minio/cmd/global-heal.go
Harshavardhana 9693c382a8
make renameData() more defensive during overwrites (#19548)
instead upon any error in renameData(), we still
preserve the existing dataDir in some form for
recoverability in strange situations such as out
of disk space type errors.

Bonus: avoid running list and heal() instead allow
versions disparity to return the actual versions,
uuid to heal. Currently limit this to 100 versions
and lesser disparate objects.

an undo now reverts back the xl.meta from xl.meta.bkp
during overwrites on such flaky setups.

Bonus: Save N depth syscalls via skipping the parents
upon overwrites and versioned updates.

Flaky setup examples are stretch clusters with regular
packet drops etc, we need to add some defensive code
around to avoid dangling objects.
2024-04-23 10:15:52 -07:00

537 lines
14 KiB
Go

// Copyright (c) 2015-2024 MinIO, Inc.
//
// This file is part of MinIO Object Storage stack
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
package cmd
import (
"context"
"errors"
"fmt"
"math/rand"
"runtime"
"sort"
"time"
"github.com/dustin/go-humanize"
"github.com/minio/madmin-go/v3"
"github.com/minio/minio/internal/color"
"github.com/minio/minio/internal/config/storageclass"
xioutil "github.com/minio/minio/internal/ioutil"
"github.com/minio/minio/internal/logger"
"github.com/minio/pkg/v2/console"
"github.com/minio/pkg/v2/wildcard"
"github.com/minio/pkg/v2/workers"
)
const (
bgHealingUUID = "0000-0000-0000-0000"
)
// NewBgHealSequence creates a background healing sequence
// operation which scans all objects and heal them.
func newBgHealSequence() *healSequence {
reqInfo := &logger.ReqInfo{API: "BackgroundHeal"}
ctx, cancelCtx := context.WithCancel(logger.SetReqInfo(GlobalContext, reqInfo))
hs := madmin.HealOpts{
// Remove objects that do not have read-quorum
Remove: healDeleteDangling,
}
return &healSequence{
startTime: UTCNow(),
clientToken: bgHealingUUID,
// run-background heal with reserved bucket
bucket: minioReservedBucket,
settings: hs,
currentStatus: healSequenceStatus{
Summary: healNotStartedStatus,
HealSettings: hs,
},
cancelCtx: cancelCtx,
ctx: ctx,
reportProgress: false,
scannedItemsMap: make(map[madmin.HealItemType]int64),
healedItemsMap: make(map[madmin.HealItemType]int64),
healFailedItemsMap: make(map[string]int64),
}
}
// getLocalBackgroundHealStatus will return the heal status of the local node
func getLocalBackgroundHealStatus(ctx context.Context, o ObjectLayer) (madmin.BgHealState, bool) {
if globalBackgroundHealState == nil {
return madmin.BgHealState{}, false
}
bgSeq, ok := globalBackgroundHealState.getHealSequenceByToken(bgHealingUUID)
if !ok {
return madmin.BgHealState{}, false
}
status := madmin.BgHealState{
ScannedItemsCount: bgSeq.getScannedItemsCount(),
}
healDisksMap := map[string]struct{}{}
for _, ep := range getLocalDisksToHeal() {
healDisksMap[ep.String()] = struct{}{}
}
if o == nil {
healing := globalBackgroundHealState.getLocalHealingDisks()
for _, disk := range healing {
status.HealDisks = append(status.HealDisks, disk.Endpoint)
}
return status, true
}
si := o.LocalStorageInfo(ctx, true)
indexed := make(map[string][]madmin.Disk)
for _, disk := range si.Disks {
setIdx := fmt.Sprintf("%d-%d", disk.PoolIndex, disk.SetIndex)
indexed[setIdx] = append(indexed[setIdx], disk)
}
for id, disks := range indexed {
ss := madmin.SetStatus{
ID: id,
SetIndex: disks[0].SetIndex,
PoolIndex: disks[0].PoolIndex,
}
for _, disk := range disks {
ss.Disks = append(ss.Disks, disk)
if disk.Healing {
ss.HealStatus = "Healing"
ss.HealPriority = "high"
status.HealDisks = append(status.HealDisks, disk.Endpoint)
}
}
sortDisks(ss.Disks)
status.Sets = append(status.Sets, ss)
}
sort.Slice(status.Sets, func(i, j int) bool {
return status.Sets[i].ID < status.Sets[j].ID
})
backendInfo := o.BackendInfo()
status.SCParity = make(map[string]int)
status.SCParity[storageclass.STANDARD] = backendInfo.StandardSCParity
status.SCParity[storageclass.RRS] = backendInfo.RRSCParity
return status, true
}
// healErasureSet lists and heals all objects in a specific erasure set
func (er *erasureObjects) healErasureSet(ctx context.Context, buckets []string, tracker *healingTracker) error {
scanMode := madmin.HealNormalScan
// Make sure to copy since `buckets slice`
// is modified in place by tracker.
