -930,7 +947,7 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 	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.
+		// verify if it has common "etag" at least.
 		if onlineMeta[i].IsValid() && onlineMeta[i].Erasure.Equal(fi.Erasure) {
 			ok := onlineMeta[i].ModTime.Equal(modTime)
 			if modTime.IsZero() || modTime.Equal(timeSentinel) {
@@ -1022,7 +1039,7 @@ 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
 			}
@@ -1049,7 +1066,7 @@ func renameData(ctx context.Context, disks []StorageAPI, srcBucket, srcEntry str
 			// 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)
+				return disks[index].DeleteVersion(context.Background(), dstBucket, dstEntry, metadata[index], false, DeleteOptions{UndoWrite: true})
 			}, index)
 		}
 		dg.Wait()
@@ -1125,10 +1142,10 @@ func (er erasureObjects) putMetacacheObject(ctx context.Context, key string, r *
 	var buffer []byte
 	switch size := data.Size(); {
 	case size == 0:
-		buffer = make([]byte, 1) // Allocate atleast a byte to reach EOF
+		buffer = make([]byte, 1) // Allocate at least 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))
@@ -1376,10 +1393,10 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 	var buffer []byte
 	switch size := data.Size(); {
 	case size == 0:
-		buffer = make([]byte, 1) // Allocate atleast a byte to reach EOF
+		buffer = make([]byte, 1) // Allocate at least 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))
@@ -1438,10 +1455,10 @@ 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
@@ -1534,7 +1551,6 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 		if errors.Is(err, errFileNotFound) {
 			return ObjectInfo{}, toObjectErr(errErasureWriteQuorum, bucket, object)
 		}
-		logger.LogOnceIf(ctx, err, "erasure-object-rename-"+bucket+"-"+object)
 		return ObjectInfo{}, toObjectErr(err, bucket, object)
 	}
 
@@ -1600,7 +1616,7 @@ 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
@@ -1713,7 +1729,7 @@ 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

cmd/erasure-object_test.go

@@ -799,7 +799,7 @@ func TestPutObjectNoQuorumSmall(t *testing.T) {
 }
 
 // Test PutObject twice, one small and another bigger
-// than small data thresold and checks reading them again
+// than small data threshold and checks reading them again
 func TestPutObjectSmallInlineData(t *testing.T) {
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()

cmd/erasure-server-pool-decom.go

@@ -716,7 +716,7 @@ func (set *erasureObjects) listObjectsToDecommission(ctx context.Context, bi dec
 		path:           bi.Prefix,
 		recursive:      true,
 		forwardTo:      "",
-		minDisks:       len(disks) / 2,
+		minDisks:       listQuorum,
 		reportNotFound: false,
 		agreed:         fn,
 		partial: func(entries metaCacheEntries, _ []error) {
@@ -936,7 +936,8 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 					bi.Name,
 					encodeDirObject(entry.name),
 					ObjectOptions{
-						DeletePrefix: true, // use prefix delete to delete all versions at once.
+						DeletePrefix:       true, // use prefix delete to delete all versions at once.
+						DeletePrefixObject: true, // use prefix delete on exact object (this is an optimization to avoid fan-out calls)
 					},
 				)
 				stopFn(err)

cmd/erasure-server-pool-rebalance.go

@@ -34,6 +34,7 @@ import (
 	"github.com/lithammer/shortuuid/v4"
 	"github.com/minio/madmin-go/v3"
 	"github.com/minio/minio/internal/hash"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/env"
 )
@@ -114,12 +115,60 @@ func (z *erasureServerPools) loadRebalanceMeta(ctx context.Context) error {
 	}
 
 	z.rebalMu.Lock()
-	z.rebalMeta = r
+	if len(r.PoolStats) == len(z.serverPools) {
+		z.rebalMeta = r
+	} else {
+		z.updateRebalanceStats(ctx)
+	}
 	z.rebalMu.Unlock()
 
 	return nil
 }
 
+// updates rebalance.bin from let's say 2 pool setup in the middle
+// of a rebalance, was expanded can cause z.rebalMeta to be outdated
+// due to a missing new pool. This function tries to handle this
+// scenario, albeit rare it seems to have occurred in the wild.
+//
+// since we do not explicitly disallow it, but it is okay for them
+// expand and then we continue to rebalance.
+func (z *erasureServerPools) updateRebalanceStats(ctx context.Context) error {
+	var ok bool
+	for i := range z.serverPools {
+		if z.findIndex(i) == -1 {
+			// Also ensure to initialize rebalanceStats to indicate
+			// its a new pool that can receive rebalanced data.
+			z.rebalMeta.PoolStats = append(z.rebalMeta.PoolStats, &rebalanceStats{})
+			ok = true
+		}
+	}
+	if ok {
+		lock := z.serverPools[0].NewNSLock(minioMetaBucket, rebalMetaName)
+		lkCtx, err := lock.GetLock(ctx, globalOperationTimeout)
+		if err != nil {
+			logger.LogIf(ctx, fmt.Errorf("failed to acquire write lock on %s/%s: %w", minioMetaBucket, rebalMetaName, err))
+			return err
+		}
+		defer lock.Unlock(lkCtx)
+
+		ctx = lkCtx.Context()
+
+		noLockOpts := ObjectOptions{NoLock: true}
+		return z.rebalMeta.saveWithOpts(ctx, z.serverPools[0], noLockOpts)
+	}
+
+	return nil
+}
+
+func (z *erasureServerPools) findIndex(index int) int {
+	for i := 0; i < len(z.rebalMeta.PoolStats); i++ {
+		if i == index {
+			return index
+		}
+	}
+	return -1
+}
+
 // initRebalanceMeta initializes rebalance metadata for a new rebalance
 // operation and saves it in the object store.
 func (z *erasureServerPools) initRebalanceMeta(ctx context.Context, buckets []string) (arn string, err error) {
@@ -129,7 +178,7 @@ 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
@@ -324,7 +373,7 @@ func (z *erasureServerPools) IsPoolRebalancing(poolIndex int) bool {
 
 func (z *erasureServerPools) rebalanceBuckets(ctx context.Context, poolIdx int) (err error) {
 	doneCh := make(chan struct{})
-	defer close(doneCh)
+	defer xioutil.SafeClose(doneCh)
 
 	// Save rebalance.bin periodically.
 	go func() {
@@ -510,7 +559,6 @@ func (z *erasureServerPools) rebalanceBucket(ctx context.Context, bucket string,
 
 				// Apply lifecycle rules on the objects that are expired.
 				if filterLifecycle(bucket, version.Name, version) {
-					rebalanced++
 					expired++
 					continue
 				}
@@ -609,7 +657,8 @@ func (z *erasureServerPools) rebalanceBucket(ctx context.Context, bucket string,
 					bucket,
 					encodeDirObject(entry.name),
 					ObjectOptions{
-						DeletePrefix: true, // use prefix delete to delete all versions at once.
+						DeletePrefix:       true, // use prefix delete to delete all versions at once.
+						DeletePrefixObject: true, // use prefix delete on exact object (this is an optimization to avoid fan-out calls)
 					},
 				)
 				stopFn(err)
@@ -624,10 +673,12 @@ 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 +686,7 @@ 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{}{}
@@ -693,8 +744,7 @@ func (z *erasureServerPools) saveRebalanceStats(ctx context.Context, poolIdx int
 	}
 	z.rebalMeta = r
 
-	err = z.rebalMeta.saveWithOpts(ctx, z.serverPools[0], noLockOpts)
-	return err
+	return z.rebalMeta.saveWithOpts(ctx, z.serverPools[0], noLockOpts)
 }
 
 func auditLogRebalance(ctx context.Context, apiName, bucket, object, versionID string, err error) {

cmd/erasure-server-pool.go

@@ -38,7 +38,9 @@ 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"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/sync/errgroup"
 	"github.com/minio/pkg/v2/wildcard"
@@ -82,6 +84,22 @@ 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
+	}
+
+	if globalIsCICD {
+		n = 256 // 256MiB for CI/CD environments is sufficient
+	}
+
+	// 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)
+	globalBytePoolCap.Populate()
+
 	var localDrives []StorageAPI
 	local := endpointServerPools.FirstLocal()
 	for i, ep := range endpointServerPools {
@@ -101,8 +119,10 @@ func newErasureServerPools(ctx context.Context, endpointServerPools EndpointServ
 			return nil, fmt.Errorf("parity validation returned an error: %w <- (%d, %d), for pool(%s)", err, commonParityDrives, ep.DrivesPerSet, humanize.Ordinal(i+1))
 		}
 
-		storageDisks[i], formats[i], err = waitForFormatErasure(local, ep.Endpoints, i+1,
-			ep.SetCount, ep.DrivesPerSet, deploymentID, distributionAlgo)
+		bootstrapTrace("waitForFormatErasure: loading disks", func() {
+			storageDisks[i], formats[i], err = waitForFormatErasure(local, ep.Endpoints, i+1,
+				ep.SetCount, ep.DrivesPerSet, deploymentID, distributionAlgo)
+		})
 		if err != nil {
 			return nil, err
 		}
@@ -121,7 +141,9 @@ func newErasureServerPools(ctx context.Context, endpointServerPools EndpointServ
 			return nil, fmt.Errorf("all pools must have same deployment ID - expected %s, got %s for pool(%s)", deploymentID, formats[i].ID, humanize.Ordinal(i+1))
 		}
 
-		z.serverPools[i], err = newErasureSets(ctx, ep, storageDisks[i], formats[i], commonParityDrives, i)
+		bootstrapTrace(fmt.Sprintf("newErasureSets: initializing %s pool", humanize.Ordinal(i+1)), func() {
+			z.serverPools[i], err = newErasureSets(ctx, ep, storageDisks[i], formats[i], commonParityDrives, i)
+		})
 		if err != nil {
 			return nil, err
 		}
@@ -159,8 +181,12 @@ func newErasureServerPools(ctx context.Context, endpointServerPools EndpointServ
 	setObjectLayer(z)
 
 	r := rand.New(rand.NewSource(time.Now().UnixNano()))
+	attempt := 1
 	for {
-		err := z.Init(ctx) // Initializes all pools.
+		var err error
+		bootstrapTrace(fmt.Sprintf("poolMeta.Init: loading pool metadata, attempt: %d", attempt), func() {
+			err = z.Init(ctx) // Initializes all pools.
+		})
 		if err != nil {
 			if !configRetriableErrors(err) {
 				logger.Fatal(err, "Unable to initialize backend")
@@ -168,6 +194,7 @@ func newErasureServerPools(ctx context.Context, endpointServerPools EndpointServ
 			retry := time.Duration(r.Float64() * float64(5*time.Second))
 			logger.LogIf(ctx, fmt.Errorf("Unable to initialize backend: %w, retrying in %s", err, retry))
 			time.Sleep(retry)
+			attempt++
 			continue
 		}
 		break
@@ -597,7 +624,7 @@ 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))
@@ -605,7 +632,7 @@ 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)
 	}
@@ -621,13 +648,13 @@ 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 {
 	// Updates must be closed before we return.
-	defer close(updates)
+	defer xioutil.SafeClose(updates)
 
 	ctx, cancel := context.WithCancel(ctx)
 	defer cancel()
@@ -654,7 +681,7 @@ func (z *erasureServerPools) NSScanner(ctx context.Context, updates chan<- DataU
 			results = append(results, dataUsageCache{})
 			go func(i int, erObj *erasureObjects) {
 				updates := make(chan dataUsageCache, 1)
-				defer close(updates)
+				defer xioutil.SafeClose(updates)
 				// Start update collector.
 				go func() {
 					defer wg.Done()
@@ -713,7 +740,7 @@ func (z *erasureServerPools) NSScanner(ctx context.Context, updates chan<- DataU
 				return
 			case v := <-updateCloser:
 				update()
-				close(v)
+				xioutil.SafeClose(v)
 				return
 			case <-updateTicker.C:
 				update()
@@ -1915,31 +1942,12 @@ func (z *erasureServerPools) HealFormat(ctx context.Context, dryRun bool) (madmi
 }
 
 func (z *erasureServerPools) HealBucket(ctx context.Context, bucket string, opts madmin.HealOpts) (madmin.HealResultItem, error) {
-	r := madmin.HealResultItem{
-		Type:   madmin.HealItemBucket,
-		Bucket: bucket,
-	}
-
 	// Attempt heal on the bucket metadata, ignore any failures
 	hopts := opts
 	hopts.Recreate = false
 	defer z.HealObject(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, bucket, bucketMetadataFile), "", hopts)
 
-	for _, pool := range z.serverPools {
-		result, err := pool.HealBucket(ctx, bucket, opts)
-		if err != nil {
-			if _, ok := err.(BucketNotFound); ok {
-				continue
-			}
-			return result, err
-		}
-		r.DiskCount += result.DiskCount
-		r.SetCount += result.SetCount
-		r.Before.Drives = append(r.Before.Drives, result.Before.Drives...)
-		r.After.Drives = append(r.After.Drives, result.After.Drives...)
-	}
-
-	return r, nil
+	return z.s3Peer.HealBucket(ctx, bucket, opts)
 }
 
