Files
seaweedfs/weed/command/backup.go
Chris Lu 1c0e24f06a fix(balance): don't move remote-tiered volumes; don't fatal on missing .idx (#9335)
* fix(volume): don't fatal on missing .idx for remote-tiered volume

A .vif left behind without its .idx (orphaned by a crashed move, partial
copy, or hand-edit) would trip glog.Fatalf in checkIdxFile and take the
whole volume server down on boot, killing every healthy volume on it
too. For remote-tiered volumes treat it as a per-volume load error so
the server can come up and the operator can clean up the stray .vif.

Refs #9331.

* fix(balance): skip remote-tiered volumes in admin balance detection

The admin/worker balance detector had no equivalent of the shell-side
guard ("does not move volume in remote storage" in
command_volume_balance.go), so it scheduled moves on remote-tiered
volumes. The "move" copies .idx/.vif to the destination and then calls
Volume.Destroy on the source, which calls backendStorage.DeleteFile —
deleting the remote object the destination's new .vif now points at.

Populate HasRemoteCopy on the metrics emitted by both the admin
maintenance scanner and the worker's master poll, then drop those
volumes at the top of Detection.

Fixes #9331.

* Apply suggestion from @gemini-code-assist[bot]

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

* fix(volume): keep remote data on volume-move-driven delete

The on-source delete after a volume move (admin/worker balance and
shell volume.move) ran Volume.Destroy with no way to opt out of the
remote-object cleanup. Volume.Destroy unconditionally calls
backendStorage.DeleteFile for remote-tiered volumes, so a successful
move would copy .idx/.vif to the destination and then nuke the cloud
object the destination's new .vif was already pointing at.

Add VolumeDeleteRequest.keep_remote_data and plumb it through
Store.DeleteVolume / DiskLocation.DeleteVolume / Volume.Destroy. The
balance task and shell volume.move set it to true; the post-tier-upload
cleanup of other replicas and the over-replication trim in
volume.fix.replication also set it to true since the remote object is
still referenced. Other real-delete callers keep the default. The
delete-before-receive path in VolumeCopy also sets it: the inbound copy
carries a .vif that may reference the same cloud object as the
existing volume.

Refs #9331.

* test(storage): in-process remote-tier integration tests

Cover the four operations the user is most likely to run against a
cloud-tiered volume — balance/move, vacuum, EC encode, EC decode — by
registering a local-disk-backed BackendStorage as the "remote" tier and
exercising the real Volume / DiskLocation / EC encoder code paths.

Locks in:
- Destroy(keepRemoteData=true) preserves the remote object (move case)
- Destroy(keepRemoteData=false) deletes it (real-delete case)
- Vacuum/compact on a remote-tier volume never deletes the remote object
- EC encode requires the local .dat (callers must download first)
- EC encode + rebuild round-trips after a tier-down

Tests run in-process and finish in under a second total — no cluster,
binary, or external storage required.

* fix(rust-volume): keep remote data on volume-move-driven delete

Mirror the Go fix in seaweed-volume: plumb keep_remote_data through
grpc volume_delete → Store.delete_volume → DiskLocation.delete_volume
→ Volume.destroy, and skip the s3-tier delete_file call when the flag
is set. The pre-receive cleanup in volume_copy passes true for the
same reason as the Go side: the inbound copy carries a .vif that may
reference the same cloud object as the existing volume.

The Rust loader already warns rather than fataling on a stray .vif
without an .idx (volume.rs load_index_inmemory / load_index_redb), so
no counterpart to the Go fatal-on-missing-idx fix is needed.

Refs #9331.

* fix(volume): preserve remote tier on IO-error eviction; fix EC test target

Two review nits:

- Store.MaybeAddVolumes' periodic cleanup pass deleted IO-errored
  volumes with keepRemoteData=false, so a transient local fault on a
  remote-tiered volume would also nuke the cloud object. Track the
  delete reason via a parallel slice and pass keepRemoteData=v.HasRemoteFile()
  for IO-error evictions; TTL-expired evictions still pass false.

