replication: improve MRF healing. (#15556)

This PR improves the replication failure healing by persisting
most recent failures to disk and re-queuing them until the replication
is successful.

While this does not eliminate the need for healing during a full scan, 
queuing MRF vastly improves the ETA to keeping replicated buckets 
in sync as it does not wait for the scanner visit to detect unreplicated 
object versions.

.github/workflows/depsreview.yaml

@@ -0,0 +1,14 @@
+name: 'Dependency Review'
+on: [pull_request]
+
+permissions:
+  contents: read
+
+jobs:
+  dependency-review:
+    runs-on: ubuntu-latest
+    steps:
+      - name: 'Checkout Repository'
+        uses: actions/checkout@v3
+      - name: 'Dependency Review'
+        uses: actions/dependency-review-action@v1

.github/workflows/go-cross.yml

@@ -20,7 +20,7 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        go-version: [1.17.x, 1.18.x]
+        go-version: [1.18.x]
         os: [ubuntu-latest]
     steps:
       - uses: actions/checkout@629c2de402a417ea7690ca6ce3f33229e27606a5 # v2

.github/workflows/go-fips.yml

@@ -20,7 +20,7 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        go-version: [1.17.11b7, 1.18.3b7]
+        go-version: [1.18.5b7]
         os: [ubuntu-latest]
     steps:
       - uses: actions/checkout@v2

.github/workflows/go-healing.yml

@@ -20,7 +20,7 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        go-version: [1.17.x]
+        go-version: [1.18.x]
         os: [ubuntu-latest]
     steps:
       - uses: actions/checkout@v2
@@ -49,3 +49,4 @@ jobs:
           make verify-healing
           make verify-healing-inconsistent-versions
           make verify-healing-with-root-disks
+          make verify-healing-with-rewrite

.github/workflows/go-lint.yml

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

.github/workflows/go.yml

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

.github/workflows/iam-integrations.yaml

@@ -61,7 +61,7 @@ jobs:
       # are turned off - i.e. if ldap="", then ldap server is not enabled for
       # the tests.
       matrix:
-        go-version: [1.17.x]
+        go-version: [1.18.x]
         ldap: ["", "localhost:389"]
         etcd: ["", "http://localhost:2379"]
         openid: ["", "http://127.0.0.1:5556/dex"]

.github/workflows/replication.yaml

@@ -21,7 +21,7 @@ jobs:
 
     strategy:
       matrix:
-        go-version: [1.17.x]
+        go-version: [1.18.x]
 
     steps:
       - uses: actions/checkout@v2

.github/workflows/upgrade-ci-cd.yaml

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

.golangci.yml

@@ -14,13 +14,9 @@ linters:
     - govet
     - revive
     - ineffassign
-    - gosimple
     - deadcode
-    - structcheck
     - gomodguard
     - gofmt
-    - unused
-    - structcheck
     - unconvert
     - varcheck
     - gocritic

.nancy-ignore

@@ -2,3 +2,5 @@ CVE-2020-26160
 CVE-2020-15136
 CVE-2020-15115
 CVE-2020-15114
+CVE-2022-2835
+CVE-2022-2837

Makefile

@@ -15,7 +15,7 @@ checks: ## check dependencies
 	@(env bash $(PWD)/buildscripts/checkdeps.sh)
 
 help: ## print this help
-	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' Makefile | awk 'BEGIN {FS = ":.*?## "}; {printf "\033[36m%-30s\033[0m %s\n", $$1, $$2}'
+	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' Makefile | awk 'BEGIN {FS = ":.*?## "}; {printf "\033[36m%-40s\033[0m %s\n", $$1, $$2}'
 
 getdeps: ## fetch necessary dependencies
 	@mkdir -p ${GOPATH}/bin
@@ -90,11 +90,16 @@ verify-healing: ## verify healing and replacing disks with minio binary
 	@(env bash $(PWD)/buildscripts/verify-healing.sh)
 	@(env bash $(PWD)/buildscripts/unaligned-healing.sh)
 
-verify-healing-with-root-disks:
-	@echo "Verify healing with root disks"
+verify-healing-with-root-disks: ## verify healing root disks
+	@echo "Verify healing with root drives"
 	@GORACE=history_size=7 CGO_ENABLED=1 go build -race -tags kqueue -trimpath --ldflags "$(LDFLAGS)" -o $(PWD)/minio 1>/dev/null
 	@(env bash $(PWD)/buildscripts/verify-healing-with-root-disks.sh)
 
+verify-healing-with-rewrite: ## verify healing to rewrite old xl.meta -> new xl.meta
+	@echo "Verify healing with rewrite"
+	@GORACE=history_size=7 CGO_ENABLED=1 go build -race -tags kqueue -trimpath --ldflags "$(LDFLAGS)" -o $(PWD)/minio 1>/dev/null
+	@(env bash $(PWD)/buildscripts/rewrite-old-new.sh)
+
 verify-healing-inconsistent-versions: ## verify resolving inconsistent versions
 	@echo "Verify resolving inconsistent versions build with race"
 	@GORACE=history_size=7 CGO_ENABLED=1 go build -race -tags kqueue -trimpath --ldflags "$(LDFLAGS)" -o $(PWD)/minio 1>/dev/null

README.md

@@ -125,7 +125,7 @@ You can also connect using any S3-compatible tool, such as the MinIO Client `mc`
 
