Files
seaweedfs/weed/shell/command_s3_versions_audit.go
T
Chris LuandGitHub a9cfbd8d3a s3: tear down the emptied .versions directory on last-version delete; drain existing residue (#10278)
* s3: routed last-version delete removes the emptied .versions directory

The routed versioned delete (routedDeleteSpecificVersion) repoints the
latest pointer and deletes the version file, but unlike the lock-path
fallback (updateLatestVersionAfterDeletion) it never tears down the
.versions/ directory it just emptied. The residue keeps the key's read
path in the self-heal rescan loop: every GET of the deleted key logs
event=surfaced plus a GetObject error until the background
EmptyFolderCleaner gets to the directory — at least two minutes away on
its delay queue, and possibly never (the queue is in-memory, bounded,
and gated on the bucket's allow-empty-folders policy). Veeam's lock
arbitration probes deleted lock keys continuously, so those windows are
always open and the log spam is chronic.

ObjectMutation DELETE gains remove_empty_parent: after the child delete,
the filer best-effort removes the parent directory in the same locked
transaction. Non-recursive on purpose — a concurrent write that lands a
new version fails the removal instead of being lost with it. The routed
last-version delete sets it on the version-file DELETE, matching the
lock-path fallback's contract.

Claude-Session: https://claude.ai/code/session_014mMYAHXZySkCCUfpRFNtSv

* s3: drain empty .versions residue on read heal and in s3.versions.audit

Directories already stranded by pre-teardown deletes (or dropped from
the EmptyFolderCleaner's bounded in-memory queue) previously re-entered
the self-heal rescan on every GET forever: the heal only cleared the
pointer and nothing ever removed the directory, and s3.versions.audit
counted the state as clean.

When the heal rescan finds no remaining version, remove the directory
outright (non-recursive, so orphan children still block and fall back to
the pointer clear) and log event=healed mode=empty_dir_removed; the next
GET takes the clean not-found path. The audit gains an empty category so
the residue is visible, and -heal removes such directories in bulk.

Claude-Session: https://claude.ai/code/session_014mMYAHXZySkCCUfpRFNtSv
2026-07-08 18:50:54 -07:00