 // Walk a bucket, optionally prefix recursively, until we have returned
@@ -1950,7 +1958,7 @@ func (z *erasureServerPools) HealBucket(ctx context.Context, bucket string, opts
 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)
+		xioutil.SafeClose(results)
 		return err
 	}
 
@@ -1959,7 +1967,7 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 	ctx, cancel := context.WithCancel(ctx)
 	go func() {
 		defer cancel()
-		defer close(results)
+		defer xioutil.SafeClose(results)
 
 		for _, erasureSet := range z.serverPools {
 			var wg sync.WaitGroup
@@ -1969,7 +1977,7 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 				go func() {
 					defer wg.Done()
 
-					disks, _ := set.getOnlineDisksWithHealing(true)
+					disks, infos, _ := set.getOnlineDisksWithHealingAndInfo(true)
 					if len(disks) == 0 {
 						cancel()
 						return
@@ -1986,7 +1994,17 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 
 					askDisks := getListQuorum(opts.AskDisks, set.setDriveCount)
 					if askDisks == -1 {
-						askDisks = getListQuorum("strict", set.setDriveCount)
+						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
@@ -2253,7 +2271,7 @@ const (
 	vmware = "VMWare"
 )
 
-// HealthOptions takes input options to return sepcific information
+// HealthOptions takes input options to return specific information
 type HealthOptions struct {
 	Maintenance    bool
 	DeploymentType string
@@ -2268,7 +2286,10 @@ type HealthResult struct {
 	ESHealth      []struct {
 		Maintenance   bool
 		PoolID, SetID int
+		Healthy       bool
 		HealthyDrives int
+		HealingDrives int
+		ReadQuorum    int
 		WriteQuorum   int
 	}
 	WriteQuorum   int
@@ -2315,55 +2336,51 @@ 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()
+	poolReadQuorums := make([]int, len(b.StandardSCData))
 	poolWriteQuorums := make([]int, len(b.StandardSCData))
 	for i, data := range b.StandardSCData {
+		poolReadQuorums[i] = data
 		poolWriteQuorums[i] = data
 		if data == b.StandardSCParity {
 			poolWriteQuorums[i] = data + 1
 		}
 	}
 
-	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))
 		}
 	}
 
@@ -2391,29 +2408,34 @@ func (z *erasureServerPools) Health(ctx context.Context, opts HealthOptions) Hea
 	for poolIdx := range erasureSetUpCount {
 		for setIdx := range erasureSetUpCount[poolIdx] {
 			result.ESHealth = append(result.ESHealth, struct {
-				Maintenance                bool
-				PoolID, SetID              int
-				HealthyDrives, WriteQuorum int
+				Maintenance                  bool
+				PoolID, SetID                int
+				Healthy                      bool
+				HealthyDrives, HealingDrives int
+				ReadQuorum, WriteQuorum      int
 			}{
 				Maintenance:   opts.Maintenance,
 				SetID:         setIdx,
 				PoolID:        poolIdx,
-				HealthyDrives: erasureSetUpCount[poolIdx][setIdx],
+				Healthy:       erasureSetUpCount[poolIdx][setIdx].online >= poolWriteQuorums[poolIdx],
+				HealthyDrives: erasureSetUpCount[poolIdx][setIdx].online,
+				HealingDrives: erasureSetUpCount[poolIdx][setIdx].healing,
+				ReadQuorum:    poolReadQuorums[poolIdx],
 				WriteQuorum:   poolWriteQuorums[poolIdx],
 			})
 
-			if erasureSetUpCount[poolIdx][setIdx] < poolWriteQuorums[poolIdx] {
+			result.Healthy = erasureSetUpCount[poolIdx][setIdx].online >= poolWriteQuorums[poolIdx]
+			if !result.Healthy {
 				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.Healthy = false
 			}
 		}
 	}
 
 	if opts.Maintenance {
-		result.Healthy = result.Healthy && len(aggHealStateResult.HealDisks) == 0
-		result.HealingDrives = len(aggHealStateResult.HealDisks)
+		result.Healthy = result.Healthy && drivesHealing == 0
+		result.HealingDrives = drivesHealing
 	}
 
 	return result
@@ -2427,7 +2449,7 @@ func (z *erasureServerPools) PutObjectMetadata(ctx context.Context, bucket, obje
 	}
 
 	opts.MetadataChg = true
-	// We don't know the size here set 1GiB atleast.
+	// We don't know the size here set 1GiB at least.
 	idx, err := z.getPoolIdxExistingWithOpts(ctx, bucket, object, opts)
 	if err != nil {
 		return ObjectInfo{}, err
@@ -2445,7 +2467,7 @@ func (z *erasureServerPools) PutObjectTags(ctx context.Context, bucket, object s
 
 	opts.MetadataChg = true
 
-	// We don't know the size here set 1GiB atleast.
+	// We don't know the size here set 1GiB at least.
 	idx, err := z.getPoolIdxExistingWithOpts(ctx, bucket, object, opts)
 	if err != nil {
 		return ObjectInfo{}, err

cmd/erasure-sets.go

@@ -36,8 +36,8 @@ 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"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/console"
 	"github.com/minio/pkg/v2/sync/errgroup"
@@ -115,24 +115,26 @@ func (s *erasureSets) getDiskMap() map[Endpoint]StorageAPI {
 
 // Initializes a new StorageAPI from the endpoint argument, returns
 // StorageAPI and also `format` which exists on the disk.
-func connectEndpoint(endpoint Endpoint) (StorageAPI, *formatErasureV3, error) {
+func connectEndpoint(endpoint Endpoint) (StorageAPI, *formatErasureV3, []byte, error) {
 	disk, err := newStorageAPI(endpoint, storageOpts{
 		cleanUp:     false,
 		healthCheck: false,
 	})
 	if err != nil {
-		return nil, nil, err
+		return nil, nil, nil, err
 	}
 
-	format, err := loadFormatErasure(disk)
+	format, formatData, err := loadFormatErasureWithData(disk, false)
 	if err != nil {
 		if errors.Is(err, errUnformattedDisk) {
-			info, derr := disk.DiskInfo(context.TODO(), false)
+			info, derr := disk.DiskInfo(context.TODO(), DiskInfoOptions{})
 			if derr != nil && info.RootDisk {
-				return nil, nil, fmt.Errorf("Drive: %s is a root drive", disk)
+				disk.Close()
+				return nil, nil, nil, fmt.Errorf("Drive: %s is a root drive", disk)
 			}
 		}
-		return nil, nil, fmt.Errorf("Drive: %s returned %w", disk, err) // make sure to '%w' to wrap the error
+		disk.Close()
+		return nil, nil, nil, fmt.Errorf("Drive: %s returned %w", disk, err) // make sure to '%w' to wrap the error
 	}
 
 	disk.Close()
@@ -141,10 +143,10 @@ func connectEndpoint(endpoint Endpoint) (StorageAPI, *formatErasureV3, error) {
 		healthCheck: true,
 	})
 	if err != nil {
-		return nil, nil, err
+		return nil, nil, nil, err
 	}
 
-	return disk, format, nil
+	return disk, format, formatData, nil
 }
 
 // findDiskIndex - returns the i,j'th position of the input `diskID` against the reference
@@ -225,7 +227,7 @@ func (s *erasureSets) connectDisks() {
 		wg.Add(1)
 		go func(endpoint Endpoint) {
 			defer wg.Done()
-			disk, format, err := connectEndpoint(endpoint)
+			disk, format, formatData, err := connectEndpoint(endpoint)
 			if err != nil {
 				if endpoint.IsLocal && errors.Is(err, errUnformattedDisk) {
 					globalBackgroundHealState.pushHealLocalDisks(endpoint)
@@ -260,12 +262,15 @@ func (s *erasureSets) connectDisks() {
 
 			disk.SetDiskID(format.Erasure.This)
 			disk.SetDiskLoc(s.poolIndex, setIndex, diskIndex)
+			disk.SetFormatData(formatData)
 			s.erasureDisks[setIndex][diskIndex] = disk
 			s.erasureDisksMu.Unlock()
 
-			globalLocalDrivesMu.Lock()
-			globalLocalSetDrives[s.poolIndex][setIndex][diskIndex] = disk
-			globalLocalDrivesMu.Unlock()
+			if disk.IsLocal() && globalIsDistErasure {
+				globalLocalDrivesMu.Lock()
+				globalLocalSetDrives[s.poolIndex][setIndex][diskIndex] = disk
+				globalLocalDrivesMu.Unlock()
+			}
 		}(endpoint)
 	}
 
@@ -369,21 +374,6 @@ 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)
 	}
@@ -395,19 +385,27 @@ func newErasureSets(ctx context.Context, endpoints PoolEndpoints, storageDisks [
 		}
 	}
 
+	var wg sync.WaitGroup
+	var lk sync.Mutex
 	for i := 0; i < setCount; i++ {
 		lockerEpSet := set.NewStringSet()
 		for j := 0; j < setDriveCount; j++ {
-			endpoint := endpoints.Endpoints[i*setDriveCount+j]
-			// Only add lockers only one per endpoint and per erasure set.
-			if locker, ok := erasureLockers[endpoint.Host]; ok && !lockerEpSet.Contains(endpoint.Host) {
-				lockerEpSet.Add(endpoint.Host)
-				s.erasureLockers[i] = append(s.erasureLockers[i], locker)
-			}
+			wg.Add(1)
+			go func(i int, endpoint Endpoint) {
+				defer wg.Done()
+
+				lk.Lock()
+				// Only add lockers only one per endpoint and per erasure set.
+				if locker, ok := erasureLockers[endpoint.Host]; ok && !lockerEpSet.Contains(endpoint.Host) {
+					lockerEpSet.Add(endpoint.Host)
+					s.erasureLockers[i] = append(s.erasureLockers[i], locker)
+				}
+				lk.Unlock()
+			}(i, endpoints.Endpoints[i*setDriveCount+j])
 		}
 	}
+	wg.Wait()
 
-	var wg sync.WaitGroup
 	for i := 0; i < setCount; i++ {
 		wg.Add(1)
 		go func(i int) {
@@ -419,6 +417,19 @@ 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.Close()
+					disk = ldisk
+					globalLocalDrivesMu.RUnlock()
+				}
+
 				innerWg.Add(1)
 				go func(disk StorageAPI, i, j int) {
 					defer innerWg.Done()
@@ -459,8 +470,6 @@ func newErasureSets(ctx context.Context, endpoints PoolEndpoints, storageDisks [
 				getLockers:         s.GetLockers(i),
 				getEndpoints:       s.GetEndpoints(i),
 				nsMutex:            mutex,
-				bp:                 bp,
-				bpOld:              bpOld,
 			}
 		}(i)
 	}
@@ -615,7 +624,7 @@ 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))
@@ -624,7 +633,7 @@ 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)
 	}
@@ -659,10 +668,10 @@ func (s *erasureSets) Shutdown(ctx context.Context) error {
 	select {
 	case _, ok := <-s.setReconnectEvent:
 		if ok {
-			close(s.setReconnectEvent)
+			xioutil.SafeClose(s.setReconnectEvent)
 		}
 	default:
-		close(s.setReconnectEvent)
+		xioutil.SafeClose(s.setReconnectEvent)
 	}
 	return nil
 }
@@ -1021,52 +1030,11 @@ func formatsToDrivesInfo(endpoints Endpoints, formats []*formatErasureV3, sErrs
 	return beforeDrives
 }
 
-func getHealDiskInfos(storageDisks []StorageAPI, errs []error) ([]DiskInfo, []error) {
-	infos := make([]DiskInfo, len(storageDisks))
-	g := errgroup.WithNErrs(len(storageDisks))
-	for index := range storageDisks {
-		index := index
-		g.Go(func() error {
-			if errs[index] != nil && errs[index] != errUnformattedDisk {
-				return errs[index]
-			}
-			if storageDisks[index] == nil {
-				return errDiskNotFound
-			}
-			var err error
-			infos[index], err = storageDisks[index].DiskInfo(context.TODO(), false)
-			return err
-		}, index)
-	}
-	return infos, g.Wait()
-}
-
-// Mark root disks as down so as not to heal them.
-func markRootDisksAsDown(storageDisks []StorageAPI, errs []error) {
-	if globalIsCICD {
-		// Do nothing
-		return
-	}
-	infos, ierrs := getHealDiskInfos(storageDisks, errs)
-	for i := range storageDisks {
-		if ierrs[i] != nil && ierrs[i] != errUnformattedDisk {
-			storageDisks[i] = nil
-			continue
-		}
-		if storageDisks[i] != nil && infos[i].RootDisk {
-			// We should not heal on root disk. i.e in a situation where the minio-administrator has unmounted a
-			// defective drive we should not heal a path on the root disk.
-			logger.LogIf(GlobalContext, fmt.Errorf("Drive `%s` is part of root drive, will not be used", storageDisks[i]))
-			storageDisks[i] = nil
-		}
-	}
-}
-
 // HealFormat - heals missing `format.json` on fresh unformatted disks.
 func (s *erasureSets) HealFormat(ctx context.Context, dryRun bool) (res madmin.HealResultItem, err error) {
 	storageDisks, _ := initStorageDisksWithErrors(s.endpoints.Endpoints, storageOpts{
-		cleanUp:     true,
-		healthCheck: true,
+		cleanUp:     false,
+		healthCheck: false,
 	})
 
 	defer func(storageDisks []StorageAPI) {
@@ -1080,9 +1048,6 @@ func (s *erasureSets) HealFormat(ctx context.Context, dryRun bool) (res madmin.H
 		return madmin.HealResultItem{}, err
 	}
 