 ## Install from Source
 
-Use the following commands to compile and run a standalone MinIO server from source. Source installation is only intended for developers and advanced users. If you do not have a working Golang environment, please follow [How to install Golang](https://golang.org/doc/install). Minimum version required is [go1.17](https://golang.org/dl/#stable)
+Use the following commands to compile and run a standalone MinIO server from source. Source installation is only intended for developers and advanced users. If you do not have a working Golang environment, please follow [How to install Golang](https://golang.org/doc/install). Minimum version required is [go1.18](https://golang.org/dl/#stable)
 
 ```sh
 GO111MODULE=on go install github.com/minio/minio@latest

buildscripts/heal-manual.go

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

buildscripts/resolve-right-versions.sh

@@ -39,7 +39,7 @@ function start_minio_5drive() {
     "${MINIO[@]}" --address ":$start_port" "${WORK_DIR}/cicd-corpus/disk{1...5}" > "${WORK_DIR}/server1.log" 2>&1 &
     pid=$!
     disown $pid
-    sleep 30
+    sleep 5
 
     if ! ps -p ${pid} 1>&2 >/dev/null; then
 	echo "server1 log:"

buildscripts/rewrite-old-new.sh

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

cmd/acl-handlers.go

@@ -80,7 +80,7 @@ func (api objectAPIHandlers) PutBucketACLHandler(w http.ResponseWriter, r *http.
 	}
 
 	// Before proceeding validate if bucket exists.
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -142,7 +142,7 @@ func (api objectAPIHandlers) GetBucketACLHandler(w http.ResponseWriter, r *http.
 	}
 
 	// Before proceeding validate if bucket exists.
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return

cmd/admin-bucket-handlers.go

@@ -69,7 +69,7 @@ func (a adminAPIHandlers) PutBucketQuotaConfigHandler(w http.ResponseWriter, r *
 	vars := mux.Vars(r)
 	bucket := pathClean(vars["bucket"])
 
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -131,7 +131,7 @@ func (a adminAPIHandlers) GetBucketQuotaConfigHandler(w http.ResponseWriter, r *
 	vars := mux.Vars(r)
 	bucket := pathClean(vars["bucket"])
 
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -173,7 +173,7 @@ func (a adminAPIHandlers) SetRemoteTargetHandler(w http.ResponseWriter, r *http.
 	}
 
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -249,7 +249,7 @@ func (a adminAPIHandlers) SetRemoteTargetHandler(w http.ResponseWriter, r *http.
 	}
 	if err = globalBucketTargetSys.SetTarget(ctx, bucket, &target, update); err != nil {
 		switch err.(type) {
-		case BucketRemoteConnectionErr:
+		case RemoteTargetConnectionErr:
 			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrReplicationRemoteConnectionError, err), r.URL)
 		default:
 			writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
@@ -301,7 +301,7 @@ func (a adminAPIHandlers) ListRemoteTargetsHandler(w http.ResponseWriter, r *htt
 	}
 	if bucket != "" {
 		// Check if bucket exists.
-		if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+		if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 			writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 			return
 		}
@@ -340,7 +340,7 @@ func (a adminAPIHandlers) RemoveRemoteTargetHandler(w http.ResponseWriter, r *ht
 	}
 
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -374,7 +374,7 @@ func (a adminAPIHandlers) ExportBucketMetadataHandler(w http.ResponseWriter, r *
 	defer logger.AuditLog(ctx, w, r, mustGetClaimsFromToken(r))
 
 	bucket := pathClean(r.Form.Get("bucket"))
-	if !globalIsErasure {
+	if globalIsGateway {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
 		return
 	}
@@ -390,13 +390,13 @@ func (a adminAPIHandlers) ExportBucketMetadataHandler(w http.ResponseWriter, r *
 	)
 	if bucket != "" {
 		// Check if bucket exists.
-		if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+		if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 			writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
 			return
 		}
 		buckets = append(buckets, BucketInfo{Name: bucket})
 	} else {
-		buckets, err = objectAPI.ListBuckets(ctx)
+		buckets, err = objectAPI.ListBuckets(ctx, BucketOptions{})
 		if err != nil {
 			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 			return
@@ -619,6 +619,39 @@ func (a adminAPIHandlers) ExportBucketMetadataHandler(w http.ResponseWriter, r *
 	}
 }
 
+type importMetaReport struct {
+	madmin.BucketMetaImportErrs
+}
+
+func (i *importMetaReport) SetStatus(bucket, fname string, err error) {
+	st := i.Buckets[bucket]
+	var errMsg string
+	if err != nil {
+		errMsg = err.Error()
+	}
+	switch fname {
+	case bucketPolicyConfig:
+		st.Policy = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketNotificationConfig:
+		st.Notification = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketLifecycleConfig:
+		st.Lifecycle = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketSSEConfig:
+		st.SSEConfig = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketTaggingConfig:
+		st.Tagging = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketQuotaConfigFile:
+		st.Quota = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case objectLockConfig:
+		st.ObjectLock = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	case bucketVersioningConfig:
+		st.Versioning = madmin.MetaStatus{IsSet: true, Err: errMsg}
+	default:
+		st.Err = errMsg
+	}
+	i.Buckets[bucket] = st
+}
+
 // ImportBucketMetadataHandler - imports all bucket metadata from a zipped file and overwrite bucket metadata config
 // There are some caveats regarding the following:
 // 1. object lock config - object lock should have been specified at time of bucket creation. Only default retention settings are imported here.
@@ -628,9 +661,8 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 	ctx := newContext(r, w, "ImportBucketMetadata")
 
 	defer logger.AuditLog(ctx, w, r, mustGetClaimsFromToken(r))
-	bucket := pathClean(r.Form.Get("bucket"))
 
-	if !globalIsErasure {
+	if globalIsGateway {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
 		return
 	}
@@ -651,47 +683,46 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 		return
 	}
 	bucketMap := make(map[string]struct{}, 1)
-
+	rpt := importMetaReport{
+		madmin.BucketMetaImportErrs{
+			Buckets: make(map[string]madmin.BucketStatus, len(zr.File)),
+		},
+	}
 	// import object lock config if any - order of import matters here.
 	for _, file := range zr.File {
 		slc := strings.Split(file.Name, slashSeparator)
 		if len(slc) != 2 { // expecting bucket/configfile in the zipfile
-			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-			return
-		}
-		b, fileName := slc[0], slc[1]
-		if bucket == "" { // use bucket requested in query parameters if specified. Otherwise default bucket name to directory name within zip
-			bucket = b
+			rpt.SetStatus(file.Name, "", fmt.Errorf("malformed zip - expecting format bucket/<config.json>"))
+			continue
 		}
+		bucket, fileName := slc[0], slc[1]
 		switch fileName {
 		case objectLockConfig:
 			reader, err := file.Open()
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			config, err := objectlock.ParseObjectLockConfig(reader)
 			if err != nil {
-				apiErr := errorCodes.ToAPIErr(ErrMalformedXML)
-				apiErr.Description = err.Error()
-				writeErrorResponse(ctx, w, apiErr, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}
 
 			configData, err := xml.Marshal(config)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			if _, ok := bucketMap[bucket]; !ok {
-				opts := BucketOptions{
+				opts := MakeBucketOptions{
 					LockEnabled: config.ObjectLockEnabled == "Enabled",
 				}
 				err = objectAPI.MakeBucketWithLocation(ctx, bucket, opts)
 				if err != nil {
 					if _, ok := err.(BucketExists); !ok {
-						writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-						return
+						rpt.SetStatus(bucket, fileName, err)
+						continue
 					}
 				}
 				bucketMap[bucket] = struct{}{}
@@ -699,15 +730,16 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 
 			// Deny object locking configuration settings on existing buckets without object lock enabled.
 			if _, _, err = globalBucketMetadataSys.GetObjectLockConfig(bucket); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, objectLockConfig, configData)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
 
 			// Call site replication hook.
 			//
@@ -720,8 +752,8 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				ObjectLockConfig: &cfgStr,
 				UpdatedAt:        updatedAt,
 			}); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 		}
 	}
@@ -730,162 +762,145 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 	for _, file := range zr.File {
 		slc := strings.Split(file.Name, slashSeparator)
 		if len(slc) != 2 { // expecting bucket/configfile in the zipfile
-			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-			return
-		}
-		b, fileName := slc[0], slc[1]
-		if bucket == "" { // use bucket requested in query parameters if specified. Otherwise default bucket name to directory name within zip
-			bucket = b
+			rpt.SetStatus(file.Name, "", fmt.Errorf("malformed zip - expecting format bucket/<config.json>"))
+			continue
 		}
+		bucket, fileName := slc[0], slc[1]
 		switch fileName {
 		case bucketVersioningConfig:
 			reader, err := file.Open()
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			v, err := versioning.ParseConfig(io.LimitReader(reader, maxBucketVersioningConfigSize))
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			if _, ok := bucketMap[bucket]; !ok {
-				err = objectAPI.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
-				if err != nil {
+				if err = objectAPI.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{}); err != nil {
 					if _, ok := err.(BucketExists); !ok {
-						writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-						return
+						rpt.SetStatus(bucket, fileName, err)
+						continue
 					}
 				}
 				bucketMap[bucket] = struct{}{}
 			}
 
 			if globalSiteReplicationSys.isEnabled() && v.Suspended() {
-				writeErrorResponse(ctx, w, APIError{
-					Code:           "InvalidBucketState",
-					Description:    "Cluster replication is enabled for this site, so the versioning state cannot be changed.",
-					HTTPStatusCode: http.StatusConflict,
-				}, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("Cluster replication is enabled for this site, so the versioning state cannot be suspended."))
+				continue
 			}
 
 			if rcfg, _ := globalBucketObjectLockSys.Get(bucket); rcfg.LockEnabled && v.Suspended() {
-				writeErrorResponse(ctx, w, APIError{
-					Code:           "InvalidBucketState",
-					Description:    "An Object Lock configuration is present on this bucket, so the versioning state cannot be changed.",
-					HTTPStatusCode: http.StatusConflict,
-				}, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("An Object Lock configuration is present on this bucket, so the versioning state cannot be suspended."))
+				continue
 			}
 			if _, err := getReplicationConfig(ctx, bucket); err == nil && v.Suspended() {
-				writeErrorResponse(ctx, w, APIError{
-					Code:           "InvalidBucketState",
-					Description:    "A replication configuration is present on this bucket, so the versioning state cannot be changed.",
-					HTTPStatusCode: http.StatusConflict,
-				}, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("A replication configuration is present on this bucket, so the versioning state cannot be suspended."))
+				continue
 			}
 
 			configData, err := xml.Marshal(v)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}
 
 			if _, err = globalBucketMetadataSys.Update(ctx, bucket, bucketVersioningConfig, configData); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
 		}
 	}
 
 	for _, file := range zr.File {
 		reader, err := file.Open()
 		if err != nil {
-			writeErrorResponse(ctx, w, importError(ctx, err, file.Name, ""), r.URL)
-			return
+			rpt.SetStatus(file.Name, "", err)
+			continue
 		}
 		sz := file.FileInfo().Size()
 		slc := strings.Split(file.Name, slashSeparator)
 		if len(slc) != 2 { // expecting bucket/configfile in the zipfile
-			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-			return
-		}
-		b, fileName := slc[0], slc[1]
-		if bucket == "" { // use bucket requested in query parameters if specified. Otherwise default bucket name to directory name within zip
-			bucket = b
+			rpt.SetStatus(file.Name, "", fmt.Errorf("malformed zip - expecting format bucket/<config.json>"))
+			continue
 		}
+		bucket, fileName := slc[0], slc[1]
 		// create bucket if it does not exist yet.
 		if _, ok := bucketMap[bucket]; !ok {
-			err = objectAPI.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+			err = objectAPI.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 			if err != nil {
 				if _, ok := err.(BucketExists); !ok {
-					writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-					return
+					rpt.SetStatus(bucket, "", err)
+					continue
 				}
 			}
 			bucketMap[bucket] = struct{}{}
 		}
+		if _, ok := bucketMap[bucket]; !ok {
+			continue
+		}
 		switch fileName {
 		case bucketNotificationConfig:
 			config, err := event.ParseConfig(io.LimitReader(reader, sz), globalSite.Region, globalNotificationSys.targetList)
 			if err != nil {
-				apiErr := errorCodes.ToAPIErr(ErrMalformedXML)
-				if event.IsEventError(err) {
-					apiErr = importError(ctx, err, file.Name, bucket)
-				}
-				writeErrorResponse(ctx, w, apiErr, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}
 
 			configData, err := xml.Marshal(config)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			if _, err = globalBucketMetadataSys.Update(ctx, bucket, bucketNotificationConfig, configData); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 			rulesMap := config.ToRulesMap()
 			globalNotificationSys.AddRulesMap(bucket, rulesMap)
+			rpt.SetStatus(bucket, fileName, nil)
 		case bucketPolicyConfig:
 			// Error out if Content-Length is beyond allowed size.
 			if sz > maxBucketPolicySize {
-				writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrPolicyTooLarge), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf(ErrPolicyTooLarge.String()))
+				continue
 			}
 
 			bucketPolicyBytes, err := ioutil.ReadAll(io.LimitReader(reader, sz))
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			bucketPolicy, err := policy.ParseConfig(bytes.NewReader(bucketPolicyBytes), bucket)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			// Version in policy must not be empty
 			if bucketPolicy.Version == "" {
-				writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrMalformedPolicy), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf(ErrMalformedPolicy.String()))
+				continue
 			}
 
 			configData, err := json.Marshal(bucketPolicy)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, bucketPolicyConfig, configData)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
 			// Call site replication hook.
 			if err = globalSiteReplicationSys.BucketMetaHook(ctx, madmin.SRBucketMeta{
 				Type:      madmin.SRBucketMetaTypePolicy,
@@ -893,56 +908,51 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				Policy:    bucketPolicyBytes,
 				UpdatedAt: updatedAt,
 			}); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
-
 		case bucketLifecycleConfig:
 			bucketLifecycle, err := lifecycle.ParseLifecycleConfig(io.LimitReader(reader, sz))
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			// Validate the received bucket policy document
 			if err = bucketLifecycle.Validate(); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			// Validate the transition storage ARNs
 			if err = validateTransitionTier(bucketLifecycle); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			configData, err := xml.Marshal(bucketLifecycle)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			if _, err = globalBucketMetadataSys.Update(ctx, bucket, bucketLifecycleConfig, configData); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
 		case bucketSSEConfig:
 			// Parse bucket encryption xml
 			encConfig, err := validateBucketSSEConfig(io.LimitReader(reader, maxBucketSSEConfigSize))
 			if err != nil {
-				apiErr := APIError{
-					Code:           "MalformedXML",
-					Description:    fmt.Sprintf("%s (%s)", errorCodes[ErrMalformedXML].Description, err),
-					HTTPStatusCode: errorCodes[ErrMalformedXML].HTTPStatusCode,
-				}
-				writeErrorResponse(ctx, w, apiErr, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}
 
 			// Return error if KMS is not initialized
 			if GlobalKMS == nil {
-				writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrKMSNotConfigured), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s", errorCodes[ErrKMSNotConfigured].Description))
+				continue
 			}
 			kmsKey := encConfig.KeyID()
 			if kmsKey != "" {
@@ -950,26 +960,27 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				_, err := GlobalKMS.GenerateKey(ctx, kmsKey, kmsContext)
 				if err != nil {
 					if errors.Is(err, kes.ErrKeyNotFound) {
-						writeErrorResponse(ctx, w, importError(ctx, errKMSKeyNotFound, file.Name, bucket), r.URL)
-						return
+						rpt.SetStatus(bucket, fileName, errKMSKeyNotFound)
+						continue
 					}
-					writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-					return
+					rpt.SetStatus(bucket, fileName, err)
+					continue
 				}
 			}
 
 			configData, err := xml.Marshal(encConfig)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			// Store the bucket encryption configuration in the object layer
 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, bucketSSEConfig, configData)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
 
 			// Call site replication hook.
 			//
@@ -982,29 +993,30 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				SSEConfig: &cfgStr,
 				UpdatedAt: updatedAt,
 			}); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 		case bucketTaggingConfig:
 			tags, err := tags.ParseBucketXML(io.LimitReader(reader, sz))
 			if err != nil {
-				apiErr := errorCodes.ToAPIErrWithErr(ErrMalformedXML, fmt.Errorf("error importing %s with %w", file.Name, err))
-				writeErrorResponse(ctx, w, apiErr, r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("%s (%s)", errorCodes[ErrMalformedXML].Description, err))
+				continue
 			}
 
 			configData, err := xml.Marshal(tags)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, bucketTaggingConfig, configData)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
+
 			// Call site replication hook.
 			//
 			// We encode the xml bytes as base64 to ensure there are no encoding
@@ -1016,32 +1028,33 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 				Tags:      &cfgStr,
 				UpdatedAt: updatedAt,
 			}); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 		case bucketQuotaConfigFile:
 			data, err := ioutil.ReadAll(reader)
 			if err != nil {
-				writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			quotaConfig, err := parseBucketQuota(bucket, data)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 
 			if quotaConfig.Type == "fifo" {
-				writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrInvalidRequest), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, fmt.Errorf("Detected older 'fifo' quota config, 'fifo' feature is removed and not supported anymore"))
+				continue
 			}
 
 			updatedAt, err := globalBucketMetadataSys.Update(ctx, bucket, bucketQuotaConfigFile, data)
 			if err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
+			rpt.SetStatus(bucket, fileName, nil)
 
 			bucketMeta := madmin.SRBucketMeta{
 				Type:      madmin.SRBucketMetaTypeQuotaConfig,
@@ -1055,11 +1068,19 @@ func (a adminAPIHandlers) ImportBucketMetadataHandler(w http.ResponseWriter, r *
 
 			// Call site replication hook.
 			if err = globalSiteReplicationSys.BucketMetaHook(ctx, bucketMeta); err != nil {
-				writeErrorResponse(ctx, w, importError(ctx, err, file.Name, bucket), r.URL)
-				return
+				rpt.SetStatus(bucket, fileName, err)
+				continue
 			}
 		}
 	}
+
+	rptData, err := json.Marshal(rpt.BucketMetaImportErrs)
+	if err != nil {
+		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
+		return
+	}
+
+	writeSuccessResponseJSON(w, rptData)
 }
 
 // ReplicationDiffHandler - POST returns info on unreplicated versions for a remote target ARN
@@ -1089,7 +1110,7 @@ func (a adminAPIHandlers) ReplicationDiffHandler(w http.ResponseWriter, r *http.
 	}
 
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}

cmd/admin-handlers-config-kv.go

@@ -189,15 +189,20 @@ func (a adminAPIHandlers) GetConfigKVHandler(w http.ResponseWriter, r *http.Requ
 
 	cfg := globalServerConfig.Clone()
 	vars := mux.Vars(r)
-	buf := &bytes.Buffer{}
-	cw := config.NewConfigWriteTo(cfg, vars["key"])
-	if _, err := cw.WriteTo(buf); err != nil {
+	subSys := vars["key"]
+	subSysConfigs, err := cfg.GetSubsysInfo(subSys)
+	if err != nil {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
 	}
 
+	var s strings.Builder
+	for _, subSysConfig := range subSysConfigs {
+		subSysConfig.AddString(&s, false)
+	}
+
 	password := cred.SecretKey
-	econfigData, err := madmin.EncryptData(password, buf.Bytes())
+	econfigData, err := madmin.EncryptData(password, []byte(s.String()))
 	if err != nil {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
@@ -427,42 +432,30 @@ func (a adminAPIHandlers) GetConfigHandler(w http.ResponseWriter, r *http.Reques
 		count += len(cfg[hkv.Key])
 	}
 	for _, hkv := range hkvs {
-		v := cfg[hkv.Key]
-		for target, kv := range v {
-			off := kv.Get(config.Enable) == config.EnableOff
+		// We ignore the error below, as we cannot get one.
+		cfgSubsysItems, _ := cfg.GetSubsysInfo(hkv.Key)
+
+		for _, item := range cfgSubsysItems {
+			off := item.Config.Get(config.Enable) == config.EnableOff
 			switch hkv.Key {
 			case config.EtcdSubSys:
-				off = !etcd.Enabled(kv)
+				off = !etcd.Enabled(item.Config)
 			case config.CacheSubSys:
-				off = !cache.Enabled(kv)
+				off = !cache.Enabled(item.Config)
 			case config.StorageClassSubSys:
-				off = !storageclass.Enabled(kv)
+				off = !storageclass.Enabled(item.Config)
 			case config.PolicyPluginSubSys:
-				off = !polplugin.Enabled(kv)
+				off = !polplugin.Enabled(item.Config)
 			case config.IdentityOpenIDSubSys:
-				off = !openid.Enabled(kv)
+				off = !openid.Enabled(item.Config)
 			case config.IdentityLDAPSubSys:
-				off = !xldap.Enabled(kv)
+				off = !xldap.Enabled(item.Config)
 			case config.IdentityTLSSubSys:
 				off = !globalSTSTLSConfig.Enabled
 			case config.IdentityPluginSubSys:
-				off = !idplugin.Enabled(kv)
-			}
-			if off {
-				s.WriteString(config.KvComment)
-				s.WriteString(config.KvSpaceSeparator)
-			}
-			s.WriteString(hkv.Key)
-			if target != config.Default {
-				s.WriteString(config.SubSystemSeparator)
-				s.WriteString(target)
-			}
-			s.WriteString(config.KvSpaceSeparator)
-			s.WriteString(kv.String())
-			count--
-			if count > 0 {
-				s.WriteString(config.KvNewline)
+				off = !idplugin.Enabled(item.Config)
 			}
+			item.AddString(&s, off)
 		}
 	}
 

cmd/admin-handlers-site-replication.go

@@ -24,6 +24,8 @@ import (
 	"io"
 	"io/ioutil"
 	"net/http"
+	"strings"
+	"time"
 
 	"github.com/gorilla/mux"
 	"github.com/minio/madmin-go"
@@ -116,19 +118,38 @@ func (a adminAPIHandlers) SRPeerBucketOps(w http.ResponseWriter, r *http.Request
 		_, isLockEnabled := r.Form["lockEnabled"]
 		_, isVersioningEnabled := r.Form["versioningEnabled"]
 		_, isForceCreate := r.Form["forceCreate"]
-		opts := BucketOptions{
+		createdAtStr := strings.TrimSpace(r.Form.Get("createdAt"))
+		createdAt, cerr := time.Parse(time.RFC3339Nano, createdAtStr)
+		if cerr != nil {
+			createdAt = timeSentinel
+		}
+
+		opts := MakeBucketOptions{
 			Location:          r.Form.Get("location"),
 			LockEnabled:       isLockEnabled,
 			VersioningEnabled: isVersioningEnabled,
 			ForceCreate:       isForceCreate,
+			CreatedAt:         createdAt,
 		}
 		err = globalSiteReplicationSys.PeerBucketMakeWithVersioningHandler(ctx, bucket, opts)
 	case madmin.ConfigureReplBktOp:
 		err = globalSiteReplicationSys.PeerBucketConfigureReplHandler(ctx, bucket)
 	case madmin.DeleteBucketBktOp:
-		err = globalSiteReplicationSys.PeerBucketDeleteHandler(ctx, bucket, false)
+		_, noRecreate := r.Form["noRecreate"]
+		err = globalSiteReplicationSys.PeerBucketDeleteHandler(ctx, bucket, DeleteBucketOptions{
+			Force:      false,
+			NoRecreate: noRecreate,
+			SRDeleteOp: getSRBucketDeleteOp(true),
+		})
 	case madmin.ForceDeleteBucketBktOp:
-		err = globalSiteReplicationSys.PeerBucketDeleteHandler(ctx, bucket, true)
+		_, noRecreate := r.Form["noRecreate"]
+		err = globalSiteReplicationSys.PeerBucketDeleteHandler(ctx, bucket, DeleteBucketOptions{
+			Force:      true,
+			NoRecreate: noRecreate,
+			SRDeleteOp: getSRBucketDeleteOp(true),
+		})
+	case madmin.PurgeDeletedBucketOp:
+		globalSiteReplicationSys.purgeDeletedBucket(ctx, objectAPI, bucket)
 	}
 	if err != nil {
 		logger.LogIf(ctx, err)
@@ -436,6 +457,7 @@ func getSRStatusOptions(r *http.Request) (opts madmin.SRStatusOptions) {
 	opts.Users = q.Get("users") == "true"
 	opts.Entity = madmin.GetSREntityType(q.Get("entity"))
 	opts.EntityValue = q.Get("entityvalue")
+	opts.ShowDeleted = q.Get("showDeleted") == "true"
 	return
 }
 

cmd/admin-handlers-users.go

@@ -1177,7 +1177,7 @@ func (a adminAPIHandlers) AccountInfoHandler(w http.ResponseWriter, r *http.Requ
 			return buckets[i].Name < buckets[j].Name
 		})
 	} else {
-		buckets, err = objectAPI.ListBuckets(ctx)
+		buckets, err = objectAPI.ListBuckets(ctx, BucketOptions{})
 		if err != nil {
 			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 			return
@@ -1189,17 +1189,32 @@ func (a adminAPIHandlers) AccountInfoHandler(w http.ResponseWriter, r *http.Requ
 		// For derived credentials, check the parent user's permissions.
 		accountName = cred.ParentUser
 	}
-	policies, err := globalIAMSys.PolicyDBGet(accountName, false, cred.Groups...)
-	if err != nil {
-		logger.LogIf(ctx, err)
-		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
-		return
-	}
 
-	buf, err := json.MarshalIndent(globalIAMSys.GetCombinedPolicy(policies...), "", " ")
-	if err != nil {
-		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
-		return
+	var buf []byte
+	if accountName == globalActiveCred.AccessKey {
+		for _, policy := range iampolicy.DefaultPolicies {
+			if policy.Name == "consoleAdmin" {
+				buf, err = json.MarshalIndent(policy.Definition, "", " ")
+				if err != nil {
+					writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+					return
+				}
+				break
+			}
+		}
+	} else {
+		policies, err := globalIAMSys.PolicyDBGet(accountName, false, cred.Groups...)
+		if err != nil {
+			logger.LogIf(ctx, err)
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+
+		buf, err = json.MarshalIndent(globalIAMSys.GetCombinedPolicy(policies...), "", " ")
+		if err != nil {
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
 	}
 
 	acctInfo := madmin.AccountInfo{

cmd/admin-handlers.go

@@ -71,16 +71,6 @@ const (
 	mgmtForceStop   = "forceStop"
 )
 
-func updateServer(u *url.URL, sha256Sum []byte, lrTime time.Time, releaseInfo string, mode string) (us madmin.ServerUpdateStatus, err error) {
-	if err = doUpdate(u, lrTime, sha256Sum, releaseInfo, mode); err != nil {
-		return us, err
-	}
-
-	us.CurrentVersion = Version
-	us.UpdatedVersion = lrTime.Format(minioReleaseTagTimeLayout)
-	return us, nil
-}
-
 // ServerUpdateHandler - POST /minio/admin/v3/update?updateURL={updateURL}
 // ----------
 // updates all minio servers and restarts them gracefully.
@@ -152,7 +142,16 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 		return
 	}
 
-	for _, nerr := range globalNotificationSys.ServerUpdate(ctx, u, sha256Sum, lrTime, releaseInfo) {
+	// Download Binary Once
+	reader, err := downloadBinary(u, mode)
+	if err != nil {
+		logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	// Push binary to other servers
+	for _, nerr := range globalNotificationSys.VerifyBinary(ctx, u, sha256Sum, releaseInfo, reader) {
 		if nerr.Err != nil {
 			err := AdminError{
 				Code:       AdminUpdateApplyFailure,
@@ -166,13 +165,39 @@ func (a adminAPIHandlers) ServerUpdateHandler(w http.ResponseWriter, r *http.Req
 		}
 	}
 
-	updateStatus, err := updateServer(u, sha256Sum, lrTime, releaseInfo, mode)
+	err = verifyBinary(u, sha256Sum, releaseInfo, mode, reader)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
 		return
 	}
 
+	for _, nerr := range globalNotificationSys.CommitBinary(ctx) {
+		if nerr.Err != nil {
+			err := AdminError{
+				Code:       AdminUpdateApplyFailure,
+				Message:    nerr.Err.Error(),
+				StatusCode: http.StatusInternalServerError,
+			}
+			logger.GetReqInfo(ctx).SetTags("peerAddress", nerr.Host.String())
+			logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
+			writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+			return
+		}
+	}
+
+	err = commitBinary()
+	if err != nil {
+		logger.LogIf(ctx, fmt.Errorf("server update failed with %w", err))
+		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
+		return
+	}
+
+	updateStatus := madmin.ServerUpdateStatus{
+		CurrentVersion: Version,
+		UpdatedVersion: lrTime.Format(minioReleaseTagTimeLayout),
+	}
+
 	// Marshal API response
 	jsonBytes, err := json.Marshal(updateStatus)
 	if err != nil {
@@ -373,8 +398,21 @@ func (a adminAPIHandlers) MetricsHandler(w http.ResponseWriter, r *http.Request)
 	} else {
 		types = madmin.MetricsAll
 	}
+
+	disks := strings.Split(r.Form.Get("disks"), ",")
+	byDisk := strings.EqualFold(r.Form.Get("by-disk"), "true")
+	var diskMap map[string]struct{}
+	if len(disks) > 0 && disks[0] != "" {
+		diskMap = make(map[string]struct{}, len(disks))
+		for _, k := range disks {
+			if k != "" {
+				diskMap[k] = struct{}{}
+			}
+		}
+	}
+
 	hosts := strings.Split(r.Form.Get("hosts"), ",")
-	byhost := strings.EqualFold(r.Form.Get("by-host"), "true")
+	byHost := strings.EqualFold(r.Form.Get("by-host"), "true")
 	var hostMap map[string]struct{}
 	if len(hosts) > 0 && hosts[0] != "" {
 		hostMap = make(map[string]struct{}, len(hosts))
@@ -384,6 +422,7 @@ func (a adminAPIHandlers) MetricsHandler(w http.ResponseWriter, r *http.Request)
 			}
 		}
 	}
+
 	done := ctx.Done()
 	ticker := time.NewTicker(interval)
 	defer ticker.Stop()
@@ -391,17 +430,20 @@ func (a adminAPIHandlers) MetricsHandler(w http.ResponseWriter, r *http.Request)
 
 	for n > 0 {
 		var m madmin.RealtimeMetrics
-		mLocal := collectLocalMetrics(types, hostMap)
+		mLocal := collectLocalMetrics(types, hostMap, diskMap)
 		m.Merge(&mLocal)
 
 		// Allow half the interval for collecting remote...
 		cctx, cancel := context.WithTimeout(ctx, interval/2)
-		mRemote := collectRemoteMetrics(cctx, types, hostMap)
+		mRemote := collectRemoteMetrics(cctx, types, hostMap, diskMap)
 		cancel()
 		m.Merge(&mRemote)
-		if !byhost {
+		if !byHost {
 			m.ByHost = nil
 		}
+		if !byDisk {
+			m.ByDisk = nil
+		}
 
 		m.Final = n <= 1
 		// Marshal API response
@@ -1190,16 +1232,14 @@ func (a adminAPIHandlers) ObjectSpeedTestHandler(w http.ResponseWriter, r *http.
 		concurrent = 32
 	}
 
-	if runtime.GOMAXPROCS(0) < concurrent {
-		concurrent = runtime.GOMAXPROCS(0)
-	}
-
 	duration, err := time.ParseDuration(durationStr)
 	if err != nil {
 		duration = time.Second * 10
 	}
 
-	sufficientCapacity, canAutotune, capacityErrMsg := validateObjPerfOptions(ctx, objectAPI, concurrent, size, autotune)
+	storageInfo, _ := objectAPI.StorageInfo(ctx)
+
+	sufficientCapacity, canAutotune, capacityErrMsg := validateObjPerfOptions(ctx, storageInfo, concurrent, size, autotune)
 	if !sufficientCapacity {
 		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, AdminError{
 			Code:       "XMinioSpeedtestInsufficientCapacity",
@@ -1249,14 +1289,24 @@ func (a adminAPIHandlers) ObjectSpeedTestHandler(w http.ResponseWriter, r *http.
 		storageClass:     storageClass,
 		bucketName:       customBucket,
 	})
+	var prevResult madmin.SpeedTestResult
 	for {
 		select {
 		case <-ctx.Done():
 			return
 		case <-keepAliveTicker.C:
-			// Write a blank entry to prevent client from disconnecting
-			if err := enc.Encode(madmin.SpeedTestResult{}); err != nil {
-				return
+			// if previous result is set keep writing the
+			// previous result back to the client
+			if prevResult.Version != "" {
+				if err := enc.Encode(prevResult); err != nil {
+					return
+				}
+			} else {
+				// first result is not yet obtained, keep writing
+				// empty entry to prevent client from disconnecting.
+				if err := enc.Encode(madmin.SpeedTestResult{}); err != nil {
+					return
+				}
 			}
 			w.(http.Flusher).Flush()
 		case result, ok := <-ch:
@@ -1266,13 +1316,14 @@ func (a adminAPIHandlers) ObjectSpeedTestHandler(w http.ResponseWriter, r *http.
 			if err := enc.Encode(result); err != nil {
 				return
 			}
+			prevResult = result
 			w.(http.Flusher).Flush()
 		}
 	}
 }
 
 func makeObjectPerfBucket(ctx context.Context, objectAPI ObjectLayer, bucketName string) (bucketExists bool, err error) {
-	if err = objectAPI.MakeBucketWithLocation(ctx, bucketName, BucketOptions{}); err != nil {
+	if err = objectAPI.MakeBucketWithLocation(ctx, bucketName, MakeBucketOptions{}); err != nil {
 		if _, ok := err.(BucketExists); !ok {
 			// Only BucketExists error can be ignored.
 			return false, err
@@ -1286,11 +1337,11 @@ func deleteObjectPerfBucket(objectAPI ObjectLayer) {
 	objectAPI.DeleteBucket(context.Background(), globalObjectPerfBucket, DeleteBucketOptions{
 		Force:      true,
 		NoRecreate: true,
+		SRDeleteOp: getSRBucketDeleteOp(globalSiteReplicationSys.isEnabled()),
 	})
 }
 
-func validateObjPerfOptions(ctx context.Context, objectAPI ObjectLayer, concurrent int, size int, autotune bool) (sufficientCapacity bool, canAutotune bool, capacityErrMsg string) {
-	storageInfo, _ := objectAPI.StorageInfo(ctx)
+func validateObjPerfOptions(ctx context.Context, storageInfo madmin.StorageInfo, concurrent int, size int, autotune bool) (bool, bool, string) {
 	capacityNeeded := uint64(concurrent * size)
 	capacity := GetTotalUsableCapacityFree(storageInfo.Disks, storageInfo)
 
@@ -2199,15 +2250,13 @@ func (a adminAPIHandlers) HealthInfoHandler(w http.ResponseWriter, r *http.Reque
 	getAndWriteObjPerfInfo := func() {
 		if query.Get(string(madmin.HealthDataTypePerfObj)) == "true" {
 			concurrent := 32
-			if runtime.GOMAXPROCS(0) < concurrent {
-				concurrent = runtime.GOMAXPROCS(0)
-			}
+
+			storageInfo, _ := objectAPI.StorageInfo(ctx)
 
 			size := 64 * humanize.MiByte
 			autotune := true
 
-			sufficientCapacity, canAutotune, capacityErrMsg := validateObjPerfOptions(ctx, objectAPI, concurrent, size, autotune)
-
+			sufficientCapacity, canAutotune, capacityErrMsg := validateObjPerfOptions(ctx, storageInfo, concurrent, size, autotune)
 			if !sufficientCapacity {
 				healthInfo.Perf.Error = capacityErrMsg
 				partialWrite(healthInfo)
@@ -2636,6 +2685,100 @@ func checkConnection(endpointStr string, timeout time.Duration) error {
 	return nil
 }
 
+type clusterInfo struct {
+	DeploymentID string `json:"deployment_id"`
+	ClusterName  string `json:"cluster_name"`
+	UsedCapacity uint64 `json:"used_capacity"`
+	Info         struct {
+		MinioVersion    string `json:"minio_version"`
+		PoolsCount      int    `json:"pools_count"`
+		ServersCount    int    `json:"servers_count"`
+		DrivesCount     int    `json:"drives_count"`
+		BucketsCount    uint64 `json:"buckets_count"`
+		ObjectsCount    uint64 `json:"objects_count"`
+		TotalDriveSpace uint64 `json:"total_drive_space"`
+		UsedDriveSpace  uint64 `json:"used_drive_space"`
+	} `json:"info"`
+}
+
+func embedFileInZip(zipWriter *zip.Writer, name string, data []byte) error {
+	// Send profiling data to zip as file
+	header, zerr := zip.FileInfoHeader(dummyFileInfo{
+		name:    name,
+		size:    int64(len(data)),
+		mode:    0o600,
+		modTime: UTCNow(),
+		isDir:   false,
+		sys:     nil,
+	})
+	if zerr != nil {
+		return zerr
+	}
+	header.Method = zip.Deflate
+	zwriter, zerr := zipWriter.CreateHeader(header)
+	if zerr != nil {
+		return zerr
+	}
+	_, err := io.Copy(zwriter, bytes.NewReader(data))
+	return err
+}
+
+// appendClusterMetaInfoToZip gets information of the current cluster and embedded
+// it in the passed zipwriter, This is not a critical function and it is allowed
+// to fail with a ten seconds timeout.
+func appendClusterMetaInfoToZip(ctx context.Context, zipWriter *zip.Writer) {
+	objectAPI := newObjectLayerFn()
+	if objectAPI == nil {
+		return
+	}
+
+	// Add a ten seconds timeout because getting profiling data
+	// is critical for debugging, in contrary to getting cluster info
+	ctx, cancel := context.WithTimeout(ctx, 10*time.Second)
+	defer cancel()
+