healBuckets := make([]string, len(buckets))
copy(healBuckets, buckets)
objAPI := newObjectLayerFn()
if objAPI == nil {
return errServerNotInitialized
}
for _, bucket := range healBuckets {
_, err := objAPI.HealBucket(ctx, bucket, madmin.HealOpts{ScanMode: scanMode})
if err != nil {
// Log bucket healing error if any, we shall retry again.
healingLogIf(ctx, err)
}
}
info, err := tracker.disk.DiskInfo(ctx, DiskInfoOptions{})
if err != nil {
return fmt.Errorf("unable to get disk information before healing it: %w", err)
}
var numHealers uint64
if numCores := uint64(runtime.GOMAXPROCS(0)); info.NRRequests > numCores {
numHealers = numCores / 4
} else {
numHealers = info.NRRequests / 4
}
if numHealers < 4 {
numHealers = 4
}
// allow overriding this value as well..
if v := globalHealConfig.GetWorkers(); v > 0 {
numHealers = uint64(v)
}
healingLogEvent(ctx, fmt.Sprintf("Healing drive '%s' - use %d parallel workers.", tracker.disk.String(), numHealers))
jt, _ := workers.New(int(numHealers))
var retErr error
// Heal all buckets with all objects
for _, bucket := range healBuckets {
if tracker.isHealed(bucket) {
continue
}
var forwardTo string
// If we resume to the same bucket, forward to last known item.
if b := tracker.getBucket(); b != "" {
if b == bucket {
forwardTo = tracker.getObject()
} else {
// Reset to where last bucket ended if resuming.
tracker.resume()
}
}
tracker.setObject("")
tracker.setBucket(bucket)
// Heal current bucket again in case if it is failed
// in the beginning of erasure set healing
if _, err := objAPI.HealBucket(ctx, bucket, madmin.HealOpts{
ScanMode: scanMode,
}); err != nil {
// Set this such that when we return this function
// we let the caller retry this disk again for the
// buckets that failed healing.
retErr = err
healingLogIf(ctx, err)
continue
}
vc, err := globalBucketVersioningSys.Get(bucket)
if err != nil {
retErr = err
healingLogIf(ctx, err)
continue
}
// Check if the current bucket has a configured lifecycle policy
lc, err := globalLifecycleSys.Get(bucket)
if err != nil && !errors.Is(err, BucketLifecycleNotFound{Bucket: bucket}) {
retErr = err
healingLogIf(ctx, err)
continue
}
// Check if bucket is object locked.