-	// Mark all root disks down
-	markRootDisksAsDown(storageDisks, sErrs)
-
 	refFormat, err := getFormatErasureInQuorum(formats)
 	if err != nil {
 		return res, err
@@ -1111,10 +1076,17 @@ func (s *erasureSets) HealFormat(ctx context.Context, dryRun bool) (res madmin.H
 		return res, errNoHealRequired
 	}
 
+	if !reflect.DeepEqual(s.format, refFormat) {
+		// Format is corrupted and unrecognized by the running instance.
+		logger.LogIf(ctx, fmt.Errorf("Unable to heal the newly replaced drives due to format.json inconsistencies, please engage MinIO support for further assistance: %w",
+			errCorruptedFormat))
+		return res, errCorruptedFormat
+	}
+
 	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)
@@ -1136,6 +1108,7 @@ func (s *erasureSets) HealFormat(ctx context.Context, dryRun bool) (res madmin.H
 			}
 			if err := saveFormatErasure(storageDisks[index], format, formatOpID); err != nil {
 				logger.LogIf(ctx, fmt.Errorf("Drive %s failed to write updated 'format.json': %v", storageDisks[index], err))
+				storageDisks[index].Close()
 				tmpNewFormats[index] = nil // this disk failed to write new format
 			}
 		}
@@ -1157,43 +1130,46 @@ 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 {
+				if disk.IsLocal() {
+					disk.SetDiskLoc(s.poolIndex, m, n)
+
+					xldisk, ok := disk.(*xlStorageDiskIDCheck)
+					if ok {
+						if driveQuorum {
+							commonWrites, commonDeletes := calcCommonWritesDeletes(currentDisksInfo[m], (s.setDriveCount+1)/2)
+							xldisk.totalWrites.Store(commonWrites)
+							xldisk.totalDeletes.Store(commonDeletes)
+							xldisk.storage.setWriteAttribute(commonWrites)
+							xldisk.storage.setDeleteAttribute(commonDeletes)
+						}
+						go xldisk.monitorDiskWritable(xldisk.diskCtx)
+					}
+				} else {
+					disk.Close() // Close the remote storage client, re-initialize with healthchecks.
+					disk, err = newStorageRESTClient(disk.Endpoint(), true, globalGrid.Load())
+					if err != nil {
+						continue
+					}
+					disk.SetDiskLoc(s.poolIndex, m, n)
+				}
+
+				s.erasureDisks[m][n] = disk
+
+				if disk.IsLocal() && globalIsDistErasure {
+					globalLocalDrivesMu.Lock()
+					globalLocalSetDrives[s.poolIndex][m][n] = disk
+					globalLocalDrivesMu.Unlock()
+				}
 			}
 		}
 
-		// Replace reference format with what was loaded from disks.
-		s.format = refFormat
-
 		s.erasureDisksMu.Unlock()
 	}
 
 	return res, nil
 }
 
-// HealBucket - heals inconsistent buckets and bucket metadata on all sets.
-func (s *erasureSets) HealBucket(ctx context.Context, bucket string, opts madmin.HealOpts) (result madmin.HealResultItem, err error) {
-	// Initialize heal result info
-	result = madmin.HealResultItem{
-		Type:      madmin.HealItemBucket,
-		Bucket:    bucket,
-		DiskCount: s.setCount * s.setDriveCount,
-		SetCount:  s.setCount,
-	}
-
-	for _, set := range s.sets {
-		healResult, err := set.HealBucket(ctx, bucket, opts)
-		if err != nil {
-			return result, toObjectErr(err, bucket)
-		}
-		result.Before.Drives = append(result.Before.Drives, healResult.Before.Drives...)
-		result.After.Drives = append(result.After.Drives, healResult.After.Drives...)
-	}
-
-	return result, nil
-}
-
 // HealObject - heals inconsistent object on a hashedSet based on object name.
 func (s *erasureSets) HealObject(ctx context.Context, bucket, object, versionID string, opts madmin.HealOpts) (madmin.HealResultItem, error) {
 	return s.getHashedSet(object).HealObject(ctx, bucket, object, versionID, opts)

cmd/erasure-utils.go

@@ -85,7 +85,7 @@ func writeDataBlocks(ctx context.Context, dst io.Writer, enBlocks [][]byte, data
 		if write < int64(len(block)) {
 			n, err := dst.Write(block[:write])
 			if err != nil {
-				// The writer will be closed incase of range queries, which will emit ErrClosedPipe.
+				// The writer will be closed in case of range queries, which will emit ErrClosedPipe.
 				// The reader pipe might be closed at ListObjects io.EOF ignore it.
 				if err != io.ErrClosedPipe && err != io.EOF {
 					logger.LogIf(ctx, err)
@@ -99,7 +99,7 @@ func writeDataBlocks(ctx context.Context, dst io.Writer, enBlocks [][]byte, data
 		// Copy the block.
 		n, err := dst.Write(block)
 		if err != nil {
-			// The writer will be closed incase of range queries, which will emit ErrClosedPipe.
+			// The writer will be closed in case of range queries, which will emit ErrClosedPipe.
 			// The reader pipe might be closed at ListObjects io.EOF ignore it.
 			if err != io.ErrClosedPipe && err != io.EOF {
 				logger.LogIf(ctx, err)

cmd/erasure.go

@@ -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.
@@ -99,7 +91,7 @@ func diskErrToDriveState(err error) (state string) {
 	switch {
 	case errors.Is(err, errDiskNotFound) || errors.Is(err, context.DeadlineExceeded):
 		state = madmin.DriveStateOffline
-	case errors.Is(err, errCorruptedFormat):
+	case errors.Is(err, errCorruptedFormat) || errors.Is(err, errCorruptedBackend):
 		state = madmin.DriveStateCorrupt
 	case errors.Is(err, errUnformattedDisk):
 		state = madmin.DriveStateUnformatted
@@ -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(), DiskInfoOptions{Metrics: metrics})
 			di.DrivePath = info.MountPath
 			di.TotalSpace = info.Total
 			di.UsedSpace = info.Used
@@ -211,6 +203,8 @@ func getDisksInfo(disks []StorageAPI, endpoints []Endpoint) (disksInfo []madmin.
 				APICalls:                make(map[string]uint64, len(info.Metrics.APICalls)),
 				TotalErrorsAvailability: info.Metrics.TotalErrorsAvailability,
 				TotalErrorsTimeout:      info.Metrics.TotalErrorsTimeout,
+				TotalTokens:             info.Metrics.TotalTokens,
+				TotalWaiting:            info.Metrics.TotalWaiting,
 			}
 			for k, v := range info.Metrics.LastMinute {
 				if v.N > 0 {
@@ -233,8 +227,8 @@ 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 +247,11 @@ 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()
 
@@ -271,15 +265,15 @@ func (er erasureObjects) LocalStorageInfo(ctx context.Context) StorageInfo {
 		}
 	}
 
-	return getStorageInfo(localDisks, localEndpoints)
+	return getStorageInfo(localDisks, localEndpoints, metrics)
 }
 
-// getOnlineDisksWithHealing - returns online disks and overall healing status.
+// 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) getOnlineDisksWithHealing(inclHealing bool) (newDisks []StorageAPI, healing bool) {
+func (er erasureObjects) getOnlineDisksWithHealingAndInfo(inclHealing bool) (newDisks []StorageAPI, newInfos []DiskInfo, healing bool) {
 	var wg sync.WaitGroup
 	disks := er.getDisks()
 	infos := make([]DiskInfo, len(disks))
@@ -292,32 +286,24 @@ func (er erasureObjects) getOnlineDisksWithHealing(inclHealing bool) (newDisks [
 
 			disk := disks[i]
 			if disk == nil {
-				infos[i].Error = "offline drive"
+				infos[i].Error = errDiskNotFound.Error()
 				return
 			}
 
-			di, err := disk.DiskInfo(context.Background(), false)
+			di, err := disk.DiskInfo(context.Background(), DiskInfoOptions{})
+			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()
 			}
-			if !inclHealing && di.Healing {
-				return
-			}
-
-			infos[i] = di
 		}()
 	}
 	wg.Wait()
 
 	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
@@ -329,23 +315,34 @@ func (er erasureObjects) getOnlineDisksWithHealing(inclHealing bool) (newDisks [
 			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 non-scanning disks over disks which are currently being scanned.
 	newDisks = append(newDisks, scanningDisks...)
+	newInfos = append(newInfos, scanningInfos...)
+
 	/// Then add healing disks.
 	newDisks = append(newDisks, healingDisks...)
+	newInfos = append(newInfos, healingInfos...)
 
-	return newDisks, healing
+	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/
@@ -426,7 +423,7 @@ func (er erasureObjects) nsScanner(ctx context.Context, buckets []BucketInfo, wa
 			bucketCh <- b
 		}
 	}
-	close(bucketCh)
+	xioutil.SafeClose(bucketCh)
 
 	bucketResults := make(chan dataUsageEntryInfo, len(disks))
 
@@ -527,7 +524,7 @@ func (er erasureObjects) nsScanner(ctx context.Context, buckets []BucketInfo, wa
 				// Calc usage
 				before := cache.Info.LastUpdate
 				var err error
-				cache, err = disk.NSScanner(ctx, cache, updates, healScanMode)
+				cache, err = disk.NSScanner(ctx, cache, updates, healScanMode, nil)
 				if err != nil {
 					if !cache.Info.LastUpdate.IsZero() && cache.Info.LastUpdate.After(before) {
 						logger.LogIf(ctx, cache.save(ctx, er, cacheName))
@@ -563,7 +560,7 @@ func (er erasureObjects) nsScanner(ctx context.Context, buckets []BucketInfo, wa
 		}(i)
 	}
 	wg.Wait()
-	close(bucketResults)
+	xioutil.SafeClose(bucketResults)
 	saverWg.Wait()
 
 	return nil

cmd/format-erasure.go

@@ -1,4 +1,4 @@
-// Copyright (c) 2015-2021 MinIO, Inc.
+// Copyright (c) 2015-2024 MinIO, Inc.
 //
 // This file is part of MinIO Object Storage stack
 //
@@ -28,6 +28,7 @@ import (
 	"sync"
 
 	"github.com/dustin/go-humanize"
+	jsoniter "github.com/json-iterator/go"
 	"github.com/minio/minio/internal/color"
 	"github.com/minio/minio/internal/config"
 	"github.com/minio/minio/internal/config/storageclass"
@@ -123,6 +124,7 @@ 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) {
@@ -324,7 +326,7 @@ func loadFormatErasureAll(storageDisks []StorageAPI, heal bool) ([]*formatErasur
 			if storageDisks[index] == nil {
 				return errDiskNotFound
 			}
-			format, err := loadFormatErasure(storageDisks[index])
+			format, formatData, err := loadFormatErasureWithData(storageDisks[index], heal)
 			if err != nil {
 				return err
 			}
@@ -333,6 +335,7 @@ func loadFormatErasureAll(storageDisks []StorageAPI, heal bool) ([]*formatErasur
 				// If no healing required, make the disks valid and
 				// online.
 				storageDisks[index].SetDiskID(format.Erasure.This)
+				storageDisks[index].SetFormatData(formatData)
 			}
 			return nil
 		}, index)
@@ -347,14 +350,8 @@ func saveFormatErasure(disk StorageAPI, format *formatErasureV3, healID string)
 		return errDiskNotFound
 	}
 
-	diskID := format.Erasure.This
-
-	if err := makeFormatErasureMetaVolumes(disk); err != nil {
-		return err
-	}
-
 	// Marshal and write to disk.
-	formatBytes, err := json.Marshal(format)
+	formatData, err := json.Marshal(format)
 	if err != nil {
 		return err
 	}
@@ -368,7 +365,7 @@ func saveFormatErasure(disk StorageAPI, format *formatErasureV3, healID string)
 	})
 
 	// write to unique file.
-	if err = disk.WriteAll(context.TODO(), minioMetaBucket, tmpFormat, formatBytes); err != nil {
+	if err = disk.WriteAll(context.TODO(), minioMetaBucket, tmpFormat, formatData); err != nil {
 		return err
 	}
 
@@ -377,7 +374,8 @@ func saveFormatErasure(disk StorageAPI, format *formatErasureV3, healID string)
 		return err
 	}
 
-	disk.SetDiskID(diskID)
+	disk.SetDiskID(format.Erasure.This)
+	disk.SetFormatData(formatData)
 	if healID != "" {
 		ctx := context.Background()
 		ht := initHealingTracker(disk, healID)
@@ -386,14 +384,38 @@ func saveFormatErasure(disk StorageAPI, format *formatErasureV3, healID string)
 	return nil
 }
 