+	resultCh := make(chan clusterInfo)
+	go func() {
+		ci := clusterInfo{}
+		ci.Info.PoolsCount = len(globalEndpoints)
+		ci.Info.ServersCount = len(globalEndpoints.Hostnames())
+		ci.Info.MinioVersion = Version
+
+		si, _ := objectAPI.StorageInfo(ctx)
+
+		ci.Info.DrivesCount = len(si.Disks)
+		for _, disk := range si.Disks {
+			ci.Info.TotalDriveSpace += disk.TotalSpace
+			ci.Info.UsedDriveSpace += disk.UsedSpace
+		}
+
+		dataUsageInfo, _ := loadDataUsageFromBackend(ctx, objectAPI)
+
+		ci.UsedCapacity = dataUsageInfo.ObjectsTotalSize
+		ci.Info.BucketsCount = dataUsageInfo.BucketsCount
+		ci.Info.ObjectsCount = dataUsageInfo.ObjectsTotalCount
+
+		ci.DeploymentID = globalDeploymentID
+		ci.ClusterName = fmt.Sprintf("%d-servers-%d-disks-%s", ci.Info.ServersCount, ci.Info.DrivesCount, ci.Info.MinioVersion)
+		resultCh <- ci
+	}()
+
+	select {
+	case <-ctx.Done():
+		return
+	case ci := <-resultCh:
+		out, err := json.MarshalIndent(ci, "", "   ")
+		if err != nil {
+			logger.LogIf(ctx, err)
+			return
+		}
+		err = embedFileInZip(zipWriter, "cluster.info", out)
+		if err != nil {
+			logger.LogIf(ctx, err)
+		}
+	}
+}
+
 // getRawDataer provides an interface for getting raw FS files.
 type getRawDataer interface {
 	GetRawData(ctx context.Context, volume, file string, fn func(r io.Reader, host string, disk string, filename string, info StatInfo) error) error
@@ -2655,7 +2798,8 @@ func (a adminAPIHandlers) InspectDataHandler(w http.ResponseWriter, r *http.Requ
 	}
 	defer logger.AuditLog(ctx, w, r, mustGetClaimsFromToken(r))
 
-	o, ok := newObjectLayerFn().(getRawDataer)
+	objLayer := newObjectLayerFn()
+	o, ok := objLayer.(getRawDataer)
 	if !ok {
 		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
 		return
@@ -2778,6 +2922,8 @@ func (a adminAPIHandlers) InspectDataHandler(w http.ResponseWriter, r *http.Requ
 	}); err != nil {
 		logger.LogIf(ctx, err)
 	}
+
+	appendClusterMetaInfoToZip(ctx, zipWriter)
 }
 
 func createHostAnonymizerForFSMode() map[string]string {

cmd/admin-heal-ops.go

@@ -822,7 +822,7 @@ func (h *healSequence) healBuckets(objAPI ObjectLayer, bucketsOnly bool) error {
 		return h.healBucket(objAPI, h.bucket, bucketsOnly)
 	}
 
-	buckets, err := objAPI.ListBuckets(h.ctx)
+	buckets, err := objAPI.ListBuckets(h.ctx, BucketOptions{})
 	if err != nil {
 		return errFnHealFromAPIErr(h.ctx, err)
 	}

cmd/api-errors.go

@@ -196,8 +196,9 @@ const (
 	ErrInvalidTagDirective
 	// Add new error codes here.
 
-	// SSE-S3 related API errors
+	// SSE-S3/SSE-KMS related API errors
 	ErrInvalidEncryptionMethod
+	ErrInvalidEncryptionKeyID
 
 	// Server-Side-Encryption (with Customer provided key) related API errors.
 	ErrInsecureSSECustomerRequest
@@ -887,7 +888,7 @@ var errorCodes = errorCodeMap{
 	},
 	ErrReplicationRemoteConnectionError: {
 		Code:           "XMinioAdminReplicationRemoteConnectionError",
-		Description:    "Remote service connection error - please check remote service credentials and target bucket",
+		Description:    "Remote service connection error",
 		HTTPStatusCode: http.StatusNotFound,
 	},
 	ErrReplicationBandwidthLimitError: {
@@ -1072,6 +1073,11 @@ var errorCodes = errorCodeMap{
 		Description:    "The encryption method specified is not supported",
 		HTTPStatusCode: http.StatusBadRequest,
 	},
+	ErrInvalidEncryptionKeyID: {
+		Code:           "InvalidRequest",
+		Description:    "The specified KMS KeyID contains unsupported characters",
+		HTTPStatusCode: http.StatusBadRequest,
+	},
 	ErrInsecureSSECustomerRequest: {
 		Code:           "InvalidRequest",
 		Description:    "Requests specifying Server Side Encryption with Customer provided keys must be made over a secure connection.",
@@ -1158,7 +1164,7 @@ var errorCodes = errorCodeMap{
 	// MinIO extensions.
 	ErrStorageFull: {
 		Code:           "XMinioStorageFull",
-		Description:    "Storage backend has reached its minimum free disk threshold. Please delete a few objects to proceed.",
+		Description:    "Storage backend has reached its minimum free drive threshold. Please delete a few objects to proceed.",
 		HTTPStatusCode: http.StatusInsufficientStorage,
 	},
 	ErrRequestBodyParse: {
@@ -1921,6 +1927,8 @@ func toAPIErrorCode(ctx context.Context, err error) (apiErr APIErrorCode) {
 		apiErr = ErrInvalidEncryptionParameters
 	case crypto.ErrInvalidEncryptionMethod:
 		apiErr = ErrInvalidEncryptionMethod
+	case crypto.ErrInvalidEncryptionKeyID:
+		apiErr = ErrInvalidEncryptionKeyID
 	case crypto.ErrInvalidCustomerAlgorithm:
 		apiErr = ErrInvalidSSECustomerAlgorithm
 	case crypto.ErrMissingCustomerKey:
@@ -2066,7 +2074,7 @@ func toAPIErrorCode(ctx context.Context, err error) (apiErr APIErrorCode) {
 		apiErr = ErrRemoteDestinationNotFoundError
 	case BucketRemoteTargetNotFound:
 		apiErr = ErrRemoteTargetNotFoundError
-	case BucketRemoteConnectionErr:
+	case RemoteTargetConnectionErr:
 		apiErr = ErrReplicationRemoteConnectionError
 	case BucketRemoteAlreadyExists:
 		apiErr = ErrBucketRemoteAlreadyExists

cmd/api-headers.go

@@ -199,5 +199,12 @@ func setObjectHeaders(w http.ResponseWriter, objInfo ObjectInfo, rs *HTTPRangeSp
 		lc.SetPredictionHeaders(w, objInfo.ToLifecycleOpts())
 	}
 
+	if v, ok := objInfo.UserDefined[ReservedMetadataPrefix+"compression"]; ok {
+		if i := strings.LastIndexByte(v, '/'); i >= 0 {
+			v = v[i+1:]
+		}
+		w.Header()[xhttp.MinIOCompressed] = []string{v}
+	}
+
 	return nil
 }

cmd/background-newdisks-heal-ops.go

@@ -87,7 +87,7 @@ type healingTracker struct {
 // The disk ID will be validated against the loaded one.
 func loadHealingTracker(ctx context.Context, disk StorageAPI) (*healingTracker, error) {
 	if disk == nil {
-		return nil, errors.New("loadHealingTracker: nil disk given")
+		return nil, errors.New("loadHealingTracker: nil drive given")
 	}
 	diskID, err := disk.GetDiskID()
 	if err != nil {
@@ -104,7 +104,7 @@ func loadHealingTracker(ctx context.Context, disk StorageAPI) (*healingTracker,
 		return nil, err
 	}
 	if h.ID != diskID && h.ID != "" {
-		return nil, fmt.Errorf("loadHealingTracker: disk id mismatch expected %s, got %s", h.ID, diskID)
+		return nil, fmt.Errorf("loadHealingTracker: drive id mismatch expected %s, got %s", h.ID, diskID)
 	}
 	h.disk = disk
 	h.ID = diskID
@@ -129,7 +129,7 @@ func newHealingTracker(disk StorageAPI) *healingTracker {
 // If the tracker has been deleted an error is returned.
 func (h *healingTracker) update(ctx context.Context) error {
 	if h.disk.Healing() == nil {
-		return fmt.Errorf("healingTracker: disk %q is not marked as healing", h.ID)
+		return fmt.Errorf("healingTracker: drive %q is not marked as healing", h.ID)
 	}
 	if h.ID == "" || h.PoolIndex < 0 || h.SetIndex < 0 || h.DiskIndex < 0 {
 		h.ID, _ = h.disk.GetDiskID()
@@ -310,7 +310,7 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	}
 
 	// Prevent parallel erasure set healing
-	locker := z.NewNSLock(minioMetaBucket, fmt.Sprintf("new-disk-healing/%s/%d/%d", endpoint, poolIdx, setIdx))
+	locker := z.NewNSLock(minioMetaBucket, fmt.Sprintf("new-drive-healing/%s/%d/%d", endpoint, poolIdx, setIdx))
 	lkctx, err := locker.GetLock(ctx, newDiskHealingTimeout)
 	if err != nil {
 		return err
@@ -318,8 +318,7 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	ctx = lkctx.Context()
 	defer locker.Unlock(lkctx.Cancel)
 
-	buckets, _ := z.ListBuckets(ctx)
-
+	buckets, _ := z.ListBuckets(ctx, BucketOptions{})
 	// Buckets data are dispersed in multiple zones/sets, make
 	// sure to heal all bucket metadata configuration.
 	buckets = append(buckets, BucketInfo{
@@ -338,14 +337,14 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	})
 
 	if serverDebugLog {
-		logger.Info("Healing disk '%v' on %s pool", disk, humanize.Ordinal(poolIdx+1))
+		logger.Info("Healing drive '%v' on %s pool", disk, humanize.Ordinal(poolIdx+1))
 	}
 
 	// Load healing tracker in this disk
 	tracker, err := loadHealingTracker(ctx, disk)
 	if err != nil {
 		// So someone changed the drives underneath, healing tracker missing.
-		logger.LogIf(ctx, fmt.Errorf("Healing tracker missing on '%s', disk was swapped again on %s pool: %w",
+		logger.LogIf(ctx, fmt.Errorf("Healing tracker missing on '%s', drive was swapped again on %s pool: %w",
 			disk, humanize.Ordinal(poolIdx+1), err))
 		tracker = newHealingTracker(disk)
 	}
@@ -370,9 +369,9 @@ func healFreshDisk(ctx context.Context, z *erasureServerPools, endpoint Endpoint
 	}
 
 	if tracker.ItemsFailed > 0 {
-		logger.Info("Healing disk '%s' failed (healed: %d, failed: %d).", disk, tracker.ItemsHealed, tracker.ItemsFailed)
+		logger.Info("Healing drive '%s' failed (healed: %d, failed: %d).", disk, tracker.ItemsHealed, tracker.ItemsFailed)
 	} else {
-		logger.Info("Healing disk '%s' complete (healed: %d, failed: %d).", disk, tracker.ItemsHealed, tracker.ItemsFailed)
+		logger.Info("Healing drive '%s' complete (healed: %d, failed: %d).", disk, tracker.ItemsHealed, tracker.ItemsFailed)
 	}
 
 	if serverDebugLog {

cmd/benchmark-utils_test.go

@@ -35,7 +35,7 @@ func runPutObjectBenchmark(b *testing.B, obj ObjectLayer, objSize int) {
 	// obtains random bucket name.
 	bucket := getRandomBucketName()
 	// create bucket.
-	err = obj.MakeBucketWithLocation(context.Background(), bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(context.Background(), bucket, MakeBucketOptions{})
 	if err != nil {
 		b.Fatal(err)
 	}
@@ -76,7 +76,7 @@ func runPutObjectPartBenchmark(b *testing.B, obj ObjectLayer, partSize int) {
 	object := getRandomObjectName()
 
 	// create bucket.
-	err = obj.MakeBucketWithLocation(context.Background(), bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(context.Background(), bucket, MakeBucketOptions{})
 	if err != nil {
 		b.Fatal(err)
 	}
@@ -196,7 +196,7 @@ func runPutObjectBenchmarkParallel(b *testing.B, obj ObjectLayer, objSize int) {
 	// obtains random bucket name.
 	bucket := getRandomBucketName()
 	// create bucket.
-	err := obj.MakeBucketWithLocation(context.Background(), bucket, BucketOptions{})
+	err := obj.MakeBucketWithLocation(context.Background(), bucket, MakeBucketOptions{})
 	if err != nil {
 		b.Fatal(err)
 	}

cmd/bitrot-streaming.go

@@ -180,7 +180,7 @@ func (b *streamingBitrotReader) ReadAt(buf []byte, offset int64) (int, error) {
 	b.h.Write(buf)
 
 	if !bytes.Equal(b.h.Sum(nil), b.hashBytes) {
-		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s  -> %s/%s - content hash does not match - expected %s, got %s",
+		logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s  -> %s/%s - content hash does not match - expected %s, got %s",
 			b.disk, b.volume, b.filePath, hex.EncodeToString(b.hashBytes), hex.EncodeToString(b.h.Sum(nil))))
 		return 0, errFileCorrupt
 	}

cmd/bitrot-whole.go

@@ -38,12 +38,12 @@ type wholeBitrotWriter struct {
 func (b *wholeBitrotWriter) Write(p []byte) (int, error) {
 	err := b.disk.AppendFile(context.TODO(), b.volume, b.filePath, p)
 	if err != nil {
-		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s returned %w", b.disk, err))
+		logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s returned %w", b.disk, err))
 		return 0, err
 	}
 	_, err = b.Hash.Write(p)
 	if err != nil {
-		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s returned %w", b.disk, err))
+		logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s returned %w", b.disk, err))
 		return 0, err
 	}
 	return len(p), nil
@@ -72,12 +72,12 @@ func (b *wholeBitrotReader) ReadAt(buf []byte, offset int64) (n int, err error)
 	if b.buf == nil {
 		b.buf = make([]byte, b.tillOffset-offset)
 		if _, err := b.disk.ReadFile(context.TODO(), b.volume, b.filePath, offset, b.buf, b.verifier); err != nil {
-			logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s -> %s/%s returned %w", b.disk, b.volume, b.filePath, err))
+			logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s -> %s/%s returned %w", b.disk, b.volume, b.filePath, err))
 			return 0, err
 		}
 	}
 	if len(b.buf) < len(buf) {
-		logger.LogIf(GlobalContext, fmt.Errorf("Disk: %s -> %s/%s returned %w", b.disk, b.volume, b.filePath, errLessData))
+		logger.LogIf(GlobalContext, fmt.Errorf("Drive: %s -> %s/%s returned %w", b.disk, b.volume, b.filePath, errLessData))
 		return 0, errLessData
 	}
 	n = copy(buf, b.buf)

cmd/bitrot_test.go

@@ -20,17 +20,11 @@ package cmd
 import (
 	"context"
 	"io"
-	"io/ioutil"
-	"os"
 	"testing"
 )
 
 func testBitrotReaderWriterAlgo(t *testing.T, bitrotAlgo BitrotAlgorithm) {
-	tmpDir, err := ioutil.TempDir("", "")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(tmpDir)
+	tmpDir := t.TempDir()
 
 	volume := "testvol"
 	filePath := "testfile"
@@ -60,7 +54,9 @@ func testBitrotReaderWriterAlgo(t *testing.T, bitrotAlgo BitrotAlgorithm) {
 	if err != nil {
 		t.Fatal(err)
 	}
-	writer.(io.Closer).Close()
+	if bw, ok := writer.(io.Closer); ok {
+		bw.Close()
+	}
 
 	reader := newBitrotReader(disk, nil, volume, filePath, 35, bitrotAlgo, bitrotWriterSum(writer), 10)
 	b := make([]byte, 10)
@@ -76,6 +72,9 @@ func testBitrotReaderWriterAlgo(t *testing.T, bitrotAlgo BitrotAlgorithm) {
 	if _, err = reader.ReadAt(b[:5], 30); err != nil {
 		t.Fatal(err)
 	}
+	if br, ok := reader.(io.Closer); ok {
+		br.Close()
+	}
 }
 
 func TestAllBitrotAlgorithms(t *testing.T) {

cmd/bucket-encryption-handlers.go

@@ -65,7 +65,7 @@ func (api objectAPIHandlers) PutBucketEncryptionHandler(w http.ResponseWriter, r
 	}
 
 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -155,7 +155,7 @@ func (api objectAPIHandlers) GetBucketEncryptionHandler(w http.ResponseWriter, r
 
 	// Check if bucket exists
 	var err error
-	if _, err = objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err = objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -198,7 +198,7 @@ func (api objectAPIHandlers) DeleteBucketEncryptionHandler(w http.ResponseWriter
 
 	// Check if bucket exists
 	var err error
-	if _, err = objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err = objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}

cmd/bucket-handlers.go

@@ -201,7 +201,7 @@ func (api objectAPIHandlers) GetBucketLocationHandler(w http.ResponseWriter, r *
 
 	getBucketInfo := objectAPI.GetBucketInfo
 
-	if _, err := getBucketInfo(ctx, bucket); err != nil {
+	if _, err := getBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -332,7 +332,7 @@ func (api objectAPIHandlers) ListBucketsHandler(w http.ResponseWriter, r *http.R
 	} else {
 		// Invoke the list buckets.
 		var err error
-		bucketsInfo, err = listBuckets(ctx)
+		bucketsInfo, err = listBuckets(ctx, BucketOptions{})
 		if err != nil {
 			writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 			return
@@ -447,7 +447,7 @@ func (api objectAPIHandlers) DeleteMultipleObjectsHandler(w http.ResponseWriter,
 	checkRequestAuthType(ctx, r, policy.DeleteObjectAction, bucket, "")
 
 	// Before proceeding validate if bucket exists.
-	_, err := objectAPI.GetBucketInfo(ctx, bucket)
+	_, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -727,11 +727,30 @@ func (api objectAPIHandlers) PutBucketHandler(w http.ResponseWriter, r *http.Req
 		}
 	}
 
-	if s3Error := checkRequestAuthType(ctx, r, policy.CreateBucketAction, bucket, ""); s3Error != ErrNone {
+	cred, owner, s3Error := checkRequestAuthTypeCredential(ctx, r, policy.CreateBucketAction, bucket, "")
+	if s3Error != ErrNone {
 		writeErrorResponse(ctx, w, errorCodes.ToAPIErr(s3Error), r.URL)
 		return
 	}
 
+	if objectLockEnabled {
+		// Creating a bucket with locking requires the user having more permissions
+		for _, action := range []iampolicy.Action{iampolicy.PutBucketObjectLockConfigurationAction, iampolicy.PutBucketVersioningAction} {
+			if !globalIAMSys.IsAllowed(iampolicy.Args{
+				AccountName:     cred.AccessKey,
+				Groups:          cred.Groups,
+				Action:          action,
+				ConditionValues: getConditionValues(r, "", cred.AccessKey, cred.Claims),
+				BucketName:      bucket,
+				IsOwner:         owner,
+				Claims:          cred.Claims,
+			}) {
+				writeErrorResponse(ctx, w, errorCodes.ToAPIErr(ErrAccessDenied), r.URL)
+				return
+			}
+		}
+	}
+
 	// Parse incoming location constraint.
 	location, s3Error := parseLocationConstraint(r)
 	if s3Error != ErrNone {
@@ -746,7 +765,7 @@ func (api objectAPIHandlers) PutBucketHandler(w http.ResponseWriter, r *http.Req
 		return
 	}
 
-	opts := BucketOptions{
+	opts := MakeBucketOptions{
 		Location:    location,
 		LockEnabled: objectLockEnabled,
 		ForceCreate: forceCreate,
@@ -765,7 +784,11 @@ func (api objectAPIHandlers) PutBucketHandler(w http.ResponseWriter, r *http.Req
 				}
 
 				if err = globalDNSConfig.Put(bucket); err != nil {
-					objectAPI.DeleteBucket(context.Background(), bucket, DeleteBucketOptions{Force: false, NoRecreate: true})
+					objectAPI.DeleteBucket(context.Background(), bucket, DeleteBucketOptions{
+						Force:      false,
+						NoRecreate: true,
+						SRDeleteOp: getSRBucketDeleteOp(globalSiteReplicationSys.isEnabled()),
+					})
 					writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 					return
 				}
@@ -1165,7 +1188,7 @@ func (api objectAPIHandlers) GetBucketPolicyStatusHandler(w http.ResponseWriter,
 	}
 
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -1228,7 +1251,7 @@ func (api objectAPIHandlers) HeadBucketHandler(w http.ResponseWriter, r *http.Re
 
 	getBucketInfo := objectAPI.GetBucketInfo
 
-	if _, err := getBucketInfo(ctx, bucket); err != nil {
+	if _, err := getBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponseHeadersOnly(w, toAPIError(ctx, err))
 		return
 	}
@@ -1305,7 +1328,10 @@ func (api objectAPIHandlers) DeleteBucketHandler(w http.ResponseWriter, r *http.
 	deleteBucket := objectAPI.DeleteBucket
 
 	// Attempt to delete bucket.
-	if err := deleteBucket(ctx, bucket, DeleteBucketOptions{Force: forceDelete}); err != nil {
+	if err := deleteBucket(ctx, bucket, DeleteBucketOptions{
+		Force:      forceDelete,
+		SRDeleteOp: getSRBucketDeleteOp(globalSiteReplicationSys.isEnabled()),
+	}); err != nil {
 		apiErr := toAPIError(ctx, err)
 		if _, ok := err.(BucketNotEmpty); ok {
 			if globalBucketVersioningSys.Enabled(bucket) || globalBucketVersioningSys.Suspended(bucket) {
@@ -1472,7 +1498,7 @@ func (api objectAPIHandlers) PutBucketTaggingHandler(w http.ResponseWriter, r *h
 	}
 
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}

cmd/bucket-lifecycle-handlers.go

@@ -62,7 +62,7 @@ func (api objectAPIHandlers) PutBucketLifecycleHandler(w http.ResponseWriter, r
 	}
 
 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -121,7 +121,7 @@ func (api objectAPIHandlers) GetBucketLifecycleHandler(w http.ResponseWriter, r
 	}
 
 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -163,7 +163,7 @@ func (api objectAPIHandlers) DeleteBucketLifecycleHandler(w http.ResponseWriter,
 	}
 
 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}

cmd/bucket-lifecycle.go

@@ -560,7 +560,7 @@ func (r *RestoreObjectRequest) validate(ctx context.Context, objAPI ObjectLayer)
 	}
 	// Check if bucket exists.
 	if !r.OutputLocation.IsEmpty() {
-		if _, err := objAPI.GetBucketInfo(ctx, r.OutputLocation.S3.BucketName); err != nil {
+		if _, err := objAPI.GetBucketInfo(ctx, r.OutputLocation.S3.BucketName, BucketOptions{}); err != nil {
 			return err
 		}
 		if r.OutputLocation.S3.Prefix == "" {

cmd/bucket-metadata.go

@@ -105,10 +105,8 @@ type BucketMetadata struct {
 
 // newBucketMetadata creates BucketMetadata with the supplied name and Created to Now.
 func newBucketMetadata(name string) BucketMetadata {
-	now := UTCNow()
 	return BucketMetadata{
-		Name:    name,
-		Created: now,
+		Name: name,
 		notificationConfig: &event.Config{
 			XMLNS: "http://s3.amazonaws.com/doc/2006-03-01/",
 		},
@@ -121,6 +119,17 @@ func newBucketMetadata(name string) BucketMetadata {
 	}
 }
 
+// SetCreatedAt preserves the CreatedAt time for bucket across sites in site replication. It defaults to
+// creation time of bucket on this cluster in all other cases.
+func (b *BucketMetadata) SetCreatedAt(createdAt time.Time) {
+	if b.Created.IsZero() {
+		b.Created = UTCNow()
+	}
+	if !createdAt.IsZero() {
+		b.Created = createdAt.UTC()
+	}
+}
+
 // Load - loads the metadata of bucket by name from ObjectLayer api.
 // If an error is returned the returned metadata will be default initialized.
 func (b *BucketMetadata) Load(ctx context.Context, api ObjectLayer, name string) error {
@@ -160,17 +169,19 @@ func loadBucketMetadata(ctx context.Context, objectAPI ObjectLayer, bucket strin
 	if err != nil && !errors.Is(err, errConfigNotFound) {
 		return b, err
 	}
-
+	if err == nil {
+		b.defaultTimestamps()
+	}
 	// Old bucket without bucket metadata. Hence we migrate existing settings.
 	if err := b.convertLegacyConfigs(ctx, objectAPI); err != nil {
 		return b, err
 	}
 
 	// migrate unencrypted remote targets
-	if err = b.migrateTargetConfig(ctx, objectAPI); err != nil {
+	if err := b.migrateTargetConfig(ctx, objectAPI); err != nil {
 		return b, err
 	}
-	b.defaultTimestamps()
+
 	return b, nil
 }
 
@@ -345,6 +356,7 @@ func (b *BucketMetadata) convertLegacyConfigs(ctx context.Context, objectAPI Obj
 			b.BucketTargetsConfigJSON = configData
 		}
 	}
+	b.defaultTimestamps()
 
 	if err := b.Save(ctx, objectAPI); err != nil {
 		return err

cmd/bucket-notification-handlers.go

@@ -60,7 +60,7 @@ func (api objectAPIHandlers) GetBucketNotificationHandler(w http.ResponseWriter,
 		return
 	}
 
-	_, err := objAPI.GetBucketInfo(ctx, bucketName)
+	_, err := objAPI.GetBucketInfo(ctx, bucketName, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -132,7 +132,7 @@ func (api objectAPIHandlers) PutBucketNotificationHandler(w http.ResponseWriter,
 		return
 	}
 
-	_, err := objectAPI.GetBucketInfo(ctx, bucketName)
+	_, err := objectAPI.GetBucketInfo(ctx, bucketName, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return

cmd/bucket-policy-handlers.go

@@ -61,7 +61,7 @@ func (api objectAPIHandlers) PutBucketPolicyHandler(w http.ResponseWriter, r *ht
 	}
 
 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -145,7 +145,7 @@ func (api objectAPIHandlers) DeleteBucketPolicyHandler(w http.ResponseWriter, r
 	}
 
 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -191,7 +191,7 @@ func (api objectAPIHandlers) GetBucketPolicyHandler(w http.ResponseWriter, r *ht
 	}
 
 	// Check if bucket exists.
-	if _, err := objAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}

cmd/bucket-policy-handlers_test.go

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

cmd/bucket-replication-handlers.go

@@ -55,7 +55,7 @@ func (api objectAPIHandlers) PutBucketReplicationConfigHandler(w http.ResponseWr
 		return
 	}
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -118,7 +118,7 @@ func (api objectAPIHandlers) GetBucketReplicationConfigHandler(w http.ResponseWr
 		return
 	}
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -157,7 +157,7 @@ func (api objectAPIHandlers) DeleteBucketReplicationConfigHandler(w http.Respons
 		return
 	}
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -198,7 +198,7 @@ func (api objectAPIHandlers) GetBucketReplicationMetricsHandler(w http.ResponseW
 	}
 
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -267,7 +267,7 @@ func (api objectAPIHandlers) ResetBucketReplicationStartHandler(w http.ResponseW
 	}
 
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}
@@ -310,7 +310,7 @@ func (api objectAPIHandlers) ResetBucketReplicationStartHandler(w http.ResponseW
 	rinfo.Targets = append(rinfo.Targets, ResyncTarget{Arn: tgtArns[0], ResetID: target.ResetID})
 	if err = globalBucketTargetSys.SetTarget(ctx, bucket, &target, true); err != nil {
 		switch err.(type) {
-		case BucketRemoteConnectionErr:
+		case RemoteTargetConnectionErr:
 			writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErrWithErr(ErrReplicationRemoteConnectionError, err), r.URL)
 		default:
 			writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
@@ -356,7 +356,7 @@ func (api objectAPIHandlers) ResetBucketReplicationStatusHandler(w http.Response
 	}
 
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}

cmd/bucket-replication-utils.go

@@ -737,3 +737,32 @@ func extractReplicateDiffOpts(q url.Values) (opts madmin.ReplDiffOpts) {
 	opts.Prefix = q.Get("prefix")
 	return
 }
+
+const (
+	replicationMRFDir = bucketMetaPrefix + SlashSeparator + replicationDir + SlashSeparator + "mrf"
+	mrfMetaFormat     = 1
+	mrfMetaVersionV1  = 1
+	mrfMetaVersion    = mrfMetaVersionV1
+)
+
+// MRFReplicateEntry mrf entry to save to disk
+type MRFReplicateEntry struct {
+	Bucket    string `json:"bucket" msg:"b"`
+	Object    string `json:"object" msg:"o"`
+	versionID string `json:"-"`
+}
+
+// MRFReplicateEntries has the map of MRF entries to save to disk
+type MRFReplicateEntries struct {
+	Entries map[string]MRFReplicateEntry `json:"entries" msg:"e"`
+	Version int                          `json:"version" msg:"v"`
+}
+
+// ToMRFEntry returns the relevant info needed by MRF
+func (ri ReplicateObjectInfo) ToMRFEntry() MRFReplicateEntry {
+	return MRFReplicateEntry{
+		Bucket:    ri.Bucket,
+		Object:    ri.Name,
+		versionID: ri.VersionID,
+	}
+}

cmd/bucket-replication-utils_gen.go

@@ -260,6 +260,413 @@ func (z *BucketReplicationResyncStatus) Msgsize() (s int) {
 	return
 }
 
+// DecodeMsg implements msgp.Decodable
+func (z *MRFReplicateEntries) DecodeMsg(dc *msgp.Reader) (err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, err = dc.ReadMapHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, err = dc.ReadMapKeyPtr()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "e":
+			var zb0002 uint32
+			zb0002, err = dc.ReadMapHeader()
+			if err != nil {
+				err = msgp.WrapError(err, "Entries")
+				return
+			}
+			if z.Entries == nil {
+				z.Entries = make(map[string]MRFReplicateEntry, zb0002)
+			} else if len(z.Entries) > 0 {
+				for key := range z.Entries {
+					delete(z.Entries, key)
+				}
+			}
+			for zb0002 > 0 {
+				zb0002--
+				var za0001 string
+				var za0002 MRFReplicateEntry
+				za0001, err = dc.ReadString()
+				if err != nil {
+					err = msgp.WrapError(err, "Entries")
+					return
+				}
+				var zb0003 uint32
+				zb0003, err = dc.ReadMapHeader()
+				if err != nil {
+					err = msgp.WrapError(err, "Entries", za0001)
+					return
+				}
+				for zb0003 > 0 {
+					zb0003--
+					field, err = dc.ReadMapKeyPtr()
+					if err != nil {
+						err = msgp.WrapError(err, "Entries", za0001)
+						return
+					}
+					switch msgp.UnsafeString(field) {
+					case "b":
+						za0002.Bucket, err = dc.ReadString()
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001, "Bucket")
+							return
+						}
+					case "o":
+						za0002.Object, err = dc.ReadString()
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001, "Object")
+							return
+						}
+					default:
+						err = dc.Skip()
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001)
+							return
+						}
+					}
+				}
+				z.Entries[za0001] = za0002
+			}
+		case "v":
+			z.Version, err = dc.ReadInt()
+			if err != nil {
+				err = msgp.WrapError(err, "Version")
+				return
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z *MRFReplicateEntries) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 2
+	// write "e"
+	err = en.Append(0x82, 0xa1, 0x65)
+	if err != nil {
+		return
+	}
+	err = en.WriteMapHeader(uint32(len(z.Entries)))
+	if err != nil {
+		err = msgp.WrapError(err, "Entries")
+		return
+	}
+	for za0001, za0002 := range z.Entries {
+		err = en.WriteString(za0001)
+		if err != nil {
+			err = msgp.WrapError(err, "Entries")
+			return
+		}
+		// map header, size 2
+		// write "b"
+		err = en.Append(0x82, 0xa1, 0x62)
+		if err != nil {
+			return
+		}
+		err = en.WriteString(za0002.Bucket)
+		if err != nil {
+			err = msgp.WrapError(err, "Entries", za0001, "Bucket")
+			return
+		}
+		// write "o"
+		err = en.Append(0xa1, 0x6f)
+		if err != nil {
+			return
+		}
+		err = en.WriteString(za0002.Object)
+		if err != nil {
+			err = msgp.WrapError(err, "Entries", za0001, "Object")
+			return
+		}
+	}
+	// write "v"
+	err = en.Append(0xa1, 0x76)
+	if err != nil {
+		return
+	}
+	err = en.WriteInt(z.Version)
+	if err != nil {
+		err = msgp.WrapError(err, "Version")
+		return
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z *MRFReplicateEntries) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 2
+	// string "e"
+	o = append(o, 0x82, 0xa1, 0x65)
+	o = msgp.AppendMapHeader(o, uint32(len(z.Entries)))
+	for za0001, za0002 := range z.Entries {
+		o = msgp.AppendString(o, za0001)
+		// map header, size 2
+		// string "b"
+		o = append(o, 0x82, 0xa1, 0x62)
+		o = msgp.AppendString(o, za0002.Bucket)
+		// string "o"
+		o = append(o, 0xa1, 0x6f)
+		o = msgp.AppendString(o, za0002.Object)
+	}
+	// string "v"
+	o = append(o, 0xa1, 0x76)
+	o = msgp.AppendInt(o, z.Version)
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *MRFReplicateEntries) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadMapHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, bts, err = msgp.ReadMapKeyZC(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "e":
+			var zb0002 uint32
+			zb0002, bts, err = msgp.ReadMapHeaderBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Entries")
+				return
+			}
+			if z.Entries == nil {
+				z.Entries = make(map[string]MRFReplicateEntry, zb0002)
+			} else if len(z.Entries) > 0 {
+				for key := range z.Entries {
+					delete(z.Entries, key)
+				}
+			}
+			for zb0002 > 0 {
+				var za0001 string
+				var za0002 MRFReplicateEntry
+				zb0002--
+				za0001, bts, err = msgp.ReadStringBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Entries")
+					return
+				}
+				var zb0003 uint32
+				zb0003, bts, err = msgp.ReadMapHeaderBytes(bts)
+				if err != nil {
+					err = msgp.WrapError(err, "Entries", za0001)
+					return
+				}
+				for zb0003 > 0 {
+					zb0003--
+					field, bts, err = msgp.ReadMapKeyZC(bts)
+					if err != nil {
+						err = msgp.WrapError(err, "Entries", za0001)
+						return
+					}
+					switch msgp.UnsafeString(field) {
+					case "b":
+						za0002.Bucket, bts, err = msgp.ReadStringBytes(bts)
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001, "Bucket")
+							return
+						}
+					case "o":
+						za0002.Object, bts, err = msgp.ReadStringBytes(bts)
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001, "Object")
+							return
+						}
+					default:
+						bts, err = msgp.Skip(bts)
+						if err != nil {
+							err = msgp.WrapError(err, "Entries", za0001)
+							return
+						}
+					}
+				}
+				z.Entries[za0001] = za0002
+			}
+		case "v":
+			z.Version, bts, err = msgp.ReadIntBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Version")
+				return
+			}
+		default:
+			bts, err = msgp.Skip(bts)
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z *MRFReplicateEntries) Msgsize() (s int) {
+	s = 1 + 2 + msgp.MapHeaderSize
+	if z.Entries != nil {
+		for za0001, za0002 := range z.Entries {
+			_ = za0002
+			s += msgp.StringPrefixSize + len(za0001) + 1 + 2 + msgp.StringPrefixSize + len(za0002.Bucket) + 2 + msgp.StringPrefixSize + len(za0002.Object)
+		}
+	}
+	s += 2 + msgp.IntSize
+	return
+}
+
+// DecodeMsg implements msgp.Decodable
+func (z *MRFReplicateEntry) DecodeMsg(dc *msgp.Reader) (err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, err = dc.ReadMapHeader()
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, err = dc.ReadMapKeyPtr()
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "b":
+			z.Bucket, err = dc.ReadString()
+			if err != nil {
+				err = msgp.WrapError(err, "Bucket")
+				return
+			}
+		case "o":
+			z.Object, err = dc.ReadString()
+			if err != nil {
+				err = msgp.WrapError(err, "Object")
+				return
+			}
+		default:
+			err = dc.Skip()
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	return
+}
+
+// EncodeMsg implements msgp.Encodable
+func (z MRFReplicateEntry) EncodeMsg(en *msgp.Writer) (err error) {
+	// map header, size 2
+	// write "b"
+	err = en.Append(0x82, 0xa1, 0x62)
+	if err != nil {
+		return
+	}
+	err = en.WriteString(z.Bucket)
+	if err != nil {
+		err = msgp.WrapError(err, "Bucket")
+		return
+	}
+	// write "o"
+	err = en.Append(0xa1, 0x6f)
+	if err != nil {
+		return
+	}
+	err = en.WriteString(z.Object)
+	if err != nil {
+		err = msgp.WrapError(err, "Object")
+		return
+	}
+	return
+}
+
+// MarshalMsg implements msgp.Marshaler
+func (z MRFReplicateEntry) MarshalMsg(b []byte) (o []byte, err error) {
+	o = msgp.Require(b, z.Msgsize())
+	// map header, size 2
+	// string "b"
+	o = append(o, 0x82, 0xa1, 0x62)
+	o = msgp.AppendString(o, z.Bucket)
+	// string "o"
+	o = append(o, 0xa1, 0x6f)
+	o = msgp.AppendString(o, z.Object)
+	return
+}
+
+// UnmarshalMsg implements msgp.Unmarshaler
+func (z *MRFReplicateEntry) UnmarshalMsg(bts []byte) (o []byte, err error) {
+	var field []byte
+	_ = field
+	var zb0001 uint32
+	zb0001, bts, err = msgp.ReadMapHeaderBytes(bts)
+	if err != nil {
+		err = msgp.WrapError(err)
+		return
+	}
+	for zb0001 > 0 {
+		zb0001--
+		field, bts, err = msgp.ReadMapKeyZC(bts)
+		if err != nil {
+			err = msgp.WrapError(err)
+			return
+		}
+		switch msgp.UnsafeString(field) {
+		case "b":
+			z.Bucket, bts, err = msgp.ReadStringBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Bucket")
+				return
+			}
+		case "o":
+			z.Object, bts, err = msgp.ReadStringBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "Object")
+				return
+			}
+		default:
+			bts, err = msgp.Skip(bts)
+			if err != nil {
+				err = msgp.WrapError(err)
+				return
+			}
+		}
+	}
+	o = bts
+	return
+}
+
+// Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
+func (z MRFReplicateEntry) Msgsize() (s int) {
+	s = 1 + 2 + msgp.StringPrefixSize + len(z.Bucket) + 2 + msgp.StringPrefixSize + len(z.Object)
+	return
+}
+
 // DecodeMsg implements msgp.Decodable
 func (z *ReplicateDecision) DecodeMsg(dc *msgp.Reader) (err error) {
 	var field []byte

cmd/bucket-replication-utils_gen_test.go

@@ -122,6 +122,232 @@ func BenchmarkDecodeBucketReplicationResyncStatus(b *testing.B) {
 	}
 }
 
+func TestMarshalUnmarshalMRFReplicateEntries(t *testing.T) {
+	v := MRFReplicateEntries{}
+	bts, err := v.MarshalMsg(nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	left, err := v.UnmarshalMsg(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after UnmarshalMsg(): %q", len(left), left)
+	}
+
+	left, err = msgp.Skip(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after Skip(): %q", len(left), left)
+	}
+}
+
+func BenchmarkMarshalMsgMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	bts := make([]byte, 0, v.Msgsize())
+	bts, _ = v.MarshalMsg(bts[0:0])
+	b.SetBytes(int64(len(bts)))
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		bts, _ = v.MarshalMsg(bts[0:0])
+	}
+}
+
+func BenchmarkUnmarshalMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	bts, _ := v.MarshalMsg(nil)
+	b.ReportAllocs()
+	b.SetBytes(int64(len(bts)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		_, err := v.UnmarshalMsg(bts)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestEncodeDecodeMRFReplicateEntries(t *testing.T) {
+	v := MRFReplicateEntries{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeMRFReplicateEntries Msgsize() is inaccurate")
+	}
+
+	vn := MRFReplicateEntries{}
+	err := msgp.Decode(&buf, &vn)
+	if err != nil {
+		t.Error(err)
+	}
+
+	buf.Reset()
+	msgp.Encode(&buf, &v)
+	err = msgp.NewReader(&buf).Skip()
+	if err != nil {
+		t.Error(err)
+	}
+}
+
+func BenchmarkEncodeMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	en := msgp.NewWriter(msgp.Nowhere)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.EncodeMsg(en)
+	}
+	en.Flush()
+}
+
+func BenchmarkDecodeMRFReplicateEntries(b *testing.B) {
+	v := MRFReplicateEntries{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	rd := msgp.NewEndlessReader(buf.Bytes(), b)
+	dc := msgp.NewReader(rd)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		err := v.DecodeMsg(dc)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestMarshalUnmarshalMRFReplicateEntry(t *testing.T) {
+	v := MRFReplicateEntry{}
+	bts, err := v.MarshalMsg(nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	left, err := v.UnmarshalMsg(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after UnmarshalMsg(): %q", len(left), left)
+	}
+
+	left, err = msgp.Skip(bts)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(left) > 0 {
+		t.Errorf("%d bytes left over after Skip(): %q", len(left), left)
+	}
+}
+
+func BenchmarkMarshalMsgMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.MarshalMsg(nil)
+	}
+}
+
+func BenchmarkAppendMsgMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	bts := make([]byte, 0, v.Msgsize())
+	bts, _ = v.MarshalMsg(bts[0:0])
+	b.SetBytes(int64(len(bts)))
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		bts, _ = v.MarshalMsg(bts[0:0])
+	}
+}
+
+func BenchmarkUnmarshalMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	bts, _ := v.MarshalMsg(nil)
+	b.ReportAllocs()
+	b.SetBytes(int64(len(bts)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		_, err := v.UnmarshalMsg(bts)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func TestEncodeDecodeMRFReplicateEntry(t *testing.T) {
+	v := MRFReplicateEntry{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+
+	m := v.Msgsize()
+	if buf.Len() > m {
+		t.Log("WARNING: TestEncodeDecodeMRFReplicateEntry Msgsize() is inaccurate")
+	}
+
+	vn := MRFReplicateEntry{}
+	err := msgp.Decode(&buf, &vn)
+	if err != nil {
+		t.Error(err)
+	}
+
+	buf.Reset()
+	msgp.Encode(&buf, &v)
+	err = msgp.NewReader(&buf).Skip()
+	if err != nil {
+		t.Error(err)
+	}
+}
+
+func BenchmarkEncodeMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	en := msgp.NewWriter(msgp.Nowhere)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		v.EncodeMsg(en)
+	}
+	en.Flush()
+}
+
+func BenchmarkDecodeMRFReplicateEntry(b *testing.B) {
+	v := MRFReplicateEntry{}
+	var buf bytes.Buffer
+	msgp.Encode(&buf, &v)
+	b.SetBytes(int64(buf.Len()))
+	rd := msgp.NewEndlessReader(buf.Bytes(), b)
+	dc := msgp.NewReader(rd)
+	b.ReportAllocs()
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		err := v.DecodeMsg(dc)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
 func TestMarshalUnmarshalReplicateDecision(t *testing.T) {
 	v := ReplicateDecision{}
 	bts, err := v.MarshalMsg(nil)

cmd/bucket-replication.go

@@ -403,6 +403,7 @@ func replicateDelete(ctx context.Context, dobj DeletedObjectReplicationInfo, obj
 	lk := objectAPI.NewNSLock(bucket, "/[replicate]/"+dobj.ObjectName)
 	lkctx, err := lk.GetLock(ctx, globalOperationTimeout)
 	if err != nil {
+		globalReplicationPool.queueMRFSave(dobj.ToMRFEntry())
 		logger.LogIf(ctx, fmt.Errorf("failed to get lock for object: %s bucket:%s arn:%s", dobj.ObjectName, bucket, rcfg.RoleArn))
 		sendEvent(eventArgs{
 			BucketName: bucket,
@@ -477,6 +478,7 @@ func replicateDelete(ctx context.Context, dobj DeletedObjectReplicationInfo, obj
 	eventName := event.ObjectReplicationComplete
 	if replicationStatus == replication.Failed {
 		eventName = event.ObjectReplicationFailed
+		globalReplicationPool.queueMRFSave(dobj.ToMRFEntry())
 	}
 	drs := getReplicationState(rinfos, dobj.ReplicationState, dobj.VersionID)
 	if replicationStatus != prevStatus {
@@ -534,7 +536,7 @@ func replicateDeleteToTarget(ctx context.Context, dobj DeletedObjectReplicationI
 	if dobj.VersionID != "" && rinfo.VersionPurgeStatus == Complete {
 		return
 	}
-	if tgt.IsOffline() {
+	if globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
 		logger.LogIf(ctx, fmt.Errorf("remote target is offline for bucket:%s arn:%s", dobj.Bucket, tgt.ARN))
 		sendEvent(eventArgs{
 			BucketName: dobj.Bucket,
@@ -910,6 +912,7 @@ func replicateObject(ctx context.Context, ri ReplicateObjectInfo, objectAPI Obje
 			Object:     objInfo,
 			Host:       "Internal: [Replication]",
 		})
+		globalReplicationPool.queueMRFSave(ri.ToMRFEntry())
 		logger.LogIf(ctx, fmt.Errorf("failed to get lock for object: %s bucket:%s arn:%s", object, bucket, cfg.RoleArn))
 		return
 	}
@@ -992,12 +995,12 @@ func replicateObject(ctx context.Context, ri ReplicateObjectInfo, objectAPI Obje
 
 	// re-queue failures once more - keep a retry count to avoid flooding the queue if
 	// the target site is down. Leave it to scanner to catch up instead.
-	if rinfos.ReplicationStatus() != replication.Completed && ri.RetryCount < 1 {
+	if rinfos.ReplicationStatus() != replication.Completed {
 		ri.OpType = replication.HealReplicationType
 		ri.EventType = ReplicateMRF
 		ri.ReplicationStatusInternal = rinfos.ReplicationStatusInternal()
 		ri.RetryCount++
-		globalReplicationPool.queueReplicaFailedTask(ri)
+		globalReplicationPool.queueMRFSave(ri.ToMRFEntry())
 	}
 }
 
@@ -1039,7 +1042,7 @@ func replicateObjectToTarget(ctx context.Context, ri ReplicateObjectInfo, object
 		rinfo.ReplicationResynced = true
 		return
 	}
-	if tgt.IsOffline() {
+	if globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
 		logger.LogIf(ctx, fmt.Errorf("remote target is offline for bucket:%s arn:%s", bucket, tgt.ARN))
 		sendEvent(eventArgs{
 			EventName:  event.ObjectReplicationNotTracked,
@@ -1059,13 +1062,15 @@ func replicateObjectToTarget(ctx context.Context, ri ReplicateObjectInfo, object
 		VersionSuspended: versionSuspended,
 	})
 	if err != nil {
-		sendEvent(eventArgs{
-			EventName:  event.ObjectReplicationNotTracked,
-			BucketName: bucket,
-			Object:     objInfo,
-			Host:       "Internal: [Replication]",
-		})
-		logger.LogIf(ctx, fmt.Errorf("Unable to update replicate for %s/%s(%s): %w", bucket, object, objInfo.VersionID, err))
+		if !isErrObjectNotFound(err) {
+			sendEvent(eventArgs{
+				EventName:  event.ObjectReplicationNotTracked,
+				BucketName: bucket,
+				Object:     objInfo,
+				Host:       "Internal: [Replication]",
+			})
+			logger.LogIf(ctx, fmt.Errorf("Unable to update replicate metadata for %s/%s(%s): %w", bucket, object, objInfo.VersionID, err))
+		}
 		return
 	}
 	defer func() {
@@ -1302,6 +1307,19 @@ type DeletedObjectReplicationInfo struct {
 	TargetArn string
 }
 
+// ToMRFEntry returns the relevant info needed by MRF
+func (di DeletedObjectReplicationInfo) ToMRFEntry() MRFReplicateEntry {
+	versionID := di.DeleteMarkerVersionID
+	if versionID == "" {
+		versionID = di.VersionID
+	}
+	return MRFReplicateEntry{
+		Bucket:    di.Bucket,
+		Object:    di.ObjectName,
+		versionID: versionID,
+	}
+}
+
 // Replication specific APIName
 const (
 	ReplicateObjectAPI = "ReplicateObject"
@@ -1346,13 +1364,16 @@ type ReplicationPool struct {
 	mrfReplicaCh            chan ReplicateObjectInfo
 	existingReplicaCh       chan ReplicateObjectInfo
 	existingReplicaDeleteCh chan DeletedObjectReplicationInfo
-	workerSize              int
-	mrfWorkerSize           int
-	resyncState             replicationResyncState
-	workerWg                sync.WaitGroup
-	mrfWorkerWg             sync.WaitGroup
-	once                    sync.Once
-	mu                      sync.Mutex
+	mrfSaveCh               chan MRFReplicateEntry
+
+	workerSize    int
+	mrfWorkerSize int
+	resyncState   replicationResyncState
+	workerWg      sync.WaitGroup
+	mrfWorkerWg   sync.WaitGroup
+	once          sync.Once
+	mu            sync.Mutex
+	mrfMutex      sync.Mutex
 }
 
 // NewReplicationPool creates a pool of replication workers of specified size
@@ -1366,6 +1387,7 @@ func NewReplicationPool(ctx context.Context, o ObjectLayer, opts replicationPool
 		existingReplicaCh:       make(chan ReplicateObjectInfo, 100000),
 		existingReplicaDeleteCh: make(chan DeletedObjectReplicationInfo, 100000),
 		resyncState:             replicationResyncState{statusMap: make(map[string]BucketReplicationResyncStatus)},
+		mrfSaveCh:               make(chan MRFReplicateEntry, 100000),
 		ctx:                     ctx,
 		objLayer:                o,
 	}
@@ -1374,6 +1396,8 @@ func NewReplicationPool(ctx context.Context, o ObjectLayer, opts replicationPool
 	pool.ResizeFailedWorkers(opts.FailedWorkers)
 	go pool.AddExistingObjectReplicateWorker()
 	go pool.updateResyncStatus(ctx, o)
+	go pool.processMRF()
+	go pool.persistMRF()
 	return pool
 }
 
@@ -1479,33 +1503,17 @@ func (p *ReplicationPool) suggestedWorkers(failQueue bool) int {
 	return int(float64(p.workerSize) * ReplicationWorkerMultiplier)
 }
 
-func (p *ReplicationPool) queueReplicaFailedTask(ri ReplicateObjectInfo) {
-	if p == nil {
-		return
-	}
-	select {
-	case <-GlobalContext.Done():
-		p.once.Do(func() {
-			close(p.replicaCh)
-			close(p.mrfReplicaCh)
-			close(p.existingReplicaCh)
-		})
-	case p.mrfReplicaCh <- ri:
-	default:
-		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to keep up retrying failed replication - we recommend increasing number of replication failed workers with `mc admin config set api replication_failed_workers=%d`", p.suggestedWorkers(true)), replicationSubsystem)
-	}
-}
-
 func (p *ReplicationPool) queueReplicaTask(ri ReplicateObjectInfo) {
 	if p == nil {
 		return
 	}
-	var ch chan ReplicateObjectInfo
+	var ch, healCh chan ReplicateObjectInfo
 	switch ri.OpType {
 	case replication.ExistingObjectReplicationType:
 		ch = p.existingReplicaCh
 	case replication.HealReplicationType:
-		fallthrough
+		ch = p.mrfReplicaCh
+		healCh = p.replicaCh
 	default:
 		ch = p.replicaCh
 	}
@@ -1516,9 +1524,10 @@ func (p *ReplicationPool) queueReplicaTask(ri ReplicateObjectInfo) {
 			close(p.mrfReplicaCh)
 			close(p.existingReplicaCh)
 		})
+	case healCh <- ri:
 	case ch <- ri:
 	default:
-		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to keep up with incoming traffic - we recommend increasing number of replicate object workers with `mc admin config set api replication_workers=%d`", p.suggestedWorkers(false)), replicationSubsystem)
+		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to keep up with incoming traffic - we recommend increasing number of replicate object workers with `mc admin config set api replication_workers=%d`", p.suggestedWorkers(false)), string(replicationSubsystem))
 	}
 }
 
@@ -1555,7 +1564,7 @@ func (p *ReplicationPool) queueReplicaDeleteTask(doi DeletedObjectReplicationInf
 		})
 	case ch <- doi:
 	default:
-		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to keep up with incoming deletes - we recommend increasing number of replicate workers with `mc admin config set api replication_workers=%d`", p.suggestedWorkers(false)), replicationSubsystem)
+		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to keep up with incoming deletes - we recommend increasing number of replicate workers with `mc admin config set api replication_workers=%d`", p.suggestedWorkers(false)), string(replicationSubsystem))
 	}
 }
 
@@ -1659,7 +1668,7 @@ func proxyHeadToRepTarget(ctx context.Context, bucket, object string, rs *HTTPRa
 	}
 	for _, t := range proxyTargets.Targets {
 		tgt = globalBucketTargetSys.GetRemoteTargetClient(ctx, t.Arn)
-		if tgt == nil || tgt.IsOffline() {
+		if tgt == nil || globalBucketTargetSys.isOffline(tgt.EndpointURL()) {
 			continue
 		}
 		// if proxying explicitly disabled on remote target
@@ -1805,6 +1814,7 @@ func (c replicationConfig) Resync(ctx context.Context, oi ObjectInfo, dsc *Repli
 	objInfo.VersionPurgeStatusInternal = ""
 	objInfo.ReplicationStatus = ""
 	objInfo.VersionPurgeStatus = ""
+	delete(objInfo.UserDefined, xhttp.AmzBucketReplicationStatus)
 	resyncdsc := mustReplicate(ctx, oi.Bucket, oi.Name, getMustReplicateOptions(objInfo, replication.ExistingObjectReplicationType, ObjectOptions{}))
 	dsc = &resyncdsc
 	return c.resync(oi, dsc, tgtStatuses)
@@ -1991,7 +2001,7 @@ func (p *ReplicationPool) updateResyncStatus(ctx context.Context, objectAPI Obje
 				if updt {
 					brs.LastUpdate = now
 					if err := saveResyncStatus(ctx, bucket, brs, objectAPI); err != nil {
-						logger.LogIf(ctx, fmt.Errorf("Could not save resync metadata to disk for %s - %w", bucket, err))
+						logger.LogIf(ctx, fmt.Errorf("Could not save resync metadata to drive for %s - %w", bucket, err))
 						continue
 					}
 				}
@@ -2052,9 +2062,9 @@ func resyncBucket(ctx context.Context, bucket, arn string, heal bool, objectAPI
 		return
 	}
 
-	// Walk through all object versions - note ascending order of walk needed to ensure delete marker replicated to
-	// target after object version is first created.
-	if err := objectAPI.Walk(ctx, bucket, "", objInfoCh, ObjectOptions{WalkAscending: true}); err != nil {
+	// Walk through all object versions - Walk() is always in ascending order needed to ensure
+	// delete marker replicated to target after object version is first created.
+	if err := objectAPI.Walk(ctx, bucket, "", objInfoCh, ObjectOptions{}); err != nil {
 		logger.LogIf(ctx, err)
 		return
 	}
@@ -2385,3 +2395,279 @@ func getReplicationDiff(ctx context.Context, objAPI ObjectLayer, bucket string,
 	}()
 	return diffCh, nil
 }
+
+// QueueReplicationHeal is a wrapper for queueReplicationHeal
+func QueueReplicationHeal(ctx context.Context, bucket string, oi ObjectInfo) {
+	// un-versioned case
+	if oi.VersionID == "" {
+		return
+	}
+	rcfg, _, _ := globalBucketMetadataSys.GetReplicationConfig(ctx, bucket)
+	tgts, _ := globalBucketTargetSys.ListBucketTargets(ctx, bucket)
+	queueReplicationHeal(ctx, bucket, oi, replicationConfig{
+		Config:  rcfg,
+		remotes: tgts,
+	})
+}
+
+// queueReplicationHeal enqueues objects that failed replication OR eligible for resyncing through
+// an ongoing resync operation or via existing objects replication configuration setting.
+func queueReplicationHeal(ctx context.Context, bucket string, oi ObjectInfo, rcfg replicationConfig) (roi ReplicateObjectInfo) {
+	// un-versioned case
+	if oi.VersionID == "" {
+		return roi
+	}
+
+	if rcfg.Config == nil || rcfg.remotes == nil {
+		return roi
+	}
+	roi = getHealReplicateObjectInfo(oi, rcfg)
+	if !roi.Dsc.ReplicateAny() {
+		return
+	}
+	// early return if replication already done, otherwise we need to determine if this
+	// version is an existing object that needs healing.
+	if oi.ReplicationStatus == replication.Completed && oi.VersionPurgeStatus.Empty() && !roi.ExistingObjResync.mustResync() {
+		return
+	}
+
+	if roi.DeleteMarker || !roi.VersionPurgeStatus.Empty() {
+		versionID := ""
+		dmVersionID := ""
+		if roi.VersionPurgeStatus.Empty() {
+			dmVersionID = roi.VersionID
+		} else {
+			versionID = roi.VersionID
+		}
+
+		dv := DeletedObjectReplicationInfo{
+			DeletedObject: DeletedObject{
+				ObjectName:            roi.Name,
+				DeleteMarkerVersionID: dmVersionID,
+				VersionID:             versionID,
+				ReplicationState:      roi.getReplicationState(roi.Dsc.String(), versionID, true),
+				DeleteMarkerMTime:     DeleteMarkerMTime{roi.ModTime},
+				DeleteMarker:          roi.DeleteMarker,
+			},
+			Bucket:    roi.Bucket,
+			OpType:    replication.HealReplicationType,
+			EventType: ReplicateHealDelete,
+		}
+		// heal delete marker replication failure or versioned delete replication failure
+		if roi.ReplicationStatus == replication.Pending ||
+			roi.ReplicationStatus == replication.Failed ||
+			roi.VersionPurgeStatus == Failed || roi.VersionPurgeStatus == Pending {
+			globalReplicationPool.queueReplicaDeleteTask(dv)
+			return
+		}
+		// if replication status is Complete on DeleteMarker and existing object resync required
+		if roi.ExistingObjResync.mustResync() && (roi.ReplicationStatus == replication.Completed || roi.ReplicationStatus.Empty()) {
+			queueReplicateDeletesWrapper(dv, roi.ExistingObjResync)
+			return
+		}
+		return
+	}
+	if roi.ExistingObjResync.mustResync() {
+		roi.OpType = replication.ExistingObjectReplicationType
+	}
+	switch roi.ReplicationStatus {
+	case replication.Pending, replication.Failed:
+		roi.EventType = ReplicateHeal
+		globalReplicationPool.queueReplicaTask(roi)
+		return
+	}
+	if roi.ExistingObjResync.mustResync() {
+		roi.EventType = ReplicateExisting
+		globalReplicationPool.queueReplicaTask(roi)
+	}
+	return
+}
+
+const mrfTimeInterval = 5 * time.Minute
+
+func (p *ReplicationPool) persistMRF() {
+	var mu sync.Mutex
+	entries := make(map[string]MRFReplicateEntry)
+	mTimer := time.NewTimer(mrfTimeInterval)
+	defer mTimer.Stop()
+	saveMRFToDisk := func(drain bool) {
+		mu.Lock()
+		defer mu.Unlock()
+		if len(entries) == 0 {
+			return
+		}
+		cctx := p.ctx
+		if drain {
+			cctx = context.Background()
+			// drain all mrf entries and save to disk
+			for e := range p.mrfSaveCh {
+				entries[e.versionID] = e
+			}
+		}
+		if err := p.saveMRFEntries(cctx, entries); err != nil {
+			logger.LogOnceIf(p.ctx, fmt.Errorf("Unable to persist replication failures to disk:%w", err), string(replicationSubsystem))
+		}
+		entries = make(map[string]MRFReplicateEntry)
+		return
+	}
+	for {
+		select {
+		case <-mTimer.C:
+			saveMRFToDisk(false)
+			mTimer.Reset(mrfTimeInterval)
+		case <-p.ctx.Done():
+			close(p.mrfSaveCh)
+			saveMRFToDisk(true)
+			return
+		case e, ok := <-p.mrfSaveCh:
+			if !ok {
+				return
+			}
+			var cnt int
+			mu.Lock()
+			entries[e.versionID] = e
+			cnt = len(entries)
+			mu.Unlock()
+			if cnt >= cap(p.mrfSaveCh) || len(p.mrfSaveCh) >= int(0.8*float32(cap(p.mrfSaveCh))) {
+				saveMRFToDisk(true)
+			}
+		}
+	}
+}
+
+func (p *ReplicationPool) queueMRFSave(entry MRFReplicateEntry) {
+	if p == nil {
+		return
+	}
+	select {
+	case <-GlobalContext.Done():
+		return
+	case p.mrfSaveCh <- entry:
+	}
+}
+
+// save mrf entries to mrf_<uuid>.bin
+func (p *ReplicationPool) saveMRFEntries(ctx context.Context, entries map[string]MRFReplicateEntry) error {
+	if len(entries) == 0 {
+		return nil
+	}
+	v := MRFReplicateEntries{
+		Entries: entries,
+		Version: mrfMetaVersionV1,
+	}
+	data := make([]byte, 4, v.Msgsize()+4)
+
+	// Initialize the resync meta header.
+	binary.LittleEndian.PutUint16(data[0:2], resyncMetaFormat)
+	binary.LittleEndian.PutUint16(data[2:4], resyncMetaVersion)
+
+	buf, err := v.MarshalMsg(data)
+	if err != nil {
+		return err
+	}
+
+	configFile := path.Join(replicationMRFDir, mustGetUUID()+".bin")
+	err = saveConfig(ctx, p.objLayer, configFile, buf)
+	return err
+}
+
+// load mrf entries from disk
+func (p *ReplicationPool) loadMRF(fileName string) (re MRFReplicateEntries, e error) {
+	data, err := readConfig(p.ctx, p.objLayer, fileName)
+	if err != nil && err != errConfigNotFound {
+		return re, err
+	}
+	if len(data) == 0 {
+		// Seems to be empty.
+		return re, nil
+	}
+	if len(data) <= 4 {
+		return re, fmt.Errorf("replication mrf: no data")
+	}
+	// Read resync meta header
+	switch binary.LittleEndian.Uint16(data[0:2]) {
+	case mrfMetaFormat:
+	default:
+		return re, fmt.Errorf("replication mrf: unknown format: %d", binary.LittleEndian.Uint16(data[0:2]))
+	}
+	switch binary.LittleEndian.Uint16(data[2:4]) {
+	case mrfMetaVersion:
+	default:
+		return re, fmt.Errorf("replication mrf: unknown version: %d", binary.LittleEndian.Uint16(data[2:4]))
+	}
+	// OK, parse data.
+	if _, err = re.UnmarshalMsg(data[4:]); err != nil {
+		return re, err
+	}
+
+	switch re.Version {
+	case mrfMetaVersionV1:
+	default:
+		return re, fmt.Errorf("unexpected mrf meta version: %d", re.Version)
+	}
+	return re, nil
+}
+
+func (p *ReplicationPool) processMRF() {
+	if p == nil || p.objLayer == nil {
+		return
+	}
+	pTimer := time.NewTimer(mrfTimeInterval)
+	defer pTimer.Stop()
+	for {
+		select {
+		case <-pTimer.C:
+			// skip healing if all targets are offline
+			var offlineCnt int
+			tgts := globalBucketTargetSys.ListTargets(p.ctx, "", "")
+			for _, tgt := range tgts {
+				if globalBucketTargetSys.isOffline(tgt.URL()) {
+					offlineCnt++
+				}
+			}
+			if len(tgts) == offlineCnt {
+				pTimer.Reset(mrfTimeInterval)
+				continue
+			}
+			objCh := make(chan ObjectInfo)
+			cctx, cancelFn := context.WithCancel(p.ctx)
+			if err := p.objLayer.Walk(cctx, minioMetaBucket, replicationMRFDir, objCh, ObjectOptions{}); err != nil {
+				pTimer.Reset(mrfTimeInterval)
+				cancelFn()
+				logger.LogIf(p.ctx, err)
+				continue
+			}
+			for item := range objCh {
+				if err := p.queueMRFHeal(item.Name); err == nil {
+					p.objLayer.DeleteObject(p.ctx, minioMetaBucket, item.Name, ObjectOptions{})
+				}
+			}
+			pTimer.Reset(mrfTimeInterval)
+			cancelFn()
+		case <-p.ctx.Done():
+			return
+		}
+	}
+}
+
+// process sends error logs to the heal channel for an attempt to heal replication.
+func (p *ReplicationPool) queueMRFHeal(file string) error {
+	if p == nil || p.objLayer == nil {
+		return errServerNotInitialized
+	}
+
+	mrfRec, err := p.loadMRF(file)
+	if err != nil {
+		return err
+	}
+	for vID, e := range mrfRec.Entries {
+		oi, err := p.objLayer.GetObjectInfo(p.ctx, e.Bucket, e.Object, ObjectOptions{
+			VersionID: vID,
+		})
+		if err != nil {
+			continue
+		}
+		QueueReplicationHeal(p.ctx, e.Bucket, oi)
+	}
+	return nil
+}

cmd/bucket-targets.go

@@ -19,6 +19,8 @@ package cmd
 
 import (
 	"context"
+	"fmt"
+	"net/url"
 	"sync"
 	"time"
 
@@ -35,7 +37,9 @@ import (
 )
 
 const (
-	defaultHealthCheckDuration = 30 * time.Second
+	defaultHealthCheckDuration = 5 * time.Second
+	// default interval for reload of all remote target endpoints
+	defaultHealthCheckReloadDuration = 30 * time.Minute
 )
 
 // BucketTargetSys represents bucket targets subsystem
@@ -43,6 +47,113 @@ type BucketTargetSys struct {
 	sync.RWMutex
 	arnRemotesMap map[string]*TargetClient
 	targetsMap    map[string][]madmin.BucketTarget
+	hMutex        sync.RWMutex
+	hc            map[string]epHealth
+	hcClient      *madmin.AnonymousClient
+}
+
+// epHealth struct represents health of a replication target endpoint.
+type epHealth struct {
+	Endpoint string
+	Scheme   string
+	Online   bool
+}
+
+// isOffline returns current liveness result of remote target. Add endpoint to
+// healthcheck map if missing and default to online status
+func (sys *BucketTargetSys) isOffline(ep *url.URL) bool {
+	sys.hMutex.RLock()
+	defer sys.hMutex.RUnlock()
+	if h, ok := sys.hc[ep.Host]; ok {
+		return !h.Online
+	}
+	go sys.initHC(ep)
+	return false
+}
+
+func (sys *BucketTargetSys) initHC(ep *url.URL) {
+	sys.hMutex.Lock()
+	sys.hc[ep.Host] = epHealth{
+		Endpoint: ep.Host,
+		Scheme:   ep.Scheme,
+		Online:   true,
+	}
+	sys.hMutex.Unlock()
+}
+
+// newHCClient initializes an anonymous client for performing health check on the remote endpoints
+func newHCClient() *madmin.AnonymousClient {
+	clnt, e := madmin.NewAnonymousClientNoEndpoint()
+	if e != nil {
+		logger.LogOnceIf(GlobalContext, fmt.Errorf("WARNING: Unable to initialize health check client"), string(replicationSubsystem))
+		return nil
+	}
+	clnt.SetCustomTransport(globalRemoteTargetTransport)
+	return clnt
+}
+
+// heartBeat performs liveness check on remote endpoints.
+func (sys *BucketTargetSys) heartBeat(ctx context.Context) {
+	hcTimer := time.NewTimer(defaultHealthCheckDuration)
+	defer hcTimer.Stop()
+	for {
+		select {
+		case <-hcTimer.C:
+			sys.hMutex.RLock()
+			var eps []madmin.ServerProperties
+			for _, ep := range sys.hc {
+				eps = append(eps, madmin.ServerProperties{Endpoint: ep.Endpoint, Scheme: ep.Scheme})
+			}
+			sys.hMutex.RUnlock()
+
+			if len(eps) > 0 {
+				cctx, cancel := context.WithTimeout(ctx, 30*time.Second)
+				defer cancel()
+				m := map[string]epHealth{}
+				for result := range sys.hcClient.Alive(cctx, madmin.AliveOpts{}, eps...) {
+					var online bool
+					if result.Error == nil {
+						online = result.Online
+					}
+					m[result.Endpoint.Host] = epHealth{
+						Endpoint: result.Endpoint.Host,
+						Scheme:   result.Endpoint.Scheme,
+						Online:   online,
+					}
+				}
+				sys.hMutex.Lock()
+				sys.hc = m
+				sys.hMutex.Unlock()
+			}
+			hcTimer.Reset(defaultHealthCheckDuration)
+		case <-ctx.Done():
+			return
+		}
+	}
+}
+
+// periodically rebuild the healthcheck map from list of targets to clear
+// out stale endpoints
+func (sys *BucketTargetSys) reloadHealthCheckers(ctx context.Context) {
+	m := make(map[string]epHealth)
+	tgts := sys.ListTargets(ctx, "", "")
+	for _, t := range tgts {
+		if _, ok := m[t.Endpoint]; !ok {
+			scheme := "http"
+			if t.Secure {
+				scheme = "https"
+			}
+			m[t.Endpoint] = epHealth{
+				Online:   true,
+				Endpoint: t.Endpoint,
+				Scheme:   scheme,
+			}
+		}
+	}
+	sys.hMutex.Lock()
+	// swap out the map
+	sys.hc = m
+	sys.hMutex.Unlock()
 }
 
 // ListTargets lists bucket targets across tenant or for individual bucket, and returns
@@ -91,9 +202,6 @@ func (sys *BucketTargetSys) Delete(bucket string) {
 		return
 	}
 	for _, t := range tgts {
-		if tgt, ok := sys.arnRemotesMap[t.Arn]; ok && tgt.healthCancelFn != nil {
-			tgt.healthCancelFn()
-		}
 		delete(sys.arnRemotesMap, t.Arn)
 	}
 	delete(sys.targetsMap, bucket)
@@ -116,7 +224,7 @@ func (sys *BucketTargetSys) SetTarget(ctx context.Context, bucket string, tgt *m
 		if minio.ToErrorResponse(err).Code == "NoSuchBucket" {
 			return BucketRemoteTargetNotFound{Bucket: tgt.TargetBucket}
 		}
-		return BucketRemoteConnectionErr{Bucket: tgt.TargetBucket, Err: err}
+		return RemoteTargetConnectionErr{Bucket: tgt.TargetBucket, Err: err}
 	}
 	if tgt.Type == madmin.ReplicationService {
 		if !globalBucketVersioningSys.Enabled(bucket) {
@@ -124,7 +232,7 @@ func (sys *BucketTargetSys) SetTarget(ctx context.Context, bucket string, tgt *m
 		}
 		vcfg, err := clnt.GetBucketVersioning(ctx, tgt.TargetBucket)
 		if err != nil {
-			return BucketRemoteConnectionErr{Bucket: tgt.TargetBucket, Err: err}
+			return RemoteTargetConnectionErr{Bucket: tgt.TargetBucket, Err: err}
 		}
 		if vcfg.Status != string(versioning.Enabled) {
 			return BucketRemoteTargetNotVersioned{Bucket: tgt.TargetBucket}
@@ -152,10 +260,7 @@ func (sys *BucketTargetSys) SetTarget(ctx context.Context, bucket string, tgt *m
 	if !found && !update {
 		newtgts = append(newtgts, *tgt)
 	}
-	// cancel health check for previous target client to avoid leak.
-	if prevClnt, ok := sys.arnRemotesMap[tgt.Arn]; ok && prevClnt.healthCancelFn != nil {
-		prevClnt.healthCancelFn()
-	}
+
 	sys.targetsMap[bucket] = newtgts
 	sys.arnRemotesMap[tgt.Arn] = clnt
 	sys.updateBandwidthLimit(bucket, tgt.BandwidthLimit)
@@ -227,9 +332,6 @@ func (sys *BucketTargetSys) RemoveTarget(ctx context.Context, bucket, arnStr str
 		return BucketRemoteTargetNotFound{Bucket: bucket}
 	}
 	sys.targetsMap[bucket] = targets
-	if tgt, ok := sys.arnRemotesMap[arnStr]; ok && tgt.healthCancelFn != nil {
-		tgt.healthCancelFn()
-	}
 	delete(sys.arnRemotesMap, arnStr)
 	sys.updateBandwidthLimit(bucket, 0)
 	return nil
@@ -258,11 +360,29 @@ func (sys *BucketTargetSys) GetRemoteBucketTargetByArn(ctx context.Context, buck
 }
 
 // NewBucketTargetSys - creates new replication system.
-func NewBucketTargetSys() *BucketTargetSys {
-	return &BucketTargetSys{
+func NewBucketTargetSys(ctx context.Context) *BucketTargetSys {
+	sys := &BucketTargetSys{
 		arnRemotesMap: make(map[string]*TargetClient),
 		targetsMap:    make(map[string][]madmin.BucketTarget),
+		hc:            make(map[string]epHealth),
+		hcClient:      newHCClient(),
 	}
+	// reload healthcheck endpoints map periodically to remove stale endpoints from the map.
+	go func() {
+		rTimer := time.NewTimer(defaultHealthCheckReloadDuration)
+		defer rTimer.Stop()
+		for {
+			select {
+			case <-rTimer.C:
+				sys.reloadHealthCheckers(ctx)
+				rTimer.Reset(defaultHealthCheckReloadDuration)
+			case <-ctx.Done():
+				return
+			}
+		}
+	}()
+	go sys.heartBeat(ctx)
+	return sys
 }
 
 // UpdateAllTargets updates target to reflect metadata updates
@@ -276,9 +396,6 @@ func (sys *BucketTargetSys) UpdateAllTargets(bucket string, tgts *madmin.BucketT
 	// Remove existingtarget and arn association
 	if tgts, ok := sys.targetsMap[bucket]; ok {
 		for _, t := range tgts {
-			if tgt, ok := sys.arnRemotesMap[t.Arn]; ok && tgt.healthCancelFn != nil {
-				tgt.healthCancelFn()
-			}
 			delete(sys.arnRemotesMap, t.Arn)
 		}
 		delete(sys.targetsMap, bucket)
@@ -345,10 +462,6 @@ func (sys *BucketTargetSys) getRemoteTargetClient(tcfg *madmin.BucketTarget) (*T
 	if tcfg.HealthCheckDuration >= 1 { // require minimum health check duration of 1 sec.
 		hcDuration = tcfg.HealthCheckDuration
 	}
-	cancelFn, err := api.HealthCheck(hcDuration)
-	if err != nil {
-		return nil, err
-	}
 	tc := &TargetClient{
 		Client:              api,
 		healthCheckDuration: hcDuration,
@@ -356,9 +469,10 @@ func (sys *BucketTargetSys) getRemoteTargetClient(tcfg *madmin.BucketTarget) (*T
 		Bucket:              tcfg.TargetBucket,
 		StorageClass:        tcfg.StorageClass,
 		disableProxy:        tcfg.DisableProxy,
-		healthCancelFn:      cancelFn,
 		ARN:                 tcfg.Arn,
 		ResetID:             tcfg.ResetID,
+		Endpoint:            tcfg.Endpoint,
+		Secure:              tcfg.Secure,
 	}
 	return tc, nil
 }
@@ -432,7 +546,8 @@ type TargetClient struct {
 	replicateSync       bool
 	StorageClass        string // storage class on remote
 	disableProxy        bool
-	healthCancelFn      context.CancelFunc // cancellation function for client healthcheck
-	ARN                 string             // ARN to uniquely identify remote target
+	ARN                 string // ARN to uniquely identify remote target
 	ResetID             string
+	Endpoint            string
+	Secure              bool
 }

cmd/bucket-versioning-handler.go

@@ -143,7 +143,7 @@ func (api objectAPIHandlers) GetBucketVersioningHandler(w http.ResponseWriter, r
 	}
 
 	// Check if bucket exists.
-	if _, err := objectAPI.GetBucketInfo(ctx, bucket); err != nil {
+	if _, err := objectAPI.GetBucketInfo(ctx, bucket, BucketOptions{}); err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
 	}

cmd/common-main.go

@@ -264,7 +264,7 @@ func initConsoleServer() (*restapi.Server, error) {
 		return nil, err
 	}
 
-	api := operations.NewConsoleAPI(swaggerSpec, buildOpenIDConsoleConfig())
+	api := operations.NewConsoleAPI(swaggerSpec)
 
 	if !serverDebugLog {
 		// Disable console logging if server debug log is not enabled
@@ -275,6 +275,12 @@ func initConsoleServer() (*restapi.Server, error) {
 		api.Logger = noLog
 	}
 
+	// Pass in console application config. This needs to happen before the
+	// ConfigureAPI() call.
+	restapi.GlobalMinIOConfig = restapi.MinIOConfig{
+		OpenIDProviders: buildOpenIDConsoleConfig(),
+	}
+
 	server := restapi.NewServer(api)
 	// register all APIs
 	server.ConfigureAPI()
@@ -502,8 +508,7 @@ func parsEnvEntry(envEntry string) (envKV, error) {
 			Skip: true,
 		}, nil
 	}
-	const envSeparator = "="
-	envTokens := strings.SplitN(strings.TrimSpace(strings.TrimPrefix(envEntry, "export")), envSeparator, 2)
+	envTokens := strings.SplitN(strings.TrimSpace(strings.TrimPrefix(envEntry, "export")), config.EnvSeparator, 2)
 	if len(envTokens) != 2 {
 		return envKV{}, fmt.Errorf("envEntry malformed; %s, expected to be of form 'KEY=value'", envEntry)
 	}
@@ -529,7 +534,7 @@ func parsEnvEntry(envEntry string) (envKV, error) {
 // the environment values from a file, in the form "key, value".
 // in a structured form.
 func minioEnvironFromFile(envConfigFile string) ([]envKV, error) {
-	f, err := os.Open(envConfigFile)
+	f, err := Open(envConfigFile)
 	if err != nil {
 		return nil, err
 	}
@@ -824,43 +829,56 @@ func handleKMSConfig() {
 				endpoints = append(endpoints, strings.Join(lbls, ""))
 			}
 		}
-		// Manually load the certificate and private key into memory.
-		// We need to check whether the private key is encrypted, and
-		// if so, decrypt it using the user-provided password.
-		certBytes, err := os.ReadFile(env.Get(config.EnvKESClientCert, ""))
-		if err != nil {
-			logger.Fatal(err, "Unable to load KES client certificate as specified by the shell environment")
-		}
-		keyBytes, err := os.ReadFile(env.Get(config.EnvKESClientKey, ""))
-		if err != nil {
-			logger.Fatal(err, "Unable to load KES client private key as specified by the shell environment")
-		}
-		privateKeyPEM, rest := pem.Decode(bytes.TrimSpace(keyBytes))
-		if len(rest) != 0 {
-			logger.Fatal(errors.New("private key contains additional data"), "Unable to load KES client private key as specified by the shell environment")
-		}
-		if x509.IsEncryptedPEMBlock(privateKeyPEM) {
-			keyBytes, err = x509.DecryptPEMBlock(privateKeyPEM, []byte(env.Get(config.EnvKESClientPassword, "")))
-			if err != nil {
-				logger.Fatal(err, "Unable to decrypt KES client private key as specified by the shell environment")
-			}
-			keyBytes = pem.EncodeToMemory(&pem.Block{Type: privateKeyPEM.Type, Bytes: keyBytes})
-		}
-		certificate, err := tls.X509KeyPair(certBytes, keyBytes)
-		if err != nil {
-			logger.Fatal(err, "Unable to load KES client certificate as specified by the shell environment")
-		}
 		rootCAs, err := certs.GetRootCAs(env.Get(config.EnvKESServerCA, globalCertsCADir.Get()))
 		if err != nil {
 			logger.Fatal(err, fmt.Sprintf("Unable to load X.509 root CAs for KES from %q", env.Get(config.EnvKESServerCA, globalCertsCADir.Get())))
 		}
 
+		loadX509KeyPair := func(certFile, keyFile string) (tls.Certificate, error) {
+			// Manually load the certificate and private key into memory.
+			// We need to check whether the private key is encrypted, and
+			// if so, decrypt it using the user-provided password.
+			certBytes, err := os.ReadFile(certFile)
+			if err != nil {
+				return tls.Certificate{}, fmt.Errorf("Unable to load KES client certificate as specified by the shell environment: %v", err)
+			}
+			keyBytes, err := os.ReadFile(keyFile)
+			if err != nil {
+				return tls.Certificate{}, fmt.Errorf("Unable to load KES client private key as specified by the shell environment: %v", err)
+			}
+			privateKeyPEM, rest := pem.Decode(bytes.TrimSpace(keyBytes))
+			if len(rest) != 0 {
+				return tls.Certificate{}, errors.New("Unable to load KES client private key as specified by the shell environment: private key contains additional data")
+			}
+			if x509.IsEncryptedPEMBlock(privateKeyPEM) {
+				keyBytes, err = x509.DecryptPEMBlock(privateKeyPEM, []byte(env.Get(config.EnvKESClientPassword, "")))
+				if err != nil {
+					return tls.Certificate{}, fmt.Errorf("Unable to decrypt KES client private key as specified by the shell environment: %v", err)
+				}
+				keyBytes = pem.EncodeToMemory(&pem.Block{Type: privateKeyPEM.Type, Bytes: keyBytes})
+			}
+			certificate, err := tls.X509KeyPair(certBytes, keyBytes)
+			if err != nil {
+				return tls.Certificate{}, fmt.Errorf("Unable to load KES client certificate as specified by the shell environment: %v", err)
+			}
+			return certificate, nil
+		}
+
+		reloadCertEvents := make(chan tls.Certificate, 1)
+		certificate, err := certs.NewCertificate(env.Get(config.EnvKESClientCert, ""), env.Get(config.EnvKESClientKey, ""), loadX509KeyPair)
+		if err != nil {
+			logger.Fatal(err, "Failed to load KES client certificate")
+		}
+		certificate.Watch(context.Background(), 15*time.Minute, syscall.SIGHUP)
+		certificate.Notify(reloadCertEvents)
+
 		defaultKeyID := env.Get(config.EnvKESKeyName, "")
 		KMS, err := kms.NewWithConfig(kms.Config{
-			Endpoints:    endpoints,
-			DefaultKeyID: defaultKeyID,
-			Certificate:  certificate,
-			RootCAs:      rootCAs,
+			Endpoints:        endpoints,
+			DefaultKeyID:     defaultKeyID,
+			Certificate:      certificate,
+			ReloadCertEvents: reloadCertEvents,
+			RootCAs:          rootCAs,
 		})
 		if err != nil {
 			logger.Fatal(err, "Unable to initialize a connection to KES as specified by the shell environment")
@@ -910,7 +928,7 @@ func getTLSConfig() (x509Certs []*x509.Certificate, manager *certs.Manager, secu
 	// Therefore, we read all filenames in the cert directory and check
 	// for each directory whether it contains a public.crt and private.key.
 	// If so, we try to add it to certificate manager.
-	root, err := os.Open(globalCertsDir.Get())
+	root, err := Open(globalCertsDir.Get())
 	if err != nil {
 		return nil, nil, false, err
 	}
@@ -929,7 +947,7 @@ func getTLSConfig() (x509Certs []*x509.Certificate, manager *certs.Manager, secu
 			continue
 		}
 		if file.Mode()&os.ModeSymlink == os.ModeSymlink {
-			file, err = os.Stat(filepath.Join(root.Name(), file.Name()))
+			file, err = Stat(filepath.Join(root.Name(), file.Name()))
 			if err != nil {
 				// not accessible ignore
 				continue

cmd/config-current.go

@@ -471,14 +471,12 @@ func lookupConfigs(s config.Config, objAPI ObjectLayer) {
 	}
 
 	if etcdCfg.Enabled {
-		if globalEtcdClient == nil {
-			globalEtcdClient, err = etcd.New(etcdCfg)
-			if err != nil {
-				if globalIsGateway {
-					logger.FatalIf(err, "Unable to initialize etcd config")
-				} else {
-					logger.LogIf(ctx, fmt.Errorf("Unable to initialize etcd config: %w", err))
-				}
+		globalEtcdClient, err = etcd.New(etcdCfg)
+		if err != nil {
+			if globalIsGateway {
+				logger.FatalIf(err, "Unable to initialize etcd config")
+			} else {
+				logger.LogIf(ctx, fmt.Errorf("Unable to initialize etcd config: %w", err))
 			}
 		}
 
@@ -486,7 +484,7 @@ func lookupConfigs(s config.Config, objAPI ObjectLayer) {
 			if globalDNSConfig != nil {
 				// if global DNS is already configured, indicate with a warning, incase
 				// users are confused.
-				logger.LogIf(ctx, fmt.Errorf("DNS store is already configured with %s, not using etcd for DNS store", globalDNSConfig))
+				logger.LogIf(ctx, fmt.Errorf("DNS store is already configured with %s, etcd is not used for DNS store", globalDNSConfig))
 			} else {
 				globalDNSConfig, err = dns.NewCoreDNS(etcdCfg.Config,
 					dns.DomainNames(globalDomainNames),
@@ -518,16 +516,6 @@ func lookupConfigs(s config.Config, objAPI ObjectLayer) {
 		logger.LogIf(ctx, fmt.Errorf("Invalid site configuration: %w", err))
 	}
 
-	apiConfig, err := api.LookupConfig(s[config.APISubSys][config.Default])