lr, err := globalBucketObjectLockSys.Get(bucket)
if err != nil {
retErr = err
healingLogIf(ctx, err)
continue
}
rcfg, err := getReplicationConfig(ctx, bucket)
if err != nil {
retErr = err
healingLogIf(ctx, err)
continue
}
if serverDebugLog {
console.Debugf(color.Green("healDrive:")+" healing bucket %s content on %s erasure set\n",
bucket, humanize.Ordinal(er.setIndex+1))
}
disks, _ := er.getOnlineDisksWithHealing(false)
if len(disks) == 0 {
// No object healing necessary
tracker.bucketDone(bucket)
healingLogIf(ctx, tracker.update(ctx))
continue
}
rand.Shuffle(len(disks), func(i, j int) {
disks[i], disks[j] = disks[j], disks[i]
})
expectedDisks := len(disks)/2 + 1
fallbackDisks := disks[expectedDisks:]
disks = disks[:expectedDisks]
type healEntryResult struct {
bytes uint64
success bool
skipped bool
entryDone bool
name string
}
healEntryDone := func(name string) healEntryResult {
return healEntryResult{
entryDone: true,
name: name,
}
}
healEntrySuccess := func(sz uint64) healEntryResult {
return healEntryResult{
bytes: sz,
success: true,
}
}
healEntryFailure := func(sz uint64) healEntryResult {
return healEntryResult{
bytes: sz,
}
}
healEntrySkipped := func(sz uint64) healEntryResult {
return healEntryResult{
bytes: sz,
skipped: true,
}
}
filterLifecycle := func(bucket, object string, fi FileInfo) bool {
if lc == nil {
return false
}
versioned := vc != nil && vc.Versioned(object)
objInfo := fi.ToObjectInfo(bucket, object, versioned)
evt := evalActionFromLifecycle(ctx, *lc, lr, rcfg, objInfo)
switch {
case evt.Action.DeleteRestored(): // if restored copy has expired,delete it synchronously
applyExpiryOnTransitionedObject(ctx, newObjectLayerFn(), objInfo, evt, lcEventSrc_Heal)
return false
case evt.Action.Delete():
globalExpiryState.enqueueByDays(objInfo, evt, lcEventSrc_Heal)
return true
default:
return false
}
}
// Collect updates to tracker from concurrent healEntry calls
results := make(chan healEntryResult, 1000)
go func() {
for res := range results {
if res.entryDone {
tracker.setObject(res.name)
if time.Since(tracker.getLastUpdate()) > time.Minute {
healingLogIf(ctx, tracker.update(ctx))
}
continue
}
tracker.updateProgress(res.success, res.skipped, res.bytes)
}
}()
send := func(result healEntryResult) bool {
select {
case <-ctx.Done():
if !contextCanceled(ctx) {
healingLogIf(ctx, ctx.Err())
}
return false
case results <- result:
return true
}
}
// Note: updates from healEntry to tracker must be sent on results channel.
healEntry := func(bucket string, entry metaCacheEntry) {
defer jt.Give()
if entry.name == "" && len(entry.metadata) == 0 {
// ignore entries that don't have metadata.
return
}
if entry.isDir() {
// ignore healing entry.name's with `/` suffix.
return
}
// We might land at .metacache, .trash, .multipart
// no need to heal them skip, only when bucket
// is '.minio.sys'
if bucket == minioMetaBucket {
if wildcard.Match("buckets/*/.metacache/*", entry.name) {
return
}
if wildcard.Match("tmp/.trash/*", entry.name) {
return
}
if wildcard.Match("multipart/*", entry.name) {
return
}
}
// erasureObjects layer needs object names to be encoded
encodedEntryName := encodeDirObject(entry.name)
var result healEntryResult
fivs, err := entry.fileInfoVersions(bucket)
if err != nil {
_, err := er.HealObject(ctx, bucket, encodedEntryName, "",
madmin.HealOpts{
ScanMode: scanMode,
Remove: healDeleteDangling,
})
if err != nil {
if isErrObjectNotFound(err) || isErrVersionNotFound(err) {
// queueing happens across namespace, ignore
// objects that are not found.