+// loadFormatErasureWithData - loads format.json from disk.
+func loadFormatErasureWithData(disk StorageAPI, heal bool) (format *formatErasureV3, data []byte, err error) {
+	data, err = disk.ReadAll(context.TODO(), minioMetaBucket, formatConfigFile)
+	if err != nil {
+		// 'file not found' and 'volume not found' as
+		// same. 'volume not found' usually means its a fresh disk.
+		if errors.Is(err, errFileNotFound) || errors.Is(err, errVolumeNotFound) {
+			return nil, nil, errUnformattedDisk
+		}
+		return nil, nil, err
+	}
+
+	// Try to decode format json into formatConfigV1 struct.
+	format = &formatErasureV3{}
+	if err = json.Unmarshal(data, format); err != nil {
+		return nil, nil, err
+	}
+
+	if heal {
+		info, err := disk.DiskInfo(context.Background(), DiskInfoOptions{NoOp: heal})
+		if err != nil {
+			return nil, nil, err
+		}
+		format.Info = info
+	}
+
+	// Success.
+	return format, data, nil
+}
+
 // 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
@@ -404,6 +426,8 @@ func loadFormatErasure(disk StorageAPI) (format *formatErasureV3, err error) {
 		return nil, err
 	}
 
+	json := jsoniter.ConfigCompatibleWithStandardLibrary
+
 	// Try to decode format json into formatConfigV1 struct.
 	format = &formatErasureV3{}
 	if err = json.Unmarshal(buf, format); err != nil {
@@ -480,9 +504,7 @@ func formatErasureGetDeploymentID(refFormat *formatErasureV3, formats []*formatE
 }
 
 // formatErasureFixDeploymentID - Add deployment id if it is not present.
-func formatErasureFixDeploymentID(endpoints Endpoints, storageDisks []StorageAPI, refFormat *formatErasureV3) (err error) {
-	// Attempt to load all `format.json` from all disks.
-	formats, _ := loadFormatErasureAll(storageDisks, false)
+func formatErasureFixDeploymentID(endpoints Endpoints, storageDisks []StorageAPI, refFormat *formatErasureV3, formats []*formatErasureV3) (err error) {
 	for index := range formats {
 		// If the Erasure sets do not match, set those formats to nil,
 		// We do not have to update the ID on those format.json file.
@@ -781,15 +803,18 @@ func initFormatErasure(ctx context.Context, storageDisks []StorageAPI, setCount,
 						logger.Info("   - Drive: %s", disk.String())
 					}
 				})
-				logger.Info(color.Yellow("WARNING:")+" Host %v has more than %v drives of set. "+
-					"A host failure will result in data becoming unavailable.", host, wantAtMost)
+				var warning string
+				if wantAtMost == 0 {
+					warning = fmt.Sprintf("Host %v has all drives of set. ", host)
+				} else {
+					warning = fmt.Sprintf("Host %v has more than %v drives of set. ", host, wantAtMost)
+				}
+				logger.Info(color.Yellow("WARNING: ") + warning +
+					"A host failure will result in data becoming unavailable.")
 			}
 		}
 	}
 
-	// Mark all root disks down
-	markRootDisksAsDown(storageDisks, sErrs)
-
 	// Save formats `format.json` across all disks.
 	if err := saveFormatErasureAll(ctx, storageDisks, formats); err != nil {
 		return nil, err
@@ -808,27 +833,16 @@ func ecDrivesNoConfig(setDriveCount int) (int, error) {
 	return sc.GetParityForSC(storageclass.STANDARD), nil
 }
 
-// Make Erasure backend meta volumes.
-func makeFormatErasureMetaVolumes(disk StorageAPI) error {
-	if disk == nil {
-		return errDiskNotFound
-	}
-	volumes := []string{
-		minioMetaTmpDeletedBucket, // creates .minio.sys/tmp as well as .minio.sys/tmp/.trash
-		minioMetaMultipartBucket,  // creates .minio.sys/multipart
-		dataUsageBucket,           // creates .minio.sys/buckets
-		minioConfigBucket,         // creates .minio.sys/config
-	}
-	// Attempt to create MinIO internal buckets.
-	return disk.MakeVolBulk(context.TODO(), volumes...)
-}
-
 // 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) {
@@ -841,7 +855,12 @@ 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) {
+				if format.Info.Endpoint != "" {
+					currentDisksInfo[i][j] = format.Info
+				}
+			}
 		}
 	}
-	return newFormats
+	return newFormats, currentDisksInfo
 }

cmd/format-erasure_test.go

@@ -483,7 +483,7 @@ func TestGetErasureID(t *testing.T) {
 
 	formats[2].ID = "bad-id"
 	if _, err = formatErasureGetDeploymentID(quorumFormat, formats); !errors.Is(err, errCorruptedFormat) {
-		t.Fatalf("Unexpect error %s", err)
+		t.Fatalf("Unexpected error %s", err)
 	}
 }
 
@@ -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")
 	}

cmd/ftp-server-driver.go

@@ -32,6 +32,7 @@ import (
 	"github.com/minio/minio-go/v7"
 	"github.com/minio/minio-go/v7/pkg/credentials"
 	"github.com/minio/minio/internal/auth"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	ftp "goftp.io/server/v2"
 )
@@ -386,7 +387,7 @@ func (driver *ftpDriver) DeleteDir(ctx *ftp.Context, path string) (err error) {
 
 	// Send object names that are needed to be removed to objectsCh
 	go func() {
-		defer close(objectsCh)
+		defer xioutil.SafeClose(objectsCh)
 		opts := minio.ListObjectsOptions{
 			Prefix:    prefix,
 			Recursive: true,

cmd/ftp-server.go

@@ -50,7 +50,7 @@ func (log *minioLogger) Printf(sessionID string, format string, v ...interface{}
 	}
 }
 
-// PrintCommand impelment Logger
+// PrintCommand implement Logger
 func (log *minioLogger) PrintCommand(sessionID string, command string, params string) {
 	if serverDebugLog {
 		if command == "PASS" {
@@ -61,7 +61,7 @@ func (log *minioLogger) PrintCommand(sessionID string, command string, params st
 	}
 }
 
-// PrintResponse impelment Logger
+// PrintResponse implement Logger
 func (log *minioLogger) PrintResponse(sessionID string, code int, message string) {
 	if serverDebugLog {
 		logger.Info("%s < %d %s", sessionID, code, message)

cmd/generic-handlers.go

@@ -263,7 +263,7 @@ func isKMSReq(r *http.Request) bool {
 
 // Supported Amz date headers.
 var amzDateHeaders = []string{
-	// Do not chane this order, x-amz-date value should be
+	// Do not change this order, x-amz-date value should be
 	// validated first.
 	"x-amz-date",
 	"date",

cmd/global-heal.go

@@ -28,6 +28,7 @@ import (
 	"github.com/minio/madmin-go/v3"
 	"github.com/minio/minio/internal/color"
 	"github.com/minio/minio/internal/config/storageclass"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/console"
 	"github.com/minio/pkg/v2/wildcard"
@@ -97,7 +98,7 @@ 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 {
@@ -143,17 +144,20 @@ func (er *erasureObjects) healErasureSet(ctx context.Context, buckets []string,
 	healBuckets := make([]string, len(buckets))
 	copy(healBuckets, buckets)
 
-	// Heal all buckets first in this erasure set - this is useful
-	// for new objects upload in different buckets to be successful
+	objAPI := newObjectLayerFn()
+	if objAPI == nil {
+		return errServerNotInitialized
+	}
+
 	for _, bucket := range healBuckets {
-		_, err := er.HealBucket(ctx, bucket, madmin.HealOpts{ScanMode: scanMode})
+		_, err := objAPI.HealBucket(ctx, bucket, madmin.HealOpts{ScanMode: scanMode})
 		if err != nil {
 			// Log bucket healing error if any, we shall retry again.
 			logger.LogIf(ctx, err)
 		}
 	}
 
-	info, err := tracker.disk.DiskInfo(ctx, false)
+	info, err := tracker.disk.DiskInfo(ctx, DiskInfoOptions{})
 	if err != nil {
 		return fmt.Errorf("unable to get disk information before healing it: %w", err)
 	}
@@ -195,10 +199,9 @@ 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{
+		if _, err := objAPI.HealBucket(ctx, bucket, madmin.HealOpts{
 			ScanMode: scanMode,
 		}); err != nil {
 			logger.LogIf(ctx, err)
@@ -431,7 +434,7 @@ func (er *erasureObjects) healErasureSet(ctx context.Context, buckets []string,
 			partial: func(entries metaCacheEntries, _ []error) {
 				entry, ok := entries.resolve(&resolver)
 				if !ok {
-					// check if we can get one entry atleast
+					// check if we can get one entry at least
 					// proceed to heal nonetheless.
 					entry, _ = entries.firstFound()
 				}
@@ -441,7 +444,7 @@ func (er *erasureObjects) healErasureSet(ctx context.Context, buckets []string,
 			finished: nil,
 		})
 		jt.Wait() // synchronize all the concurrent heal jobs
-		close(results)
+		xioutil.SafeClose(results)
 		if err != nil {
 			// Set this such that when we return this function
 			// we let the caller retry this disk again for the

cmd/globals.go

@@ -25,13 +25,15 @@ 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"
@@ -189,6 +191,9 @@ 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
 
@@ -225,6 +230,7 @@ var (
 	globalBucketMonitor     *bandwidth.Monitor
 	globalPolicySys         *PolicySys
 	globalIAMSys            *IAMSys
+	globalBytePoolCap       *bpool.BytePoolCap
 
 	globalLifecycleSys       *LifecycleSys
 	globalBucketSSEConfigSys *BucketSSEConfigSys
@@ -286,7 +292,7 @@ var (
 	// Global server's network statistics
 	globalConnStats = newConnStats()
 
-	// Global HTTP request statisitics
+	// Global HTTP request statistics
 	globalHTTPStats = newHTTPStats()
 
 	// Global bucket network and API statistics
@@ -392,7 +398,7 @@ var (
 
 	globalTierJournal *TierJournal
 
-	globalConsoleSrv *restapi.Server
+	globalConsoleSrv *consoleapi.Server
 
 	// handles service freeze or un-freeze S3 API calls.
 	globalServiceFreeze atomic.Value

cmd/grid.go

@@ -22,7 +22,6 @@ import (
 	"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"
@@ -36,28 +35,11 @@ var globalGrid atomic.Pointer[grid.Manager]
 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
 	}
+	hosts, local := eps.GridHosts()
 	g, err := grid.NewManager(ctx, grid.ManagerOptions{
 		Dialer:       grid.ContextDialer(xhttp.DialContextWithLookupHost(lookupHost, xhttp.NewInternodeDialContext(rest.DefaultTimeout, globalTCPOptions))),
 		Local:        local,

cmd/handler-api.go

@@ -133,8 +133,8 @@ func (t *apiConfig) init(cfg api.Config, setDriveCounts []int) {
 		}
 	} else {
 		apiRequestsMaxPerNode = cfg.RequestsMax
-		if len(globalEndpoints.Hostnames()) > 0 {
-			apiRequestsMaxPerNode /= len(globalEndpoints.Hostnames())
+		if n := totalNodeCount(); n > 0 {
+			apiRequestsMaxPerNode /= n
 		}
 	}
 

cmd/handler-utils.go

@@ -206,7 +206,7 @@ func getReqAccessCred(r *http.Request, region string) (cred auth.Credentials) {
 	return cred
 }
 
-// Extract request params to be sent with event notifiation.
+// Extract request params to be sent with event notification.
 func extractReqParams(r *http.Request) map[string]string {
 	if r == nil {
 		return nil
@@ -237,7 +237,7 @@ func extractReqParams(r *http.Request) map[string]string {
 	return m
 }
 
-// Extract response elements to be sent with event notifiation.
+// Extract response elements to be sent with event notification.
 func extractRespElements(w http.ResponseWriter) map[string]string {
 	if w == nil {
 		return map[string]string{}

cmd/healthcheck-handler.go

@@ -152,7 +152,7 @@ func ReadinessCheckHandler(w http.ResponseWriter, r *http.Request) {
 
 // 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.
+// in K8S environments. 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

cmd/iam-object-store.go

@@ -31,6 +31,7 @@ import (
 	"github.com/minio/madmin-go/v3"
 	"github.com/minio/minio-go/v7/pkg/set"
 	"github.com/minio/minio/internal/config"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/kms"
 	"github.com/minio/minio/internal/logger"
 )
@@ -575,7 +576,7 @@ func listIAMConfigItems(ctx context.Context, objAPI ObjectLayer, pathPrefix stri
 	ch := make(chan itemOrErr)
 
 	go func() {
-		defer close(ch)
+		defer xioutil.SafeClose(ch)
 
 		// Allocate new results channel to receive ObjectInfo.
 		objInfoCh := make(chan ObjectInfo)

cmd/iam-store.go

@@ -1017,18 +1017,37 @@ func (store *IAMStoreSys) PolicyDBUpdate(ctx context.Context, name string, isGro
 	newPolicyMapping.UpdatedAt = UTCNow()
 	addedOrRemoved = policiesToUpdate.ToSlice()
 