- TestRemoteTier_ECEncodeDecode_AfterDownload deleted shards 0..3 but
  called them "parity" — by the klauspost/reedsolomon convention shards
  0..DataShardsCount-1 are data and DataShardsCount..TotalShardsCount-1
  are parity. Switch the loop to delete the parity range so the
  intent matches the indices.

---------

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2026-05-06 15:19:43 -07:00

236 lines
7.6 KiB
Go

package command
import (
"context"
"fmt"
"google.golang.org/grpc"
"github.com/seaweedfs/seaweedfs/weed/pb"
"github.com/seaweedfs/seaweedfs/weed/security"
"github.com/seaweedfs/seaweedfs/weed/storage/needle"
"github.com/seaweedfs/seaweedfs/weed/storage/super_block"
"github.com/seaweedfs/seaweedfs/weed/util"
"github.com/seaweedfs/seaweedfs/weed/operation"
"github.com/seaweedfs/seaweedfs/weed/storage"
)
var (
s BackupOptions
)
type BackupOptions struct {
master *string
server *string // deprecated, for backward compatibility
collection *string
dir *string
volumeId *int
ttl *string
replication *string
}
func init() {
cmdBackup.Run = runBackup // break init cycle
s.master = cmdBackup.Flag.String("master", "localhost:9333", "SeaweedFS master location")
s.server = cmdBackup.Flag.String("server", "", "SeaweedFS master location (deprecated, use -master instead)")
s.collection = cmdBackup.Flag.String("collection", "", "collection name")
s.dir = cmdBackup.Flag.String("dir", ".", "directory to store volume data files")
s.volumeId = cmdBackup.Flag.Int("volumeId", -1, "a volume id. The volume .dat and .idx files should already exist in the dir.")
s.ttl = cmdBackup.Flag.String("ttl", "", `backup volume's time to live, format:
3m: 3 minutes
4h: 4 hours
5d: 5 days
6w: 6 weeks
7M: 7 months
8y: 8 years
default is the same with origin`)
s.replication = cmdBackup.Flag.String("replication", "", "backup volume's replication, default is the same with origin")
}
var cmdBackup = &Command{
UsageLine: "backup -dir=. -volumeId=234 -master=localhost:9333",
Short: "incrementally backup a volume to local folder",
Long: `Incrementally backup volume data.
It is expected that you use this inside a script, to loop through
all possible volume ids that needs to be backup to local folder.
The volume id does not need to exist locally or even remotely.
This will help to backup future new volumes.
Usually backing up is just copying the .dat (and .idx) files.
But it's tricky to incrementally copy the differences.
The complexity comes when there are multiple addition, deletion and compaction.
This tool will handle them correctly and efficiently, avoiding unnecessary data transportation.
`,
}
// parseTTL parses the TTL from user input or volume stats.
// Returns (ttl, error, isFatal) where isFatal=true for invalid user input.
func parseTTL(userTTL string, statsTTL string) (*needle.TTL, error, bool) {
if userTTL != "" {
ttl, err := needle.ReadTTL(userTTL)
if err != nil {
// User-provided TTL is invalid - this is fatal
return nil, fmt.Errorf("invalid user-provided ttl %s: %w", userTTL, err), true
}
return ttl, nil, false
}
ttl, err := needle.ReadTTL(statsTTL)
if err != nil {
return nil, fmt.Errorf("parsing ttl %s from stats: %w", statsTTL, err), false
}
return ttl, nil, false
}
// parseReplication parses the replication from user input or volume stats.
// Returns (replication, error, isFatal) where isFatal=true for invalid user input.
func parseReplication(userReplication string, statsReplication string) (*super_block.ReplicaPlacement, error, bool) {
if userReplication != "" {
replication, err := super_block.NewReplicaPlacementFromString(userReplication)
if err != nil {
// User-provided replication is invalid - this is fatal
return nil, fmt.Errorf("invalid user-provided replication %s: %w", userReplication, err), true
}
return replication, nil, false
}
replication, err := super_block.NewReplicaPlacementFromString(statsReplication)
if err != nil {
return nil, fmt.Errorf("parsing replication %s from stats: %w", statsReplication, err), false