return
}
result = healEntryFailure(0)
healingLogIf(ctx, fmt.Errorf("unable to heal object %s/%s: %w", bucket, entry.name, err))
} else {
result = healEntrySuccess(0)
}
send(result)
return
}
var versionNotFound int
for _, version := range fivs.Versions {
// Ignore a version with a modtime newer than healing start time.
if version.ModTime.After(tracker.Started) {
continue
}
// Apply lifecycle rules on the objects that are expired.
if filterLifecycle(bucket, version.Name, version) {
versionNotFound++
if !send(healEntrySkipped(uint64(version.Size))) {
return
}
continue
}
if _, err := er.HealObject(ctx, bucket, encodedEntryName,
version.VersionID, madmin.HealOpts{
ScanMode: scanMode,
Remove: healDeleteDangling,
}); err != nil {
if isErrObjectNotFound(err) || isErrVersionNotFound(err) {
// queueing happens across namespace, ignore
// objects that are not found.
versionNotFound++
continue
}
// If not deleted, assume they failed.
result = healEntryFailure(uint64(version.Size))
if version.VersionID != "" {
healingLogIf(ctx, fmt.Errorf("unable to heal object %s/%s-v(%s): %w", bucket, version.Name, version.VersionID, err))
} else {
healingLogIf(ctx, fmt.Errorf("unable to heal object %s/%s: %w", bucket, version.Name, err))
}
} else {
result = healEntrySuccess(uint64(version.Size))
}
if !send(result) {
return
}
}
// All versions resulted in 'ObjectNotFound/VersionNotFound'
if versionNotFound == len(fivs.Versions) {
return
}
select {
case <-ctx.Done():
return
case results <- healEntryDone(entry.name):
}
// Wait and proceed if there are active requests
waitForLowHTTPReq()
}
actualBucket, prefix := path2BucketObject(bucket)
// How to resolve partial results.
resolver := metadataResolutionParams{
dirQuorum: 1,
objQuorum: 1,
bucket: actualBucket,
}
err = listPathRaw(ctx, listPathRawOptions{
disks: disks,
fallbackDisks: fallbackDisks,
bucket: actualBucket,
path: prefix,
recursive: true,
forwardTo: forwardTo,
minDisks: 1,
reportNotFound: false,
agreed: func(entry metaCacheEntry) {
jt.Take()
go healEntry(actualBucket, entry)
},
partial: func(entries metaCacheEntries, _ []error) {
entry, ok := entries.resolve(&resolver)
if !ok {
// check if we can get one entry at least
// proceed to heal nonetheless.
entry, _ = entries.firstFound()
}
jt.Take()
go healEntry(actualBucket, *entry)
},
finished: nil,
})
jt.Wait() // synchronize all the concurrent heal jobs
xioutil.SafeClose(results)
if err != nil {
// Set this such that when we return this function
// we let the caller retry this disk again for the
// buckets it failed to list.
retErr = err
healingLogIf(ctx, err)
continue
}
select {
// If context is canceled don't mark as done...
case <-ctx.Done():
return ctx.Err()
default:
tracker.bucketDone(bucket)
healingLogIf(ctx, tracker.update(ctx))
}
}
tracker.setObject("")
tracker.setBucket("")
return retErr
}
func healBucket(bucket string, scan madmin.HealScanMode) error {
// Get background heal sequence to send elements to heal
bgSeq, ok := globalBackgroundHealState.getHealSequenceByToken(bgHealingUUID)
if ok {
return bgSeq.queueHealTask(healSource{bucket: bucket}, madmin.HealItemBucket)
}
return nil
}
// healObject sends the given object/version to the background healing workers
func healObject(bucket, object, versionID string, scan madmin.HealScanMode) error {
// Get background heal sequence to send elements to heal
bgSeq, ok := globalBackgroundHealState.getHealSequenceByToken(bgHealingUUID)
if ok {
return bgSeq.healObject(bucket, object, versionID, scan)
}
return nil
}