-	if err = store.saveMappedPolicy(ctx, name, userType, isGroup, newPolicyMapping); err != nil {
-		return
-	}
-	if !isGroup {
-		if userType == stsUser {
-			cache.iamSTSPolicyMap[name] = newPolicyMapping
+	// In case of detach operation, it is possible that no policies are mapped -
+	// in this case, we delete the mapping from the store.
+	if len(newPolicies) == 0 {
+		if err = store.deleteMappedPolicy(ctx, name, userType, isGroup); err != nil && !errors.Is(err, errNoSuchPolicy) {
+			return
+		}
+		if !isGroup {
+			if userType == stsUser {
+				delete(cache.iamSTSPolicyMap, name)
+			} else {
+				delete(cache.iamUserPolicyMap, name)
+			}
 		} else {
-			cache.iamUserPolicyMap[name] = newPolicyMapping
+			delete(cache.iamGroupPolicyMap, name)
 		}
 	} else {
-		cache.iamGroupPolicyMap[name] = newPolicyMapping
+
+		if err = store.saveMappedPolicy(ctx, name, userType, isGroup, newPolicyMapping); err != nil {
+			return
+		}
+		if !isGroup {
+			if userType == stsUser {
+				cache.iamSTSPolicyMap[name] = newPolicyMapping
+			} else {
+				cache.iamUserPolicyMap[name] = newPolicyMapping
+			}
+		} else {
+			cache.iamGroupPolicyMap[name] = newPolicyMapping
+		}
 	}
+
 	cache.updatedAt = UTCNow()
 	return cache.updatedAt, addedOrRemoved, newPolicies, nil
 }
@@ -2319,6 +2338,27 @@ 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()

cmd/iam.go

@@ -45,6 +45,7 @@ import (
 	"github.com/minio/minio/internal/config/policy/opa"
 	polplugin "github.com/minio/minio/internal/config/policy/plugin"
 	xhttp "github.com/minio/minio/internal/http"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/jwt"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/policy"
@@ -207,7 +208,7 @@ func (sys *IAMSys) Load(ctx context.Context, firstTime bool) error {
 	select {
 	case <-sys.configLoaded:
 	default:
-		close(sys.configLoaded)
+		xioutil.SafeClose(sys.configLoaded)
 	}
 	return nil
 }
@@ -1041,7 +1042,7 @@ 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 +1056,7 @@ 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 +1070,20 @@ 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)
@@ -1344,9 +1359,15 @@ func (sys *IAMSys) updateGroupMembershipsForLDAP(ctx context.Context) {
 	// DN to ldap username mapping for each LDAP user
 	parentUserToLDAPUsernameMap := make(map[string]string)
 	for _, cred := range allCreds {
+		// Expired credentials don't need parent user updates.
+		if cred.IsExpired() {
+			continue
+		}
+
 		if !sys.LDAPConfig.IsLDAPUserDN(cred.ParentUser) {
 			continue
 		}
+
 		// Check if this is the first time we are
 		// encountering this LDAP user.
 		if _, ok := parentUserToCredsMap[cred.ParentUser]; !ok {
@@ -1411,6 +1432,11 @@ func (sys *IAMSys) updateGroupMembershipsForLDAP(ctx context.Context) {
 				continue
 			}
 
+			// Expired credentials don't need group membership updates.
+			if cred.IsExpired() {
+				continue
+			}
+
 			cred.Groups = currGroups
 			if err := sys.store.UpdateUserIdentity(ctx, cred); err != nil {
 				// Log and continue error - perhaps it'll work the next time.

cmd/leak-detect_test.go

@@ -140,7 +140,7 @@ func pickRelevantGoroutines() (gs []string) {
 	buf := debug.Stack()
 	// runtime stack of go routines will be listed with 2 blank spaces between each of them, so split on "\n\n" .
 	for _, g := range strings.Split(string(buf), "\n\n") {
-		// Again split on a new line, the first line of the second half contaisn the info about the go routine.
+		// Again split on a new line, the first line of the second half contains the info about the go routine.
 		sl := strings.SplitN(g, "\n", 2)
 		if len(sl) != 2 {
 			continue

cmd/license-update.go

@@ -47,7 +47,7 @@ func initLicenseUpdateJob(ctx context.Context, objAPI ObjectLayer) {
 			// sleep for some time and try again.
 			duration := time.Duration(r.Float64() * float64(time.Hour))
 			if duration < time.Second {
-				// Make sure to sleep atleast a second to avoid high CPU ticks.
+				// Make sure to sleep at least a second to avoid high CPU ticks.
 				duration = time.Second
 			}
 			time.Sleep(duration)
@@ -89,7 +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

cmd/local-locker.go

@@ -34,7 +34,7 @@ type lockRequesterInfo struct {
 	UID             string    // UID to uniquely identify request of client.
 	Timestamp       time.Time // Timestamp set at the time of initialization.
 	TimeLastRefresh time.Time // Timestamp for last lock refresh.
-	Source          string    // Contains line, function and filename reqesting the lock.
+	Source          string    // Contains line, function and filename requesting the lock.
 	Group           bool      // indicates if it was a group lock.
 	// Owner represents the UUID of the owner who originally requested the lock
 	// useful in expiry.

cmd/metacache-entries.go

@@ -26,6 +26,7 @@ import (
 	"sort"
 	"strings"
 
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/console"
 )
@@ -659,7 +660,7 @@ func (m *metaCacheEntriesSorted) forwardPast(s string) {
 // If the context is canceled the function will return the error,
 // otherwise the function will return nil.
 func mergeEntryChannels(ctx context.Context, in []chan metaCacheEntry, out chan<- metaCacheEntry, readQuorum int) error {
-	defer close(out)
+	defer xioutil.SafeClose(out)
 	top := make([]*metaCacheEntry, len(in))
 	nDone := 0
 	ctxDone := ctx.Done()

cmd/metacache-server-pool.go

@@ -28,6 +28,7 @@ import (
 	"sync"
 	"time"
 
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 )
 
@@ -339,7 +340,7 @@ func (z *erasureServerPools) listMerged(ctx context.Context, o listPathOptions,
 // When 'in' is closed or the context is canceled the
 // function closes 'out' and exits.
 func applyBucketActions(ctx context.Context, o listPathOptions, in <-chan metaCacheEntry, out chan<- metaCacheEntry) {
-	defer close(out)
+	defer xioutil.SafeClose(out)
 
 	for {
 		var obj metaCacheEntry
@@ -472,16 +473,16 @@ func (z *erasureServerPools) listAndSave(ctx context.Context, o *listPathOptions
 				funcReturnedMu.Unlock()
 				outCh <- entry
 				if returned {
-					close(outCh)
+					xioutil.SafeClose(outCh)
 				}
 			}
 			entry.reusable = returned
 			saveCh <- entry
 		}
 		if !returned {
-			close(outCh)
+			xioutil.SafeClose(outCh)
 		}
-		close(saveCh)
+		xioutil.SafeClose(saveCh)
 	}()
 
 	return filteredResults()

cmd/metacache-set.go

@@ -37,6 +37,7 @@ import (
 	"github.com/minio/minio/internal/bucket/versioning"
 	"github.com/minio/minio/internal/color"
 	"github.com/minio/minio/internal/hash"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/pkg/v2/console"
 )
@@ -590,25 +591,115 @@ func (er *erasureObjects) streamMetadataParts(ctx context.Context, o listPathOpt
 func getListQuorum(quorum string, driveCount int) int {
 	switch quorum {
 	case "disk":
-		// smallest possible value, generally meant for testing.
 		return 1
 	case "reduced":
 		return 2
 	case "optimal":
 		return (driveCount + 1) / 2
+	case "auto":
+		return -1
 	}
 	// defaults to 'strict'
 	return driveCount
 }
 
+func calcCommonWritesDeletes(infos []DiskInfo, readQuorum int) (commonWrite, commonDelete uint64) {
+	deletes := make([]uint64, len(infos))
+	writes := make([]uint64, len(infos))
+	for index, di := range infos {
+		deletes[index] = di.Metrics.TotalDeletes
+		writes[index] = di.Metrics.TotalWrites
+	}
+
+	filter := func(list []uint64) (commonCount uint64) {
+		max := 0
+		signatureMap := map[uint64]int{}
+		for _, v := range list {
+			signatureMap[v]++
+		}
+		for ops, count := range signatureMap {
+			if max < count && commonCount < ops {
+				max = count
+				commonCount = ops
+			}
+		}
+		if max < readQuorum {
+			return 0
+		}
+		return commonCount
+	}
+
+	commonWrite = filter(writes)
+	commonDelete = filter(deletes)
+	return
+}
+
+func calcCommonCounter(infos []DiskInfo, readQuorum int) (commonCount uint64) {
+	filter := func() (commonCount uint64) {
+		max := 0
+		signatureMap := map[uint64]int{}
+		for _, info := range infos {
+			if info.Error != "" {
+				continue
+			}
+			mutations := info.Metrics.TotalDeletes + info.Metrics.TotalWrites
+			signatureMap[mutations]++
+		}
+		for ops, count := range signatureMap {
+			if max < count && commonCount < ops {
+				max = count
+				commonCount = ops
+			}
+		}
+		if max < readQuorum {
+			return 0
+		}
+		return commonCount
+	}
+
+	return filter()
+}
+
+func getQuorumDiskInfos(disks []StorageAPI, infos []DiskInfo, readQuorum int) (newDisks []StorageAPI, newInfos []DiskInfo) {
+	commonMutations := calcCommonCounter(infos, readQuorum)
+	for i, info := range infos {
+		mutations := info.Metrics.TotalDeletes + info.Metrics.TotalWrites
+		if mutations >= commonMutations {
+			newDisks = append(newDisks, disks[i])
+			newInfos = append(newInfos, infos[i])
+		}
+	}
+
+	return newDisks, newInfos
+}
+
+func getQuorumDisks(disks []StorageAPI, infos []DiskInfo, readQuorum int) (newDisks []StorageAPI) {
+	newDisks, _ = getQuorumDiskInfos(disks, infos, readQuorum)
+	return newDisks
+}
+
 // Will return io.EOF if continuing would not yield more results.
 func (er *erasureObjects) listPath(ctx context.Context, o listPathOptions, results chan<- metaCacheEntry) (err error) {
-	defer close(results)
+	defer xioutil.SafeClose(results)
 	o.debugf(color.Green("listPath:")+" with options: %#v", o)
 
 	// get prioritized non-healing disks for listing
-	disks, _ := er.getOnlineDisksWithHealing(true)
+	disks, infos, _ := er.getOnlineDisksWithHealingAndInfo(true)
 	askDisks := getListQuorum(o.AskDisks, er.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", er.setDriveCount)
+		}
+	}
+
 	var fallbackDisks []StorageAPI
 
 	// Special case: ask all disks if the drive count is 4

cmd/metacache-stream.go

@@ -27,6 +27,7 @@ import (
 
 	jsoniter "github.com/json-iterator/go"
 	"github.com/klauspost/compress/s2"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/tinylib/msgp/msgp"
 	"github.com/valyala/bytebufferpool"
@@ -550,7 +551,7 @@ func (r *metacacheReader) readAll(ctx context.Context, dst chan<- metaCacheEntry
 	if r.err != nil {
 		return r.err
 	}
-	defer close(dst)
+	defer xioutil.SafeClose(dst)
 	if r.current.name != "" {
 		select {
 		case <-ctx.Done():

cmd/metacache-walk.go

@@ -94,7 +94,7 @@ func (s *xlStorage) WalkDir(ctx context.Context, opts WalkDirOptions, wr io.Writ
 	if err != nil {
 		return err
 	}
-	defer close(out)
+	defer xioutil.SafeClose(out)
 	var objsReturned int
 
 	objReturned := func(metadata []byte) {
@@ -407,12 +407,12 @@ func (client *storageRESTClient) WalkDir(ctx context.Context, opts WalkDirOption
 	opts.DiskID = client.diskID
 	b, err := opts.MarshalMsg(grid.GetByteBuffer()[:0])
 	if err != nil {
-		return err
+		return toStorageErr(err)
 	}
 
 	st, err := client.gridConn.NewStream(ctx, grid.HandlerWalkDir, b)
 	if err != nil {
-		return err
+		return toStorageErr(err)
 	}
 	return toStorageErr(st.Results(func(in []byte) error {
 		_, err := wr.Write(in)

cmd/metrics-realtime.go

@@ -136,7 +136,7 @@ func collectLocalDisksMetrics(disks map[string]struct{}) map[string]madmin.DiskM
 		return metrics
 	}
 