}
return replication, nil, false
}
// backupFromLocation attempts to backup a volume from a specific volume server location.
// Returns (error, isFatal) where isFatal=true means the error is due to invalid user input
// and should not be retried with other locations.
func backupFromLocation(volumeServer pb.ServerAddress, grpcDialOption grpc.DialOption, vid needle.VolumeId) (error, bool) {
stats, err := operation.GetVolumeSyncStatus(volumeServer, grpcDialOption, uint32(vid))
if err != nil {
return fmt.Errorf("getting volume status: %w", err), false
}
// Parse TTL
ttl, err, isFatal := parseTTL(*s.ttl, stats.Ttl)
if err != nil {
return err, isFatal
}
// Parse replication
replication, err, isFatal := parseReplication(*s.replication, stats.Replication)
if err != nil {
return err, isFatal
}
ver := needle.Version(stats.Version)
// Create or load the volume
v, err := storage.NewVolume(*s.dir, *s.dir, *s.collection, vid, storage.NeedleMapInMemory, replication, ttl, 0, ver, 0, 0)
if err != nil {
return fmt.Errorf("creating or reading volume: %w", err), false
}
// Handle compaction if needed
if v.SuperBlock.CompactionRevision < uint16(stats.CompactRevision) {
if err = v.CompactByIndex(nil); err != nil {
v.Close()
return fmt.Errorf("compacting volume: %w", err), false
}
if err = v.CommitCompact(); err != nil {
v.Close()
return fmt.Errorf("committing compaction: %w", err), false
}
v.SuperBlock.CompactionRevision = uint16(stats.CompactRevision)
if _, err = v.DataBackend.WriteAt(v.SuperBlock.Bytes(), 0); err != nil {
v.Close()
return fmt.Errorf("writing superblock: %w", err), false
}
}
datSize, _, _ := v.FileStat()
// If local volume is larger than remote, recreate it
if datSize > stats.TailOffset {
if err := v.Destroy(false, false); err != nil {
v.Close()
return fmt.Errorf("destroying volume: %w", err), false
}
v.Close() // Close the destroyed volume
// recreate an empty volume
v, err = storage.NewVolume(*s.dir, *s.dir, *s.collection, vid, storage.NeedleMapInMemory, replication, ttl, 0, ver, 0, 0)
if err != nil {
return fmt.Errorf("recreating volume: %w", err), false
}
}
// Perform the incremental backup
if err := v.IncrementalBackup(volumeServer, grpcDialOption); err != nil {
v.Close()
return fmt.Errorf("incremental backup: %w", err), false
}
v.Close()
return nil, false
}
func runBackup(cmd *Command, args []string) bool {
*s.dir = util.ResolvePath(*s.dir)
util.LoadSecurityConfiguration()
grpcDialOption := security.LoadClientTLS(util.GetViper(), "grpc.client")
// Backward compatibility: if -server is provided, use it
masterServer := *s.master
if *s.server != "" {
masterServer = *s.server
}
if *s.volumeId == -1 {
return false
}
vid := needle.VolumeId(*s.volumeId)
// find volume location, replication, ttl info
lookup, err := operation.LookupVolumeId(func(_ context.Context) pb.ServerAddress { return pb.ServerAddress(masterServer) }, grpcDialOption, vid.String())
if err != nil {
fmt.Printf("Error looking up volume %d: %v\n", vid, err)
return true
}
if len(lookup.Locations) == 0 {
fmt.Printf("Error: volume %d has no locations available\n", vid)
return true
}
// Try each available location until one succeeds
var lastErr error
for i, location := range lookup.Locations {
volumeServer := location.ServerAddress()
fmt.Printf("Attempting to backup volume %d from location %d/%d: %s\n", vid, i+1, len(lookup.Locations), volumeServer)
err, isFatal := backupFromLocation(volumeServer, grpcDialOption, vid)
if err != nil {
fmt.Printf("Error backing up volume %d from %s: %v\n", vid, volumeServer, err)
lastErr = err
// Check if this is a fatal user-input error
if isFatal {
return true
}
continue
}
// Success!
fmt.Printf("Successfully backed up volume %d from %s\n", vid, volumeServer)
return true
}
// All locations failed
fmt.Printf("Failed to backup volume %d after trying all %d locations. Last error: %v\n", vid, len(lookup.Locations), lastErr)
return true
}