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Invalid api configuration: %w", err))
-	}
-
-	// Initialize remote instance transport once.
-	getRemoteInstanceTransportOnce.Do(func() {
-		getRemoteInstanceTransport = newGatewayHTTPTransport(apiConfig.RemoteTransportDeadline)
-	})
-
 	globalCacheConfig, err = cache.LookupConfig(s[config.CacheSubSys][config.Default])
 	if err != nil {
 		if globalIsGateway {
@@ -560,56 +548,19 @@ func lookupConfigs(s config.Config, objAPI ObjectLayer) {
 		logger.LogIf(ctx, fmt.Errorf("CRITICAL: enabling %s is not recommended in a production environment", xtls.EnvIdentityTLSSkipVerify))
 	}
 
-	globalOpenIDConfig, err = openid.LookupConfig(s,
-		NewGatewayHTTPTransport(), xhttp.DrainBody, globalSite.Region)
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Unable to initialize OpenID: %w", err))
-	}
-
-	globalLDAPConfig, err = xldap.Lookup(s[config.IdentityLDAPSubSys][config.Default],
-		globalRootCAs)
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Unable to parse LDAP configuration: %w", err))
-	}
-
-	authNPluginCfg, err := idplugin.LookupConfig(s[config.IdentityPluginSubSys][config.Default],
-		NewGatewayHTTPTransport(), xhttp.DrainBody, globalSite.Region)
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Unable to initialize AuthNPlugin: %w", err))
-	}
-	globalAuthNPlugin = idplugin.New(authNPluginCfg)
-
-	authZPluginCfg, err := polplugin.LookupConfig(s[config.PolicyPluginSubSys][config.Default],
-		NewGatewayHTTPTransport(), xhttp.DrainBody)
-	if err != nil {
-		logger.LogIf(ctx, fmt.Errorf("Unable to initialize AuthZPlugin: %w", err))
-	}
-	if authZPluginCfg.URL == nil {
-		opaCfg, err := opa.LookupConfig(s[config.PolicyOPASubSys][config.Default],
-			NewGatewayHTTPTransport(), xhttp.DrainBody)
-		if err != nil {
-			logger.LogIf(ctx, fmt.Errorf("Unable to initialize AuthZPlugin from legacy OPA config: %w", err))
-		} else {
-			authZPluginCfg.URL = opaCfg.URL
-			authZPluginCfg.AuthToken = opaCfg.AuthToken
-			authZPluginCfg.Transport = opaCfg.Transport
-			authZPluginCfg.CloseRespFn = opaCfg.CloseRespFn
-		}
-	}
-
-	setGlobalAuthZPlugin(polplugin.New(authZPluginCfg))
-
 	globalSubnetConfig, err = subnet.LookupConfig(s[config.SubnetSubSys][config.Default], globalProxyTransport)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("Unable to parse subnet configuration: %w", err))
 	}
 
-	globalConfigTargetList, err = notify.GetNotificationTargets(GlobalContext, s, NewGatewayHTTPTransport(), false)
+	transport := NewGatewayHTTPTransport()
+
+	globalConfigTargetList, err = notify.GetNotificationTargets(GlobalContext, s, transport, false)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("Unable to initialize notification target(s): %w", err))
 	}
 
-	globalEnvTargetList, err = notify.GetNotificationTargets(GlobalContext, newServerConfig(), NewGatewayHTTPTransport(), true)
+	globalEnvTargetList, err = notify.GetNotificationTargets(GlobalContext, newServerConfig(), transport, true)
 	if err != nil {
 		logger.LogIf(ctx, fmt.Errorf("Unable to initialize notification target(s): %w", err))
 	}
@@ -636,6 +587,11 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 			setDriveCounts = objAPI.SetDriveCounts()
 		}
 		globalAPIConfig.init(apiConfig, setDriveCounts)
+
+		// Initialize remote instance transport once.
+		getRemoteInstanceTransportOnce.Do(func() {
+			getRemoteInstanceTransport = newGatewayHTTPTransport(apiConfig.RemoteTransportDeadline)
+		})
 	case config.CompressionSubSys:
 		cmpCfg, err := compress.LookupConfig(s[config.CompressionSubSys][config.Default])
 		if err != nil {
@@ -678,8 +634,7 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 				loggerCfg.HTTP[n] = l
 			}
 		}
-		err = logger.UpdateSystemTargets(loggerCfg)
-		if err != nil {
+		if err = logger.UpdateSystemTargets(loggerCfg); err != nil {
 			logger.LogIf(ctx, fmt.Errorf("Unable to update logger webhook config: %w", err))
 		}
 	case config.AuditWebhookSubSys:
@@ -697,8 +652,7 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 			}
 		}
 
-		err = logger.UpdateAuditWebhookTargets(loggerCfg)
-		if err != nil {
+		if err = logger.UpdateAuditWebhookTargets(loggerCfg); err != nil {
 			logger.LogIf(ctx, fmt.Errorf("Unable to update audit webhook targets: %w", err))
 		}
 	case config.AuditKafkaSubSys:
@@ -712,8 +666,7 @@ func applyDynamicConfigForSubSys(ctx context.Context, objAPI ObjectLayer, s conf
 				loggerCfg.AuditKafka[n] = l
 			}
 		}
-		err = logger.UpdateAuditKafkaTargets(loggerCfg)
-		if err != nil {
+		if err = logger.UpdateAuditKafkaTargets(loggerCfg); err != nil {
 			logger.LogIf(ctx, fmt.Errorf("Unable to update audit kafka targets: %w", err))
 		}
 	case config.StorageClassSubSys:
@@ -852,7 +805,7 @@ func newSrvConfig(objAPI ObjectLayer) error {
 	globalServerConfigMu.Unlock()
 
 	// Save config into file.
-	return saveServerConfig(GlobalContext, objAPI, globalServerConfig)
+	return saveServerConfig(GlobalContext, objAPI, srvCfg)
 }
 
 func getValidConfig(objAPI ObjectLayer) (config.Config, error) {

cmd/config-encrypted.go

@@ -69,6 +69,12 @@ func migrateIAMConfigsEtcdToEncrypted(ctx context.Context, client *etcd.Client)
 		return err
 	}
 
+	// If backend doesn't have this file means we have already
+	// attempted then migration
+	if !encrypted {
+		return nil
+	}
+
 	if encrypted && GlobalKMS != nil {
 		stat, err := GlobalKMS.Stat(ctx)
 		if err != nil {
@@ -109,6 +115,8 @@ func migrateIAMConfigsEtcdToEncrypted(ctx context.Context, client *etcd.Client)
 							return fmt.Errorf("Decrypting IAM config failed %w, possibly credentials are incorrect", err)
 						}
 					}
+				} else {
+					return fmt.Errorf("Decrypting IAM config failed %w, possibly credentials are incorrect", err)
 				}
 			}
 			data = pdata
@@ -131,6 +139,7 @@ func migrateIAMConfigsEtcdToEncrypted(ctx context.Context, client *etcd.Client)
 	if encrypted && GlobalKMS != nil {
 		logger.Info("Migration of encrypted IAM config data completed. All data is now encrypted on etcd.")
 	}
+
 	return deleteKeyEtcd(ctx, client, backendEncryptedFile)
 }
 
@@ -146,44 +155,56 @@ func migrateConfigPrefixToEncrypted(objAPI ObjectLayer, encrypted bool) error {
 		logger.Info(fmt.Sprintf("Attempting to re-encrypt config, IAM users and policies on MinIO with %q (%s)", stat.DefaultKey, stat.Name))
 	}
 
-	var marker string
-	for {
-		res, err := objAPI.ListObjects(GlobalContext, minioMetaBucket, minioConfigPrefix, marker, "", maxObjectList)
+	results := make(chan ObjectInfo)
+	if err := objAPI.Walk(GlobalContext, minioMetaBucket, minioConfigPrefix, results, ObjectOptions{}); err != nil {
+		return err
+	}
+
+	for obj := range results {
+		data, err := readConfig(GlobalContext, objAPI, obj.Name)
 		if err != nil {
 			return err
 		}
-		for _, obj := range res.Objects {
-			data, err := readConfig(GlobalContext, objAPI, obj.Name)
+
+		if !utf8.Valid(data) {
+			pdata, err := madmin.DecryptData(globalActiveCred.String(), bytes.NewReader(data))
+			if err != nil {
+				if GlobalKMS != nil {
+					pdata, err = config.DecryptBytes(GlobalKMS, data, kms.Context{
+						minioMetaBucket: path.Join(minioMetaBucket, obj.Name),
+					})
+					if err != nil {
+						pdata, err = config.DecryptBytes(GlobalKMS, data, kms.Context{
+							minioMetaBucket: obj.Name,
+						})
+						if err != nil {
+							return fmt.Errorf("Decrypting IAM config failed %w, possibly credentials are incorrect", err)
+						}
+					}
+				} else {
+					return fmt.Errorf("Decrypting IAM config failed %w, possibly credentials are incorrect", err)
+				}
+			}
+			data = pdata
+		}
+
+		if GlobalKMS != nil {
+			data, err = config.EncryptBytes(GlobalKMS, data, kms.Context{
+				obj.Bucket: path.Join(obj.Bucket, obj.Name),
+			})
 			if err != nil {
 				return err
 			}
-
-			if !utf8.Valid(data) {
-				data, err = madmin.DecryptData(globalActiveCred.String(), bytes.NewReader(data))
-				if err != nil {
-					return fmt.Errorf("Decrypting config failed %w, possibly credentials are incorrect", err)
-				}
-			}
-			if GlobalKMS != nil {
-				data, err = config.EncryptBytes(GlobalKMS, data, kms.Context{
-					obj.Bucket: path.Join(obj.Bucket, obj.Name),
-				})
-				if err != nil {
-					return err
-				}
-			}
-			if err = saveConfig(GlobalContext, objAPI, obj.Name, data); err != nil {
-				return err
-			}
 		}
 
-		if !res.IsTruncated {
-			break
+		if err = saveConfig(GlobalContext, objAPI, obj.Name, data); err != nil {
+			return err
 		}
-		marker = res.NextMarker
 	}
+
 	if encrypted && GlobalKMS != nil {
 		logger.Info("Migration of encrypted config data completed. All config data is now encrypted.")
 	}
+
 	return deleteConfig(GlobalContext, objAPI, backendEncryptedFile)
 }

cmd/config-migrate_test.go

@@ -37,11 +37,7 @@ func TestServerConfigMigrateV1(t *testing.T) {
 	if err != nil {
 		t.Fatalf("Init Test config failed")
 	}
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
+	rootPath := t.TempDir()
 	globalConfigDir = &ConfigDir{path: rootPath}
 
 	globalObjLayerMutex.Lock()
@@ -74,13 +70,7 @@ func TestServerConfigMigrateV1(t *testing.T) {
 // Test if all migrate code returns nil when config file does not
 // exist
 func TestServerConfigMigrateInexistentConfig(t *testing.T) {
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
-
-	globalConfigDir = &ConfigDir{path: rootPath}
+	globalConfigDir = &ConfigDir{path: t.TempDir()}
 
 	if err := migrateV2ToV3(); err != nil {
 		t.Fatal("migrate v2 to v3 should succeed when no config file is found")
@@ -164,11 +154,7 @@ func TestServerConfigMigrateInexistentConfig(t *testing.T) {
 
 // Test if a config migration from v2 to v33 is successfully done
 func TestServerConfigMigrateV2toV33(t *testing.T) {
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
+	rootPath := t.TempDir()
 
 	globalConfigDir = &ConfigDir{path: rootPath}
 
@@ -234,11 +220,7 @@ func TestServerConfigMigrateV2toV33(t *testing.T) {
 
 // Test if all migrate code returns error with corrupted config files
 func TestServerConfigMigrateFaultyConfig(t *testing.T) {
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
+	rootPath := t.TempDir()
 
 	globalConfigDir = &ConfigDir{path: rootPath}
 	configPath := rootPath + SlashSeparator + minioConfigFile
@@ -331,35 +313,31 @@ func TestServerConfigMigrateFaultyConfig(t *testing.T) {
 
 // Test if all migrate code returns error with corrupted config files
 func TestServerConfigMigrateCorruptedConfig(t *testing.T) {
-	rootPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
+	rootPath := t.TempDir()
 
 	globalConfigDir = &ConfigDir{path: rootPath}
 	configPath := rootPath + SlashSeparator + minioConfigFile
 
 	for i := 3; i <= 17; i++ {
 		// Create a corrupted config file
-		if err = ioutil.WriteFile(configPath, []byte(fmt.Sprintf("{ \"version\":\"%d\", \"credential\": { \"accessKey\": 1 } }", i)),
+		if err := ioutil.WriteFile(configPath, []byte(fmt.Sprintf("{ \"version\":\"%d\", \"credential\": { \"accessKey\": 1 } }", i)),
 			0o644); err != nil {
 			t.Fatal("Unexpected error: ", err)
 		}
 
 		// Test different migrate versions and be sure they are returning an error
-		if err = migrateConfig(); err == nil {
+		if err := migrateConfig(); err == nil {
 			t.Fatal("migrateConfig() should fail with a corrupted json")
 		}
 	}
 
 	// Create a corrupted config file for version '2'.
-	if err = ioutil.WriteFile(configPath, []byte("{ \"version\":\"2\", \"credentials\": { \"accessKeyId\": 1 } }"), 0o644); err != nil {
+	if err := ioutil.WriteFile(configPath, []byte("{ \"version\":\"2\", \"credentials\": { \"accessKeyId\": 1 } }"), 0o644); err != nil {
 		t.Fatal("Unexpected error: ", err)
 	}
 
 	// Test different migrate versions and be sure they are returning an error
-	if err = migrateConfig(); err == nil {
+	if err := migrateConfig(); err == nil {
 		t.Fatal("migrateConfig() should fail with a corrupted json")
 	}
 }

cmd/data-scanner.go

@@ -64,7 +64,11 @@ const (
 var (
 	globalHealConfig heal.Config
 
-	dataScannerLeaderLockTimeout = newDynamicTimeout(30*time.Second, 10*time.Second)
+	dataScannerLeaderLockTimeout = newDynamicTimeoutWithOpts(dynamicTimeoutOpts{
+		timeout:       30 * time.Second,
+		minimum:       10 * time.Second,
+		retryInterval: time.Second,
+	})
 	// Sleeper values are updated when config is loaded.
 	scannerSleeper = newDynamicSleeper(10, 10*time.Second, true)
 	scannerCycle   = uatomic.NewDuration(dataScannerStartDelay)
@@ -1225,29 +1229,16 @@ func (i *scannerItem) objectPath() string {
 
 // healReplication will heal a scanned item that has failed replication.
 func (i *scannerItem) healReplication(ctx context.Context, o ObjectLayer, oi ObjectInfo, sizeS *sizeSummary) {
-	roi := getHealReplicateObjectInfo(oi, i.replication)
-	if !roi.Dsc.ReplicateAny() {
+	if oi.VersionID == "" {
 		return
 	}
-
+	if i.replication.Config == nil {
+		return
+	}
+	roi := queueReplicationHeal(ctx, oi.Bucket, oi, i.replication)
 	if oi.DeleteMarker || !oi.VersionPurgeStatus.Empty() {
-		// heal delete marker replication failure or versioned delete replication failure
-		if oi.ReplicationStatus == replication.Pending ||
-			oi.ReplicationStatus == replication.Failed ||
-			oi.VersionPurgeStatus == Failed || oi.VersionPurgeStatus == Pending {
-			i.healReplicationDeletes(ctx, o, roi)
-			return
-		}
-		// if replication status is Complete on DeleteMarker and existing object resync required
-		if roi.ExistingObjResync.mustResync() && (oi.ReplicationStatus == replication.Completed || oi.ReplicationStatus.Empty()) {
-			i.healReplicationDeletes(ctx, o, roi)
-			return
-		}
 		return
 	}
-	if roi.ExistingObjResync.mustResync() {
-		roi.OpType = replication.ExistingObjectReplicationType
-	}
 
 	if sizeS.replTargetStats == nil && len(roi.TargetStatuses) > 0 {
 		sizeS.replTargetStats = make(map[string]replTargetSizeSummary)
@@ -1277,52 +1268,9 @@ func (i *scannerItem) healReplication(ctx context.Context, o ObjectLayer, oi Obj
 	}
 
 	switch oi.ReplicationStatus {
-	case replication.Pending, replication.Failed:
-		roi.EventType = ReplicateHeal
-		globalReplicationPool.queueReplicaTask(roi)
-		return
 	case replication.Replica:
 		sizeS.replicaSize += oi.Size
 	}
-	if roi.ExistingObjResync.mustResync() {
-		roi.EventType = ReplicateExisting
-		globalReplicationPool.queueReplicaTask(roi)
-	}
-}
-
-// healReplicationDeletes will heal a scanned deleted item that failed to replicate deletes.
-func (i *scannerItem) healReplicationDeletes(ctx context.Context, o ObjectLayer, roi ReplicateObjectInfo) {
-	// handle soft delete and permanent delete failures here.
-	if roi.DeleteMarker || !roi.VersionPurgeStatus.Empty() {
-		versionID := ""
-		dmVersionID := ""
-		if roi.VersionPurgeStatus.Empty() {
-			dmVersionID = roi.VersionID
-		} else {
-			versionID = roi.VersionID
-		}
-
-		doi := DeletedObjectReplicationInfo{
-			DeletedObject: DeletedObject{
-				ObjectName:            roi.Name,
-				DeleteMarkerVersionID: dmVersionID,
-				VersionID:             versionID,
-				ReplicationState:      roi.getReplicationState(roi.Dsc.String(), versionID, true),
-				DeleteMarkerMTime:     DeleteMarkerMTime{roi.ModTime},
-				DeleteMarker:          roi.DeleteMarker,
-			},
-			Bucket:    roi.Bucket,
-			OpType:    replication.HealReplicationType,
-			EventType: ReplicateHealDelete,
-		}
-		if roi.ExistingObjResync.mustResync() {
-			doi.OpType = replication.ExistingObjectReplicationType
-			doi.EventType = ReplicateExistingDelete
-			queueReplicateDeletesWrapper(doi, roi.ExistingObjResync)
-			return
-		}
-		globalReplicationPool.queueReplicaDeleteTask(doi)
-	}
 }
 
 type dynamicSleeper struct {

cmd/data-update-tracker.go

@@ -227,7 +227,7 @@ func (d *dataUpdateTracker) load(ctx context.Context, drives ...string) {
 	for _, drive := range drives {
 
 		cacheFormatPath := pathJoin(drive, dataUpdateTrackerFilename)
-		f, err := os.Open(cacheFormatPath)
+		f, err := OpenFile(cacheFormatPath, readMode, 0o666)
 		if err != nil {
 			if osIsNotExist(err) {
 				continue

cmd/data-update-tracker_test.go

@@ -21,7 +21,6 @@ import (
 	"bytes"
 	"context"
 	"fmt"
-	"io/ioutil"
 	"math/rand"
 	"os"
 	"path"
@@ -100,15 +99,11 @@ func TestDataUpdateTracker(t *testing.T) {
 
 	dut.Current.bf = dut.newBloomFilter()
 
-	tmpDir, err := ioutil.TempDir("", "TestDataUpdateTracker")
+	tmpDir := t.TempDir()
+	err := os.MkdirAll(filepath.Dir(filepath.Join(tmpDir, dataUpdateTrackerFilename)), os.ModePerm)
 	if err != nil {
 		t.Fatal(err)
 	}
-	err = os.MkdirAll(filepath.Dir(filepath.Join(tmpDir, dataUpdateTrackerFilename)), os.ModePerm)
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(tmpDir)
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()
 	dut.start(ctx, tmpDir)

cmd/data-usage_test.go

@@ -35,12 +35,8 @@ type usageTestFile struct {
 }
 
 func TestDataUsageUpdate(t *testing.T) {
-	base, err := ioutil.TempDir("", "TestDataUsageUpdate")
-	if err != nil {
-		t.Skip(err)
-	}
+	base := t.TempDir()
 	const bucket = "bucket"
-	defer os.RemoveAll(base)
 	files := []usageTestFile{
 		{name: "rootfile", size: 10000},
 		{name: "rootfile2", size: 10000},
@@ -251,12 +247,8 @@ func TestDataUsageUpdate(t *testing.T) {
 }
 
 func TestDataUsageUpdatePrefix(t *testing.T) {
-	base, err := ioutil.TempDir("", "TestDataUpdateUsagePrefix")
-	if err != nil {
-		t.Skip(err)
-	}
+	base := t.TempDir()
 	scannerSleeper.Update(0, 0)
-	defer os.RemoveAll(base)
 	files := []usageTestFile{
 		{name: "bucket/rootfile", size: 10000},
 		{name: "bucket/rootfile2", size: 10000},
@@ -537,12 +529,8 @@ func generateUsageTestFiles(t *testing.T, base, bucket string, nFolders, nFiles,
 }
 
 func TestDataUsageCacheSerialize(t *testing.T) {
-	base, err := ioutil.TempDir("", "TestDataUsageCacheSerialize")
-	if err != nil {
-		t.Skip(err)
-	}
+	base := t.TempDir()
 	const bucket = "abucket"
-	defer os.RemoveAll(base)
 	files := []usageTestFile{
 		{name: "rootfile", size: 10000},
 		{name: "rootfile2", size: 10000},

cmd/decommetric_string.go

@@ -0,0 +1,25 @@
+// Code generated by "stringer -type=decomMetric -trimprefix=decomMetric erasure-server-pool-decom.go"; DO NOT EDIT.
+
+package cmd
+
+import "strconv"
+
+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[decomMetricDecommissionBucket-0]
+	_ = x[decomMetricDecommissionObject-1]
+	_ = x[decomMetricDecommissionRemoveObject-2]
+}
+
+const _decomMetric_name = "DecommissionBucketDecommissionObjectDecommissionRemoveObject"
+
+var _decomMetric_index = [...]uint8{0, 18, 36, 60}
+
+func (i decomMetric) String() string {
+	if i >= decomMetric(len(_decomMetric_index)-1) {
+		return "decomMetric(" + strconv.FormatInt(int64(i), 10) + ")"
+	}
+	return _decomMetric_name[_decomMetric_index[i]:_decomMetric_index[i+1]]
+}

cmd/disk-cache-backend.go

@@ -624,7 +624,7 @@ func (c *diskCache) saveMetadata(ctx context.Context, bucket, object string, met
 	if err := os.MkdirAll(cachedPath, 0o777); err != nil {
 		return err
 	}
-	f, err := os.OpenFile(metaPath, os.O_RDWR|os.O_CREATE, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|os.O_CREATE|writeMode, 0o666)
 	if err != nil {
 		return err
 	}
@@ -687,7 +687,7 @@ func (c *diskCache) updateMetadata(ctx context.Context, bucket, object, etag str
 	if err := os.MkdirAll(cachedPath, 0o777); err != nil {
 		return err
 	}
-	f, err := os.OpenFile(metaPath, os.O_RDWR, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|writeMode, 0o666)
 	if err != nil {
 		return err
 	}
@@ -1275,12 +1275,12 @@ func (c *diskCache) NewMultipartUpload(ctx context.Context, bucket, object, uID
 
 	cachePath := getMultipartCacheSHADir(c.dir, bucket, object)
 	uploadIDDir := path.Join(cachePath, uploadID)
-	if err := os.MkdirAll(uploadIDDir, 0o777); err != nil {
+	if err := mkdirAll(uploadIDDir, 0o777); err != nil {
 		return uploadID, err
 	}
 	metaPath := pathJoin(uploadIDDir, cacheMetaJSONFile)
 
-	f, err := os.OpenFile(metaPath, os.O_RDWR|os.O_CREATE, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|os.O_CREATE|writeMode, 0o666)
 	if err != nil {
 		return uploadID, err
 	}
@@ -1386,7 +1386,7 @@ func (c *diskCache) SavePartMetadata(ctx context.Context, bucket, object, upload
 	defer uploadLock.Unlock(ulkctx.Cancel)
 
 	metaPath := pathJoin(uploadDir, cacheMetaJSONFile)
-	f, err := os.OpenFile(metaPath, os.O_RDWR, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|writeMode, 0o666)
 	if err != nil {
 		return err
 	}
@@ -1465,7 +1465,7 @@ func newCachePartEncryptReader(ctx context.Context, bucket, object string, partI
 func (c *diskCache) uploadIDExists(bucket, object, uploadID string) (err error) {
 	mpartCachePath := getMultipartCacheSHADir(c.dir, bucket, object)
 	uploadIDDir := path.Join(mpartCachePath, uploadID)
-	if _, err := os.Stat(uploadIDDir); err != nil {
+	if _, err := Stat(uploadIDDir); err != nil {
 		return err
 	}
 	return nil
@@ -1564,7 +1564,7 @@ func (c *diskCache) CompleteMultipartUpload(ctx context.Context, bucket, object,
 	uploadMeta.Hits++
 	metaPath := pathJoin(uploadIDDir, cacheMetaJSONFile)
 
-	f, err := os.OpenFile(metaPath, os.O_RDWR|os.O_CREATE, 0o666)
+	f, err := OpenFile(metaPath, os.O_RDWR|os.O_CREATE|writeMode, 0o666)
 	if err != nil {
 		return oi, err
 	}
@@ -1634,7 +1634,7 @@ func (c *diskCache) cleanupStaleUploads(ctx context.Context) {
 			readDirFn(pathJoin(c.dir, minioMetaBucket, cacheMultipartDir), func(shaDir string, typ os.FileMode) error {
 				return readDirFn(pathJoin(c.dir, minioMetaBucket, cacheMultipartDir, shaDir), func(uploadIDDir string, typ os.FileMode) error {
 					uploadIDPath := pathJoin(c.dir, minioMetaBucket, cacheMultipartDir, shaDir, uploadIDDir)
-					fi, err := os.Stat(uploadIDPath)
+					fi, err := Stat(uploadIDPath)
 					if err != nil {
 						return nil
 					}
@@ -1649,8 +1649,8 @@ func (c *diskCache) cleanupStaleUploads(ctx context.Context) {
 			readDirFn(pathJoin(c.dir, minioMetaBucket, cacheWritebackDir), func(shaDir string, typ os.FileMode) error {
 				wbdir := pathJoin(c.dir, minioMetaBucket, cacheWritebackDir, shaDir)
 				cachedir := pathJoin(c.dir, shaDir)
-				if _, err := os.Stat(cachedir); os.IsNotExist(err) {
-					fi, err := os.Stat(wbdir)
+				if _, err := Stat(cachedir); os.IsNotExist(err) {
+					fi, err := Stat(wbdir)
 					if err != nil {
 						return nil
 					}

cmd/disk-cache.go

@@ -607,12 +607,12 @@ func newCache(config cache.Config) ([]*diskCache, bool, error) {
 				warningMsg = fmt.Sprintf("Invalid cache dir %s err : %s", dir, err.Error())
 			}
 			if rootDsk {
-				warningMsg = fmt.Sprintf("cache dir cannot be part of root disk: %s", dir)
+				warningMsg = fmt.Sprintf("cache dir cannot be part of root drive: %s", dir)
 			}
 		}
 
 		if err := checkAtimeSupport(dir); err != nil {
-			return nil, false, fmt.Errorf("Atime support required for disk caching, atime check failed with %w", err)
+			return nil, false, fmt.Errorf("Atime support required for drive caching, atime check failed with %w", err)
 		}
 
 		cache, err := newDiskCache(ctx, dir, config)
@@ -622,7 +622,7 @@ func newCache(config cache.Config) ([]*diskCache, bool, error) {
 		caches = append(caches, cache)
 	}
 	if warningMsg != "" {
-		logger.Info(color.Yellow(fmt.Sprintf("WARNING: Usage of root disk for disk caching is deprecated: %s", warningMsg)))
+		logger.Info(color.Yellow(fmt.Sprintf("WARNING: Usage of root drive for drive caching is deprecated: %s", warningMsg)))
 	}
 	return caches, migrating, nil
 }

cmd/dummy-handlers.go

@@ -52,7 +52,7 @@ func (api objectAPIHandlers) GetBucketWebsiteHandler(w http.ResponseWriter, r *h
 	}
 
 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -84,7 +84,7 @@ func (api objectAPIHandlers) GetBucketAccelerateHandler(w http.ResponseWriter, r
 	}
 
 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -117,7 +117,7 @@ func (api objectAPIHandlers) GetBucketRequestPaymentHandler(w http.ResponseWrite
 	}
 
 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -151,7 +151,7 @@ func (api objectAPIHandlers) GetBucketLoggingHandler(w http.ResponseWriter, r *h
 	}
 
 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return
@@ -189,7 +189,7 @@ func (api objectAPIHandlers) GetBucketCorsHandler(w http.ResponseWriter, r *http
 	}
 
 	// Validate if bucket exists, before proceeding further...
-	_, err := objAPI.GetBucketInfo(ctx, bucket)
+	_, err := objAPI.GetBucketInfo(ctx, bucket, BucketOptions{})
 	if err != nil {
 		writeErrorResponse(ctx, w, toAPIError(ctx, err), r.URL)
 		return

cmd/dynamic-timeouts.go

@@ -34,11 +34,24 @@ const (
 
 // timeouts that are dynamically adapted based on actual usage results
 type dynamicTimeout struct {
-	timeout int64
-	minimum int64
-	entries int64
-	log     [dynamicTimeoutLogSize]time.Duration
-	mutex   sync.Mutex
+	timeout       int64
+	minimum       int64
+	entries       int64
+	log           [dynamicTimeoutLogSize]time.Duration
+	mutex         sync.Mutex
+	retryInterval time.Duration
+}
+
+type dynamicTimeoutOpts struct {
+	timeout       time.Duration
+	minimum       time.Duration
+	retryInterval time.Duration
+}
+
+func newDynamicTimeoutWithOpts(opts dynamicTimeoutOpts) *dynamicTimeout {
+	dt := newDynamicTimeout(opts.timeout, opts.minimum)
+	dt.retryInterval = opts.retryInterval
+	return dt
 }
 
 // newDynamicTimeout returns a new dynamic timeout initialized with timeout value
@@ -57,6 +70,10 @@ func (dt *dynamicTimeout) Timeout() time.Duration {
 	return time.Duration(atomic.LoadInt64(&dt.timeout))
 }
 
+func (dt *dynamicTimeout) RetryInterval() time.Duration {
+	return dt.retryInterval
+}
+
 // LogSuccess logs the duration of a successful action that
 // did not hit the timeout
 func (dt *dynamicTimeout) LogSuccess(duration time.Duration) {

cmd/encryption-v1.go

@@ -82,7 +82,7 @@ func (o *MultipartInfo) KMSKeyID() string { return kmsKeyIDFromMetadata(o.UserDe
 // metadata, if any. It returns an empty ID if no key ID is
 // present.
 func kmsKeyIDFromMetadata(metadata map[string]string) string {
-	const ARNPrefix = "arn:aws:kms:"
+	const ARNPrefix = crypto.ARNPrefix
 	if len(metadata) == 0 {
 		return ""
 	}

cmd/endpoint-ellipses.go

@@ -156,7 +156,7 @@ func getSetIndexes(args []string, totalSizes []uint64, customSetDriveCount uint6
 
 	setCounts := possibleSetCounts(commonSize)
 	if len(setCounts) == 0 {
-		msg := fmt.Sprintf("Incorrect number of endpoints provided %s, number of disks %d is not divisible by any supported erasure set sizes %d", args, commonSize, setSizes)
+		msg := fmt.Sprintf("Incorrect number of endpoints provided %s, number of drives %d is not divisible by any supported erasure set sizes %d", args, commonSize, setSizes)
 		return nil, config.ErrInvalidNumberOfErasureEndpoints(nil).Msg(msg)
 	}
 
@@ -183,7 +183,7 @@ func getSetIndexes(args []string, totalSizes []uint64, customSetDriveCount uint6
 		setCounts = possibleSetCountsWithSymmetry(setCounts, argPatterns)
 
 		if len(setCounts) == 0 {
-			msg := fmt.Sprintf("No symmetric distribution detected with input endpoints provided %s, disks %d cannot be spread symmetrically by any supported erasure set sizes %d", args, commonSize, setSizes)
+			msg := fmt.Sprintf("No symmetric distribution detected with input endpoints provided %s, drives %d cannot be spread symmetrically by any supported erasure set sizes %d", args, commonSize, setSizes)
 			return nil, config.ErrInvalidNumberOfErasureEndpoints(nil).Msg(msg)
 		}
 
@@ -193,7 +193,7 @@ func getSetIndexes(args []string, totalSizes []uint64, customSetDriveCount uint6
 
 	// Check whether setSize is with the supported range.
 	if !isValidSetSize(setSize) {
-		msg := fmt.Sprintf("Incorrect number of endpoints provided %s, number of disks %d is not divisible by any supported erasure set sizes %d", args, commonSize, setSizes)
+		msg := fmt.Sprintf("Incorrect number of endpoints provided %s, number of drives %d is not divisible by any supported erasure set sizes %d", args, commonSize, setSizes)
 		return nil, config.ErrInvalidNumberOfErasureEndpoints(nil).Msg(msg)
 	}
 

cmd/endpoint.go

@@ -292,6 +292,19 @@ func (l EndpointServerPools) LocalDisksPaths() []string {
 	return disks
 }
 
+// NLocalDisksPathsPerPool returns the disk paths of the local disks per pool
+func (l EndpointServerPools) NLocalDisksPathsPerPool() []int {
+	localDisksCount := make([]int, len(l))
+	for i, ep := range l {
+		for _, endpoint := range ep.Endpoints {
+			if endpoint.IsLocal {
+				localDisksCount[i]++
+			}
+		}
+	}
+	return localDisksCount
+}
+
 // FirstLocal returns true if the first endpoint is local.
 func (l EndpointServerPools) FirstLocal() bool {
 	return l[0].Endpoints[0].IsLocal
@@ -302,7 +315,7 @@ func (l EndpointServerPools) HTTPS() bool {
 	return l[0].Endpoints.HTTPS()
 }
 
-// NEndpoints - returns all nodes count
+// NEndpoints - returns number of endpoints
 func (l EndpointServerPools) NEndpoints() (count int) {
 	for _, ep := range l {
 		count += len(ep.Endpoints)

cmd/erasure-bucket.go

@@ -35,7 +35,7 @@ var bucketMetadataOpIgnoredErrs = append(bucketOpIgnoredErrs, errVolumeNotFound)
 // Bucket operations
 
 // MakeBucket - make a bucket.
-func (er erasureObjects) MakeBucketWithLocation(ctx context.Context, bucket string, opts BucketOptions) error {
+func (er erasureObjects) MakeBucketWithLocation(ctx context.Context, bucket string, opts MakeBucketOptions) error {
 	defer NSUpdated(bucket, slashSeparator)
 
 	// Verify if bucket is valid.
@@ -94,7 +94,7 @@ func undoDeleteBucket(storageDisks []StorageAPI, bucket string) {
 }
 
 // getBucketInfo - returns the BucketInfo from one of the load balanced disks.
-func (er erasureObjects) getBucketInfo(ctx context.Context, bucketName string) (bucketInfo BucketInfo, err error) {
+func (er erasureObjects) getBucketInfo(ctx context.Context, bucketName string, opts BucketOptions) (bucketInfo BucketInfo, err error) {
 	storageDisks := er.getDisks()
 
 	g := errgroup.WithNErrs(len(storageDisks))
@@ -108,9 +108,17 @@ func (er erasureObjects) getBucketInfo(ctx context.Context, bucketName string) (
 			}
 			volInfo, err := storageDisks[index].StatVol(ctx, bucketName)
 			if err != nil {
+				if opts.Deleted {
+					dvi, derr := storageDisks[index].StatVol(ctx, pathJoin(minioMetaBucket, bucketMetaPrefix, deletedBucketsPrefix, bucketName))
+					if derr != nil {
+						return err
+					}
+					bucketsInfo[index] = BucketInfo{Name: bucketName, Deleted: dvi.Created}
+					return nil
+				}
 				return err
 			}
-			bucketsInfo[index] = BucketInfo(volInfo)
+			bucketsInfo[index] = BucketInfo{Name: volInfo.Name, Created: volInfo.Created}
 			return nil
 		}, index)
 	}
@@ -131,8 +139,8 @@ func (er erasureObjects) getBucketInfo(ctx context.Context, bucketName string) (
 }
 
 // GetBucketInfo - returns BucketInfo for a bucket.
-func (er erasureObjects) GetBucketInfo(ctx context.Context, bucket string) (bi BucketInfo, e error) {
-	bucketInfo, err := er.getBucketInfo(ctx, bucket)
+func (er erasureObjects) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (bi BucketInfo, e error) {
+	bucketInfo, err := er.getBucketInfo(ctx, bucket, opts)
 	if err != nil {
 		return bi, toObjectErr(err, bucket)
 	}
@@ -191,12 +199,58 @@ func (er erasureObjects) DeleteBucket(ctx context.Context, bucket string, opts D
 		if purgedDangling {
 			err = nil
 		}
-
+		if opts.SRDeleteOp == MarkDelete {
+			er.markDelete(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, deletedBucketsPrefix, bucket))
+		}
 	}
 
 	return toObjectErr(err, bucket)
 }
 
+// markDelete creates a vol entry in .minio.sys/buckets/.deleted until site replication
+// syncs the delete to peers
+func (er erasureObjects) markDelete(ctx context.Context, bucket, prefix string) error {
+	storageDisks := er.getDisks()
+	g := errgroup.WithNErrs(len(storageDisks))
+	// Make a volume entry on all underlying storage disks.
+	for index := range storageDisks {
+		index := index
+		if storageDisks[index] == nil {
+			continue
+		}
+		g.Go(func() error {
+			if err := storageDisks[index].MakeVol(ctx, pathJoin(bucket, prefix)); err != nil {
+				if errors.Is(err, errVolumeExists) {
+					return nil
+				}
+				return err
+			}
+			return nil
+		}, index)
+	}
+	err := reduceWriteQuorumErrs(ctx, g.Wait(), bucketOpIgnoredErrs, er.defaultWQuorum())
+	return toObjectErr(err, bucket)
+}
+
+// purgeDelete deletes vol entry in .minio.sys/buckets/.deleted after site replication
+// syncs the delete to peers OR on a new MakeBucket call.
+func (er erasureObjects) purgeDelete(ctx context.Context, bucket, prefix string) error {
+	storageDisks := er.getDisks()
+	g := errgroup.WithNErrs(len(storageDisks))
+	// Make a volume entry on all underlying storage disks.
+	for index := range storageDisks {
+		index := index
+		g.Go(func() error {
+			if storageDisks[index] != nil {
+				return storageDisks[index].DeleteVol(ctx, pathJoin(bucket, prefix), true)
+			}
+			return errDiskNotFound
+		}, index)
+	}
+	err := reduceWriteQuorumErrs(ctx, g.Wait(), bucketOpIgnoredErrs, er.defaultWQuorum())
+	return toObjectErr(err, bucket)
+}
+
 // IsNotificationSupported returns whether bucket notification is applicable for this layer.
 func (er erasureObjects) IsNotificationSupported() bool {
 	return true

cmd/erasure-decode_test.go

@@ -85,19 +85,17 @@ var erasureDecodeTests = []struct {
 
 func TestErasureDecode(t *testing.T) {
 	for i, test := range erasureDecodeTests {
-		setup, err := newErasureTestSetup(test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
+		setup, err := newErasureTestSetup(t, test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
 		if err != nil {
 			t.Fatalf("Test %d: failed to create test setup: %v", i, err)
 		}
 		erasure, err := NewErasure(context.Background(), test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create ErasureStorage: %v", i, err)
 		}
 		disks := setup.disks
 		data := make([]byte, test.data)
 		if _, err = io.ReadFull(crand.Reader, data); err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to generate random test data: %v", i, err)
 		}
 
@@ -113,11 +111,9 @@ func TestErasureDecode(t *testing.T) {
 		n, err := erasure.Encode(context.Background(), bytes.NewReader(data), writers, buffer, erasure.dataBlocks+1)
 		closeBitrotWriters(writers)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create erasure test file: %v", i, err)
 		}
 		if n != test.data {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create erasure test file", i)
 		}
 		for i, w := range writers {
@@ -195,7 +191,6 @@ func TestErasureDecode(t *testing.T) {
 				}
 			}
 		}
-		setup.Remove()
 	}
 }
 
@@ -210,12 +205,11 @@ func TestErasureDecodeRandomOffsetLength(t *testing.T) {
 	dataBlocks := 7
 	parityBlocks := 7
 	blockSize := int64(1 * humanize.MiByte)
-	setup, err := newErasureTestSetup(dataBlocks, parityBlocks, blockSize)
+	setup, err := newErasureTestSetup(t, dataBlocks, parityBlocks, blockSize)
 	if err != nil {
 		t.Error(err)
 		return
 	}
-	defer setup.Remove()
 	disks := setup.disks
 	erasure, err := NewErasure(context.Background(), dataBlocks, parityBlocks, blockSize)
 	if err != nil {
@@ -288,11 +282,10 @@ func TestErasureDecodeRandomOffsetLength(t *testing.T) {
 // Benchmarks
 
 func benchmarkErasureDecode(data, parity, dataDown, parityDown int, size int64, b *testing.B) {
-	setup, err := newErasureTestSetup(data, parity, blockSizeV2)
+	setup, err := newErasureTestSetup(b, data, parity, blockSizeV2)
 	if err != nil {
 		b.Fatalf("failed to create test setup: %v", err)
 	}
-	defer setup.Remove()
 	disks := setup.disks
 	erasure, err := NewErasure(context.Background(), data, parity, blockSizeV2)
 	if err != nil {

cmd/erasure-encode_test.go

@@ -87,21 +87,19 @@ var erasureEncodeTests = []struct {
 
 func TestErasureEncode(t *testing.T) {
 	for i, test := range erasureEncodeTests {
-		setup, err := newErasureTestSetup(test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
+		setup, err := newErasureTestSetup(t, test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
 		if err != nil {
 			t.Fatalf("Test %d: failed to create test setup: %v", i, err)
 		}
 		disks := setup.disks
 		erasure, err := NewErasure(context.Background(), test.dataBlocks, test.onDisks-test.dataBlocks, test.blocksize)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create ErasureStorage: %v", i, err)
 		}
 		buffer := make([]byte, test.blocksize, 2*test.blocksize)
 
 		data := make([]byte, test.data)
 		if _, err = io.ReadFull(rand.Reader, data); err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to generate random test data: %v", i, err)
 		}
 		writers := make([]io.Writer, len(disks))
@@ -160,18 +158,16 @@ func TestErasureEncode(t *testing.T) {
 				}
 			}
 		}
-		setup.Remove()
 	}
 }
 
 // Benchmarks
 
 func benchmarkErasureEncode(data, parity, dataDown, parityDown int, size int64, b *testing.B) {
-	setup, err := newErasureTestSetup(data, parity, blockSizeV2)