283 lines
9.2 KiB
Go

package shell
import (
"context"
"errors"
"flag"
"fmt"
"io"
"strings"
"sync/atomic"
"time"
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
s3_constants "github.com/seaweedfs/seaweedfs/weed/s3api/s3_constants"
"github.com/seaweedfs/seaweedfs/weed/util"
)
func init() {
Commands = append(Commands, &commandS3VersionsAudit{})
}
type commandS3VersionsAudit struct{}
func (c *commandS3VersionsAudit) Name() string {
return "s3.versions.audit"
}
func (c *commandS3VersionsAudit) Help() string {
return `audit .versions/ directories under a prefix for stranded pointer/missing-file state
Walks every entry under the given prefix and, for each directory whose
name ends in ".versions", checks whether its extended-attr latest-version
pointer references a file that actually exists in the directory.
Reports counts for:
- directories scanned
- clean (no pointer, or pointer matches existing file)
- stranded (pointer set but file is missing) — the symptom seen by
Veeam/etc. as "Storage not found" on the next GET
- orphan (directory has files lacking the version-id extended attr,
which the post-delete cleanup path will refuse to rm)
- empty (no pointer and no children — residue of a fully-drained key;
every GET of the key replays the read-side self-heal rescan)
Example:
# Audit a whole bucket
s3.versions.audit -prefix /buckets/mybucket
# Audit a specific client subtree, print each finding
s3.versions.audit -prefix /buckets/mybucket/Veeam/Backup/groupsoftware/Clients/<uuid>/ -v
# Dry run (default) — read-only, prints what would be healed
# Add -heal to clear stranded pointers in place (calls the same path
# the read-side self-heal uses)
s3.versions.audit -prefix /buckets/mybucket -heal
This command is read-only by default. With -heal, it clears the stale
latest-version pointer on stranded directories and removes empty ones;
the blob is already gone, so reads then return NoSuchKey via the
clean-miss path instead of replaying the 10-retry self-heal loop on
every request.
`
}
func (c *commandS3VersionsAudit) HasTag(CommandTag) bool {
return false
}
func (c *commandS3VersionsAudit) Do(args []string, commandEnv *CommandEnv, writer io.Writer) error {
cmd := flag.NewFlagSet(c.Name(), flag.ContinueOnError)
prefix := cmd.String("prefix", "", "filer path to audit recursively (e.g. /buckets/mybucket)")
verbose := cmd.Bool("v", false, "print each stranded/orphan directory as it's found")
doHeal := cmd.Bool("heal", false, "clear the latest-version pointer on stranded directories (default: read-only)")
if err := cmd.Parse(args); err != nil {
return err
}
if *prefix == "" {
return fmt.Errorf("-prefix is required")
}
// Counters
var (
dirsScanned uint64
versionsDirs uint64
clean uint64
stranded uint64
orphanOnly uint64
empty uint64
healed uint64
healFailed uint64
)
start := time.Now()
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
err := commandEnv.WithFilerClient(false, func(client filer_pb.SeaweedFilerClient) error {
return filer_pb.TraverseBfs(ctx, &filerClientWrapper{client: client}, util.FullPath(*prefix), func(parentPath util.FullPath, entry *filer_pb.Entry) error {
atomic.AddUint64(&dirsScanned, 1)
if !entry.IsDirectory {
return nil
}
if !strings.HasSuffix(entry.Name, ".versions") {
return nil
}
atomic.AddUint64(&versionsDirs, 1)
// What does the pointer name?
var pointerFile string
if entry.Extended != nil {
if v, ok := entry.Extended[s3_constants.ExtLatestVersionFileNameKey]; ok {
pointerFile = string(v)
}
}
// List the children to see if the pointer's file exists and to
// count entries without an ExtVersionIdKey (orphans that block
// non-recursive teardown). filer_pb.List returns one 1024-entry
// page; walk all pages so a large .versions/ directory doesn't
// produce a false positive "stranded" report from only seeing
// the first page.
versionsPath := string(parentPath) + "/" + entry.Name
pointerSeen := false
hasOrphan := false
hasChild := false
const auditPageSize = 1024
var startName string
for {
pageEntries := 0
var lastEntryName string
lookupErr := filer_pb.List(ctx, &filerClientWrapper{client: client}, versionsPath, "", func(child *filer_pb.Entry, isLast bool) error {
if child == nil {
return nil
}
pageEntries++
lastEntryName = child.Name
hasChild = true
hasVersionId := false
if child.Extended != nil {
if _, ok := child.Extended[s3_constants.ExtVersionIdKey]; ok {