-	storageInfo := objLayer.LocalStorageInfo(GlobalContext)
+	storageInfo := objLayer.LocalStorageInfo(GlobalContext, true)
 	for _, d := range storageInfo.Disks {
 		if len(disks) != 0 {
 			_, ok := disks[d.Endpoint]

cmd/metrics-resource.go

@@ -257,7 +257,7 @@ func collectDriveMetrics(m madmin.RealtimeMetrics) {
 
 	for _, d := range localDrives {
 		labels := map[string]string{"drive": d.Endpoint().RawPath}
-		di, err := d.DiskInfo(GlobalContext, false)
+		di, err := d.DiskInfo(GlobalContext, DiskInfoOptions{})
 		if err == nil {
 			updateResourceMetrics(driveSubsystem, usedBytes, float64(di.Used), labels, false)
 			updateResourceMetrics(driveSubsystem, totalBytes, float64(di.Total), labels, false)

cmd/metrics-v2.go

@@ -1,4 +1,4 @@
-// Copyright (c) 2015-2023 MinIO, Inc.
+// Copyright (c) 2015-2024 MinIO, Inc.
 //
 // This file is part of MinIO Object Storage stack
 //
@@ -20,6 +20,7 @@ package cmd
 import (
 	"context"
 	"fmt"
+	"math"
 	"net/http"
 	"runtime"
 	"strconv"
@@ -31,6 +32,7 @@ import (
 	"github.com/minio/kes-go"
 	"github.com/minio/madmin-go/v3"
 	"github.com/minio/minio/internal/bucket/lifecycle"
+	xioutil "github.com/minio/minio/internal/ioutil"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/minio/internal/mcontext"
 	"github.com/minio/minio/internal/rest"
@@ -50,32 +52,33 @@ var (
 
 func init() {
 	clusterMetricsGroups := []*MetricsGroup{
-		getNodeHealthMetrics(),
-		getClusterStorageMetrics(),
-		getClusterTierMetrics(),
-		getClusterUsageMetrics(),
-		getKMSMetrics(),
-		getClusterHealthMetrics(),
-		getIAMNodeMetrics(),
-		getReplicationSiteMetrics(),
+		getNodeHealthMetrics(MetricsGroupOpts{dependGlobalNotificationSys: true}),
+		getClusterStorageMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true}),
+		getClusterTierMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true}),
+		getClusterUsageMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true}),
+		getKMSMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true, dependGlobalKMS: true}),
+		getClusterHealthMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true}),
+		getIAMNodeMetrics(MetricsGroupOpts{dependGlobalAuthNPlugin: true, dependGlobalIAMSys: true}),
+		getReplicationSiteMetrics(MetricsGroupOpts{dependGlobalSiteReplicationSys: true}),
+		getBatchJobsMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true}),
 	}
 
 	peerMetricsGroups = []*MetricsGroup{
 		getGoMetrics(),
-		getHTTPMetrics(false),
-		getNotificationMetrics(),
-		getLocalStorageMetrics(),
+		getHTTPMetrics(MetricsGroupOpts{}),
+		getNotificationMetrics(MetricsGroupOpts{dependGlobalLambdaTargetList: true}),
 		getMinioProcMetrics(),
 		getMinioVersionMetrics(),
 		getNetworkMetrics(),
 		getS3TTFBMetric(),
 		getILMNodeMetrics(),
 		getScannerNodeMetrics(),
-		getIAMNodeMetrics(),
-		getKMSNodeMetrics(),
-		getMinioHealingMetrics(),
+		getIAMNodeMetrics(MetricsGroupOpts{dependGlobalAuthNPlugin: true, dependGlobalIAMSys: true}),
+		getKMSNodeMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true, dependGlobalKMS: true}),
+		getMinioHealingMetrics(MetricsGroupOpts{dependGlobalBackgroundHealState: true}),
 		getWebhookMetrics(),
-		getReplicationClusterMetrics(),
+		getTierMetrics(),
+		getReplicationNodeMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true, dependBucketTargetSys: true}),
 	}
 
 	allMetricsGroups := func() (allMetrics []*MetricsGroup) {
@@ -85,25 +88,27 @@ func init() {
 	}()
 
 	nodeGroups := []*MetricsGroup{
-		getNodeHealthMetrics(),
-		getHTTPMetrics(false),
+		getNodeHealthMetrics(MetricsGroupOpts{dependGlobalNotificationSys: true}),
+		getHTTPMetrics(MetricsGroupOpts{}),
 		getNetworkMetrics(),
 		getMinioVersionMetrics(),
 		getS3TTFBMetric(),
-		getNotificationMetrics(),
-		getDistLockMetrics(),
-		getIAMNodeMetrics(),
+		getTierMetrics(),
+		getNotificationMetrics(MetricsGroupOpts{dependGlobalLambdaTargetList: true}),
+		getDistLockMetrics(MetricsGroupOpts{dependGlobalIsDistErasure: true, dependGlobalLockServer: true}),
+		getIAMNodeMetrics(MetricsGroupOpts{dependGlobalAuthNPlugin: true, dependGlobalIAMSys: true}),
+		getLocalStorageMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true}),
+		getReplicationNodeMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true, dependBucketTargetSys: true}),
 	}
 
 	bucketMetricsGroups := []*MetricsGroup{
-		getBucketUsageMetrics(),
-		getHTTPMetrics(true),
+		getBucketUsageMetrics(MetricsGroupOpts{dependGlobalObjectAPI: true}),
+		getHTTPMetrics(MetricsGroupOpts{bucketOnly: true}),
 		getBucketTTFBMetric(),
-		getBatchJobsMetrics(),
 	}
 
 	bucketPeerMetricsGroups = []*MetricsGroup{
-		getHTTPMetrics(true),
+		getHTTPMetrics(MetricsGroupOpts{bucketOnly: true}),
 		getBucketTTFBMetric(),
 	}
 
@@ -155,6 +160,7 @@ const (
 	usageSubsystem            MetricSubsystem = "usage"
 	quotaSubsystem            MetricSubsystem = "quota"
 	ilmSubsystem              MetricSubsystem = "ilm"
+	tierSubsystem             MetricSubsystem = "tier"
 	scannerSubsystem          MetricSubsystem = "scanner"
 	iamSubsystem              MetricSubsystem = "iam"
 	kmsSubsystem              MetricSubsystem = "kms"
@@ -246,11 +252,13 @@ const (
 	sizeDistribution    = "size_distribution"
 	versionDistribution = "version_distribution"
 	ttfbDistribution    = "seconds_distribution"
+	ttlbDistribution    = "ttlb_seconds_distribution"
 
 	lastActivityTime = "last_activity_nano_seconds"
 	startTime        = "starttime_seconds"
 	upTime           = "uptime_seconds"
 	memory           = "resident_memory_bytes"
+	vmemory          = "virtual_memory_bytes"
 	cpu              = "cpu_total_seconds"
 
 	expiryPendingTasks     MetricName = "expiry_pending_tasks"
@@ -262,6 +270,9 @@ const (
 	transitionedObjects  MetricName = "transitioned_objects"
 	transitionedVersions MetricName = "transitioned_versions"
 
+	tierRequestsSuccess MetricName = "requests_success"
+	tierRequestsFailure MetricName = "requests_failure"
+
 	kmsOnline          = "online"
 	kmsRequestsSuccess = "request_success"
 	kmsRequestsError   = "request_error"
@@ -308,8 +319,25 @@ type Metric struct {
 
 // MetricsGroup are a group of metrics that are initialized together.
 type MetricsGroup struct {
-	metricsCache  timedValue
-	cacheInterval time.Duration
+	metricsCache     timedValue
+	cacheInterval    time.Duration
+	metricsGroupOpts MetricsGroupOpts
+}
+
+// MetricsGroupOpts are a group of metrics opts to be used to initialize the metrics group.
+type MetricsGroupOpts struct {
+	dependGlobalObjectAPI           bool
+	dependGlobalAuthNPlugin         bool
+	dependGlobalSiteReplicationSys  bool
+	dependGlobalNotificationSys     bool
+	dependGlobalKMS                 bool
+	bucketOnly                      bool
+	dependGlobalLambdaTargetList    bool
+	dependGlobalIAMSys              bool
+	dependGlobalLockServer          bool
+	dependGlobalIsDistErasure       bool
+	dependGlobalBackgroundHealState bool
+	dependBucketTargetSys           bool
 }
 
 // RegisterRead register the metrics populator function to be used
@@ -319,6 +347,63 @@ func (g *MetricsGroup) RegisterRead(read func(ctx context.Context) []Metric) {
 		g.metricsCache.Relax = true
 		g.metricsCache.TTL = g.cacheInterval
 		g.metricsCache.Update = func() (interface{}, error) {
+			if g.metricsGroupOpts.dependGlobalObjectAPI {
+				objLayer := newObjectLayerFn()
+				// Service not initialized yet
+				if objLayer == nil {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependGlobalAuthNPlugin {
+				if globalAuthNPlugin == nil {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependGlobalSiteReplicationSys {
+				if !globalSiteReplicationSys.isEnabled() {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependGlobalNotificationSys {
+				if globalNotificationSys == nil {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependGlobalKMS {
+				if GlobalKMS == nil {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependGlobalLambdaTargetList {
+				if globalLambdaTargetList == nil {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependGlobalIAMSys {
+				if globalIAMSys == nil {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependGlobalLockServer {
+				if globalLockServer == nil {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependGlobalIsDistErasure {
+				if !globalIsDistErasure {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependGlobalBackgroundHealState {
+				if globalBackgroundHealState == nil {
+					return []Metric{}, nil
+				}
+			}
+			if g.metricsGroupOpts.dependBucketTargetSys {
+				if globalBucketTargetSys == nil {
+					return []Metric{}, nil
+				}
+			}
 			return read(GlobalContext), nil
 		}
 	})
@@ -437,7 +522,7 @@ func getNodeDriveTimeoutErrorsMD() MetricDescription {
 		Namespace: nodeMetricNamespace,
 		Subsystem: driveSubsystem,
 		Name:      "errors_timeout",
-		Help:      "Total number of timeout errors since server start",
+		Help:      "Total number of drive timeout errors since server start",
 		Type:      counterMetric,
 	}
 }
@@ -447,7 +532,27 @@ func getNodeDriveAvailablityErrorsMD() MetricDescription {
 		Namespace: nodeMetricNamespace,
 		Subsystem: driveSubsystem,
 		Name:      "errors_availability",
-		Help:      "Total number of I/O errors, permission denied and timeouts since server start",
+		Help:      "Total number of drive I/O errors, permission denied and timeouts since server start",
+		Type:      counterMetric,
+	}
+}
+
+func getNodeDriveWaitingIOMD() MetricDescription {
+	return MetricDescription{
+		Namespace: nodeMetricNamespace,
+		Subsystem: driveSubsystem,
+		Name:      "io_waiting",
+		Help:      "Total number I/O operations waiting on drive",
+		Type:      counterMetric,
+	}
+}
+
+func getNodeDriveTokensIOMD() MetricDescription {
+	return MetricDescription{
+		Namespace: nodeMetricNamespace,
+		Subsystem: driveSubsystem,
+		Name:      "io_tokens",
+		Help:      "Total number concurrent I/O operations configured on drive",
 		Type:      counterMetric,
 	}
 }
@@ -1450,6 +1555,16 @@ func getMinIOProcessResidentMemory() MetricDescription {
 	}
 }
 
+func getMinIOProcessVirtualMemory() MetricDescription {
+	return MetricDescription{
+		Namespace: nodeMetricNamespace,
+		Subsystem: processSubsystem,
+		Name:      memory,
+		Help:      "Virtual memory size in bytes",
+		Type:      gaugeMetric,
+	}
+}
+
 func getMinIOProcessCPUTime() MetricDescription {
 	return MetricDescription{
 		Namespace: nodeMetricNamespace,
@@ -1572,6 +1687,14 @@ func getMinioProcMetrics() *MetricsGroup {
 				})
 		}
 
+		if stat.VirtualMemory() > 0 {
+			metrics = append(metrics,
+				Metric{
+					Description: getMinIOProcessVirtualMemory(),
+					Value:       float64(stat.VirtualMemory()),
+				})
+		}
+
 		if stat.CPUTime() > 0 {
 			metrics = append(metrics,
 				Metric{
@@ -1598,47 +1721,63 @@ func getGoMetrics() *MetricsGroup {
 	return mg
 }
 