+	setup, err := newErasureTestSetup(b, data, parity, blockSizeV2)
 	if err != nil {
 		b.Fatalf("failed to create test setup: %v", err)
 	}
-	defer setup.Remove()
 	erasure, err := NewErasure(context.Background(), data, parity, blockSizeV2)
 	if err != nil {
 		b.Fatalf("failed to create ErasureStorage: %v", err)

cmd/erasure-errors.go

@@ -20,10 +20,10 @@ package cmd
 import "errors"
 
 // errErasureReadQuorum - did not meet read quorum.
-var errErasureReadQuorum = errors.New("Read failed. Insufficient number of disks online")
+var errErasureReadQuorum = errors.New("Read failed. Insufficient number of drives online")
 
 // errErasureWriteQuorum - did not meet write quorum.
-var errErasureWriteQuorum = errors.New("Write failed. Insufficient number of disks online")
+var errErasureWriteQuorum = errors.New("Write failed. Insufficient number of drives online")
 
 // errNoHealRequired - returned when healing is attempted on a previously healed disks.
 var errNoHealRequired = errors.New("No healing is required")

cmd/erasure-heal_test.go

@@ -66,23 +66,21 @@ func TestErasureHeal(t *testing.T) {
 	for i, test := range erasureHealTests {
 		if test.offDisks < test.badStaleDisks {
 			// test case sanity check
-			t.Fatalf("Test %d: Bad test case - number of stale disks cannot be less than number of badstale disks", i)
+			t.Fatalf("Test %d: Bad test case - number of stale drives cannot be less than number of badstale drives", i)
 		}
 
 		// create some test data
-		setup, err := newErasureTestSetup(test.dataBlocks, test.disks-test.dataBlocks, test.blocksize)
+		setup, err := newErasureTestSetup(t, test.dataBlocks, test.disks-test.dataBlocks, test.blocksize)
 		if err != nil {
 			t.Fatalf("Test %d: failed to setup Erasure environment: %v", i, err)
 		}
 		disks := setup.disks
 		erasure, err := NewErasure(context.Background(), test.dataBlocks, test.disks-test.dataBlocks, test.blocksize)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create ErasureStorage: %v", i, err)
 		}
 		data := make([]byte, test.size)
 		if _, err = io.ReadFull(rand.Reader, data); err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create random test data: %v", i, err)
 		}
 		buffer := make([]byte, test.blocksize, 2*test.blocksize)
@@ -93,7 +91,6 @@ func TestErasureHeal(t *testing.T) {
 		_, err = erasure.Encode(context.Background(), bytes.NewReader(data), writers, buffer, erasure.dataBlocks+1)
 		closeBitrotWriters(writers)
 		if err != nil {
-			setup.Remove()
 			t.Fatalf("Test %d: failed to create random test data: %v", i, err)
 		}
 
@@ -156,6 +153,5 @@ func TestErasureHeal(t *testing.T) {
 				}
 			}
 		}
-		setup.Remove()
 	}
 }

cmd/erasure-healing-common_test.go

@@ -176,7 +176,7 @@ func TestListOnlineDisks(t *testing.T) {
 	}
 
 	bucket := "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket %v", err)
 	}
@@ -260,7 +260,7 @@ func TestListOnlineDisks(t *testing.T) {
 
 			if test._tamperBackend != noTamper {
 				if tamperedIndex != -1 && availableDisks[tamperedIndex] != nil {
-					t.Fatalf("disk (%v) with part.1 missing is not a disk with available data",
+					t.Fatalf("Drive (%v) with part.1 missing is not a drive with available data",
 						erasureDisks[tamperedIndex])
 				}
 			}
@@ -351,7 +351,7 @@ func TestListOnlineDisksSmallObjects(t *testing.T) {
 	}
 
 	bucket := "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket %v", err)
 	}
@@ -446,7 +446,7 @@ func TestListOnlineDisksSmallObjects(t *testing.T) {
 
 			if test._tamperBackend != noTamper {
 				if tamperedIndex != -1 && availableDisks[tamperedIndex] != nil {
-					t.Fatalf("disk (%v) with part.1 missing is not a disk with available data",
+					t.Fatalf("Drive (%v) with part.1 missing is not a drive with available data",
 						erasureDisks[tamperedIndex])
 				}
 			}
@@ -472,7 +472,7 @@ func TestDisksWithAllParts(t *testing.T) {
 	z := obj.(*erasureServerPools)
 	s := z.serverPools[0].sets[0]
 	erasureDisks := s.getDisks()
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket %v", err)
 	}
@@ -506,7 +506,7 @@ func TestDisksWithAllParts(t *testing.T) {
 		errs, fi, bucket, object, madmin.HealDeepScan)
 
 	if len(filteredDisks) != len(erasureDisks) {
-		t.Errorf("Unexpected number of disks: %d", len(filteredDisks))
+		t.Errorf("Unexpected number of drives: %d", len(filteredDisks))
 	}
 
 	for diskIndex, disk := range filteredDisks {
@@ -515,7 +515,7 @@ func TestDisksWithAllParts(t *testing.T) {
 		}
 
 		if disk == nil {
-			t.Errorf("Disk erroneously filtered, diskIndex: %d", diskIndex)
+			t.Errorf("Drive erroneously filtered, driveIndex: %d", diskIndex)
 		}
 	}
 
@@ -528,14 +528,14 @@ func TestDisksWithAllParts(t *testing.T) {
 		errs, fi, bucket, object, madmin.HealDeepScan)
 
 	if len(filteredDisks) != len(erasureDisks) {
-		t.Errorf("Unexpected number of disks: %d", len(filteredDisks))
+		t.Errorf("Unexpected number of drives: %d", len(filteredDisks))
 	}
 	for diskIndex, disk := range filteredDisks {
 		if diskIndex == 0 && disk != nil {
-			t.Errorf("Disk not filtered as expected, disk: %d", diskIndex)
+			t.Errorf("Drive not filtered as expected, drive: %d", diskIndex)
 		}
 		if diskIndex != 0 && disk == nil {
-			t.Errorf("Disk erroneously filtered, diskIndex: %d", diskIndex)
+			t.Errorf("Drive erroneously filtered, driveIndex: %d", diskIndex)
 		}
 	}
 	partsMetadata[0] = partsMetadataBackup // Revert before going to the next test
@@ -549,14 +549,14 @@ func TestDisksWithAllParts(t *testing.T) {
 		errs, fi, bucket, object, madmin.HealDeepScan)
 
 	if len(filteredDisks) != len(erasureDisks) {
-		t.Errorf("Unexpected number of disks: %d", len(filteredDisks))
+		t.Errorf("Unexpected number of drives: %d", len(filteredDisks))
 	}
 	for diskIndex, disk := range filteredDisks {
 		if diskIndex == 1 && disk != nil {
-			t.Errorf("Disk not filtered as expected, disk: %d", diskIndex)
+			t.Errorf("Drive not filtered as expected, drive: %d", diskIndex)
 		}
 		if diskIndex != 1 && disk == nil {
-			t.Errorf("Disk erroneously filtered, diskIndex: %d", diskIndex)
+			t.Errorf("Drive erroneously filtered, driveIndex: %d", diskIndex)
 		}
 	}
 	partsMetadata[1] = partsMetadataBackup // Revert before going to the next test
@@ -586,23 +586,23 @@ func TestDisksWithAllParts(t *testing.T) {
 		errs, fi, bucket, object, madmin.HealDeepScan)
 
 	if len(filteredDisks) != len(erasureDisks) {
-		t.Errorf("Unexpected number of disks: %d", len(filteredDisks))
+		t.Errorf("Unexpected number of drives: %d", len(filteredDisks))
 	}
 
 	for diskIndex, disk := range filteredDisks {
 		if _, ok := diskFailures[diskIndex]; ok {
 			if disk != nil {
-				t.Errorf("Disk not filtered as expected, disk: %d", diskIndex)
+				t.Errorf("Drive not filtered as expected, drive: %d", diskIndex)
 			}
 			if errs[diskIndex] == nil {
-				t.Errorf("Expected error not received, diskIndex: %d", diskIndex)
+				t.Errorf("Expected error not received, driveIndex: %d", diskIndex)
 			}
 		} else {
 			if disk == nil {
-				t.Errorf("Disk erroneously filtered, diskIndex: %d", diskIndex)
+				t.Errorf("Drive erroneously filtered, driveIndex: %d", diskIndex)
 			}
 			if errs[diskIndex] != nil {
-				t.Errorf("Unexpected error, %s, diskIndex: %d", errs[diskIndex], diskIndex)
+				t.Errorf("Unexpected error, %s, driveIndex: %d", errs[diskIndex], diskIndex)
 			}
 
 		}

cmd/erasure-healing.go

@@ -243,7 +243,7 @@ func listAllBuckets(ctx context.Context, storageDisks []StorageAPI, healBuckets
 
 // Only heal on disks where we are sure that healing is needed. We can expand
 // this list as and when we figure out more errors can be added to this list safely.
-func shouldHealObjectOnDisk(erErr, dataErr error, meta FileInfo, latestMeta FileInfo) bool {
+func shouldHealObjectOnDisk(erErr, dataErr error, meta FileInfo, latestMeta FileInfo, doinline bool) bool {
 	switch {
 	case errors.Is(erErr, errFileNotFound) || errors.Is(erErr, errFileVersionNotFound):
 		return true
@@ -256,6 +256,10 @@ func shouldHealObjectOnDisk(erErr, dataErr error, meta FileInfo, latestMeta File
 			// always check first.
 			return true
 		}
+		if doinline {
+			// convert small files to 'inline'
+			return true
+		}
 		if !meta.Deleted && !meta.IsRemote() {
 			// If xl.meta was read fine but there may be problem with the part.N files.
 			if IsErr(dataErr, []error{
@@ -275,10 +279,6 @@ func shouldHealObjectOnDisk(erErr, dataErr error, meta FileInfo, latestMeta File
 
 // Heals an object by re-writing corrupt/missing erasure blocks.
 func (er erasureObjects) healObject(ctx context.Context, bucket string, object string, versionID string, opts madmin.HealOpts) (result madmin.HealResultItem, err error) {
-	if !opts.DryRun {
-		defer NSUpdated(bucket, object)
-	}
-
 	dryRun := opts.DryRun
 	scanMode := opts.ScanMode
 
@@ -346,6 +346,26 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 	availableDisks, dataErrs, diskMTime := disksWithAllParts(ctx, onlineDisks, partsMetadata,
 		errs, latestMeta, bucket, object, scanMode)
 
+	var erasure Erasure
+	var recreate bool
+	if !latestMeta.Deleted && !latestMeta.IsRemote() {
+		// Initialize erasure coding
+		erasure, err = NewErasure(ctx, latestMeta.Erasure.DataBlocks,
+			latestMeta.Erasure.ParityBlocks, latestMeta.Erasure.BlockSize)
+		if err != nil {
+			return result, err
+		}
+
+		// Is only 'true' if the opts.Recreate is true and
+		// the object shardSize < smallFileThreshold do not
+		// set this to 'true' arbitrarily and must be only
+		// 'true' with caller ask.
+		recreate = (opts.Recreate &&
+			!latestMeta.InlineData() &&
+			len(latestMeta.Parts) == 1 &&
+			erasure.ShardFileSize(latestMeta.Parts[0].ActualSize) < smallFileThreshold)
+	}
+
 	// Loop to find number of disks with valid data, per-drive
 	// data state and a list of outdated disks on which data needs
 	// to be healed.
@@ -372,7 +392,7 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 			driveState = madmin.DriveStateCorrupt
 		}
 
-		if shouldHealObjectOnDisk(errs[i], dataErrs[i], partsMetadata[i], latestMeta) {
+		if shouldHealObjectOnDisk(errs[i], dataErrs[i], partsMetadata[i], latestMeta, recreate) {
 			outDatedDisks[i] = storageDisks[i]
 			disksToHealCount++
 			result.Before.Drives = append(result.Before.Drives, madmin.HealDriveInfo{
@@ -426,9 +446,9 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 		return result, nil
 	}
 
-	if !latestMeta.XLV1 && !latestMeta.Deleted && disksToHealCount > latestMeta.Erasure.ParityBlocks {
+	if !latestMeta.XLV1 && !latestMeta.Deleted && !recreate && disksToHealCount > latestMeta.Erasure.ParityBlocks {
 		// When disk to heal count is greater than parity blocks we should simply error out.
-		err := fmt.Errorf("more disks are expected to heal than parity, returned errors: %v (dataErrs %v) -> %s/%s(%s)", errs, dataErrs, bucket, object, versionID)
+		err := fmt.Errorf("more drives are expected to heal than parity, returned errors: %v (dataErrs %v) -> %s/%s(%s)", errs, dataErrs, bucket, object, versionID)
 		logger.LogIf(ctx, err)
 		return er.defaultHealResult(latestMeta, storageDisks, storageEndpoints, errs,
 			bucket, object, versionID), err
@@ -482,18 +502,9 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 	}
 
 	var inlineBuffers []*bytes.Buffer
-	if latestMeta.InlineData() {
-		inlineBuffers = make([]*bytes.Buffer, len(outDatedDisks))
-	}
-
 	if !latestMeta.Deleted && !latestMeta.IsRemote() {
-		// Heal each part. erasureHealFile() will write the healed
-		// part to .minio/tmp/uuid/ which needs to be renamed later to
-		// the final location.
-		erasure, err := NewErasure(ctx, latestMeta.Erasure.DataBlocks,
-			latestMeta.Erasure.ParityBlocks, latestMeta.Erasure.BlockSize)
-		if err != nil {
-			return result, err
+		if latestMeta.InlineData() || recreate {
+			inlineBuffers = make([]*bytes.Buffer, len(outDatedDisks))
 		}
 
 		erasureInfo := latestMeta.Erasure
@@ -529,6 +540,10 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 						tillOffset, DefaultBitrotAlgorithm, erasure.ShardSize())
 				}
 			}
+
+			// Heal each part. erasure.Heal() will write the healed
+			// part to .minio/tmp/uuid/ which needs to be renamed
+			// later to the final location.
 			err = erasure.Heal(ctx, writers, readers, partSize)
 			closeBitrotReaders(readers)
 			closeBitrotWriters(writers)
@@ -560,6 +575,7 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 				})
 				if len(inlineBuffers) > 0 && inlineBuffers[i] != nil {
 					partsMetadata[i].Data = inlineBuffers[i].Bytes()
+					partsMetadata[i].SetInlineData()
 				} else {
 					partsMetadata[i].Data = nil
 				}
@@ -567,7 +583,7 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 
 			// If all disks are having errors, we give up.
 			if disksToHealCount == 0 {
-				return result, fmt.Errorf("all disks had write errors, unable to heal %s/%s", bucket, object)
+				return result, fmt.Errorf("all drives had write errors, unable to heal %s/%s", bucket, object)
 			}
 
 		}
@@ -591,8 +607,9 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 			return result, err
 		}
 
-		// Remove any remaining parts from outdated disks from before transition.
-		if partsMetadata[i].IsRemote() {
+		// - Remove any parts from healed disks after its been inlined.
+		// - Remove any remaining parts from outdated disks from before transition.
+		if recreate || partsMetadata[i].IsRemote() {
 			rmDataDir := partsMetadata[i].DataDir
 			disk.DeleteVol(ctx, pathJoin(bucket, encodeDirObject(object), rmDataDir), true)
 		}

cmd/erasure-healing_test.go

@@ -236,7 +236,7 @@ func TestHealing(t *testing.T) {
 	er := z.serverPools[0].sets[0]
 
 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -359,7 +359,7 @@ func TestHealingDanglingObject(t *testing.T) {
 	object := getRandomObjectName()
 	data := bytes.Repeat([]byte("a"), 128*1024)
 
-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -554,7 +554,7 @@ func TestHealCorrectQuorum(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions
 
-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -685,7 +685,7 @@ func TestHealObjectCorruptedPools(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions
 
-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -861,7 +861,7 @@ func TestHealObjectCorruptedXLMeta(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions
 
-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -1004,7 +1004,7 @@ func TestHealObjectCorruptedParts(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions
 
-	err = objLayer.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -1161,7 +1161,7 @@ func TestHealObjectErasure(t *testing.T) {
 	data := bytes.Repeat([]byte("a"), 5*1024*1024)
 	var opts ObjectOptions
 
-	err = obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -1258,7 +1258,7 @@ func TestHealEmptyDirectoryErasure(t *testing.T) {
 	object := "empty-dir/"
 	var opts ObjectOptions
 
-	err = obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket - %v", err)
 	}
@@ -1359,7 +1359,7 @@ func TestHealLastDataShard(t *testing.T) {
 			}
 			var opts ObjectOptions
 
-			err = obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+			err = obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 			if err != nil {
 				t.Fatalf("Failed to make a bucket - %v", err)
 			}

cmd/erasure-metadata-utils.go

@@ -275,7 +275,7 @@ func shuffleDisks(disks []StorageAPI, distribution []int) (shuffledDisks []Stora
 // the corresponding error in errs slice is not nil
 func evalDisks(disks []StorageAPI, errs []error) []StorageAPI {
 	if len(errs) != len(disks) {
-		logger.LogIf(GlobalContext, errors.New("unexpected disks/errors slice length"))
+		logger.LogIf(GlobalContext, errors.New("unexpected drives/errors slice length"))
 		return nil
 	}
 	newDisks := make([]StorageAPI, len(disks))

cmd/erasure-multipart.go

@@ -1159,6 +1159,8 @@ func (er erasureObjects) CompleteMultipartUpload(ctx context.Context, bucket str
 		return oi, toObjectErr(err, bucket, object)
 	}
 
+	defer NSUpdated(bucket, object)
+
 	// Check if there is any offline disk and add it to the MRF list
 	for _, disk := range onlineDisks {
 		if disk != nil && disk.IsOnline() {

cmd/erasure-object.go

@@ -448,15 +448,26 @@ func (er erasureObjects) deleteIfDangling(ctx context.Context, bucket, object st
 		}
 		defer NSUpdated(bucket, object)
 
-		if opts.VersionID != "" {
-			er.deleteObjectVersion(ctx, bucket, object, 1, FileInfo{
-				VersionID: opts.VersionID,
-			}, false)
-		} else {
-			er.deleteObjectVersion(ctx, bucket, object, 1, FileInfo{
-				VersionID: m.VersionID,
-			}, false)
+		fi := FileInfo{
+			VersionID: m.VersionID,
 		}
+		if opts.VersionID != "" {
+			fi.VersionID = opts.VersionID
+		}
+
+		disks := er.getDisks()
+		g := errgroup.WithNErrs(len(disks))
+		for index := range disks {
+			index := index
+			g.Go(func() error {
+				if disks[index] == nil {
+					return errDiskNotFound
+				}
+				return disks[index].DeleteVersion(ctx, bucket, object, fi, false)
+			}, index)
+		}
+
+		g.Wait()
 	}
 	return m, err
 }
@@ -534,6 +545,11 @@ func readAllXL(ctx context.Context, disks []StorageAPI, bucket, object string, r
 		return metaFileInfos, errs
 	}
 
+	versionID := lfi.VersionID
+	if versionID == "" {
+		versionID = nullVersionID
+	}
+
 	for index := range metadataArray {
 		if metadataArray[index] == nil {
 			continue
@@ -541,21 +557,15 @@ func readAllXL(ctx context.Context, disks []StorageAPI, bucket, object string, r
 
 		// make sure to preserve this for diskmtime based healing bugfix.
 		diskMTime := metaFileInfos[index].DiskMTime
-		metaFileInfos[index], errs[index] = metadataArray[index].ToFileInfo(bucket, object, "")
+		metaFileInfos[index], errs[index] = metadataArray[index].ToFileInfo(bucket, object, versionID)
 		if errs[index] != nil {
 			continue
 		}
 
-		if metaFileInfos[index].IsValid() && metaFileInfos[index].ModTime.Equal(lfi.ModTime) {
-			versionID := metaFileInfos[index].VersionID
-			if versionID == "" {
-				versionID = nullVersionID
-			}
-			if readData {
-				metaFileInfos[index].Data = metadataArray[index].data.find(versionID)
-			}
-			metaFileInfos[index].DiskMTime = diskMTime
+		if readData {
+			metaFileInfos[index].Data = metadataArray[index].data.find(versionID)
 		}
+		metaFileInfos[index].DiskMTime = diskMTime
 	}
 
 	// Return all the metadata.
@@ -687,8 +697,6 @@ func (er erasureObjects) getObjectInfoAndQuorum(ctx context.Context, bucket, obj
 
 // Similar to rename but renames data from srcEntry to dstEntry at dataDir
 func renameData(ctx context.Context, disks []StorageAPI, srcBucket, srcEntry string, metadata []FileInfo, dstBucket, dstEntry string, writeQuorum int) ([]StorageAPI, error) {
-	defer NSUpdated(dstBucket, dstEntry)
-
 	g := errgroup.WithNErrs(len(disks))
 
 	fvID := mustGetUUID()
@@ -1141,6 +1149,8 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 		return ObjectInfo{}, toObjectErr(err, bucket, object)
 	}
 
+	defer NSUpdated(bucket, object)
+
 	for i := 0; i < len(onlineDisks); i++ {
 		if onlineDisks[i] != nil && onlineDisks[i].IsOnline() {
 			// Object info is the same in all disks, so we can pick
@@ -1173,8 +1183,16 @@ func (er erasureObjects) putObject(ctx context.Context, bucket string, object st
 	return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
 }
 
-func (er erasureObjects) deleteObjectVersion(ctx context.Context, bucket, object string, writeQuorum int, fi FileInfo, forceDelMarker bool) error {
+func (er erasureObjects) deleteObjectVersion(ctx context.Context, bucket, object string, fi FileInfo, forceDelMarker bool) error {
 	disks := er.getDisks()
+	// Assume (N/2 + 1) quorum for Delete()
+	// this is a theoretical assumption such that
+	// for delete's we do not need to honor storage
+	// class for objects that have reduced quorum
+	// due to storage class - this only needs to be honored
+	// for Read() requests alone that we already do.
+	writeQuorum := len(disks)/2 + 1
+
 	g := errgroup.WithNErrs(len(disks))
 	for index := range disks {
 		index := index
@@ -1424,69 +1442,49 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		}
 	}
 
-	// Acquire a write lock before deleting the object.
-	lk := er.NewNSLock(bucket, object)
-	lkctx, err := lk.GetLock(ctx, globalDeleteOperationTimeout)
-	if err != nil {
-		return ObjectInfo{}, err
-	}
-	ctx = lkctx.Context()
-	defer lk.Unlock(lkctx.Cancel)
-
-	versionFound := true
 	objInfo = ObjectInfo{VersionID: opts.VersionID} // version id needed in Delete API response.
-	goi, writeQuorum, gerr := er.getObjectInfoAndQuorum(ctx, bucket, object, opts)
-	if gerr != nil && goi.Name == "" {
-		switch gerr.(type) {
-		case InsufficientReadQuorum:
-			return objInfo, InsufficientWriteQuorum{}
-		}
-		// For delete marker replication, versionID being replicated will not exist on disk
-		if opts.DeleteMarker {
-			versionFound = false
-		} else {
-			return objInfo, gerr
-		}
-	}
-
-	if opts.Expiration.Expire {
-		action := evalActionFromLifecycle(ctx, *lc, rcfg, goi, false)
-		var isErr bool
-		switch action {
-		case lifecycle.NoneAction:
-			isErr = true
-		case lifecycle.TransitionAction, lifecycle.TransitionVersionAction:
-			isErr = true
-		}
-		if isErr {
-			if goi.VersionID != "" {
-				return goi, VersionNotFound{
-					Bucket:    bucket,
-					Object:    object,
-					VersionID: goi.VersionID,
-				}
-			}
-			return goi, ObjectNotFound{
-				Bucket: bucket,
-				Object: object,
-			}
-		}
-	}
-
 	defer NSUpdated(bucket, object)
 
 	storageDisks := er.getDisks()
 
-	var markDelete bool
-	// Determine whether to mark object deleted for replication
-	if goi.VersionID != "" {
-		markDelete = true
+	if opts.Expiration.Expire {
+		goi, _, err := er.getObjectInfoAndQuorum(ctx, bucket, object, opts)
+		if err == nil {
+			action := evalActionFromLifecycle(ctx, *lc, rcfg, goi, false)
+			var isErr bool
+			switch action {
+			case lifecycle.NoneAction:
+				isErr = true
+			case lifecycle.TransitionAction, lifecycle.TransitionVersionAction:
+				isErr = true
+			}
+			if isErr {
+				if goi.VersionID != "" {
+					return goi, VersionNotFound{
+						Bucket:    bucket,
+						Object:    object,
+						VersionID: goi.VersionID,
+					}
+				}
+				return goi, ObjectNotFound{
+					Bucket: bucket,
+					Object: object,
+				}
+			}
+		}
 	}
 
-	// Default deleteMarker to true if object is under versioning
-	deleteMarker := opts.Versioned
+	versionFound := !(opts.DeleteMarker && opts.VersionID != "")
 
-	if opts.VersionID != "" {
+	// Determine whether to mark object deleted for replication
+	markDelete := !opts.DeleteMarker && opts.VersionID != ""
+
+	// Default deleteMarker to true if object is under versioning
+	// versioning suspended means we add `null` version as
+	// delete marker, if its not decided already.
+	deleteMarker := (opts.Versioned || opts.VersionSuspended) && opts.VersionID == "" || (opts.DeleteMarker && opts.VersionID != "")
+
+	if markDelete {
 		// case where replica version needs to be deleted on target cluster
 		if versionFound && opts.DeleteMarkerReplicationStatus() == replication.Replica {
 			markDelete = false
@@ -1497,22 +1495,6 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		if opts.VersionPurgeStatus() == Complete {
 			markDelete = false
 		}
-
-		// Version is found but we do not wish to create more delete markers
-		// now, since VersionPurgeStatus() is already set, we can let the
-		// lower layers decide this. This fixes a regression that was introduced
-		// in PR #14555 where !VersionPurgeStatus.Empty() is automatically
-		// considered as Delete marker true to avoid listing such objects by
-		// regular ListObjects() calls. However for delete replication this
-		// ends up being a problem because "upon" a successful delete this
-		// ends up creating a new delete marker that is spurious and unnecessary.
-		if versionFound {
-			if !goi.VersionPurgeStatus.Empty() {
-				deleteMarker = false
-			} else if !goi.DeleteMarker { // implies a versioned delete of object
-				deleteMarker = false
-			}
-		}
 	}
 
 	modTime := opts.MTime
@@ -1520,38 +1502,32 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		modTime = UTCNow()
 	}
 	fvID := mustGetUUID()
-	if markDelete {
-		if opts.Versioned || opts.VersionSuspended {
-			if !deleteMarker {
-				// versioning suspended means we add `null` version as
-				// delete marker, if its not decided already.
-				deleteMarker = opts.VersionSuspended && opts.VersionID == ""
-			}
-			fi := FileInfo{
-				Name:             object,
-				Deleted:          deleteMarker,
-				MarkDeleted:      markDelete,
-				ModTime:          modTime,
-				ReplicationState: opts.DeleteReplication,
-				TransitionStatus: opts.Transition.Status,
-				ExpireRestored:   opts.Transition.ExpireRestored,
-			}
-			fi.SetTierFreeVersionID(fvID)
-			if opts.Versioned {
-				fi.VersionID = mustGetUUID()
-				if opts.VersionID != "" {
-					fi.VersionID = opts.VersionID
-				}
-			}
-			// versioning suspended means we add `null` version as
-			// delete marker. Add delete marker, since we don't have
-			// any version specified explicitly. Or if a particular
-			// version id needs to be replicated.
-			if err = er.deleteObjectVersion(ctx, bucket, object, writeQuorum, fi, opts.DeleteMarker); err != nil {
-				return objInfo, toObjectErr(err, bucket, object)
-			}
-			return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
+
+	if markDelete && (opts.Versioned || opts.VersionSuspended) {
+		fi := FileInfo{
+			Name:             object,
+			Deleted:          true,
+			MarkDeleted:      markDelete,
+			ModTime:          modTime,
+			ReplicationState: opts.DeleteReplication,
+			TransitionStatus: opts.Transition.Status,
+			ExpireRestored:   opts.Transition.ExpireRestored,
 		}
+		fi.SetTierFreeVersionID(fvID)
+		if opts.Versioned {
+			fi.VersionID = mustGetUUID()
+			if opts.VersionID != "" {
+				fi.VersionID = opts.VersionID
+			}
+		}
+		// versioning suspended means we add `null` version as
+		// delete marker. Add delete marker, since we don't have
+		// any version specified explicitly. Or if a particular
+		// version id needs to be replicated.
+		if err = er.deleteObjectVersion(ctx, bucket, object, fi, opts.DeleteMarker); err != nil {
+			return objInfo, toObjectErr(err, bucket, object)
+		}
+		return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
 	}
 
 	// Delete the object version on all disks.
@@ -1566,7 +1542,7 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		ExpireRestored:   opts.Transition.ExpireRestored,
 	}
 	dfi.SetTierFreeVersionID(fvID)
-	if err = er.deleteObjectVersion(ctx, bucket, object, writeQuorum, dfi, opts.DeleteMarker); err != nil {
+	if err = er.deleteObjectVersion(ctx, bucket, object, dfi, opts.DeleteMarker); err != nil {
 		return objInfo, toObjectErr(err, bucket, object)
 	}
 
@@ -1578,13 +1554,7 @@ func (er erasureObjects) DeleteObject(ctx context.Context, bucket, object string
 		break
 	}
 
-	return ObjectInfo{
-		Bucket:                     bucket,
-		Name:                       object,
-		VersionID:                  opts.VersionID,
-		VersionPurgeStatusInternal: opts.DeleteReplication.VersionPurgeStatusInternal,
-		ReplicationStatusInternal:  opts.DeleteReplication.ReplicationStatusInternal,
-	}, nil
+	return dfi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
 }
 
 // Send the successful but partial upload/delete, however ignore
@@ -1854,14 +1824,8 @@ func (er erasureObjects) TransitionObject(ctx context.Context, bucket, object st
 	eventName := event.ObjectTransitionComplete
 
 	storageDisks := er.getDisks()
-	// we now know the number of blocks this object needs for data and parity.
-	// writeQuorum is dataBlocks + 1
-	writeQuorum := fi.Erasure.DataBlocks
-	if fi.Erasure.DataBlocks == fi.Erasure.ParityBlocks {
-		writeQuorum++
-	}
 
-	if err = er.deleteObjectVersion(ctx, bucket, object, writeQuorum, fi, false); err != nil {
+	if err = er.deleteObjectVersion(ctx, bucket, object, fi, false); err != nil {
 		eventName = event.ObjectTransitionFailed
 	}
 

cmd/erasure-object_test.go

@@ -53,7 +53,7 @@ func TestRepeatPutObjectPart(t *testing.T) {
 	defer objLayer.Shutdown(context.Background())
 	defer removeRoots(disks)
 
-	err = objLayer.MakeBucketWithLocation(ctx, "bucket1", BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, "bucket1", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -101,7 +101,7 @@ func TestErasureDeleteObjectBasic(t *testing.T) {
 	}
 	defer xl.Shutdown(context.Background())
 
-	err = xl.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = xl.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -155,7 +155,7 @@ func TestDeleteObjectsVersioned(t *testing.T) {
 		{bucketName, "dir/obj1"},
 	}
 
-	err = obj.MakeBucketWithLocation(ctx, bucketName, BucketOptions{
+	err = obj.MakeBucketWithLocation(ctx, bucketName, MakeBucketOptions{
 		VersioningEnabled: true,
 	})
 	if err != nil {
@@ -244,7 +244,7 @@ func TestErasureDeleteObjectsErasureSet(t *testing.T) {
 		{bucketName, "obj_4"},
 	}
 
-	err := erasureSets.MakeBucketWithLocation(ctx, bucketName, BucketOptions{})
+	err := erasureSets.MakeBucketWithLocation(ctx, bucketName, MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -305,7 +305,7 @@ func TestErasureDeleteObjectDiskNotFound(t *testing.T) {
 	xl := z.serverPools[0].sets[0]
 
 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -374,7 +374,7 @@ func TestErasureDeleteObjectDiskNotFoundErasure4(t *testing.T) {
 	xl := z.serverPools[0].sets[0]
 
 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -434,7 +434,7 @@ func TestErasureDeleteObjectDiskNotFoundErr(t *testing.T) {
 	xl := z.serverPools[0].sets[0]
 
 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -505,7 +505,7 @@ func TestGetObjectNoQuorum(t *testing.T) {
 	xl := z.serverPools[0].sets[0]
 
 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -614,7 +614,7 @@ func TestHeadObjectNoQuorum(t *testing.T) {
 	xl := z.serverPools[0].sets[0]
 
 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -692,7 +692,7 @@ func TestPutObjectNoQuorum(t *testing.T) {
 	xl := z.serverPools[0].sets[0]
 
 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -755,7 +755,7 @@ func TestPutObjectNoQuorumSmall(t *testing.T) {
 	xl := z.serverPools[0].sets[0]
 
 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, "bucket", BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, "bucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -824,7 +824,7 @@ func TestPutObjectSmallInlineData(t *testing.T) {
 	object := "object"
 
 	// Create "bucket"
-	err = obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err = obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -892,7 +892,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 	ctx, cancel := context.WithCancel(GlobalContext)
 	defer cancel()
 
-	err := obj.MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+	err := obj.MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 	if err != nil {
 		t.Fatalf("Failed to make a bucket %v", err)
 	}
@@ -1073,10 +1073,7 @@ func TestGetObjectInlineNotInline(t *testing.T) {
 	// Create a backend with 4 disks named disk{1...4}, this name convention
 	// because we will unzip some object data from a sample archive.
 	const numDisks = 4
-	path, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatal(err)
-	}
+	path := t.TempDir()
 
 	var fsDirs []string
 	for i := 1; i <= numDisks; i++ {
@@ -1094,7 +1091,7 @@ func TestGetObjectInlineNotInline(t *testing.T) {
 	defer removeRoots(fsDirs)
 
 	// Create a testbucket
-	err = objLayer.MakeBucketWithLocation(ctx, "testbucket", BucketOptions{})
+	err = objLayer.MakeBucketWithLocation(ctx, "testbucket", MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -1160,7 +1157,7 @@ func TestGetObjectWithOutdatedDisks(t *testing.T) {
 
 	for i, testCase := range testCases {
 		// Step 1: create a bucket
-		err = z.MakeBucketWithLocation(ctx, testCase.bucket, BucketOptions{VersioningEnabled: testCase.versioned})
+		err = z.MakeBucketWithLocation(ctx, testCase.bucket, MakeBucketOptions{VersioningEnabled: testCase.versioned})
 		if err != nil {
 			t.Fatalf("Test %d: Failed to create a bucket: %v", i+1, err)
 		}

cmd/erasure-server-pool-decom.go

@@ -457,7 +457,7 @@ func (p *poolMeta) updateAfter(ctx context.Context, idx int, pools []*erasureSet
 	now := UTCNow()
 	if now.Sub(p.Pools[idx].LastUpdate) >= duration {
 		if serverDebugLog {
-			console.Debugf("decommission: persisting poolMeta on disk: threshold:%s, poolMeta:%#v\n", now.Sub(p.Pools[idx].LastUpdate), p.Pools[idx])
+			console.Debugf("decommission: persisting poolMeta on drive: threshold:%s, poolMeta:%#v\n", now.Sub(p.Pools[idx].LastUpdate), p.Pools[idx])
 		}
 		p.Pools[idx].LastUpdate = now
 		if err := p.save(ctx, pools); err != nil {
@@ -677,7 +677,7 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 		set := set
 		disks := set.getOnlineDisks()
 		if len(disks) == 0 {
-			logger.LogIf(GlobalContext, fmt.Errorf("no online disks found for set with endpoints %s",
+			logger.LogIf(GlobalContext, fmt.Errorf("no online drives found for set with endpoints %s",
 				set.getEndpoints()))
 			continue
 		}
@@ -800,11 +800,14 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 						logger.LogIf(ctx, err)
 						continue
 					}
+					stopFn := globalDecommissionMetrics.log(decomMetricDecommissionObject, idx, bi.Name, version.Name, version.VersionID)
 					if err = z.decommissionObject(ctx, bi.Name, gr); err != nil {
+						stopFn(err)
 						failure = true
 						logger.LogIf(ctx, err)
 						continue
 					}
+					stopFn(nil)
 					failure = false
 					break
 				}
@@ -819,6 +822,7 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 
 			// if all versions were decommissioned, then we can delete the object versions.
 			if decommissionedCount == len(fivs.Versions) {
+				stopFn := globalDecommissionMetrics.log(decomMetricDecommissionRemoveObject, idx, bi.Name, entry.name)
 				_, err := set.DeleteObject(ctx,
 					bi.Name,
 					encodeDirObject(entry.name),
@@ -826,6 +830,7 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 						DeletePrefix: true, // use prefix delete to delete all versions at once.
 					},
 				)
+				stopFn(err)
 				auditLogDecom(ctx, "DecomDeleteObject", bi.Name, entry.name, "", err)
 				if err != nil {
 					logger.LogIf(ctx, err)
@@ -881,6 +886,47 @@ func (z *erasureServerPools) decommissionPool(ctx context.Context, idx int, pool
 	return nil
 }
 
+//msgp:ignore decomMetrics
+type decomMetrics struct{}
+
+var globalDecommissionMetrics decomMetrics
+
+//msgp:ignore decomMetric
+//go:generate stringer -type=decomMetric -trimprefix=decomMetric $GOFILE
+type decomMetric uint8
+
+const (
+	decomMetricDecommissionBucket decomMetric = iota
+	decomMetricDecommissionObject
+	decomMetricDecommissionRemoveObject
+)
+
+func decomTrace(d decomMetric, poolIdx int, startTime time.Time, duration time.Duration, path string, err error) madmin.TraceInfo {
+	var errStr string
+	if err != nil {
+		errStr = err.Error()
+	}
+	return madmin.TraceInfo{
+		TraceType: madmin.TraceDecommission,
+		Time:      startTime,
+		NodeName:  globalLocalNodeName,
+		FuncName:  fmt.Sprintf("decommission.%s (pool-id=%d)", d.String(), poolIdx),
+		Duration:  duration,
+		Path:      path,
+		Error:     errStr,
+	}
+}
+
+func (m *decomMetrics) log(d decomMetric, poolIdx int, paths ...string) func(err error) {
+	startTime := time.Now()
+	return func(err error) {
+		duration := time.Since(startTime)
+		if globalTrace.NumSubscribers(madmin.TraceDecommission) > 0 {
+			globalTrace.Publish(decomTrace(d, poolIdx, startTime, duration, strings.Join(paths, " "), err))
+		}
+	}
+}
+
 func (z *erasureServerPools) decommissionInBackground(ctx context.Context, idx int) error {
 	pool := z.serverPools[idx]
 	for _, bucket := range z.poolMeta.PendingBuckets(idx) {
@@ -898,9 +944,13 @@ func (z *erasureServerPools) decommissionInBackground(ctx context.Context, idx i
 		if serverDebugLog {
 			console.Debugln("decommission: currently on bucket", bucket.Name)
 		}
+		stopFn := globalDecommissionMetrics.log(decomMetricDecommissionBucket, idx, bucket.Name)
 		if err := z.decommissionPool(ctx, idx, pool, bucket); err != nil {
+			stopFn(err)
 			return err
 		}
+		stopFn(nil)
+
 		z.poolMetaMutex.Lock()
 		z.poolMeta.BucketDone(idx, bucket)
 		z.poolMeta.save(ctx, z.serverPools)
@@ -1119,7 +1169,7 @@ func (z *erasureServerPools) StartDecommission(ctx context.Context, idx int) (er
 		return errInvalidArgument
 	}
 
-	buckets, err := z.ListBuckets(ctx)
+	buckets, err := z.ListBuckets(ctx, BucketOptions{})
 	if err != nil {
 		return err
 	}
@@ -1157,7 +1207,7 @@ func (z *erasureServerPools) StartDecommission(ctx context.Context, idx int) (er
 		pathJoin(minioMetaBucket, bucketMetaPrefix),
 	} {
 		var bucketExists BucketExists
