milvus/internal/datacoord/sync_segments_scheduler.go
congqixia 582d2eec79
enhance: Move datanode/indexnode manager to session pkg (#35634)
Related to #28861

Move session manager, worker manager to session package. Also renaming
each manager to corresponding node name(datanode, indexnode).

---------

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2024-08-22 16:02:56 +08:00

159 lines
5.1 KiB
Go

// Licensed to the LF AI & Data foundation under one
// or more contributor license agreements. See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership. The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License. You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package datacoord
import (
"sync"
"time"
"github.com/samber/lo"
"go.uber.org/zap"
"github.com/milvus-io/milvus/internal/datacoord/session"
"github.com/milvus-io/milvus/internal/proto/datapb"
"github.com/milvus-io/milvus/pkg/log"
"github.com/milvus-io/milvus/pkg/util/logutil"
"github.com/milvus-io/milvus/pkg/util/typeutil"
)
type SyncSegmentsScheduler struct {
quit chan struct{}
wg sync.WaitGroup
meta *meta
channelManager ChannelManager
sessions session.DataNodeManager
}
func newSyncSegmentsScheduler(m *meta, channelManager ChannelManager, sessions session.DataNodeManager) *SyncSegmentsScheduler {
return &SyncSegmentsScheduler{
quit: make(chan struct{}),
wg: sync.WaitGroup{},
meta: m,
channelManager: channelManager,
sessions: sessions,
}
}
func (sss *SyncSegmentsScheduler) Start() {
sss.quit = make(chan struct{})
sss.wg.Add(1)
go func() {
defer logutil.LogPanic()
ticker := time.NewTicker(Params.DataCoordCfg.SyncSegmentsInterval.GetAsDuration(time.Second))
defer sss.wg.Done()
for {
select {
case <-sss.quit:
log.Info("sync segments scheduler quit")
ticker.Stop()
return
case <-ticker.C:
sss.SyncSegmentsForCollections()
}
}
}()
log.Info("SyncSegmentsScheduler started...")
}
func (sss *SyncSegmentsScheduler) Stop() {
close(sss.quit)
sss.wg.Wait()
}
func (sss *SyncSegmentsScheduler) SyncSegmentsForCollections() {
collIDs := sss.meta.ListCollections()
for _, collID := range collIDs {
collInfo := sss.meta.GetCollection(collID)
if collInfo == nil {
log.Warn("collection info is nil, skip it", zap.Int64("collectionID", collID))
continue
}
pkField, err := typeutil.GetPrimaryFieldSchema(collInfo.Schema)
if err != nil {
log.Warn("get primary field from schema failed", zap.Int64("collectionID", collID),
zap.Error(err))
continue
}
for _, channelName := range collInfo.VChannelNames {
nodeID, err := sss.channelManager.FindWatcher(channelName)
if err != nil {
log.Warn("find watcher for channel failed", zap.Int64("collectionID", collID),
zap.String("channelName", channelName), zap.Error(err))
continue
}
for _, partitionID := range collInfo.Partitions {
if err := sss.SyncSegments(collID, partitionID, channelName, nodeID, pkField.GetFieldID()); err != nil {
log.Warn("sync segment with channel failed, retry next ticker",
zap.Int64("collectionID", collID),
zap.Int64("partitionID", partitionID),
zap.String("channel", channelName),
zap.Error(err))
continue
}
}
}
}
}
func (sss *SyncSegmentsScheduler) SyncSegments(collectionID, partitionID int64, channelName string, nodeID, pkFieldID int64) error {
log := log.With(zap.Int64("collectionID", collectionID), zap.Int64("partitionID", partitionID),
zap.String("channelName", channelName), zap.Int64("nodeID", nodeID))
// sync all healthy segments, but only check flushed segments on datanode. Because L0 growing segments may not in datacoord's meta.
// upon receiving the SyncSegments request, the datanode's segment state may have already transitioned from Growing/Flushing
// to Flushed, so the view must include this segment.
segments := sss.meta.SelectSegments(WithChannel(channelName), SegmentFilterFunc(func(info *SegmentInfo) bool {
return info.GetPartitionID() == partitionID && info.GetLevel() != datapb.SegmentLevel_L0 && isSegmentHealthy(info)
}))
req := &datapb.SyncSegmentsRequest{
ChannelName: channelName,
PartitionId: partitionID,
CollectionId: collectionID,
SegmentInfos: make(map[int64]*datapb.SyncSegmentInfo),
}
for _, seg := range segments {
req.SegmentInfos[seg.ID] = &datapb.SyncSegmentInfo{
SegmentId: seg.GetID(),
State: seg.GetState(),
Level: seg.GetLevel(),
NumOfRows: seg.GetNumOfRows(),
}
statsLogs := make([]*datapb.Binlog, 0)
for _, statsLog := range seg.GetStatslogs() {
if statsLog.GetFieldID() == pkFieldID {
statsLogs = append(statsLogs, statsLog.GetBinlogs()...)
}
}
req.SegmentInfos[seg.ID].PkStatsLog = &datapb.FieldBinlog{
FieldID: pkFieldID,
Binlogs: statsLogs,
}
}
if err := sss.sessions.SyncSegments(nodeID, req); err != nil {
log.Warn("fail to sync segments with node", zap.Error(err))
return err
}
log.Info("sync segments success", zap.Int64s("segments", lo.Map(segments, func(t *SegmentInfo, i int) int64 {
return t.GetID()
})))
return nil
}