Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

restore_data: remove wait apply phase (#50316) #50473

Merged
merged 2 commits into from
Feb 21, 2024
Merged

restore_data: remove wait apply phase (#50316) #50473

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c41478b
This is an automated cherry-pick of #50316
YuJuncen Jan 16, 2024
fc08a38
resolved conflicts
YuJuncen Feb 21, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
42 changes: 1 addition & 41 deletions br/pkg/restore/data.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -140,10 +140,6 @@ func doRecoveryData(ctx context.Context, resolveTS uint64, allStores []*metapb.S
return totalRegions, recoveryError{error: err, atStage: StageRecovering}
}

if err := recovery.WaitApply(ctx); err != nil {
return totalRegions, recoveryError{error: err, atStage: StageRecovering}
}

if err := recovery.PrepareFlashbackToVersion(ctx, resolveTS, restoreTS-1); err != nil {
return totalRegions, recoveryError{error: err, atStage: StageFlashback}
}
Expand Down Expand Up @@ -228,7 +224,7 @@ func (recovery *Recovery) ReadRegionMeta(ctx context.Context) error {
totalStores := len(recovery.allStores)
workers := utils.NewWorkerPool(uint(mathutil.Min(totalStores, common.MaxStoreConcurrency)), "Collect Region Meta") // TODO: int overflow?

// TODO: optimize the ErroGroup when TiKV is panic
// TODO: optimize the ErrorGroup when TiKV is panic
metaChan := make(chan StoreMeta, 1024)
defer close(metaChan)

Expand Down Expand Up @@ -399,42 +395,6 @@ func (recovery *Recovery) SpawnTiKVShutDownWatchers(ctx context.Context) {
go mainLoop()
}

// WaitApply send wait apply to all tikv ensure all region peer apply log into the last
func (recovery *Recovery) WaitApply(ctx context.Context) (err error) {
eg, ectx := errgroup.WithContext(ctx)
totalStores := len(recovery.allStores)
workers := utils.NewWorkerPool(uint(mathutil.Min(totalStores, common.MaxStoreConcurrency)), "wait apply")

for _, store := range recovery.allStores {
if err := ectx.Err(); err != nil {
break
}
storeAddr := getStoreAddress(recovery.allStores, store.Id)
storeId := store.Id

workers.ApplyOnErrorGroup(eg, func() error {
recoveryClient, conn, err := recovery.newRecoveryClient(ectx, storeAddr)
if err != nil {
return errors.Trace(err)
}
defer conn.Close()
log.Info("send wait apply to tikv", zap.String("tikv address", storeAddr), zap.Uint64("store id", storeId))
req := &recovpb.WaitApplyRequest{StoreId: storeId}
_, err = recoveryClient.WaitApply(ectx, req)
if err != nil {
log.Error("wait apply failed", zap.Uint64("store id", storeId))
return errors.Trace(err)
}

recovery.progress.Inc()
log.Info("wait apply execution success", zap.Uint64("store id", storeId))
return nil
})
}
// Wait for all TiKV instances force leader and wait apply to last log.
return eg.Wait()
}

// prepare the region for flashback the data, the purpose is to stop region service, put region in flashback state
func (recovery *Recovery) PrepareFlashbackToVersion(ctx context.Context, resolveTS uint64, startTS uint64) (err error) {
retryState := utils.InitialRetryState(utils.FlashbackRetryTime, utils.FlashbackWaitInterval, utils.FlashbackMaxWaitInterval)
Expand Down
Loading