-		if err = z.MakeBucketWithLocation(ctx, metaBucket, BucketOptions{}); err != nil {
+		if err = z.MakeBucketWithLocation(ctx, metaBucket, MakeBucketOptions{}); err != nil {
 			if !errors.As(err, &bucketExists) {
 				return err
 			}

cmd/erasure-server-pool.go

@@ -380,19 +380,15 @@ func (z *erasureServerPools) getServerPoolsAvailableSpace(ctx context.Context, b
 	return serverPools
 }
 
-// poolObjInfo represents the state of an object per pool
-type poolObjInfo struct {
-	PoolIndex int
-	ObjInfo   ObjectInfo
-	Err       error
+// PoolObjInfo represents the state of current object version per pool
+type PoolObjInfo struct {
+	Index   int
+	ObjInfo ObjectInfo
+	Err     error
 }
 
-func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, bucket, object string, opts ObjectOptions) (idx int, err error) {
-	if z.SinglePool() {
-		return 0, nil
-	}
-
-	poolObjInfos := make([]poolObjInfo, len(z.serverPools))
+func (z *erasureServerPools) getPoolInfoExistingWithOpts(ctx context.Context, bucket, object string, opts ObjectOptions) (PoolObjInfo, error) {
+	poolObjInfos := make([]PoolObjInfo, len(z.serverPools))
 	poolOpts := make([]ObjectOptions, len(z.serverPools))
 	for i := range z.serverPools {
 		poolOpts[i] = opts
@@ -404,8 +400,8 @@ func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, buc
 		go func(i int, pool *erasureSets, opts ObjectOptions) {
 			defer wg.Done()
 			// remember the pool index, we may sort the slice original index might be lost.
-			pinfo := poolObjInfo{
-				PoolIndex: i,
+			pinfo := PoolObjInfo{
+				Index: i,
 			}
 			// do not remove this check as it can lead to inconsistencies
 			// for all callers of bucket replication.
@@ -429,19 +425,19 @@ func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, buc
 	for _, pinfo := range poolObjInfos {
 		// skip all objects from suspended pools if asked by the
 		// caller.
-		if z.IsSuspended(pinfo.PoolIndex) && opts.SkipDecommissioned {
+		if z.IsSuspended(pinfo.Index) && opts.SkipDecommissioned {
 			continue
 		}
 
 		if pinfo.Err != nil && !isErrObjectNotFound(pinfo.Err) {
-			return -1, pinfo.Err
+			return pinfo, pinfo.Err
 		}
 
 		if isErrObjectNotFound(pinfo.Err) {
 			// No object exists or its a delete marker,
 			// check objInfo to confirm.
 			if pinfo.ObjInfo.DeleteMarker && pinfo.ObjInfo.Name != "" {
-				return pinfo.PoolIndex, nil
+				return pinfo, nil
 			}
 
 			// objInfo is not valid, truly the object doesn't
@@ -449,10 +445,18 @@ func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, buc
 			continue
 		}
 
-		return pinfo.PoolIndex, nil
+		return pinfo, nil
 	}
 
-	return -1, toObjectErr(errFileNotFound, bucket, object)
+	return PoolObjInfo{}, toObjectErr(errFileNotFound, bucket, object)
+}
+
+func (z *erasureServerPools) getPoolIdxExistingWithOpts(ctx context.Context, bucket, object string, opts ObjectOptions) (idx int, err error) {
+	pinfo, err := z.getPoolInfoExistingWithOpts(ctx, bucket, object, opts)
+	if err != nil {
+		return -1, err
+	}
+	return pinfo.Index, nil
 }
 
 // getPoolIdxExistingNoLock returns the (first) found object pool index containing an object.
@@ -533,9 +537,11 @@ func (z *erasureServerPools) BackendInfo() (b madmin.BackendInfo) {
 	rrSCParity := globalStorageClass.GetParityForSC(storageclass.RRS)
 
 	// Data blocks can vary per pool, but parity is same.
-	for _, setDriveCount := range z.SetDriveCounts() {
+	for i, setDriveCount := range z.SetDriveCounts() {
 		b.StandardSCData = append(b.StandardSCData, setDriveCount-scParity)
 		b.RRSCData = append(b.RRSCData, setDriveCount-rrSCParity)
+		b.DrivesPerSet = append(b.DrivesPerSet, setDriveCount)
+		b.TotalSets = append(b.TotalSets, z.serverPools[i].setCount)
 	}
 
 	b.StandardSCParity = scParity
@@ -613,7 +619,7 @@ func (z *erasureServerPools) NSScanner(ctx context.Context, bf *bloomFilter, upd
 	var results []dataUsageCache
 	var firstErr error
 
-	allBuckets, err := z.ListBuckets(ctx)
+	allBuckets, err := z.ListBuckets(ctx, BucketOptions{})
 	if err != nil {
 		return err
 	}
@@ -718,7 +724,7 @@ func (z *erasureServerPools) NSScanner(ctx context.Context, bf *bloomFilter, upd
 // MakeBucketWithLocation - creates a new bucket across all serverPools simultaneously
 // even if one of the sets fail to create buckets, we proceed all the successful
 // operations.
-func (z *erasureServerPools) MakeBucketWithLocation(ctx context.Context, bucket string, opts BucketOptions) error {
+func (z *erasureServerPools) MakeBucketWithLocation(ctx context.Context, bucket string, opts MakeBucketOptions) error {
 	g := errgroup.WithNErrs(len(z.serverPools))
 
 	// Lock the bucket name before creating.
@@ -758,6 +764,7 @@ func (z *erasureServerPools) MakeBucketWithLocation(ctx context.Context, bucket
 
 	// If it doesn't exist we get a new, so ignore errors
 	meta := newBucketMetadata(bucket)
+	meta.SetCreatedAt(opts.CreatedAt)
 	if opts.LockEnabled {
 		meta.VersioningConfigXML = enabledBucketVersioningConfig
 		meta.ObjectLockConfigXML = enabledBucketObjectLockConfig
@@ -996,16 +1003,36 @@ func (z *erasureServerPools) DeleteObject(ctx context.Context, bucket string, ob
 	}
 
 	object = encodeDirObject(object)
-	if z.SinglePool() {
-		return z.serverPools[0].DeleteObject(ctx, bucket, object, opts)
-	}
 
-	idx, err := z.getPoolIdxExistingWithOpts(ctx, bucket, object, opts)
+	// Acquire a write lock before deleting the object.
+	lk := z.NewNSLock(bucket, object)
+	lkctx, err := lk.GetLock(ctx, globalDeleteOperationTimeout)
 	if err != nil {
+		return ObjectInfo{}, err
+	}
+	ctx = lkctx.Context()
+	defer lk.Unlock(lkctx.Cancel)
+
+	gopts := opts
+	gopts.NoLock = true
+	pinfo, err := z.getPoolInfoExistingWithOpts(ctx, bucket, object, gopts)
+	if err != nil {
+		switch err.(type) {
+		case InsufficientReadQuorum:
+			return objInfo, InsufficientWriteQuorum{}
+		}
 		return objInfo, err
 	}
 
-	return z.serverPools[idx].DeleteObject(ctx, bucket, object, opts)
+	// Delete marker already present we are not going to create new delete markers.
+	if pinfo.ObjInfo.DeleteMarker && opts.VersionID == "" {
+		pinfo.ObjInfo.Name = decodeDirObject(object)
+		return pinfo.ObjInfo, nil
+	}
+
+	objInfo, err = z.serverPools[pinfo.Index].DeleteObject(ctx, bucket, object, opts)
+	objInfo.Name = decodeDirObject(object)
+	return objInfo, err
 }
 
 func (z *erasureServerPools) DeleteObjects(ctx context.Context, bucket string, objects []ObjectToDelete, opts ObjectOptions) ([]DeletedObject, []error) {
@@ -1031,64 +1058,84 @@ func (z *erasureServerPools) DeleteObjects(ctx context.Context, bucket string, o
 	ctx = lkctx.Context()
 	defer multiDeleteLock.Unlock(lkctx.Cancel)
 
-	if z.SinglePool() {
-		deleteObjects, dErrs := z.serverPools[0].DeleteObjects(ctx, bucket, objects, opts)
-		for i := range deleteObjects {
-			deleteObjects[i].ObjectName = decodeDirObject(deleteObjects[i].ObjectName)
-		}
-		return deleteObjects, dErrs
-	}
-
 	// Fetch location of up to 10 objects concurrently.
 	poolObjIdxMap := map[int][]ObjectToDelete{}
 	origIndexMap := map[int][]int{}
 
+	// Always perform 1/10th of the number of objects per delete
+	concurrent := len(objects) / 10
+	if concurrent <= 10 {
+		// if we cannot get 1/10th then choose the number of
+		// objects as concurrent.
+		concurrent = len(objects)
+	}
+
 	var mu sync.Mutex
-	eg := errgroup.WithNErrs(len(objects)).WithConcurrency(10)
+	eg := errgroup.WithNErrs(len(objects)).WithConcurrency(concurrent)
 	for j, obj := range objects {
 		j := j
 		obj := obj
 		eg.Go(func() error {
-			idx, err := z.getPoolIdxExistingWithOpts(ctx, bucket, obj.ObjectName, ObjectOptions{
+			pinfo, err := z.getPoolInfoExistingWithOpts(ctx, bucket, obj.ObjectName, ObjectOptions{
 				NoLock: true,
 			})
 			if err != nil {
 				derrs[j] = err
+				dobjects[j] = DeletedObject{
+					ObjectName: obj.ObjectName,
+				}
 				return nil
 			}
+
+			// Delete marker already present we are not going to create new delete markers.
+			if pinfo.ObjInfo.DeleteMarker && obj.VersionID == "" {
+				dobjects[j] = DeletedObject{
+					DeleteMarker:          pinfo.ObjInfo.DeleteMarker,
+					DeleteMarkerVersionID: pinfo.ObjInfo.VersionID,
+					DeleteMarkerMTime:     DeleteMarkerMTime{pinfo.ObjInfo.ModTime},
+					ObjectName:            pinfo.ObjInfo.Name,
+				}
+				return nil
+			}
+
+			idx := pinfo.Index
+
 			mu.Lock()
+			defer mu.Unlock()
+
 			poolObjIdxMap[idx] = append(poolObjIdxMap[idx], obj)
 			origIndexMap[idx] = append(origIndexMap[idx], j)
-			mu.Unlock()
 			return nil
 		}, j)
 	}
 
 	eg.Wait() // wait to check all the pools.
 
-	// Delete concurrently in all server pools.
-	var wg sync.WaitGroup
-	wg.Add(len(z.serverPools))
-	for idx, pool := range z.serverPools {
-		go func(idx int, pool *erasureSets) {
-			defer wg.Done()
-			objs := poolObjIdxMap[idx]
-			if len(objs) > 0 {
-				orgIndexes := origIndexMap[idx]
-				deletedObjects, errs := pool.DeleteObjects(ctx, bucket, objs, opts)
-				mu.Lock()
-				for i, derr := range errs {
-					if derr != nil {
-						derrs[orgIndexes[i]] = derr
+	if len(poolObjIdxMap) > 0 {
+		// Delete concurrently in all server pools.
+		var wg sync.WaitGroup
+		wg.Add(len(z.serverPools))
+		for idx, pool := range z.serverPools {
+			go func(idx int, pool *erasureSets) {
+				defer wg.Done()
+				objs := poolObjIdxMap[idx]
+				if len(objs) > 0 {
+					orgIndexes := origIndexMap[idx]
+					deletedObjects, errs := pool.DeleteObjects(ctx, bucket, objs, opts)
+					mu.Lock()
+					for i, derr := range errs {
+						if derr != nil {
+							derrs[orgIndexes[i]] = derr
+						}
+						deletedObjects[i].ObjectName = decodeDirObject(deletedObjects[i].ObjectName)
+						dobjects[orgIndexes[i]] = deletedObjects[i]
 					}
-					deletedObjects[i].ObjectName = decodeDirObject(deletedObjects[i].ObjectName)
-					dobjects[orgIndexes[i]] = deletedObjects[i]
+					mu.Unlock()
 				}
-				mu.Unlock()
-			}
-		}(idx, pool)
+			}(idx, pool)
+		}
+		wg.Wait()
 	}
-	wg.Wait()
 
 	return dobjects, derrs
 }
@@ -1175,7 +1222,6 @@ func (z *erasureServerPools) ListObjectVersions(ctx context.Context, bucket, pre
 	if marker == "" && versionMarker != "" {
 		return loi, NotImplemented{}
 	}
-
 	opts := listPathOptions{
 		Bucket:      bucket,
 		Prefix:      prefix,
@@ -1186,6 +1232,8 @@ func (z *erasureServerPools) ListObjectVersions(ctx context.Context, bucket, pre
 		AskDisks:    globalAPIConfig.getListQuorum(),
 		Versioned:   true,
 	}
+	// set bucket metadata in opts
+	opts.setBucketMeta(ctx)
 
 	merged, err := z.listPath(ctx, &opts)
 	if err != nil && err != io.EOF {
@@ -1232,12 +1280,16 @@ func maxKeysPlusOne(maxKeys int, addOne bool) int {
 
 func (z *erasureServerPools) ListObjects(ctx context.Context, bucket, prefix, marker, delimiter string, maxKeys int) (ListObjectsInfo, error) {
 	var loi ListObjectsInfo
-
-	// Automatically remove the object/version is an expiry lifecycle rule can be applied
-	lc, _ := globalLifecycleSys.Get(bucket)
-
-	// Check if bucket is object locked.
-	rcfg, _ := globalBucketObjectLockSys.Get(bucket)
+	opts := listPathOptions{
+		Bucket:      bucket,
+		Prefix:      prefix,
+		Separator:   delimiter,
+		Limit:       maxKeysPlusOne(maxKeys, marker != ""),
+		Marker:      marker,
+		InclDeleted: false,
+		AskDisks:    globalAPIConfig.getListQuorum(),
+	}
+	opts.setBucketMeta(ctx)
 
 	if len(prefix) > 0 && maxKeys == 1 && delimiter == "" && marker == "" {
 		// Optimization for certain applications like
@@ -1249,8 +1301,8 @@ func (z *erasureServerPools) ListObjects(ctx context.Context, bucket, prefix, ma
 		// to avoid the need for ListObjects().
 		objInfo, err := z.GetObjectInfo(ctx, bucket, prefix, ObjectOptions{NoLock: true})
 		if err == nil {
-			if lc != nil {
-				action := evalActionFromLifecycle(ctx, *lc, rcfg, objInfo, false)
+			if opts.Lifecycle != nil {
+				action := evalActionFromLifecycle(ctx, *opts.Lifecycle, opts.Retention, objInfo, false)
 				switch action {
 				case lifecycle.DeleteVersionAction, lifecycle.DeleteAction:
 					fallthrough
@@ -1263,18 +1315,6 @@ func (z *erasureServerPools) ListObjects(ctx context.Context, bucket, prefix, ma
 		}
 	}
 
-	opts := listPathOptions{
-		Bucket:      bucket,
-		Prefix:      prefix,
-		Separator:   delimiter,
-		Limit:       maxKeysPlusOne(maxKeys, marker != ""),
-		Marker:      marker,
-		InclDeleted: false,
-		AskDisks:    globalAPIConfig.getListQuorum(),
-		Lifecycle:   lc,
-		Retention:   rcfg,
-	}
-
 	merged, err := z.listPath(ctx, &opts)
 	if err != nil && err != io.EOF {
 		if !isErrBucketNotFound(err) {
@@ -1539,9 +1579,9 @@ func (z *erasureServerPools) CompleteMultipartUpload(ctx context.Context, bucket
 }
 
 // GetBucketInfo - returns bucket info from one of the erasure coded serverPools.
-func (z *erasureServerPools) GetBucketInfo(ctx context.Context, bucket string) (bucketInfo BucketInfo, err error) {
+func (z *erasureServerPools) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (bucketInfo BucketInfo, err error) {
 	if z.SinglePool() {
-		bucketInfo, err = z.serverPools[0].GetBucketInfo(ctx, bucket)
+		bucketInfo, err = z.serverPools[0].GetBucketInfo(ctx, bucket, opts)
 		if err != nil {
 			return bucketInfo, err
 		}
@@ -1552,7 +1592,7 @@ func (z *erasureServerPools) GetBucketInfo(ctx context.Context, bucket string) (
 		return bucketInfo, nil
 	}
 	for _, pool := range z.serverPools {
-		bucketInfo, err = pool.GetBucketInfo(ctx, bucket)
+		bucketInfo, err = pool.GetBucketInfo(ctx, bucket, opts)
 		if err != nil {
 			if isErrBucketNotFound(err) {
 				continue
@@ -1626,7 +1666,11 @@ func (z *erasureServerPools) DeleteBucket(ctx context.Context, bucket string, op
 
 	// Purge the entire bucket metadata entirely.
 	z.renameAll(context.Background(), minioMetaBucket, pathJoin(bucketMetaPrefix, bucket))
-
+	// If site replication is configured, hold on to deleted bucket state until sites sync
+	switch opts.SRDeleteOp {
+	case MarkDelete:
+		z.markDelete(context.Background(), minioMetaBucket, pathJoin(bucketMetaPrefix, deletedBucketsPrefix, bucket))
+	}
 	// Success.
 	return nil
 }
@@ -1644,6 +1688,27 @@ func (z *erasureServerPools) renameAll(ctx context.Context, bucket, prefix strin
 	}
 }
 
+// markDelete will create a directory of deleted bucket in .minio.sys/buckets/.deleted across all disks
+// in situations where the deleted bucket needs to be held on to until all sites are in sync for
+// site replication
+func (z *erasureServerPools) markDelete(ctx context.Context, bucket, prefix string) {
+	for _, servers := range z.serverPools {
+		for _, set := range servers.sets {
+			set.markDelete(ctx, bucket, prefix)
+		}
+	}
+}
+
+// purgeDelete deletes vol entry in .minio.sys/buckets/.deleted after site replication
+// syncs the delete to peers.
+func (z *erasureServerPools) purgeDelete(ctx context.Context, bucket, prefix string) {
+	for _, servers := range z.serverPools {
+		for _, set := range servers.sets {
+			set.purgeDelete(ctx, bucket, prefix)
+		}
+	}
+}
+
 // This function is used to undo a successful DeleteBucket operation.
 func undoDeleteBucketServerPools(ctx context.Context, bucket string, serverPools []*erasureSets, errs []error) {
 	g := errgroup.WithNErrs(len(serverPools))
@@ -1653,7 +1718,7 @@ func undoDeleteBucketServerPools(ctx context.Context, bucket string, serverPools
 		index := index
 		g.Go(func() error {
 			if errs[index] == nil {
-				return serverPools[index].MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+				return serverPools[index].MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 			}
 			return nil
 		}, index)
@@ -1665,15 +1730,15 @@ func undoDeleteBucketServerPools(ctx context.Context, bucket string, serverPools
 // List all buckets from one of the serverPools, we are not doing merge
 // sort here just for simplification. As per design it is assumed
 // that all buckets are present on all serverPools.
-func (z *erasureServerPools) ListBuckets(ctx context.Context) (buckets []BucketInfo, err error) {
+func (z *erasureServerPools) ListBuckets(ctx context.Context, opts BucketOptions) (buckets []BucketInfo, err error) {
 	if z.SinglePool() {
-		buckets, err = z.serverPools[0].ListBuckets(ctx)
+		buckets, err = z.serverPools[0].ListBuckets(ctx, opts)
 	} else {
 		for idx, pool := range z.serverPools {
 			if z.IsSuspended(idx) {
 				continue
 			}
-			buckets, err = pool.ListBuckets(ctx)
+			buckets, err = pool.ListBuckets(ctx, opts)
 			if err != nil {
 				logger.LogIf(ctx, err)
 				continue
@@ -1685,9 +1750,9 @@ func (z *erasureServerPools) ListBuckets(ctx context.Context) (buckets []BucketI
 		return nil, err
 	}
 	for i := range buckets {
-		meta, err := globalBucketMetadataSys.Get(buckets[i].Name)
+		createdAt, err := globalBucketMetadataSys.CreatedAt(buckets[i].Name)
 		if err == nil {
-			buckets[i].Created = meta.Created
+			buckets[i].Created = createdAt
 		}
 	}
 	return buckets, nil
@@ -1741,7 +1806,9 @@ func (z *erasureServerPools) HealBucket(ctx context.Context, bucket string, opts
 	}
 
 	// Attempt heal on the bucket metadata, ignore any failures
-	defer z.HealObject(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, bucket, bucketMetadataFile), "", opts)
+	hopts := opts
+	hopts.Recreate = false
+	defer z.HealObject(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, bucket, bucketMetadataFile), "", hopts)
 
 	for _, pool := range z.serverPools {
 		result, err := pool.HealBucket(ctx, bucket, opts)
@@ -1804,18 +1871,17 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 							cancel()
 							return
 						}
-						if opts.WalkAscending {
-							for i := len(fivs.Versions) - 1; i >= 0; i-- {
-								version := fivs.Versions[i]
-								versioned := vcfg != nil && vcfg.Versioned(version.Name)
 
-								results <- version.ToObjectInfo(bucket, version.Name, versioned)
-							}
-							return
-						}
+						versionsSorter(fivs.Versions).reverse()
+
 						for _, version := range fivs.Versions {
 							versioned := vcfg != nil && vcfg.Versioned(version.Name)
-							results <- version.ToObjectInfo(bucket, version.Name, versioned)
+
+							select {
+							case <-ctx.Done():
+								return
+							case results <- version.ToObjectInfo(bucket, version.Name, versioned):
+							}
 						}
 					}
 
@@ -1857,6 +1923,7 @@ func (z *erasureServerPools) Walk(ctx context.Context, bucket, prefix string, re
 
 					if err := listPathRaw(ctx, lopts); err != nil {
 						logger.LogIf(ctx, fmt.Errorf("listPathRaw returned %w: opts(%#v)", err, lopts))
+						cancel()
 						return
 					}
 				}()
@@ -1877,7 +1944,7 @@ func listAndHeal(ctx context.Context, bucket, prefix string, set *erasureObjects
 
 	disks, _ := set.getOnlineDisksWithHealing()
 	if len(disks) == 0 {
-		return errors.New("listAndHeal: No non-healing disks found")
+		return errors.New("listAndHeal: No non-healing drives found")
 	}
 
 	// How to resolve partial results.
@@ -2068,7 +2135,7 @@ func (z *erasureServerPools) getPoolAndSet(id string) (poolIdx, setIdx, diskIdx
 			}
 		}
 	}
-	return -1, -1, -1, fmt.Errorf("DiskID(%s) %w", id, errDiskNotFound)
+	return -1, -1, -1, fmt.Errorf("DriveID(%s) %w", id, errDiskNotFound)
 }
 
 // HealthOptions takes input options to return sepcific information

cmd/erasure-sets.go

@@ -27,6 +27,7 @@ import (
 	"net/http"
 	"reflect"
 	"sort"
+	"strings"
 	"sync"
 	"time"
 
@@ -128,10 +129,10 @@ func connectEndpoint(endpoint Endpoint) (StorageAPI, *formatErasureV3, error) {
 		if errors.Is(err, errUnformattedDisk) {
 			info, derr := disk.DiskInfo(context.TODO())
 			if derr != nil && info.RootDisk {
-				return nil, nil, fmt.Errorf("Disk: %s is a root disk", disk)
+				return nil, nil, fmt.Errorf("Drive: %s is a root drive", disk)
 			}
 		}
-		return nil, nil, fmt.Errorf("Disk: %s returned %w", disk, err) // make sure to '%w' to wrap the error
+		return nil, nil, fmt.Errorf("Drive: %s returned %w", disk, err) // make sure to '%w' to wrap the error
 	}
 
 	return disk, format, nil
@@ -146,7 +147,7 @@ func findDiskIndexByDiskID(refFormat *formatErasureV3, diskID string) (int, int,
 		return -1, -1, errDiskNotFound
 	}
 	if diskID == offlineDiskUUID {
-		return -1, -1, fmt.Errorf("diskID: %s is offline", diskID)
+		return -1, -1, fmt.Errorf("DriveID: %s is offline", diskID)
 	}
 	for i := 0; i < len(refFormat.Erasure.Sets); i++ {
 		for j := 0; j < len(refFormat.Erasure.Sets[0]); j++ {
@@ -156,7 +157,7 @@ func findDiskIndexByDiskID(refFormat *formatErasureV3, diskID string) (int, int,
 		}
 	}
 
-	return -1, -1, fmt.Errorf("diskID: %s not found", diskID)
+	return -1, -1, fmt.Errorf("DriveID: %s not found", diskID)
 }
 
 // findDiskIndex - returns the i,j'th position of the input `format` against the reference
@@ -169,7 +170,7 @@ func findDiskIndex(refFormat, format *formatErasureV3) (int, int, error) {
 	}
 
 	if format.Erasure.This == offlineDiskUUID {
-		return -1, -1, fmt.Errorf("diskID: %s is offline", format.Erasure.This)
+		return -1, -1, fmt.Errorf("DriveID: %s is offline", format.Erasure.This)
 	}
 
 	for i := 0; i < len(refFormat.Erasure.Sets); i++ {
@@ -180,7 +181,7 @@ func findDiskIndex(refFormat, format *formatErasureV3) (int, int, error) {
 		}
 	}
 
-	return -1, -1, fmt.Errorf("diskID: %s not found", format.Erasure.This)
+	return -1, -1, fmt.Errorf("DriveID: %s not found", format.Erasure.This)
 }
 
 // connectDisks - attempt to connect all the endpoints, loads format
@@ -238,7 +239,7 @@ func (s *erasureSets) connectDisks() {
 			s.erasureDisksMu.Lock()
 			if currentDisk := s.erasureDisks[setIndex][diskIndex]; currentDisk != nil {
 				if !reflect.DeepEqual(currentDisk.Endpoint(), disk.Endpoint()) {
-					err = fmt.Errorf("Detected unexpected disk ordering refusing to use the disk: expecting %s, found %s, refusing to use the disk",
+					err = fmt.Errorf("Detected unexpected drive ordering refusing to use the drive: expecting %s, found %s, refusing to use the drive",
 						currentDisk.Endpoint(), disk.Endpoint())
 					printEndpointError(endpoint, err, false)
 					disk.Close()
@@ -299,7 +300,7 @@ func (s *erasureSets) monitorAndConnectEndpoints(ctx context.Context, monitorInt
 			return
 		case <-monitor.C:
 			if serverDebugLog {
-				console.Debugln("running disk monitoring")
+				console.Debugln("running drive monitoring")
 			}
 
 			s.connectDisks()
@@ -445,7 +446,7 @@ func newErasureSets(ctx context.Context, endpoints PoolEndpoints, storageDisks [
 						return
 					}
 					if m != i || n != j {
-						logger.LogIf(ctx, fmt.Errorf("Detected unexpected disk ordering refusing to use the disk - poolID: %s, found disk mounted at (set=%s, disk=%s) expected mount at (set=%s, disk=%s): %s(%s)", humanize.Ordinal(poolIdx+1), humanize.Ordinal(m+1), humanize.Ordinal(n+1), humanize.Ordinal(i+1), humanize.Ordinal(j+1), disk, diskID))
+						logger.LogIf(ctx, fmt.Errorf("Detected unexpected drive ordering refusing to use the drive - poolID: %s, found drive mounted at (set=%s, drive=%s) expected mount at (set=%s, drive=%s): %s(%s)", humanize.Ordinal(poolIdx+1), humanize.Ordinal(m+1), humanize.Ordinal(n+1), humanize.Ordinal(i+1), humanize.Ordinal(j+1), disk, diskID))
 						s.erasureDisks[i][j] = &unrecognizedDisk{storage: disk}
 						return
 					}
@@ -689,7 +690,7 @@ func (s *erasureSets) Shutdown(ctx context.Context) error {
 
 // MakeBucketLocation - creates a new bucket across all sets simultaneously,
 // then return the first encountered error
-func (s *erasureSets) MakeBucketWithLocation(ctx context.Context, bucket string, opts BucketOptions) error {
+func (s *erasureSets) MakeBucketWithLocation(ctx context.Context, bucket string, opts MakeBucketOptions) error {
 	g := errgroup.WithNErrs(len(s.sets))
 
 	// Create buckets in parallel across all sets.
@@ -760,8 +761,8 @@ func (s *erasureSets) getHashedSet(input string) (set *erasureObjects) {
 }
 
 // GetBucketInfo - returns bucket info from one of the erasure coded set.
-func (s *erasureSets) GetBucketInfo(ctx context.Context, bucket string) (bucketInfo BucketInfo, err error) {
-	return s.getHashedSet("").GetBucketInfo(ctx, bucket)
+func (s *erasureSets) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (bucketInfo BucketInfo, err error) {
+	return s.getHashedSet("").GetBucketInfo(ctx, bucket, opts)
 }
 
 // IsNotificationSupported returns whether bucket notification is applicable for this layer.
@@ -825,7 +826,7 @@ func undoDeleteBucketSets(ctx context.Context, bucket string, sets []*erasureObj
 		index := index
 		g.Go(func() error {
 			if errs[index] == nil {
-				return sets[index].MakeBucketWithLocation(ctx, bucket, BucketOptions{})
+				return sets[index].MakeBucketWithLocation(ctx, bucket, MakeBucketOptions{})
 			}
 			return nil
 		}, index)
@@ -837,7 +838,7 @@ func undoDeleteBucketSets(ctx context.Context, bucket string, sets []*erasureObj
 // List all buckets from one of the set, we are not doing merge
 // sort here just for simplification. As per design it is assumed
 // that all buckets are present on all sets.
-func (s *erasureSets) ListBuckets(ctx context.Context) (buckets []BucketInfo, err error) {
+func (s *erasureSets) ListBuckets(ctx context.Context, opts BucketOptions) (buckets []BucketInfo, err error) {
 	var listBuckets []BucketInfo
 	healBuckets := map[string]VolInfo{}
 	for _, set := range s.sets {
@@ -847,8 +848,34 @@ func (s *erasureSets) ListBuckets(ctx context.Context) (buckets []BucketInfo, er
 		}
 	}
 
+	// include deleted buckets in listBuckets output
+	deletedBuckets := map[string]VolInfo{}
+
+	if opts.Deleted {
+		for _, set := range s.sets {
+			// lists all unique buckets across drives.
+			if err := listDeletedBuckets(ctx, set.getDisks(), deletedBuckets, s.defaultParityCount); err != nil {
+				return nil, err
+			}
+		}
+	}
 	for _, v := range healBuckets {
-		listBuckets = append(listBuckets, BucketInfo(v))
+		bi := BucketInfo{
+			Name:    v.Name,
+			Created: v.Created,
+		}
+		if vi, ok := deletedBuckets[v.Name]; ok {
+			bi.Deleted = vi.Created
+		}
+		listBuckets = append(listBuckets, bi)
+	}
+	for _, v := range deletedBuckets {
+		if _, ok := healBuckets[v.Name]; !ok {
+			listBuckets = append(listBuckets, BucketInfo{
+				Name:    v.Name,
+				Deleted: v.Created,
+			})
+		}
 	}
 
 	sort.Slice(listBuckets, func(i, j int) bool {
@@ -858,6 +885,45 @@ func (s *erasureSets) ListBuckets(ctx context.Context) (buckets []BucketInfo, er
 	return listBuckets, nil
 }
 
+// listDeletedBuckets lists deleted buckets from all disks.
+func listDeletedBuckets(ctx context.Context, storageDisks []StorageAPI, delBuckets map[string]VolInfo, readQuorum int) error {
+	g := errgroup.WithNErrs(len(storageDisks))
+	var mu sync.Mutex
+	for index := range storageDisks {
+		index := index
+		g.Go(func() error {
+			if storageDisks[index] == nil {
+				// we ignore disk not found errors
+				return nil
+			}
+			volsInfo, err := storageDisks[index].ListDir(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, deletedBucketsPrefix), -1)
+			if err != nil {
+				if err == errFileNotFound {
+					return nil
+				}
+				return err
+			}
+			for _, volName := range volsInfo {
+				mu.Lock()
+				if _, ok := delBuckets[volName]; !ok {
+					vi, err := storageDisks[index].StatVol(ctx, pathJoin(minioMetaBucket, bucketMetaPrefix, deletedBucketsPrefix, volName))
+					if err != nil {
+						return err
+					}
+					bkt := strings.TrimPrefix(vi.Name, pathJoin(minioMetaBucket, bucketMetaPrefix, deletedBucketsPrefix))
+					bkt = strings.TrimPrefix(bkt, slashSeparator)
+					bkt = strings.TrimSuffix(bkt, slashSeparator)
+					vi.Name = bkt
+					delBuckets[bkt] = vi
+				}
+				mu.Unlock()
+			}
+			return nil
+		}, index)
+	}
+	return reduceReadQuorumErrs(ctx, g.Wait(), bucketMetadataOpIgnoredErrs, readQuorum)
+}
+
 // --- Object Operations ---
 
 // GetObjectNInfo - returns object info and locked object ReadCloser
@@ -1174,7 +1240,7 @@ func markRootDisksAsDown(storageDisks []StorageAPI, errs []error) {
 		if storageDisks[i] != nil && infos[i].RootDisk {
 			// We should not heal on root disk. i.e in a situation where the minio-administrator has unmounted a
 			// defective drive we should not heal a path on the root disk.
-			logger.LogIf(GlobalContext, fmt.Errorf("Disk `%s` is part of root disk, will not be used", storageDisks[i]))
+			logger.LogIf(GlobalContext, fmt.Errorf("Drive `%s` is part of root drive, will not be used", storageDisks[i]))
 			storageDisks[i] = nil
 		}
 	}
@@ -1248,7 +1314,7 @@ func (s *erasureSets) HealFormat(ctx context.Context, dryRun bool) (res madmin.H
 				continue
 			}
 			if err := saveFormatErasure(storageDisks[index], format, true); err != nil {
-				logger.LogIf(ctx, fmt.Errorf("Disk %s failed to write updated 'format.json': %v", storageDisks[index], err))
+				logger.LogIf(ctx, fmt.Errorf("Drive %s failed to write updated 'format.json': %v", storageDisks[index], err))
 				tmpNewFormats[index] = nil // this disk failed to write new format
 			}
 		}

cmd/erasure-sets_test.go

@@ -187,7 +187,7 @@ func TestNewErasureSets(t *testing.T) {
 	// Initializes all erasure disks
 	storageDisks, format, err := waitForFormatErasure(true, endpoints, 1, 1, 16, "", "")
 	if err != nil {
-		t.Fatalf("Unable to format disks for erasure, %s", err)
+		t.Fatalf("Unable to format drives for erasure, %s", err)
 	}
 
 	ep := PoolEndpoints{Endpoints: endpoints}

cmd/erasure-single-drive.go

@@ -108,16 +108,40 @@ func newErasureSingle(ctx context.Context, storageDisk StorageAPI, format *forma
 // List all buckets from one of the set, we are not doing merge
 // sort here just for simplification. As per design it is assumed
 // that all buckets are present on all sets.
-func (es *erasureSingle) ListBuckets(ctx context.Context) (buckets []BucketInfo, err error) {
+func (es *erasureSingle) ListBuckets(ctx context.Context, opts BucketOptions) (buckets []BucketInfo, err error) {
 	var listBuckets []BucketInfo
 	healBuckets := map[string]VolInfo{}
 	// lists all unique buckets across drives.
 	if err := listAllBuckets(ctx, []StorageAPI{es.disk}, healBuckets, 0); err != nil {
 		return nil, err
 	}
+	// include deleted buckets in listBuckets output
+	deletedBuckets := map[string]VolInfo{}
 
+	if opts.Deleted {
+		// lists all deleted buckets across drives.
+		if err := listDeletedBuckets(ctx, []StorageAPI{es.disk}, deletedBuckets, 0); err != nil {
+			return nil, err
+		}
+	}
 	for _, v := range healBuckets {
-		listBuckets = append(listBuckets, BucketInfo(v))
+		bi := BucketInfo{
+			Name:    v.Name,
+			Created: v.Created,
+		}
+		if vi, ok := deletedBuckets[v.Name]; ok {
+			bi.Deleted = vi.Created
+		}
+		listBuckets = append(listBuckets, bi)
+	}
+
+	for _, v := range deletedBuckets {
+		if _, ok := healBuckets[v.Name]; !ok {
+			listBuckets = append(listBuckets, BucketInfo{
+				Name:    v.Name,
+				Deleted: v.Created,
+			})
+		}
 	}
 
 	sort.Slice(listBuckets, func(i, j int) bool {
@@ -256,7 +280,7 @@ type renameAllStorager interface {
 
 // Bucket operations
 // MakeBucket - make a bucket.
-func (es *erasureSingle) MakeBucketWithLocation(ctx context.Context, bucket string, opts BucketOptions) error {
+func (es *erasureSingle) MakeBucketWithLocation(ctx context.Context, bucket string, opts MakeBucketOptions) error {
 	defer NSUpdated(bucket, slashSeparator)
 
 	// Lock the bucket name before creating.
@@ -289,6 +313,7 @@ func (es *erasureSingle) MakeBucketWithLocation(ctx context.Context, bucket stri
 
 	// If it doesn't exist we get a new, so ignore errors
 	meta := newBucketMetadata(bucket)
+	meta.SetCreatedAt(opts.CreatedAt)
 	if opts.LockEnabled {
 		meta.VersioningConfigXML = enabledBucketVersioningConfig
 		meta.ObjectLockConfigXML = enabledBucketObjectLockConfig
@@ -308,12 +333,17 @@ func (es *erasureSingle) MakeBucketWithLocation(ctx context.Context, bucket stri
 }
 
 // GetBucketInfo - returns BucketInfo for a bucket.
-func (es *erasureSingle) GetBucketInfo(ctx context.Context, bucket string) (bi BucketInfo, e error) {
+func (es *erasureSingle) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (bi BucketInfo, e error) {
 	volInfo, err := es.disk.StatVol(ctx, bucket)
 	if err != nil {
+		if opts.Deleted {
+			if dvi, derr := es.disk.StatVol(ctx, pathJoin(minioMetaBucket, bucketMetaPrefix, deletedBucketsPrefix, bucket)); derr == nil {
+				return BucketInfo{Name: bucket, Deleted: dvi.Created}, nil
+			}
+		}
 		return bi, toObjectErr(err, bucket)
 	}
-	return BucketInfo(volInfo), nil
+	return BucketInfo{Name: volInfo.Name, Created: volInfo.Created}, nil
 }
 
 // DeleteBucket - deletes a bucket.
@@ -326,6 +356,28 @@ func (es *erasureSingle) DeleteBucket(ctx context.Context, bucket string, opts D
 	deleteBucketMetadata(ctx, es, bucket)
 	globalBucketMetadataSys.Remove(bucket)
 
+	if err == nil || errors.Is(err, errVolumeNotFound) {
+		if opts.SRDeleteOp == MarkDelete {
+			es.markDelete(ctx, minioMetaBucket, pathJoin(bucketMetaPrefix, deletedBucketsPrefix, bucket))
+		}
+	}
+	return toObjectErr(err, bucket)
+}
+
+// markDelete creates a vol entry in .minio.sys/buckets/.deleted until site replication
+// syncs the delete to peers
+func (es *erasureSingle) markDelete(ctx context.Context, bucket, prefix string) error {
+	err := es.disk.MakeVol(ctx, pathJoin(bucket, prefix))
+	if err != nil && errors.Is(err, errVolumeExists) {
+		return nil
+	}
+	return toObjectErr(err, bucket)
+}
+
+// purgeDelete deletes vol entry in .minio.sys/buckets/.deleted after site replication
+// syncs the delete to peers OR on a new MakeBucket call.
+func (es *erasureSingle) purgeDelete(ctx context.Context, bucket, prefix string) error {
+	err := es.disk.DeleteVol(ctx, pathJoin(bucket, prefix), true)
 	return toObjectErr(err, bucket)
 }
 
@@ -1151,6 +1203,8 @@ func (es *erasureSingle) putObject(ctx context.Context, bucket string, object st
 		return ObjectInfo{}, toObjectErr(err, bucket, object)
 	}
 
+	defer NSUpdated(bucket, object)
+
 	for i := 0; i < len(onlineDisks); i++ {
 		if onlineDisks[i] != nil && onlineDisks[i].IsOnline() {
 			// Object info is the same in all disks, so we can pick
@@ -1169,7 +1223,7 @@ func (es *erasureSingle) putObject(ctx context.Context, bucket string, object st
 	return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
 }
 
-func (es *erasureSingle) deleteObjectVersion(ctx context.Context, bucket, object string, writeQuorum int, fi FileInfo, forceDelMarker bool) error {
+func (es *erasureSingle) deleteObjectVersion(ctx context.Context, bucket, object string, fi FileInfo, forceDelMarker bool) error {
 	return es.disk.DeleteVersion(ctx, bucket, object, fi, forceDelMarker)
 }
 
@@ -1256,12 +1310,12 @@ func (es *erasureSingle) DeleteObjects(ctx context.Context, bucket string, objec
 				DeleteMarker:          vr.Deleted,
 				DeleteMarkerVersionID: vr.VersionID,
 				DeleteMarkerMTime:     DeleteMarkerMTime{vr.ModTime},
-				ObjectName:            vr.Name,
+				ObjectName:            decodeDirObject(vr.Name),
 				ReplicationState:      vr.ReplicationState,
 			}
 		} else {
 			dobjects[i] = DeletedObject{
-				ObjectName:       vr.Name,
+				ObjectName:       decodeDirObject(vr.Name),
 				VersionID:        vr.VersionID,
 				ReplicationState: vr.ReplicationState,
 			}
@@ -1393,7 +1447,7 @@ func (es *erasureSingle) DeleteObject(ctx context.Context, bucket, object string
 
 	versionFound := true
 	objInfo = ObjectInfo{VersionID: opts.VersionID} // version id needed in Delete API response.
-	goi, writeQuorum, gerr := es.getObjectInfoAndQuorum(ctx, bucket, object, opts)
+	goi, _, gerr := es.getObjectInfoAndQuorum(ctx, bucket, object, opts)
 	if gerr != nil && goi.Name == "" {
 		switch gerr.(type) {
 		case InsufficientReadQuorum:
@@ -1407,6 +1461,11 @@ func (es *erasureSingle) DeleteObject(ctx context.Context, bucket, object string
 		}
 	}
 
+	// Do not create new delete markers.
+	if goi.DeleteMarker && opts.VersionID == "" {
+		return goi, nil
+	}
+
 	if opts.Expiration.Expire {
 		action := evalActionFromLifecycle(ctx, *lc, rcfg, goi, false)
 		var isErr bool
@@ -1503,7 +1562,7 @@ func (es *erasureSingle) DeleteObject(ctx context.Context, bucket, object string
 			// delete marker. Add delete marker, since we don't have
 			// any version specified explicitly. Or if a particular
 			// version id needs to be replicated.
-			if err = es.deleteObjectVersion(ctx, bucket, object, writeQuorum, fi, opts.DeleteMarker); err != nil {
+			if err = es.deleteObjectVersion(ctx, bucket, object, fi, opts.DeleteMarker); err != nil {
 				return objInfo, toObjectErr(err, bucket, object)
 			}
 			return fi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
@@ -1522,7 +1581,7 @@ func (es *erasureSingle) DeleteObject(ctx context.Context, bucket, object string
 		ExpireRestored:   opts.Transition.ExpireRestored,
 	}
 	dfi.SetTierFreeVersionID(fvID)
-	if err = es.deleteObjectVersion(ctx, bucket, object, writeQuorum, dfi, opts.DeleteMarker); err != nil {