+// getHistogramMetrics fetches histogram metrics and returns it in a []Metric
+// Note: Typically used in MetricGroup.RegisterRead
+func getHistogramMetrics(hist *prometheus.HistogramVec, desc MetricDescription) []Metric {
+	ch := make(chan prometheus.Metric)
+	go func() {
+		defer xioutil.SafeClose(ch)
+		// Collects prometheus metrics from hist and sends it over ch
+		hist.Collect(ch)
+	}()
+
+	// Converts metrics received into internal []Metric type
+	var metrics []Metric
+	for promMetric := range ch {
+		dtoMetric := &dto.Metric{}
+		err := promMetric.Write(dtoMetric)
+		if err != nil {
+			// Log error and continue to receive other metric
+			// values
+			logger.LogIf(GlobalContext, err)
+			continue
+		}
+
+		h := dtoMetric.GetHistogram()
+		for _, b := range h.Bucket {
+			labels := make(map[string]string)
+			for _, lp := range dtoMetric.GetLabel() {
+				labels[*lp.Name] = *lp.Value
+			}
+			labels["le"] = fmt.Sprintf("%.3f", *b.UpperBound)
+			metric := Metric{
+				Description:    desc,
+				VariableLabels: labels,
+				Value:          float64(b.GetCumulativeCount()),
+			}
+			metrics = append(metrics, metric)
+		}
+		// add metrics with +Inf label
+		labels1 := make(map[string]string)
+		for _, lp := range dtoMetric.GetLabel() {
+			labels1[*lp.Name] = *lp.Value
+		}
+		labels1["le"] = fmt.Sprintf("%.3f", math.Inf(+1))
+		metrics = append(metrics, Metric{
+			Description:    desc,
+			VariableLabels: labels1,
+			Value:          dtoMetric.Counter.GetValue(),
+		})
+	}
+	return metrics
+}
+
 func getBucketTTFBMetric() *MetricsGroup {
 	mg := &MetricsGroup{
 		cacheInterval: 10 * time.Second,
 	}
-	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
-		// Read prometheus metric on this channel
-		ch := make(chan prometheus.Metric)
-		var wg sync.WaitGroup
-		wg.Add(1)
-
-		// Read prometheus histogram data and convert it to internal metric data
-		go func() {
-			defer wg.Done()
-			for promMetric := range ch {
-				dtoMetric := &dto.Metric{}
-				err := promMetric.Write(dtoMetric)
-				if err != nil {
-					logger.LogIf(GlobalContext, err)
-					return
-				}
-				h := dtoMetric.GetHistogram()
-				for _, b := range h.Bucket {
-					labels := make(map[string]string)
-					for _, lp := range dtoMetric.GetLabel() {
-						labels[*lp.Name] = *lp.Value
-					}
-					labels["le"] = fmt.Sprintf("%.3f", *b.UpperBound)
-					metric := Metric{
-						Description:    getBucketTTFBDistributionMD(),
-						VariableLabels: labels,
-						Value:          float64(b.GetCumulativeCount()),
-					}
-					metrics = append(metrics, metric)
-				}
-			}
-		}()
-
-		bucketHTTPRequestsDuration.Collect(ch)
-		close(ch)
-		wg.Wait()
-		return
+	mg.RegisterRead(func(ctx context.Context) []Metric {
+		return getHistogramMetrics(bucketHTTPRequestsDuration, getBucketTTFBDistributionMD())
 	})
 	return mg
 }
@@ -1647,43 +1786,18 @@ func getS3TTFBMetric() *MetricsGroup {
 	mg := &MetricsGroup{
 		cacheInterval: 10 * time.Second,
 	}
-	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
-		// Read prometheus metric on this channel
-		ch := make(chan prometheus.Metric)
-		var wg sync.WaitGroup
-		wg.Add(1)
+	mg.RegisterRead(func(ctx context.Context) []Metric {
+		return getHistogramMetrics(httpRequestsDuration, getS3TTFBDistributionMD())
+	})
+	return mg
+}
 
-		// Read prometheus histogram data and convert it to internal metric data
-		go func() {
-			defer wg.Done()
-			for promMetric := range ch {
-				dtoMetric := &dto.Metric{}
-				err := promMetric.Write(dtoMetric)
-				if err != nil {
-					logger.LogIf(GlobalContext, err)
-					return
-				}
-				h := dtoMetric.GetHistogram()
-				for _, b := range h.Bucket {
-					labels := make(map[string]string)
-					for _, lp := range dtoMetric.GetLabel() {
-						labels[*lp.Name] = *lp.Value
-					}
-					labels["le"] = fmt.Sprintf("%.3f", *b.UpperBound)
-					metric := Metric{
-						Description:    getS3TTFBDistributionMD(),
-						VariableLabels: labels,
-						Value:          float64(b.GetCumulativeCount()),
-					}
-					metrics = append(metrics, metric)
-				}
-			}
-		}()
-
-		httpRequestsDuration.Collect(ch)
-		close(ch)
-		wg.Wait()
-		return
+func getTierMetrics() *MetricsGroup {
+	mg := &MetricsGroup{
+		cacheInterval: 10 * time.Second,
+	}
+	mg.RegisterRead(func(ctx context.Context) []Metric {
+		return globalTierMetrics.Report()
 	})
 	return mg
 }
@@ -1902,9 +2016,10 @@ func getScannerNodeMetrics() *MetricsGroup {
 	return mg
 }
 
-func getIAMNodeMetrics() *MetricsGroup {
+func getIAMNodeMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 10 * time.Second,
+		cacheInterval:    10 * time.Second,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(_ context.Context) (metrics []Metric) {
 		lastSyncTime := atomic.LoadUint64(&globalIAMSys.LastRefreshTimeUnixNano)
@@ -2023,9 +2138,10 @@ func getIAMNodeMetrics() *MetricsGroup {
 }
 
 // replication metrics for each node - published to the cluster endpoint with nodename as label
-func getReplicationClusterMetrics() *MetricsGroup {
+func getReplicationNodeMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 1 * time.Minute,
+		cacheInterval:    1 * time.Minute,
+		metricsGroupOpts: opts,
 	}
 	const (
 		Online  = 1
@@ -2210,9 +2326,10 @@ func getReplicationClusterMetrics() *MetricsGroup {
 }
 
 // replication metrics for site replication
-func getReplicationSiteMetrics() *MetricsGroup {
+func getReplicationSiteMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 1 * time.Minute,
+		cacheInterval:    1 * time.Minute,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(_ context.Context) []Metric {
 		ml := []Metric{}
@@ -2309,9 +2426,10 @@ func getMinioVersionMetrics() *MetricsGroup {
 	return mg
 }
 
-func getNodeHealthMetrics() *MetricsGroup {
+func getNodeHealthMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 1 * time.Minute,
+		cacheInterval:    1 * time.Minute,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(_ context.Context) (metrics []Metric) {
 		metrics = make([]Metric, 0, 16)
@@ -2329,9 +2447,10 @@ func getNodeHealthMetrics() *MetricsGroup {
 	return mg
 }
 
-func getMinioHealingMetrics() *MetricsGroup {
+func getMinioHealingMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 10 * time.Second,
+		cacheInterval:    10 * time.Second,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(_ context.Context) (metrics []Metric) {
 		bgSeq, exists := globalBackgroundHealState.getHealSequenceByToken(bgHealingUUID)
@@ -2399,9 +2518,10 @@ func getObjectsScanned(seq *healSequence) (m []Metric) {
 	return
 }
 
-func getDistLockMetrics() *MetricsGroup {
+func getDistLockMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 1 * time.Second,
+		cacheInterval:    1 * time.Second,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(ctx context.Context) []Metric {
 		if !globalIsDistErasure {
@@ -2446,9 +2566,10 @@ func getDistLockMetrics() *MetricsGroup {
 	return mg
 }
 
-func getNotificationMetrics() *MetricsGroup {
+func getNotificationMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 10 * time.Second,
+		cacheInterval:    10 * time.Second,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(ctx context.Context) []Metric {
 		metrics := make([]Metric, 0, 3)
@@ -2621,12 +2742,13 @@ func getNotificationMetrics() *MetricsGroup {
 	return mg
 }
 
-func getHTTPMetrics(bucketOnly bool) *MetricsGroup {
+func getHTTPMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 10 * time.Second,
+		cacheInterval:    10 * time.Second,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
-		if !bucketOnly {
+		if !mg.metricsGroupOpts.bucketOnly {
 			httpStats := globalHTTPStats.toServerHTTPStats()
 			metrics = make([]Metric, 0, 3+
 				len(httpStats.CurrentS3Requests.APIStats)+
@@ -2812,13 +2934,13 @@ func getNetworkMetrics() *MetricsGroup {
 	return mg
 }
 
-func getClusterUsageMetrics() *MetricsGroup {
+func getClusterUsageMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 1 * time.Minute,
+		cacheInterval:    1 * time.Minute,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
 		objLayer := newObjectLayerFn()
-		// Service not initialized yet
 		if objLayer == nil {
 			return
 		}
@@ -2918,16 +3040,13 @@ func getClusterUsageMetrics() *MetricsGroup {
 	return mg
 }
 
-func getBucketUsageMetrics() *MetricsGroup {
+func getBucketUsageMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 1 * time.Minute,
+		cacheInterval:    1 * time.Minute,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
 		objLayer := newObjectLayerFn()
-		// Service not initialized yet
-		if objLayer == nil {
-			return
-		}
 
 		metrics = make([]Metric, 0, 50)
 		dataUsageInfo, err := loadDataUsageFromBackend(ctx, objLayer)
@@ -3057,20 +3176,20 @@ func getBucketUsageMetrics() *MetricsGroup {
 						}
 					}
 				}
-				metrics = append(metrics, Metric{
-					Description:          getBucketObjectDistributionMD(),
-					Histogram:            usage.ObjectSizesHistogram,
-					HistogramBucketLabel: "range",
-					VariableLabels:       map[string]string{"bucket": bucket},
-				})
-
-				metrics = append(metrics, Metric{
-					Description:          getBucketObjectVersionsMD(),
-					Histogram:            usage.ObjectVersionsHistogram,
-					HistogramBucketLabel: "range",
-					VariableLabels:       map[string]string{"bucket": bucket},
-				})
 			}
+			metrics = append(metrics, Metric{
+				Description:          getBucketObjectDistributionMD(),
+				Histogram:            usage.ObjectSizesHistogram,
+				HistogramBucketLabel: "range",
+				VariableLabels:       map[string]string{"bucket": bucket},
+			})
+
+			metrics = append(metrics, Metric{
+				Description:          getBucketObjectVersionsMD(),
+				Histogram:            usage.ObjectVersionsHistogram,
+				HistogramBucketLabel: "range",
+				VariableLabels:       map[string]string{"bucket": bucket},
+			})
 		}
 		return
 	})
@@ -3107,15 +3226,14 @@ func getClusterTransitionedVersionsMD() MetricDescription {
 	}
 }
 
-func getClusterTierMetrics() *MetricsGroup {
+func getClusterTierMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 1 * time.Minute,
+		cacheInterval:    1 * time.Minute,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
 		objLayer := newObjectLayerFn()
-		if objLayer == nil {
-			return
-		}
+
 		if globalTierConfigMgr.Empty() {
 			return
 		}
@@ -3135,19 +3253,16 @@ func getClusterTierMetrics() *MetricsGroup {
 	return mg
 }
 
-func getLocalStorageMetrics() *MetricsGroup {
+func getLocalStorageMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 1 * time.Minute,
+		cacheInterval:    1 * time.Minute,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
 		objLayer := newObjectLayerFn()
-		// Service not initialized yet
-		if objLayer == nil {
-			return
-		}
 
 		metrics = make([]Metric, 0, 50)
-		storageInfo := objLayer.LocalStorageInfo(ctx)
+		storageInfo := objLayer.LocalStorageInfo(ctx, true)
 		onlineDrives, offlineDrives := getOnlineOfflineDisksStats(storageInfo.Disks)
 		totalDrives := onlineDrives.Merge(offlineDrives)
 
@@ -3189,6 +3304,18 @@ func getLocalStorageMetrics() *MetricsGroup {
 					VariableLabels: map[string]string{"drive": disk.DrivePath},
 				})
 
+				metrics = append(metrics, Metric{
+					Description:    getNodeDriveWaitingIOMD(),
+					Value:          float64(disk.Metrics.TotalWaiting),
+					VariableLabels: map[string]string{"drive": disk.DrivePath},
+				})
+
+				metrics = append(metrics, Metric{
+					Description:    getNodeDriveTokensIOMD(),
+					Value:          float64(disk.Metrics.TotalTokens),
+					VariableLabels: map[string]string{"drive": disk.DrivePath},
+				})
+
 				for apiName, latency := range disk.Metrics.LastMinute {
 					metrics = append(metrics, Metric{
 						Description:    getNodeDriveAPILatencyMD(),
@@ -3249,32 +3376,69 @@ func getClusterHealthStatusMD() MetricDescription {
 	}
 }
 
-func getClusterErasureSetToleranceMD() MetricDescription {
+func getClusterErasureSetHealthStatusMD() MetricDescription {
 	return MetricDescription{
 		Namespace: clusterMetricNamespace,
 		Subsystem: "health",
-		Name:      "erasure_set_tolerance",
-		Help:      "Get erasure set tolerance status",
+		Name:      "erasure_set_status",
+		Help:      "Get current health status for this erasure set",
 		Type:      gaugeMetric,
 	}
 }
 
-func getClusterHealthMetrics() *MetricsGroup {
+func getClusterErasureSetReadQuorumMD() MetricDescription {
+	return MetricDescription{
+		Namespace: clusterMetricNamespace,
+		Subsystem: "health",
+		Name:      "erasure_set_read_quorum",
+		Help:      "Get the read quorum for this erasure set",
+		Type:      gaugeMetric,
+	}
+}
+
+func getClusterErasureSetWriteQuorumMD() MetricDescription {
+	return MetricDescription{
+		Namespace: clusterMetricNamespace,
+		Subsystem: "health",
+		Name:      "erasure_set_write_quorum",
+		Help:      "Get the write quorum for this erasure set",
+		Type:      gaugeMetric,
+	}
+}
+
+func getClusterErasureSetOnlineDrivesMD() MetricDescription {
+	return MetricDescription{
+		Namespace: clusterMetricNamespace,
+		Subsystem: "health",
+		Name:      "erasure_set_online_drives",
+		Help:      "Get the count of the online drives in this erasure set",
+		Type:      gaugeMetric,
+	}
+}
+
+func getClusterErasureSetHealingDrivesMD() MetricDescription {
+	return MetricDescription{
+		Namespace: clusterMetricNamespace,
+		Subsystem: "health",
+		Name:      "erasure_set_healing_drives",
+		Help:      "Get the count of healing drives of this erasure set",
+		Type:      gaugeMetric,
+	}
+}
+
+func getClusterHealthMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 10 * time.Second,
+		cacheInterval:    10 * time.Second,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
 		objLayer := newObjectLayerFn()
-		// Service not initialized yet
-		if objLayer == nil {
-			return
-		}
-
-		metrics = make([]Metric, 0, 2)
 
 		opts := HealthOptions{}
 		result := objLayer.Health(ctx, opts)
 
