Files
seaweedfs/weed/storage/store_vacuum.go
T
9575032b4c volume: forward fsync=true to replicas in ReplicatedWrite (#10805)
* volume: forward fsync=true to replicas in ReplicatedWrite

When a write request carries fsync=true, only the primary volume server
flushed to disk: the replica fan-out URL in ReplicatedWrite only carried
type/ttl/ts/cm, so replicas always wrote without fsync even when the
client explicitly requested a durable write.

Forward the fsync request parameter to the replica volume servers so a
durable write means every replica has flushed to disk, not just the
primary. Replicas without fsync are untouched (zero behavior change).

* storage: flush a durable write inline while stopping

The fsync flag on the write path really selects the async batch worker,
and it was switched off once the store is stopping. So a fsync=true write
landing during the pre-stop drain got acked without ever being flushed -
and now that ReplicatedWrite forwards fsync, that covers replicas too.

Flush it inline instead of queueing it. The drain keeps accepting writes,
which is the whole point of preStopSeconds, and the ack still means the
.dat is on disk. If the fsync fails, the append comes back off the .dat
and the needle map goes back to what it pointed at before, so nothing
resolves to an offset past the truncated end.

* storage: make the store's stopping flag atomic

SetStopping runs on the signal handler goroutine while the write and
vacuum paths read the flag, so every read of it was racy. Nothing about
the shutdown ordering changes; only the flag itself is now safe to read.

* topology: check the errors the replication test was dropping

The mock replica ignored its response write and the mock master ignored
whatever Serve returned, so a broken mock would have shown up as a
confusing timeout rather than a failure. Also drops the explicit listener
close: grpc.Server.Stop already closes the listener it was given.

---------

Co-authored-by: hzsunchao <hzsunchao@corp.netease.com>
Co-authored-by: Chris Lu <chris.lu@gmail.com>
2026-08-18 17:17:50 -07:00

124 lines
4.1 KiB
Go

package storage
import (
"fmt"
"github.com/seaweedfs/seaweedfs/weed/stats"
"github.com/seaweedfs/seaweedfs/weed/glog"
"github.com/seaweedfs/seaweedfs/weed/storage/needle"
)
func (s *Store) CheckCompactVolume(volumeId needle.VolumeId) (float64, error) {
if v := s.findVolume(volumeId); v != nil {
glog.V(3).Infof("volume %d garbage level: %f", volumeId, v.garbageLevel())
return v.garbageLevel(), nil
}
return 0, fmt.Errorf("volume id %d is not found during check compact", volumeId)
}
func (s *Store) CompactVolume(vid needle.VolumeId, preallocate int64, compactionBytePerSecond int64, progressFn ProgressFunc) error {
if v := s.findVolume(vid); v != nil {
if err := ensureCompactVolumeSpace(v, preallocate); err != nil {
return err
}
return v.CompactByIndex(&CompactOptions{
PreallocateBytes: preallocate,
MaxBytesPerSecond: compactionBytePerSecond,
ProgressCallback: progressFn,
})
}
return fmt.Errorf("volume id %d is not found during compact", vid)
}
func (s *Store) CommitCompactVolume(vid needle.VolumeId) (bool, int64, error) {
if s.isStopping.Load() {
return false, 0, fmt.Errorf("volume id %d skips compact because volume is stopping", vid)
}
if v := s.findVolume(vid); v != nil {
isReadOnly := v.IsReadOnly()
err := v.CommitCompact()
var volumeSize int64 = 0
if err == nil && v.DataBackend != nil {
volumeSize, _, _ = v.DataBackend.GetStat()
}
return isReadOnly, volumeSize, err
}
return false, 0, fmt.Errorf("volume id %d is not found during commit compact", vid)
}
func (s *Store) CommitCleanupVolume(vid needle.VolumeId) error {
if v := s.findVolume(vid); v != nil {
return v.cleanupCompact()
}
return fmt.Errorf("volume id %d is not found during cleaning up", vid)
}
func ensureCompactVolumeSpace(v *Volume, preallocate int64) error {
// Get current volume size for space calculation
volumeSize, indexSize, _ := v.FileStat()
// Calculate space needed for compaction:
// 1. Space for the new compacted volume (approximately same as current volume size)
// 2. Use the larger of preallocate or estimated volume size
estimatedCompactSize := int64(volumeSize + indexSize)
spaceNeeded := preallocate
if estimatedCompactSize > preallocate {
spaceNeeded = estimatedCompactSize
}
diskStatus := stats.NewDiskStatus(v.dir)
if int64(diskStatus.Free) < spaceNeeded {
return fmt.Errorf("insufficient free space for compaction: need %d bytes (volume: %d, index: %d), but only %d bytes available",
spaceNeeded, volumeSize, indexSize, diskStatus.Free)
}
glog.V(1).Infof("volume %d compaction space check: volume=%d, index=%d, space_needed=%d, free_space=%d",
v.Id, volumeSize, indexSize, spaceNeeded, diskStatus.Free)
return nil
}
func (s *Store) CompactVolumeFiles(vid needle.VolumeId, collection string, location *DiskLocation, needleMapKind NeedleMapKind, ldbTimeout int64, preallocate int64, compactionBytePerSecond int64) (err error) {
if location == nil {
return fmt.Errorf("volume %d compaction location is nil", vid)
}
tempVolume, err := loadVolumeWithoutWorker(location.Directory, location.IdxDirectory, collection, vid, needleMapKind, ldbTimeout)
if err != nil {
return fmt.Errorf("load volume %d for offline compaction: %w", vid, err)
}
tempVolume.location = location
defer func() {
if tempVolume.tmpNm != nil {
tempVolume.tmpNm.Close()
tempVolume.tmpNm = nil
}
tempVolume.doClose()
}()
if err := ensureCompactVolumeSpace(tempVolume, preallocate); err != nil {
return err
}
if err := tempVolume.CompactByIndex(&CompactOptions{
PreallocateBytes: preallocate,
MaxBytesPerSecond: compactionBytePerSecond,
}); err != nil {
if cleanupErr := tempVolume.cleanupCompact(); cleanupErr != nil {
return fmt.Errorf("compact volume %d: %v (cleanup failed: %v)", vid, err, cleanupErr)
}
return fmt.Errorf("compact volume %d: %w", vid, err)
}
if err := tempVolume.CommitCompact(); err != nil {
if cleanupErr := tempVolume.cleanupCompact(); cleanupErr != nil {
return fmt.Errorf("commit compact volume %d: %v (cleanup failed: %v)", vid, err, cleanupErr)
}
return fmt.Errorf("commit compact volume %d: %w", vid, err)
}
return nil
}