+	if err = es.deleteObjectVersion(ctx, bucket, object, dfi, opts.DeleteMarker); err != nil {
 		return objInfo, toObjectErr(err, bucket, object)
 	}
 
@@ -1759,14 +1818,7 @@ func (es *erasureSingle) TransitionObject(ctx context.Context, bucket, object st
 	fi.TransitionVersionID = string(rv)
 	eventName := event.ObjectTransitionComplete
 
-	// we now know the number of blocks this object needs for data and parity.
-	// writeQuorum is dataBlocks + 1
-	writeQuorum := fi.Erasure.DataBlocks
-	if fi.Erasure.DataBlocks == fi.Erasure.ParityBlocks {
-		writeQuorum++
-	}
-
-	if err = es.deleteObjectVersion(ctx, bucket, object, writeQuorum, fi, false); err != nil {
+	if err = es.deleteObjectVersion(ctx, bucket, object, fi, false); err != nil {
 		eventName = event.ObjectTransitionFailed
 	}
 
@@ -2752,6 +2804,8 @@ func (es *erasureSingle) CompleteMultipartUpload(ctx context.Context, bucket str
 		return oi, toObjectErr(err, bucket, object)
 	}
 
+	defer NSUpdated(bucket, object)
+
 	for i := 0; i < len(onlineDisks); i++ {
 		if onlineDisks[i] != nil && onlineDisks[i].IsOnline() {
 			// Object info is the same in all disks, so we can pick
@@ -2803,11 +2857,16 @@ func (es *erasureSingle) AbortMultipartUpload(ctx context.Context, bucket, objec
 func (es *erasureSingle) ListObjects(ctx context.Context, bucket, prefix, marker, delimiter string, maxKeys int) (ListObjectsInfo, error) {
 	var loi ListObjectsInfo
 
-	// Automatically remove the object/version is an expiry lifecycle rule can be applied
-	lc, _ := globalLifecycleSys.Get(bucket)
-
-	// Check if bucket is object locked.
-	rcfg, _ := globalBucketObjectLockSys.Get(bucket)
+	opts := listPathOptions{
+		Bucket:      bucket,
+		Prefix:      prefix,
+		Separator:   delimiter,
+		Limit:       maxKeysPlusOne(maxKeys, marker != ""),
+		Marker:      marker,
+		InclDeleted: false,
+		AskDisks:    globalAPIConfig.getListQuorum(),
+	}
+	opts.setBucketMeta(ctx)
 
 	if len(prefix) > 0 && maxKeys == 1 && delimiter == "" && marker == "" {
 		// Optimization for certain applications like
@@ -2819,8 +2878,8 @@ func (es *erasureSingle) ListObjects(ctx context.Context, bucket, prefix, marker
 		// to avoid the need for ListObjects().
 		objInfo, err := es.GetObjectInfo(ctx, bucket, prefix, ObjectOptions{NoLock: true})
 		if err == nil {
-			if lc != nil {
-				action := evalActionFromLifecycle(ctx, *lc, rcfg, objInfo, false)
+			if opts.Lifecycle != nil {
+				action := evalActionFromLifecycle(ctx, *opts.Lifecycle, opts.Retention, objInfo, false)
 				switch action {
 				case lifecycle.DeleteVersionAction, lifecycle.DeleteAction:
 					fallthrough
@@ -2833,18 +2892,6 @@ func (es *erasureSingle) ListObjects(ctx context.Context, bucket, prefix, marker
 		}
 	}
 
-	opts := listPathOptions{
-		Bucket:      bucket,
-		Prefix:      prefix,
-		Separator:   delimiter,
-		Limit:       maxKeysPlusOne(maxKeys, marker != ""),
-		Marker:      marker,
-		InclDeleted: false,
-		AskDisks:    globalAPIConfig.getListQuorum(),
-		Lifecycle:   lc,
-		Retention:   rcfg,
-	}
-
 	merged, err := es.listPath(ctx, &opts)
 	if err != nil && err != io.EOF {
 		if !isErrBucketNotFound(err) {
@@ -2903,7 +2950,6 @@ func (es *erasureSingle) ListObjectVersions(ctx context.Context, bucket, prefix,
 	if marker == "" && versionMarker != "" {
 		return loi, NotImplemented{}
 	}
-
 	opts := listPathOptions{
 		Bucket:      bucket,
 		Prefix:      prefix,
@@ -2914,6 +2960,7 @@ func (es *erasureSingle) ListObjectVersions(ctx context.Context, bucket, prefix,
 		AskDisks:    "strict",
 		Versioned:   true,
 	}
+	opts.setBucketMeta(ctx)
 
 	merged, err := es.listPath(ctx, &opts)
 	if err != nil && err != io.EOF {
@@ -2960,13 +3007,13 @@ func (es *erasureSingle) Walk(ctx context.Context, bucket, prefix string, result
 		return err
 	}
 
+	vcfg, _ := globalBucketVersioningSys.Get(bucket)
+
 	ctx, cancel := context.WithCancel(ctx)
 	go func() {
 		defer cancel()
 		defer close(results)
 
-		versioned := opts.Versioned || opts.VersionSuspended
-
 		var wg sync.WaitGroup
 		wg.Add(1)
 		go func() {
@@ -2981,15 +3028,17 @@ func (es *erasureSingle) Walk(ctx context.Context, bucket, prefix string, result
 					cancel()
 					return
 				}
-				if opts.WalkAscending {
-					for i := len(fivs.Versions) - 1; i >= 0; i-- {
-						version := fivs.Versions[i]
-						results <- version.ToObjectInfo(bucket, version.Name, versioned)
-					}
-					return
-				}
+
+				versionsSorter(fivs.Versions).reverse()
+
 				for _, version := range fivs.Versions {
-					results <- version.ToObjectInfo(bucket, version.Name, versioned)
+					versioned := vcfg != nil && vcfg.Versioned(version.Name)
+
+					select {
+					case <-ctx.Done():
+						return
+					case results <- version.ToObjectInfo(bucket, version.Name, versioned):
+					}
 				}
 			}
 
@@ -3031,6 +3080,7 @@ func (es *erasureSingle) Walk(ctx context.Context, bucket, prefix string, result
 
 			if err := listPathRaw(ctx, lopts); err != nil {
 				logger.LogIf(ctx, fmt.Errorf("listPathRaw returned %w: opts(%#v)", err, lopts))
+				cancel()
 				return
 			}
 		}()
@@ -3040,6 +3090,33 @@ func (es *erasureSingle) Walk(ctx context.Context, bucket, prefix string, result
 	return nil
 }
 
+// Health - returns current status of the object layer health, for single drive
+// its as simple as returning healthy as long as drive is accessible.
+func (es *erasureSingle) Health(ctx context.Context, opts HealthOptions) HealthResult {
+	_, err := es.disk.DiskInfo(ctx)
+	if err != nil {
+		return HealthResult{}
+	}
+	if opts.Maintenance {
+		// Single drive cannot be put under maintenance.
+		return HealthResult{
+			Healthy:     false,
+			WriteQuorum: 1,
+		}
+	}
+	return HealthResult{
+		Healthy:     true,
+		WriteQuorum: 1,
+	}
+}
+
+// ReadHealth - returns current status of the object layer health for reads,
+// for single drive its as simple as returning healthy as long as drive is accessible.
+func (es *erasureSingle) ReadHealth(ctx context.Context) bool {
+	res := es.Health(ctx, HealthOptions{})
+	return res.Healthy
+}
+
 // nsScanner will start scanning buckets and send updated totals as they are traversed.
 // Updates are sent on a regular basis and the caller *must* consume them.
 func (es *erasureSingle) nsScanner(ctx context.Context, buckets []BucketInfo, bf *bloomFilter, wantCycle uint32, updates chan<- dataUsageCache, healScanMode madmin.HealScanMode) error {
@@ -3235,7 +3312,7 @@ func (es *erasureSingle) NSScanner(ctx context.Context, bf *bloomFilter, updates
 	results := make([]dataUsageCache, 1)
 	var firstErr error
 
-	allBuckets, err := es.ListBuckets(ctx)
+	allBuckets, err := es.ListBuckets(ctx, BucketOptions{})
 	if err != nil {
 		return err
 	}

cmd/erasure.go

@@ -23,6 +23,7 @@ import (
 	"fmt"
 	"math/rand"
 	"os"
+	"runtime"
 	"sort"
 	"sync"
 	"time"
@@ -200,8 +201,11 @@ func getDisksInfo(disks []StorageAPI, endpoints []Endpoint) (disksInfo []madmin.
 				UsedSpace:      info.Used,
 				AvailableSpace: info.Free,
 				UUID:           info.ID,
+				Major:          info.Major,
+				Minor:          info.Minor,
 				RootDisk:       info.RootDisk,
 				Healing:        info.Healing,
+				Scanning:       info.Scanning,
 				State:          diskErrToDriveState(err),
 				FreeInodes:     info.FreeInodes,
 			}
@@ -288,7 +292,7 @@ func (er erasureObjects) getOnlineDisksWithHealing() (newDisks []StorageAPI, hea
 			disk := disks[i-1]
 
 			if disk == nil {
-				infos[i-1].Error = "nil disk"
+				infos[i-1].Error = "nil drive"
 				return
 			}
 
@@ -354,7 +358,7 @@ func (er erasureObjects) nsScanner(ctx context.Context, buckets []BucketInfo, bf
 	// Collect disks we can use.
 	disks, healing := er.getOnlineDisksWithHealing()
 	if len(disks) == 0 {
-		logger.LogIf(ctx, errors.New("data-scanner: all disks are offline or being healed, skipping scanner cycle"))
+		logger.LogIf(ctx, errors.New("data-scanner: all drives are offline or being healed, skipping scanner cycle"))
 		return nil
 	}
 
@@ -439,6 +443,13 @@ func (er erasureObjects) nsScanner(ctx context.Context, buckets []BucketInfo, bf
 	r := rand.New(rand.NewSource(time.Now().UnixNano()))
 	r.Shuffle(len(disks), func(i, j int) { disks[i], disks[j] = disks[j], disks[i] })
 
+	// Restrict parallelism for disk usage scanner
+	// upto GOMAXPROCS if GOMAXPROCS is < len(disks)
+	maxProcs := runtime.GOMAXPROCS(0)
+	if maxProcs < len(disks) {
+		disks = disks[:maxProcs]
+	}
+
 	// Start one scanner per disk
 	var wg sync.WaitGroup
 	wg.Add(len(disks))

cmd/erasure_test.go

@@ -22,7 +22,6 @@ import (
 	"context"
 	"crypto/rand"
 	"io"
-	"os"
 	"testing"
 )
 
@@ -118,20 +117,13 @@ type erasureTestSetup struct {
 	disks        []StorageAPI
 }
 
-// Removes the temporary disk directories.
-func (e erasureTestSetup) Remove() {
-	for _, path := range e.diskPaths {
-		os.RemoveAll(path)
-	}
-}
-
 // Returns an initialized setup for erasure tests.
-func newErasureTestSetup(dataBlocks int, parityBlocks int, blockSize int64) (*erasureTestSetup, error) {
+func newErasureTestSetup(tb testing.TB, dataBlocks int, parityBlocks int, blockSize int64) (*erasureTestSetup, error) {
 	diskPaths := make([]string, dataBlocks+parityBlocks)
 	disks := make([]StorageAPI, len(diskPaths))
 	var err error
 	for i := range diskPaths {
-		disks[i], diskPaths[i], err = newXLStorageTestSetup()
+		disks[i], diskPaths[i], err = newXLStorageTestSetup(tb)
 		if err != nil {
 			return nil, err
 		}

cmd/format-disk-cache.go

@@ -110,7 +110,7 @@ func formatCacheGetVersion(r io.ReadSeeker) (string, error) {
 // Creates a new cache format.json if unformatted.
 func createFormatCache(fsFormatPath string, format *formatCacheV1) error {
 	// open file using READ & WRITE permission
-	file, err := os.OpenFile(fsFormatPath, os.O_RDWR|os.O_CREATE, 0o600)
+	file, err := os.OpenFile(fsFormatPath, os.O_RDWR|os.O_CREATE, 0o666)
 	if err != nil {
 		return err
 	}
@@ -155,7 +155,7 @@ func loadFormatCache(ctx context.Context, drives []string) ([]*formatCacheV2, bo
 	migrating := false
 	for i, drive := range drives {
 		cacheFormatPath := pathJoin(drive, minioMetaBucket, formatConfigFile)
-		f, err := os.OpenFile(cacheFormatPath, os.O_RDWR, 0)
+		f, err := os.OpenFile(cacheFormatPath, os.O_RDWR, 0o666)
 		if err != nil {
 			if osIsNotExist(err) {
 				continue
@@ -478,7 +478,7 @@ func migrateOldCache(ctx context.Context, c *diskCache) error {
 
 func migrateCacheFormatJSON(cacheFormatPath string) error {
 	// now migrate format.json
-	f, err := os.OpenFile(cacheFormatPath, os.O_RDWR, 0)
+	f, err := os.OpenFile(cacheFormatPath, os.O_RDWR, 0o666)
 	if err != nil {
 		return err
 	}

cmd/format-erasure.go

@@ -79,7 +79,7 @@ type formatErasureV1 struct {
 	formatMetaV1
 	Erasure struct {
 		Version string `json:"version"` // Version of 'xl' format.
-		Disk    string `json:"disk"`    // Disk field carries assigned disk uuid.
+		Disk    string `json:"drive"`   // Disk field carries assigned disk uuid.
 		// JBOD field carries the input disk order generated the first
 		// time when fresh disks were supplied.
 		JBOD []string `json:"jbod"`
@@ -199,7 +199,7 @@ func formatErasureMigrate(export string) ([]byte, fs.FileInfo, error) {
 
 	version, err := formatGetBackendErasureVersion(formatData)
 	if err != nil {
-		return nil, nil, fmt.Errorf("Disk %s: %w", export, err)
+		return nil, nil, fmt.Errorf("Drive %s: %w", export, err)
 	}
 
 	migrate := func(formatPath string, formatData []byte) ([]byte, fs.FileInfo, error) {
@@ -217,7 +217,7 @@ func formatErasureMigrate(export string) ([]byte, fs.FileInfo, error) {
 	case formatErasureVersionV1:
 		formatData, err = formatErasureMigrateV1ToV2(formatData, version)
 		if err != nil {
-			return nil, nil, fmt.Errorf("Disk %s: %w", export, err)
+			return nil, nil, fmt.Errorf("Drive %s: %w", export, err)
 		}
 		// Migrate successful v1 => v2, proceed to v2 => v3
 		version = formatErasureVersionV2
@@ -225,7 +225,7 @@ func formatErasureMigrate(export string) ([]byte, fs.FileInfo, error) {
 	case formatErasureVersionV2:
 		formatData, err = formatErasureMigrateV2ToV3(formatData, export, version)
 		if err != nil {
-			return nil, nil, fmt.Errorf("Disk %s: %w", export, err)
+			return nil, nil, fmt.Errorf("Drive %s: %w", export, err)
 		}
 		// Migrate successful v2 => v3, v3 is latest
 		// version = formatXLVersionV3
@@ -438,14 +438,14 @@ func checkFormatErasureValues(formats []*formatErasureV3, disks []StorageAPI, se
 			return err
 		}
 		if len(formats) != len(formatErasure.Erasure.Sets)*len(formatErasure.Erasure.Sets[0]) {
-			return fmt.Errorf("%s disk is already being used in another erasure deployment. (Number of disks specified: %d but the number of disks found in the %s disk's format.json: %d)",
+			return fmt.Errorf("%s drive is already being used in another erasure deployment. (Number of drives specified: %d but the number of drives found in the %s drive's format.json: %d)",
 				disks[i], len(formats), humanize.Ordinal(i+1), len(formatErasure.Erasure.Sets)*len(formatErasure.Erasure.Sets[0]))
 		}
 		// Only if custom erasure drive count is set, verify if the
 		// set_drive_count was manually set - we need to honor what is
 		// present on the drives.
 		if globalCustomErasureDriveCount && len(formatErasure.Erasure.Sets[0]) != setDriveCount {
-			return fmt.Errorf("%s disk is already formatted with %d drives per erasure set. This cannot be changed to %d, please revert your MINIO_ERASURE_SET_DRIVE_COUNT setting", disks[i], len(formatErasure.Erasure.Sets[0]), setDriveCount)
+			return fmt.Errorf("%s drive is already formatted with %d drives per erasure set. This cannot be changed to %d, please revert your MINIO_ERASURE_SET_DRIVE_COUNT setting", disks[i], len(formatErasure.Erasure.Sets[0]), setDriveCount)
 		}
 	}
 	return nil
@@ -628,7 +628,7 @@ func formatErasureV3Check(reference *formatErasureV3, format *formatErasureV3) e
 			}
 		}
 	}
-	return fmt.Errorf("Disk ID %s not found in any disk sets %s", this, format.Erasure.Sets)
+	return fmt.Errorf("DriveID %s not found in any drive sets %s", this, format.Erasure.Sets)
 }
 
 // saveFormatErasureAll - populates `format.json` on disks in its order.

cmd/format-erasure_test.go

@@ -104,11 +104,7 @@ func TestFormatErasureEmpty(t *testing.T) {
 // Tests xl format migration.
 func TestFormatErasureMigrate(t *testing.T) {
 	// Get test root.
-	rootPath, err := getTestRoot()
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.RemoveAll(rootPath)
+	rootPath := t.TempDir()
 
 	m := &formatErasureV1{}
 	m.Format = formatBackendErasure
@@ -153,13 +149,13 @@ func TestFormatErasureMigrate(t *testing.T) {
 		t.Fatal(err)
 	}
 	if formatV3.Erasure.This != m.Erasure.Disk {
-		t.Fatalf("expected disk uuid: %s, got: %s", m.Erasure.Disk, formatV3.Erasure.This)
+		t.Fatalf("expected drive uuid: %s, got: %s", m.Erasure.Disk, formatV3.Erasure.This)
 	}
 	if len(formatV3.Erasure.Sets) != 1 {
 		t.Fatalf("expected single set after migrating from v1 to v3, but found %d", len(formatV3.Erasure.Sets))
 	}
 	if !reflect.DeepEqual(formatV3.Erasure.Sets[0], m.Erasure.JBOD) {
-		t.Fatalf("expected disk uuid: %v, got: %v", m.Erasure.JBOD, formatV3.Erasure.Sets[0])
+		t.Fatalf("expected drive uuid: %v, got: %v", m.Erasure.JBOD, formatV3.Erasure.Sets[0])
 	}
 
 	m = &formatErasureV1{}

cmd/format-fs.go

@@ -165,7 +165,7 @@ func formatFSMigrate(ctx context.Context, wlk *lock.LockedFile, fsPath string) e
 func createFormatFS(fsFormatPath string) error {
 	// Attempt a write lock on formatConfigFile `format.json`
 	// file stored in minioMetaBucket(.minio.sys) directory.
-	lk, err := lock.TryLockedOpenFile(fsFormatPath, os.O_RDWR|os.O_CREATE, 0o600)
+	lk, err := lock.TryLockedOpenFile(fsFormatPath, os.O_RDWR|os.O_CREATE, 0o666)
 	if err != nil {
 		return err
 	}
@@ -256,7 +256,7 @@ func initFormatFS(ctx context.Context, fsPath string) (rlk *lock.RLockedFile, er
 			// Hold write lock during migration so that we do not disturb any
 			// minio processes running in parallel.
 			var wlk *lock.LockedFile
-			wlk, err = lock.TryLockedOpenFile(fsFormatPath, os.O_RDWR, 0)
+			wlk, err = lock.TryLockedOpenFile(fsFormatPath, os.O_RDWR, 0o666)
 			if err == lock.ErrAlreadyLocked {
 				// Lock already present, sleep and attempt again.
 				time.Sleep(100 * time.Millisecond)
@@ -357,7 +357,7 @@ func formatFSFixDeploymentID(ctx context.Context, fsFormatPath string) error {
 		case <-ctx.Done():
 			return fmt.Errorf("Initializing FS format stopped gracefully")
 		default:
-			wlk, err = lock.TryLockedOpenFile(fsFormatPath, os.O_RDWR, 0)
+			wlk, err = lock.TryLockedOpenFile(fsFormatPath, os.O_RDWR, 0o666)
 			if err == lock.ErrAlreadyLocked {
 				// Lock already present, sleep and attempt again
 				logger.Info("Another minio process(es) might be holding a lock to the file %s. Please kill that minio process(es) (elapsed %s)\n", fsFormatPath, getElapsedTime())

cmd/fs-v1-helpers_test.go

@@ -30,11 +30,10 @@ import (
 
 func TestFSRenameFile(t *testing.T) {
 	// create xlStorage test setup
-	_, path, err := newXLStorageTestSetup()
+	_, path, err := newXLStorageTestSetup(t)
 	if err != nil {
 		t.Fatalf("Unable to create xlStorage test setup, %s", err)
 	}
-	defer os.RemoveAll(path)
 
 	if err = fsMkdir(GlobalContext, pathJoin(path, "testvolume1")); err != nil {
 		t.Fatal(err)
@@ -55,11 +54,10 @@ func TestFSRenameFile(t *testing.T) {
 
 func TestFSStats(t *testing.T) {
 	// create xlStorage test setup
-	_, path, err := newXLStorageTestSetup()
+	_, path, err := newXLStorageTestSetup(t)
 	if err != nil {
 		t.Fatalf("Unable to create xlStorage test setup, %s", err)
 	}
-	defer os.RemoveAll(path)
 
 	// Setup test environment.
 
@@ -183,11 +181,10 @@ func TestFSStats(t *testing.T) {
 
 func TestFSCreateAndOpen(t *testing.T) {
 	// Setup test environment.
-	_, path, err := newXLStorageTestSetup()
+	_, path, err := newXLStorageTestSetup(t)
 	if err != nil {
 		t.Fatalf("Unable to create xlStorage test setup, %s", err)
 	}
-	defer os.RemoveAll(path)
 
 	if err = fsMkdir(GlobalContext, pathJoin(path, "success-vol")); err != nil {
 		t.Fatalf("Unable to create directory, %s", err)
@@ -248,11 +245,10 @@ func TestFSCreateAndOpen(t *testing.T) {
 
 func TestFSDeletes(t *testing.T) {
 	// create xlStorage test setup
-	_, path, err := newXLStorageTestSetup()
+	_, path, err := newXLStorageTestSetup(t)
 	if err != nil {
 		t.Fatalf("Unable to create xlStorage test setup, %s", err)
 	}
-	defer os.RemoveAll(path)
 
 	// Setup test environment.
 	if err = fsMkdir(GlobalContext, pathJoin(path, "success-vol")); err != nil {
@@ -350,11 +346,10 @@ func TestFSDeletes(t *testing.T) {
 
 func BenchmarkFSDeleteFile(b *testing.B) {
 	// create xlStorage test setup
-	_, path, err := newXLStorageTestSetup()
+	_, path, err := newXLStorageTestSetup(b)
 	if err != nil {
 		b.Fatalf("Unable to create xlStorage test setup, %s", err)
 	}
-	defer os.RemoveAll(path)
 
 	// Setup test environment.
 	if err = fsMkdir(GlobalContext, pathJoin(path, "benchmark")); err != nil {
@@ -384,11 +379,10 @@ func BenchmarkFSDeleteFile(b *testing.B) {
 // Tests fs removes.
 func TestFSRemoves(t *testing.T) {
 	// create xlStorage test setup
-	_, path, err := newXLStorageTestSetup()
+	_, path, err := newXLStorageTestSetup(t)
 	if err != nil {
 		t.Fatalf("Unable to create xlStorage test setup, %s", err)
 	}
-	defer os.RemoveAll(path)
 
 	// Setup test environment.
 	if err = fsMkdir(GlobalContext, pathJoin(path, "success-vol")); err != nil {
@@ -501,11 +495,10 @@ func TestFSRemoves(t *testing.T) {
 
 func TestFSRemoveMeta(t *testing.T) {
 	// create xlStorage test setup
-	_, fsPath, err := newXLStorageTestSetup()
+	_, fsPath, err := newXLStorageTestSetup(t)
 	if err != nil {
 		t.Fatalf("Unable to create xlStorage test setup, %s", err)
 	}
-	defer os.RemoveAll(fsPath)
 
 	// Setup test environment.
 	if err = fsMkdir(GlobalContext, pathJoin(fsPath, "success-vol")); err != nil {
@@ -529,10 +522,7 @@ func TestFSRemoveMeta(t *testing.T) {
 
 	defer rwPool.Close(filePath)
 
-	tmpDir, tmpErr := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if tmpErr != nil {
-		t.Fatal(tmpErr)
-	}
+	tmpDir := t.TempDir()
 
 	if err := fsRemoveMeta(GlobalContext, fsPath, filePath, tmpDir); err != nil {
 		t.Fatalf("Unable to remove file, %s", err)
@@ -548,15 +538,9 @@ func TestFSRemoveMeta(t *testing.T) {
 }
 
 func TestFSIsFile(t *testing.T) {
-	dirPath, err := ioutil.TempDir(globalTestTmpDir, "minio-")
-	if err != nil {
-		t.Fatalf("Unable to create tmp directory %s", err)
-	}
-	defer os.RemoveAll(dirPath)
+	filePath := pathJoin(t.TempDir(), "tmpfile")
 
-	filePath := pathJoin(dirPath, "tmpfile")
-
-	if err = ioutil.WriteFile(filePath, nil, 0o777); err != nil {
+	if err := ioutil.WriteFile(filePath, nil, 0o777); err != nil {
 		t.Fatalf("Unable to create file %s", filePath)
 	}
 

cmd/fs-v1-metadata_test.go

@@ -56,7 +56,7 @@ func TestReadFSMetadata(t *testing.T) {
 	bucketName := "bucket"
 	objectName := "object"
 
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal("Unexpected err: ", err)
 	}
 	if _, err := obj.PutObject(GlobalContext, bucketName, objectName, mustGetPutObjReader(t, bytes.NewReader([]byte("abcd")), int64(len("abcd")), "", ""), ObjectOptions{}); err != nil {
@@ -92,7 +92,7 @@ func TestWriteFSMetadata(t *testing.T) {
 	bucketName := "bucket"
 	objectName := "object"
 
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal("Unexpected err: ", err)
 	}
 	if _, err := obj.PutObject(GlobalContext, bucketName, objectName, mustGetPutObjReader(t, bytes.NewReader([]byte("abcd")), int64(len("abcd")), "", ""), ObjectOptions{}); err != nil {

cmd/fs-v1-multipart.go

@@ -534,7 +534,7 @@ func (fs *FSObjects) ListObjectParts(ctx context.Context, bucket, object, upload
 	var fsMeta fsMetaV1
 	json := jsoniter.ConfigCompatibleWithStandardLibrary
 	if err = json.Unmarshal(fsMetaBytes, &fsMeta); err != nil {
-		return result, err
+		return result, toObjectErr(fmt.Errorf("unable to parse %s: error %w", pathJoin(uploadIDDir, fs.metaJSONFile), err), bucket, object)
 	}
 
 	result.UserDefined = fsMeta.Meta

cmd/fs-v1-multipart_test.go

@@ -45,7 +45,7 @@ func TestFSCleanupMultipartUploadsInRoutine(t *testing.T) {
 
 	// Create a context we can cancel.
 	ctx, cancel := context.WithCancel(GlobalContext)
-	obj.MakeBucketWithLocation(ctx, bucketName, BucketOptions{})
+	obj.MakeBucketWithLocation(ctx, bucketName, MakeBucketOptions{})
 
 	uploadID, err := obj.NewMultipartUpload(ctx, bucketName, objectName, ObjectOptions{})
 	if err != nil {
@@ -99,7 +99,7 @@ func TestNewMultipartUploadFaultyDisk(t *testing.T) {
 	bucketName := "bucket"
 	objectName := "object"
 
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal("Cannot create bucket, err: ", err)
 	}
 
@@ -124,7 +124,7 @@ func TestPutObjectPartFaultyDisk(t *testing.T) {
 	data := []byte("12345")
 	dataLen := int64(len(data))
 
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal("Cannot create bucket, err: ", err)
 	}
 
@@ -157,7 +157,7 @@ func TestCompleteMultipartUploadFaultyDisk(t *testing.T) {
 	objectName := "object"
 	data := []byte("12345")
 
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal("Cannot create bucket, err: ", err)
 	}
 
@@ -190,7 +190,7 @@ func TestCompleteMultipartUpload(t *testing.T) {
 	objectName := "object"
 	data := []byte("12345")
 
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal("Cannot create bucket, err: ", err)
 	}
 
@@ -227,7 +227,7 @@ func TestAbortMultipartUpload(t *testing.T) {
 	objectName := "object"
 	data := []byte("12345")
 
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal("Cannot create bucket, err: ", err)
 	}
 
@@ -258,7 +258,7 @@ func TestListMultipartUploadsFaultyDisk(t *testing.T) {
 	bucketName := "bucket"
 	objectName := "object"
 
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal("Cannot create bucket, err: ", err)
 	}
 

cmd/fs-v1-rwpool.go

@@ -187,6 +187,8 @@ func (fsi *fsIOPool) Create(path string) (wlk *lock.LockedFile, err error) {
 			return nil, errFileAccessDenied
 		case isSysErrIsDir(err):
 			return nil, errIsNotRegular
+		case isSysErrNotDir(err):
+			return nil, errFileAccessDenied
 		case isSysErrPathNotFound(err):
 			return nil, errFileAccessDenied
 		default:

cmd/fs-v1-rwpool_test.go

@@ -18,7 +18,6 @@
 package cmd
 
 import (
-	"os"
 	"runtime"
 	"testing"
 
@@ -48,11 +47,10 @@ func TestRWPoolLongPath(t *testing.T) {
 // Tests all RWPool methods.
 func TestRWPool(t *testing.T) {
 	// create xlStorage test setup
-	_, path, err := newXLStorageTestSetup()
+	_, path, err := newXLStorageTestSetup(t)
 	if err != nil {
 		t.Fatalf("Unable to create xlStorage test setup, %s", err)
 	}
-	defer os.RemoveAll(path)
 
 	rwPool := &fsIOPool{
 		readersMap: make(map[string]*lock.RLockedFile),

cmd/fs-v1.go

@@ -27,7 +27,6 @@ import (
 	"os"
 	"os/user"
 	"path"
-	"path/filepath"
 	"sort"
 	"strings"
 	"sync"
@@ -254,7 +253,7 @@ func (fs *FSObjects) NSScanner(ctx context.Context, bf *bloomFilter, updates cha
 		return err
 	}
 	totalCache.Info.Name = dataUsageRoot
-	buckets, err := fs.ListBuckets(ctx)
+	buckets, err := fs.ListBuckets(ctx, BucketOptions{})
 	if err != nil {
 		return err
 	}
@@ -453,7 +452,7 @@ func (fs *FSObjects) statBucketDir(ctx context.Context, bucket string) (os.FileI
 }
 
 // MakeBucketWithLocation - create a new bucket, returns if it already exists.
-func (fs *FSObjects) MakeBucketWithLocation(ctx context.Context, bucket string, opts BucketOptions) error {
+func (fs *FSObjects) MakeBucketWithLocation(ctx context.Context, bucket string, opts MakeBucketOptions) error {
 	if opts.LockEnabled || opts.VersioningEnabled {
 		return NotImplemented{}
 	}
@@ -524,7 +523,7 @@ func (fs *FSObjects) DeleteBucketPolicy(ctx context.Context, bucket string) erro
 }
 
 // GetBucketInfo - fetch bucket metadata info.
-func (fs *FSObjects) GetBucketInfo(ctx context.Context, bucket string) (bi BucketInfo, e error) {
+func (fs *FSObjects) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (bi BucketInfo, e error) {
 	st, err := fs.statBucketDir(ctx, bucket)
 	if err != nil {
 		return bi, toObjectErr(err, bucket)
@@ -543,7 +542,7 @@ func (fs *FSObjects) GetBucketInfo(ctx context.Context, bucket string) (bi Bucke
 }
 
 // ListBuckets - list all s3 compatible buckets (directories) at fsPath.
-func (fs *FSObjects) ListBuckets(ctx context.Context) ([]BucketInfo, error) {
+func (fs *FSObjects) ListBuckets(ctx context.Context, opts BucketOptions) ([]BucketInfo, error) {
 	if err := checkPathLength(fs.fsPath); err != nil {
 		logger.LogIf(ctx, err)
 		return nil, err
@@ -1493,19 +1492,3 @@ func (fs *FSObjects) TransitionObject(ctx context.Context, bucket, object string
 func (fs *FSObjects) RestoreTransitionedObject(ctx context.Context, bucket, object string, opts ObjectOptions) error {
 	return NotImplemented{}
 }
-
-// GetRawData returns raw file data to the callback.
-// Errors are ignored, only errors from the callback are returned.
-// For now only direct file paths are supported.
-func (fs *FSObjects) GetRawData(ctx context.Context, volume, file string, fn func(r io.Reader, host string, disk string, filename string, size int64, modtime time.Time, isDir bool) error) error {
-	f, err := os.Open(filepath.Join(fs.fsPath, volume, file))
-	if err != nil {
-		return nil
-	}
-	defer f.Close()
-	st, err := f.Stat()
-	if err != nil || st.IsDir() {
-		return nil
-	}
-	return fn(f, "fs", fs.fsUUID, file, st.Size(), st.ModTime(), st.IsDir())
-}

cmd/fs-v1_test.go

@@ -41,7 +41,7 @@ func TestNewFS(t *testing.T) {
 	}
 	_, err = NewFSObjectLayer(disk)
 	if err != nil {
-		errMsg := "Unable to recognize backend format, Disk is not in FS format."
+		errMsg := "Unable to recognize backend format, Drive is not in FS format."
 		if err.Error() == errMsg {
 			t.Errorf("Expecting %s, got %s", errMsg, err)
 		}
@@ -62,7 +62,7 @@ func TestFSShutdown(t *testing.T) {
 		fs := obj.(*FSObjects)
 
 		objectContent := "12345"
-		obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{})
+		obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{})
 		obj.PutObject(GlobalContext, bucketName, objectName, mustGetPutObjReader(t, bytes.NewReader([]byte(objectContent)), int64(len(objectContent)), "", ""), ObjectOptions{})
 		return fs, disk
 	}
@@ -95,18 +95,18 @@ func TestFSGetBucketInfo(t *testing.T) {
 	fs := obj.(*FSObjects)
 	bucketName := "bucket"
 
-	err := obj.MakeBucketWithLocation(GlobalContext, "a", BucketOptions{})
+	err := obj.MakeBucketWithLocation(GlobalContext, "a", MakeBucketOptions{})
 	if !isSameType(err, BucketNameInvalid{}) {
 		t.Fatal("BucketNameInvalid error not returned")
 	}
 
-	err = obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{})
+	err = obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
 
 	// Test with valid parameters
-	info, err := fs.GetBucketInfo(GlobalContext, bucketName)
+	info, err := fs.GetBucketInfo(GlobalContext, bucketName, BucketOptions{})
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -115,7 +115,7 @@ func TestFSGetBucketInfo(t *testing.T) {
 	}
 
 	// Test with non-existent bucket
-	_, err = fs.GetBucketInfo(GlobalContext, "a")
+	_, err = fs.GetBucketInfo(GlobalContext, "a", BucketOptions{})
 	if !isSameType(err, BucketNotFound{}) {
 		t.Fatal("BucketNotFound error not returned")
 	}
@@ -123,7 +123,7 @@ func TestFSGetBucketInfo(t *testing.T) {
 	// Check for buckets and should get disk not found.
 	os.RemoveAll(disk)
 
-	if _, err = fs.GetBucketInfo(GlobalContext, bucketName); err != nil {
+	if _, err = fs.GetBucketInfo(GlobalContext, bucketName, BucketOptions{}); err != nil {
 		if !isSameType(err, BucketNotFound{}) {
 			t.Fatal("BucketNotFound error not returned")
 		}
@@ -139,7 +139,7 @@ func TestFSPutObject(t *testing.T) {
 	bucketName := "bucket"
 	objectName := "1/2/3/4/object"
 
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal(err)
 	}
 
@@ -179,7 +179,7 @@ func TestFSDeleteObject(t *testing.T) {
 	bucketName := "bucket"
 	objectName := "object"
 
-	obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{})
+	obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{})
 	obj.PutObject(GlobalContext, bucketName, objectName, mustGetPutObjReader(t, bytes.NewReader([]byte("abcd")), int64(len("abcd")), "", ""), ObjectOptions{})
 
 	// Test with invalid bucket name
@@ -223,7 +223,7 @@ func TestFSDeleteBucket(t *testing.T) {
 	fs := obj.(*FSObjects)
 	bucketName := "bucket"
 
-	err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{})
+	err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{})
 	if err != nil {
 		t.Fatal("Unexpected error: ", err)
 	}
@@ -242,7 +242,7 @@ func TestFSDeleteBucket(t *testing.T) {
 		t.Fatal("Unexpected error: ", err)
 	}
 
-	obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{})
+	obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{})
 
 	// Delete bucket should get error disk not found.
 	os.RemoveAll(disk)
@@ -264,7 +264,7 @@ func TestFSListBuckets(t *testing.T) {
 	fs := obj.(*FSObjects)
 
 	bucketName := "bucket"
-	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, BucketOptions{}); err != nil {
+	if err := obj.MakeBucketWithLocation(GlobalContext, bucketName, MakeBucketOptions{}); err != nil {
 		t.Fatal("Unexpected error: ", err)
 	}
 
@@ -279,7 +279,7 @@ func TestFSListBuckets(t *testing.T) {
 	f.Close()
 
 	// Test list buckets to have only one entry.
-	buckets, err := fs.ListBuckets(GlobalContext)
+	buckets, err := fs.ListBuckets(GlobalContext, BucketOptions{})
 	if err != nil {
 		t.Fatal("Unexpected error: ", err)
 	}
@@ -289,7 +289,7 @@ func TestFSListBuckets(t *testing.T) {
 
 	// Test ListBuckets with disk not found.
 	os.RemoveAll(disk)
-	if _, err := fs.ListBuckets(GlobalContext); err != nil {
+	if _, err := fs.ListBuckets(GlobalContext, BucketOptions{}); err != nil {
 		if err != errDiskNotFound {
 			t.Fatal("Unexpected error: ", err)
 		}

cmd/gateway-main.go

@@ -310,7 +310,7 @@ func StartGateway(ctx *cli.Context, gw Gateway) {
 	go globalIAMSys.Init(GlobalContext, newObject, globalEtcdClient, globalRefreshIAMInterval)
 
 	if gatewayName == NASBackendGateway {
-		buckets, err := newObject.ListBuckets(GlobalContext)
+		buckets, err := newObject.ListBuckets(GlobalContext, BucketOptions{})
 		if err != nil {
 			logger.Fatal(err, "Unable to list buckets")
 		}
@@ -323,7 +323,7 @@ func StartGateway(ctx *cli.Context, gw Gateway) {
 		// initialize the new disk cache objects.
 		var cacheAPI CacheObjectLayer
 		cacheAPI, err = newServerCacheObjects(GlobalContext, globalCacheConfig)
-		logger.FatalIf(err, "Unable to initialize disk caching")
+		logger.FatalIf(err, "Unable to initialize drive caching")
 
 		globalObjLayerMutex.Lock()
 		globalCacheObjectAPI = cacheAPI
@@ -332,7 +332,7 @@ func StartGateway(ctx *cli.Context, gw Gateway) {
 
 	// Populate existing buckets to the etcd backend
 	if globalDNSConfig != nil {
-		buckets, err := newObject.ListBuckets(GlobalContext)
+		buckets, err := newObject.ListBuckets(GlobalContext, BucketOptions{})
 		if err != nil {
 			logger.Fatal(err, "Unable to list buckets")
 		}

cmd/gateway/s3/gateway-s3-sse.go

@@ -734,7 +734,7 @@ func (l *s3EncObjects) cleanupStaleEncMultipartUploads(ctx context.Context, clea
 
 // cleanupStaleUploads removes old custom encryption multipart uploads on backend
 func (l *s3EncObjects) cleanupStaleUploads(ctx context.Context, expiry time.Duration) {
-	buckets, err := l.s3Objects.ListBuckets(ctx)
+	buckets, err := l.s3Objects.ListBuckets(ctx, minio.BucketOptions{})
 	if err != nil {
 		logger.LogIf(ctx, err)
 		return

cmd/gateway/s3/gateway-s3.go

@@ -325,7 +325,7 @@ func (l *s3Objects) StorageInfo(ctx context.Context) (si minio.StorageInfo, _ []
 }
 
 // MakeBucket creates a new container on S3 backend.
-func (l *s3Objects) MakeBucketWithLocation(ctx context.Context, bucket string, opts minio.BucketOptions) error {
+func (l *s3Objects) MakeBucketWithLocation(ctx context.Context, bucket string, opts minio.MakeBucketOptions) error {
 	if opts.LockEnabled || opts.VersioningEnabled {
 		return minio.NotImplemented{}
 	}
@@ -348,7 +348,7 @@ func (l *s3Objects) MakeBucketWithLocation(ctx context.Context, bucket string, o
 }
 
 // GetBucketInfo gets bucket metadata..
-func (l *s3Objects) GetBucketInfo(ctx context.Context, bucket string) (bi minio.BucketInfo, e error) {
+func (l *s3Objects) GetBucketInfo(ctx context.Context, bucket string, opts minio.BucketOptions) (bi minio.BucketInfo, e error) {
 	buckets, err := l.Client.ListBuckets(ctx)
 	if err != nil {
 		// Listbuckets may be disallowed, proceed to check if
@@ -381,7 +381,7 @@ func (l *s3Objects) GetBucketInfo(ctx context.Context, bucket string) (bi minio.
 }
 
 // ListBuckets lists all S3 buckets
-func (l *s3Objects) ListBuckets(ctx context.Context) ([]minio.BucketInfo, error) {
+func (l *s3Objects) ListBuckets(ctx context.Context, opts minio.BucketOptions) ([]minio.BucketInfo, error) {
 	buckets, err := l.Client.ListBuckets(ctx)
 	if err != nil {
 		return nil, minio.ErrorRespToObjectError(err)

cmd/global-heal.go

@@ -209,7 +209,7 @@ func (er *erasureObjects) healErasureSet(ctx context.Context, buckets []string,
 		}
 
 		if serverDebugLog {
-			console.Debugf(color.Green("healDisk:")+" healing bucket %s content on %s erasure set\n",
+			console.Debugf(color.Green("healDrive:")+" healing bucket %s content on %s erasure set\n",
 				bucket, humanize.Ordinal(tracker.SetIndex+1))
 		}
 

cmd/globals.go

@@ -383,18 +383,30 @@ var (
 	// Add new variable global values here.
 )
 
-var globalAuthZPluginMutex sync.Mutex
+var globalAuthPluginMutex sync.Mutex
+
+func newGlobalAuthNPluginFn() *idplugin.AuthNPlugin {
+	globalAuthPluginMutex.Lock()
+	defer globalAuthPluginMutex.Unlock()
+	return globalAuthNPlugin
+}
 
 func newGlobalAuthZPluginFn() *polplugin.AuthZPlugin {
-	globalAuthZPluginMutex.Lock()
-	defer globalAuthZPluginMutex.Unlock()
+	globalAuthPluginMutex.Lock()
+	defer globalAuthPluginMutex.Unlock()
 	return globalAuthZPlugin
 }
 
+func setGlobalAuthNPlugin(authn *idplugin.AuthNPlugin) {
+	globalAuthPluginMutex.Lock()
+	globalAuthNPlugin = authn
+	globalAuthPluginMutex.Unlock()
+}
+
 func setGlobalAuthZPlugin(authz *polplugin.AuthZPlugin) {
-	globalAuthZPluginMutex.Lock()
+	globalAuthPluginMutex.Lock()
 	globalAuthZPlugin = authz
-	globalAuthZPluginMutex.Unlock()
+	globalAuthPluginMutex.Unlock()
 }
 
 var errSelfTestFailure = errors.New("self test failed. unsafe to start server")

cmd/iam-dummy-store.go

@@ -57,10 +57,6 @@ func (ids *iamDummyStore) getUsersSysType() UsersSysType {
 	return ids.usersSysType
 }
 
-func (ids *iamDummyStore) migrateBackendFormat(context.Context) error {
-	return nil
-}
-
 func (ids *iamDummyStore) loadPolicyDoc(ctx context.Context, policy string, m map[string]PolicyDoc) error {
 	v, ok := ids.iamPolicyDocsMap[policy]
 	if !ok {