+		metrics = make([]Metric, 0, 2+4*len(result.ESHealth))
+
 		metrics = append(metrics, Metric{
 			Description: getClusterWriteQuorumMD(),
 			Value:       float64(result.WriteQuorum),
@@ -3296,9 +3460,35 @@ func getClusterHealthMetrics() *MetricsGroup {
 				"set":  strconv.Itoa(h.SetID),
 			}
 			metrics = append(metrics, Metric{
-				Description:    getClusterErasureSetToleranceMD(),
+				Description:    getClusterErasureSetReadQuorumMD(),
 				VariableLabels: labels,
-				Value:          float64(h.HealthyDrives - h.WriteQuorum),
+				Value:          float64(h.ReadQuorum),
+			})
+			metrics = append(metrics, Metric{
+				Description:    getClusterErasureSetWriteQuorumMD(),
+				VariableLabels: labels,
+				Value:          float64(h.WriteQuorum),
+			})
+			metrics = append(metrics, Metric{
+				Description:    getClusterErasureSetOnlineDrivesMD(),
+				VariableLabels: labels,
+				Value:          float64(h.HealthyDrives),
+			})
+			metrics = append(metrics, Metric{
+				Description:    getClusterErasureSetHealingDrivesMD(),
+				VariableLabels: labels,
+				Value:          float64(h.HealingDrives),
+			})
+
+			health := 1
+			if !h.Healthy {
+				health = 0
+			}
+
+			metrics = append(metrics, Metric{
+				Description:    getClusterErasureSetHealthStatusMD(),
+				VariableLabels: labels,
+				Value:          float64(health),
 			})
 		}
 
@@ -3308,18 +3498,13 @@ func getClusterHealthMetrics() *MetricsGroup {
 	return mg
 }
 
-func getBatchJobsMetrics() *MetricsGroup {
+func getBatchJobsMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 10 * time.Second,
+		cacheInterval:    10 * time.Second,
+		metricsGroupOpts: opts,
 	}
 
 	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
-		objLayer := newObjectLayerFn()
-		// Service not initialized yet
-		if objLayer == nil {
-			return
-		}
-
 		var m madmin.RealtimeMetrics
 		mLocal := collectLocalMetrics(madmin.MetricsBatchJobs, collectMetricsOpts{})
 		m.Merge(&mLocal)
@@ -3379,20 +3564,17 @@ func getBatchJobsMetrics() *MetricsGroup {
 	return mg
 }
 
-func getClusterStorageMetrics() *MetricsGroup {
+func getClusterStorageMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 1 * time.Minute,
+		cacheInterval:    1 * time.Minute,
+		metricsGroupOpts: opts,
 	}
 	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
 		objLayer := newObjectLayerFn()
-		// Service not initialized yet
-		if objLayer == nil {
-			return
-		}
 
 		// Fetch disk space info, ignore errors
 		metrics = make([]Metric, 0, 10)
-		storageInfo := objLayer.StorageInfo(ctx)
+		storageInfo := objLayer.StorageInfo(ctx, true)
 		onlineDrives, offlineDrives := getOnlineOfflineDisksStats(storageInfo.Disks)
 		totalDrives := onlineDrives.Merge(offlineDrives)
 
@@ -3435,18 +3617,13 @@ func getClusterStorageMetrics() *MetricsGroup {
 	return mg
 }
 
-func getKMSNodeMetrics() *MetricsGroup {
+func getKMSNodeMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 10 * time.Second,
+		cacheInterval:    10 * time.Second,
+		metricsGroupOpts: opts,
 	}
 
 	mg.RegisterRead(func(ctx context.Context) (metrics []Metric) {
-		objLayer := newObjectLayerFn()
-		// Service not initialized yet
-		if objLayer == nil || GlobalKMS == nil {
-			return
-		}
-
 		const (
 			Online  = 1
 			Offline = 0
@@ -3540,18 +3717,13 @@ func getWebhookMetrics() *MetricsGroup {
 	return mg
 }
 
-func getKMSMetrics() *MetricsGroup {
+func getKMSMetrics(opts MetricsGroupOpts) *MetricsGroup {
 	mg := &MetricsGroup{
-		cacheInterval: 10 * time.Second,
+		cacheInterval:    10 * time.Second,
+		metricsGroupOpts: opts,
 	}
 
 	mg.RegisterRead(func(ctx context.Context) []Metric {
-		objLayer := newObjectLayerFn()
-		// Service not initialized yet
-		if objLayer == nil || GlobalKMS == nil {
-			return []Metric{}
-		}
-
 		metrics := make([]Metric, 0, 4)
 		metric, err := GlobalKMS.Metrics(ctx)
 		if err != nil {
@@ -3732,7 +3904,7 @@ func (c *minioClusterCollector) Collect(out chan<- prometheus.Metric) {
 func ReportMetrics(ctx context.Context, metricsGroups []*MetricsGroup) <-chan Metric {
 	ch := make(chan Metric)
 	go func() {
-		defer close(ch)
+		defer xioutil.SafeClose(ch)
 		populateAndPublish(metricsGroups, func(m Metric) bool {
 			if m.VariableLabels == nil {
 				m.VariableLabels = make(map[string]string)

cmd/metrics-v2_test.go

@@ -0,0 +1,88 @@
+// 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"
+	"time"
+
+	"github.com/prometheus/client_golang/prometheus"
+)
+
+func TestGetHistogramMetrics(t *testing.T) {
+	histBuckets := []float64{0.05, 0.1, 0.25, 0.5, 0.75}
+	labels := []string{"GetObject", "PutObject", "CopyObject", "CompleteMultipartUpload"}
+	ttfbHist := prometheus.NewHistogramVec(
+		prometheus.HistogramOpts{
+			Name:    "s3_ttfb_seconds",
+			Help:    "Time taken by requests served by current MinIO server instance",
+			Buckets: histBuckets,
+		},
+		[]string{"api"},
+	)
+	observations := []struct {
+		val   float64
+		label string
+	}{
+		{
+			val:   0.02,
+			label: labels[0],
+		},
+		{
+			val:   0.07,
+			label: labels[1],
+		},
+		{
+			val:   0.11,
+			label: labels[1],
+		},
+		{
+			val:   0.19,
+			label: labels[1],
+		},
+		{
+			val:   0.31,
+			label: labels[1],
+		},
+		{
+			val:   0.61,
+			label: labels[3],
+		},
+		{
+			val:   0.79,
+			label: labels[2],
+		},
+	}
+	ticker := time.NewTicker(1 * time.Millisecond)
+	defer ticker.Stop()
+	for _, obs := range observations {
+		// Send observations once every 1ms, to simulate delay between
+		// observations. This is to test the channel based
+		// synchronization used internally.
+		select {
+		case <-ticker.C:
+			ttfbHist.With(prometheus.Labels{"api": obs.label}).Observe(obs.val)
+		}
+	}
+
+	metrics := getHistogramMetrics(ttfbHist, getBucketTTFBDistributionMD())
+	// additional labels for +Inf for all histogram metrics
+	if expPoints := len(labels) * (len(histBuckets) + 1); expPoints != len(metrics) {
+		t.Fatalf("Expected %v data points but got %v", expPoints, len(metrics))
+	}
+}

cmd/metrics.go

@@ -406,7 +406,7 @@ func storageMetricsPrometheus(ch chan<- prometheus.Metric) {
 		float64(GetTotalCapacityFree(server.Disks)),
 	)
 
-	sinfo := objLayer.StorageInfo(GlobalContext)
+	sinfo := objLayer.StorageInfo(GlobalContext, true)
 
 	// Report total usable capacity
 	ch <- prometheus.MustNewConstMetric(

cmd/naughty-disk_test.go

@@ -1,4 +1,4 @@
-// Copyright (c) 2015-2021 MinIO, Inc.
+// Copyright (c) 2015-2024 MinIO, Inc.
 //
 // This file is part of MinIO Object Storage stack
 //
@@ -108,19 +108,26 @@ func (d *naughtyDisk) GetDiskID() (string, error) {
 	return d.disk.GetDiskID()
 }
 
+func (d *naughtyDisk) SetFormatData(b []byte) {
+	d.disk.SetFormatData(b)
+}
+
 func (d *naughtyDisk) SetDiskID(id string) {
 	d.disk.SetDiskID(id)
 }
 
-func (d *naughtyDisk) NSScanner(ctx context.Context, cache dataUsageCache, updates chan<- dataUsageEntry, scanMode madmin.HealScanMode) (info dataUsageCache, err error) {
-	return d.disk.NSScanner(ctx, cache, updates, scanMode)
-}
-
-func (d *naughtyDisk) DiskInfo(ctx context.Context, metrics bool) (info DiskInfo, err error) {
+func (d *naughtyDisk) NSScanner(ctx context.Context, cache dataUsageCache, updates chan<- dataUsageEntry, scanMode madmin.HealScanMode, weSleep func() bool) (info dataUsageCache, err error) {
 	if err := d.calcError(); err != nil {
 		return info, err
 	}
-	return d.disk.DiskInfo(ctx, metrics)
+	return d.disk.NSScanner(ctx, cache, updates, scanMode, weSleep)
+}
+
+func (d *naughtyDisk) DiskInfo(ctx context.Context, opts DiskInfoOptions) (info DiskInfo, err error) {
+	if err := d.calcError(); err != nil {
+		return info, err
+	}
+	return d.disk.DiskInfo(ctx, opts)
 }
 
 func (d *naughtyDisk) MakeVolBulk(ctx context.Context, volumes ...string) (err error) {
@@ -200,11 +207,11 @@ func (d *naughtyDisk) AppendFile(ctx context.Context, volume string, path string
 	return d.disk.AppendFile(ctx, volume, path, buf)
 }
 
-func (d *naughtyDisk) RenameData(ctx context.Context, srcVolume, srcPath string, fi FileInfo, dstVolume, dstPath string) (uint64, error) {
+func (d *naughtyDisk) RenameData(ctx context.Context, srcVolume, srcPath string, fi FileInfo, dstVolume, dstPath string, opts RenameOptions) (uint64, error) {
 	if err := d.calcError(); err != nil {
 		return 0, err
 	}
-	return d.disk.RenameData(ctx, srcVolume, srcPath, fi, dstVolume, dstPath)
+	return d.disk.RenameData(ctx, srcVolume, srcPath, fi, dstVolume, dstPath, opts)
 }
 
 func (d *naughtyDisk) RenameFile(ctx context.Context, srcVolume, srcPath, dstVolume, dstPath string) error {
@@ -228,7 +235,7 @@ func (d *naughtyDisk) Delete(ctx context.Context, volume string, path string, de
 	return d.disk.Delete(ctx, volume, path, deleteOpts)
 }
 
-func (d *naughtyDisk) DeleteVersions(ctx context.Context, volume string, versions []FileInfoVersions) []error {
+func (d *naughtyDisk) DeleteVersions(ctx context.Context, volume string, versions []FileInfoVersions, opts DeleteOptions) []error {
 	if err := d.calcError(); err != nil {
 		errs := make([]error, len(versions))
 		for i := range errs {
@@ -236,7 +243,7 @@ func (d *naughtyDisk) DeleteVersions(ctx context.Context, volume string, version
 		}
 		return errs
 	}
-	return d.disk.DeleteVersions(ctx, volume, versions)
+	return d.disk.DeleteVersions(ctx, volume, versions, opts)
 }
 
 func (d *naughtyDisk) WriteMetadata(ctx context.Context, volume, path string, fi FileInfo) (err error) {
@@ -253,11 +260,11 @@ func (d *naughtyDisk) UpdateMetadata(ctx context.Context, volume, path string, f
 	return d.disk.UpdateMetadata(ctx, volume, path, fi, opts)
 }
 
-func (d *naughtyDisk) DeleteVersion(ctx context.Context, volume, path string, fi FileInfo, forceDelMarker bool) (err error) {
+func (d *naughtyDisk) DeleteVersion(ctx context.Context, volume, path string, fi FileInfo, forceDelMarker bool, opts DeleteOptions) (err error) {
 	if err := d.calcError(); err != nil {
 		return err
 	}
-	return d.disk.DeleteVersion(ctx, volume, path, fi, forceDelMarker)
+	return d.disk.DeleteVersion(ctx, volume, path, fi, forceDelMarker, opts)
 }
 
 func (d *naughtyDisk) ReadVersion(ctx context.Context, volume, path, versionID string, opts ReadOptions) (fi FileInfo, err error) {