hasVersionId = true
}
}
if pointerFile != "" && child.Name == pointerFile {
pointerSeen = true
}
if !hasVersionId {
hasOrphan = true
}
return nil
}, startName, startName != "", auditPageSize)
if lookupErr != nil {
fmt.Fprintf(writer, "list %s: %v\n", versionsPath, lookupErr)
return nil
}
if pageEntries < auditPageSize {
break
}
startName = lastEntryName
}
switch {
case pointerFile == "":
// No pointer set. Empty, orphan-only and clean are mutually
// exclusive so the final report's category counts sum to
// versionsDirs.
switch {
case !hasChild:
atomic.AddUint64(&empty, 1)
if *verbose {
fmt.Fprintf(writer, "empty: %s\n", versionsPath)
}
if *doHeal {
if err := removeEmptyVersionsDir(ctx, client, parentPath, entry); err != nil {
atomic.AddUint64(&healFailed, 1)
fmt.Fprintf(writer, "heal failed: %s: %v\n", versionsPath, err)
} else {
atomic.AddUint64(&healed, 1)
}
}
case hasOrphan:
atomic.AddUint64(&orphanOnly, 1)
default:
atomic.AddUint64(&clean, 1)
}
case pointerSeen:
atomic.AddUint64(&clean, 1)
default:
// Pointer names a file that the listing does NOT contain.
atomic.AddUint64(&stranded, 1)
if *verbose {
fmt.Fprintf(writer, "stranded: %s pointer=%s orphan=%v\n", versionsPath, pointerFile, hasOrphan)
}
if *doHeal {
if err := healStrandedPointer(ctx, client, parentPath, entry); err != nil {
atomic.AddUint64(&healFailed, 1)
fmt.Fprintf(writer, "heal failed: %s: %v\n", versionsPath, err)
} else {
atomic.AddUint64(&healed, 1)
}
}
}
return nil
})
})
elapsed := time.Since(start)
fmt.Fprintf(writer, "audit complete in %s\n", elapsed)
fmt.Fprintf(writer, " total entries scanned : %d\n", dirsScanned)
fmt.Fprintf(writer, " .versions/ directories: %d\n", versionsDirs)
fmt.Fprintf(writer, " clean : %d\n", clean)
fmt.Fprintf(writer, " stranded : %d\n", stranded)
fmt.Fprintf(writer, " orphan-only : %d\n", orphanOnly)
fmt.Fprintf(writer, " empty : %d\n", empty)
if *doHeal {
fmt.Fprintf(writer, " healed : %d\n", healed)
fmt.Fprintf(writer, " heal failed : %d\n", healFailed)
}
return err
}
// healStrandedPointer clears the latest-version pointer extended attrs
// on a stranded .versions/ directory. The blob the pointer names is
// already gone; clearing the pointer makes subsequent reads return
// NoSuchKey via the clean-miss path instead of replaying the read-side
// self-heal on every request.
func healStrandedPointer(ctx context.Context, client filer_pb.SeaweedFilerClient, parentPath util.FullPath, entry *filer_pb.Entry) error {
if entry.Extended == nil {
return nil
}
delete(entry.Extended, s3_constants.ExtLatestVersionIdKey)
delete(entry.Extended, s3_constants.ExtLatestVersionFileNameKey)
// Also clear the cached list metadata so a stale size/mtime/etag
// can't be served back; those will be repopulated on the next PUT.
delete(entry.Extended, s3_constants.ExtLatestVersionSizeKey)
delete(entry.Extended, s3_constants.ExtLatestVersionMtimeKey)
delete(entry.Extended, s3_constants.ExtLatestVersionETagKey)
delete(entry.Extended, s3_constants.ExtLatestVersionOwnerKey)
delete(entry.Extended, s3_constants.ExtLatestVersionIsDeleteMarker)
_, err := client.UpdateEntry(ctx, &filer_pb.UpdateEntryRequest{
Directory: string(parentPath),
Entry: entry,
})
return err
}
// removeEmptyVersionsDir deletes a .versions/ directory that holds no
// children: residue of a fully-drained key. Non-recursive on purpose so a
// concurrent write landing a new version fails the delete instead of being
// lost with it.
func removeEmptyVersionsDir(ctx context.Context, client filer_pb.SeaweedFilerClient, parentPath util.FullPath, entry *filer_pb.Entry) error {
resp, err := client.DeleteEntry(ctx, &filer_pb.DeleteEntryRequest{
Directory: string(parentPath),
Name: entry.Name,
IsDeleteData: true,
})
if err != nil {
return err
}
if resp.Error != "" {
return errors.New(resp.Error)
}
return nil
}
// filerClientWrapper adapts a raw SeaweedFilerClient to the
// filer_pb.FilerClient interface that List / TraverseBfs expect.
type filerClientWrapper struct {
client filer_pb.SeaweedFilerClient
}
func (w *filerClientWrapper) WithFilerClient(streamingMode bool, fn func(filer_pb.SeaweedFilerClient) error) error {
return fn(w.client)
}
func (w *filerClientWrapper) AdjustedUrl(location *filer_pb.Location) string {
return location.Url
}
func (w *filerClientWrapper) GetDataCenter() string {
return ""
}