Remove redundant logs (#19412)

Signed-off-by: zhenshan.cao <zhenshan.cao@zilliz.com>

Signed-off-by: zhenshan.cao <zhenshan.cao@zilliz.com>
pull/18982/head
zhenshan.cao 2022-09-26 15:40:53 +08:00 committed by GitHub
parent dff4d1042e
commit c15b880f0e
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
10 changed files with 16 additions and 65 deletions

View File

@ -22,10 +22,8 @@ import (
"time"
"github.com/milvus-io/milvus/api/commonpb"
"github.com/milvus-io/milvus/internal/log"
"github.com/milvus-io/milvus/internal/proto/rootcoordpb"
"github.com/milvus-io/milvus/internal/util/typeutil"
"go.uber.org/zap"
)
const (
@ -137,10 +135,6 @@ func (ia *IDAllocator) pickCanDoFunc() {
}
}
ia.ToDoReqs = ia.ToDoReqs[idx:]
log.Debug("IDAllocator pickCanDoFunc",
zap.Uint32("need", need),
zap.Uint32("total", total),
zap.Int("remainReqCnt", len(ia.ToDoReqs)))
}
func (ia *IDAllocator) processFunc(req Request) error {

View File

@ -63,7 +63,7 @@ func (c *channelStateTimer) loadAllChannels(nodeID UniqueID) ([]*datapb.ChannelW
return nil, err
}
ret := []*datapb.ChannelWatchInfo{}
var ret []*datapb.ChannelWatchInfo
for i, k := range keys {
watchInfo, err := parseWatchInfo(k, []byte(values[i]))
@ -101,7 +101,7 @@ func (c *channelStateTimer) startOne(watchState datapb.ChannelWatchState, channe
zap.Time("timeout time", timeoutT))
select {
case <-time.NewTimer(time.Until(timeoutT)).C:
log.Info("timeout and stop timer: wait for channel ACK timeout",
log.Debug("timeout and stop timer: wait for channel ACK timeout",
zap.String("watch state", watchState.String()),
zap.Int64("nodeID", nodeID),
zap.String("channel name", channelName),

View File

@ -742,11 +742,8 @@ func (c *ChannelManager) Reassign(nodeID UniqueID, channelName string) error {
if err := c.remove(nodeID, ch); err != nil {
return fmt.Errorf("failed to remove watch info: %v,%s", ch, err.Error())
}
log.Debug("try to cleanup removal flag ", zap.String("channel name", channelName))
c.h.FinishDropChannel(channelName)
log.Info("removed channel assignment", zap.Any("channel", ch))
log.Info("removed channel assignment", zap.String("channel name", channelName))
return nil
}

View File

@ -1044,11 +1044,9 @@ func (s *Server) WatchChannels(ctx context.Context, req *datapb.WatchChannelsReq
// GetFlushState gets the flush state of multiple segments
func (s *Server) GetFlushState(ctx context.Context, req *milvuspb.GetFlushStateRequest) (*milvuspb.GetFlushStateResponse, error) {
log.Info("DataCoord receive get flush state request", zap.Int64s("segmentIDs", req.GetSegmentIDs()), zap.Int("len", len(req.GetSegmentIDs())))
resp := &milvuspb.GetFlushStateResponse{Status: &commonpb.Status{ErrorCode: commonpb.ErrorCode_UnexpectedError}}
if s.isClosed() {
log.Warn("failed to get flush state because of closed server",
log.Warn("DataCoord receive GetFlushState request, server closed",
zap.Int64s("segmentIDs", req.GetSegmentIDs()), zap.Int("len", len(req.GetSegmentIDs())))
resp.Status.Reason = msgDataCoordIsUnhealthy(Params.DataCoordCfg.GetNodeID())
return resp, nil
@ -1062,18 +1060,16 @@ func (s *Server) GetFlushState(ctx context.Context, req *milvuspb.GetFlushStateR
segment.GetState() == commonpb.SegmentState_Flushed {
continue
}
unflushed = append(unflushed, sid)
}
if len(unflushed) != 0 {
log.Info("[flush state] unflushed segment ids", zap.Int64s("segmentIDs", unflushed), zap.Int("len", len(unflushed)))
log.Info("DataCoord receive GetFlushState request, Flushed is false", zap.Int64s("segmentIDs", unflushed), zap.Int("len", len(unflushed)))
resp.Flushed = false
} else {
log.Info("[flush state] all segment is flushed", zap.Int64s("segment ids", req.GetSegmentIDs()))
log.Info("DataCoord receive GetFlushState request, Flushed is true", zap.Int64s("segmentIDs", req.GetSegmentIDs()), zap.Int("len", len(req.GetSegmentIDs())))
resp.Flushed = true
}
resp.Status.ErrorCode = commonpb.ErrorCode_Success
return resp, nil
}

View File

@ -393,15 +393,12 @@ func parseDeleteEventKey(key string) string {
func (node *DataNode) handlePutEvent(watchInfo *datapb.ChannelWatchInfo, version int64) (err error) {
vChanName := watchInfo.GetVchan().GetChannelName()
log.Info("handle put event", zap.String("watch state", watchInfo.State.String()), zap.String("vChanName", vChanName))
switch watchInfo.State {
case datapb.ChannelWatchState_Uncomplete, datapb.ChannelWatchState_ToWatch:
if err := node.flowgraphManager.addAndStart(node, watchInfo.GetVchan()); err != nil {
return fmt.Errorf("fail to add and start flowgraph for vChanName: %s, err: %v", vChanName, err)
}
log.Debug("handle put event: new data sync service success", zap.String("vChanName", vChanName))
log.Info("handle put event: new data sync service success", zap.String("vChanName", vChanName))
watchInfo.State = datapb.ChannelWatchState_WatchSuccess
case datapb.ChannelWatchState_ToRelease:
@ -429,12 +426,14 @@ func (node *DataNode) handlePutEvent(watchInfo *datapb.ChannelWatchInfo, version
// etcd valid but the states updated.
if !success {
log.Info("handle put event: failed to compare version and swap, release flowgraph",
zap.String("key", key), zap.String("state", watchInfo.State.String()))
zap.String("key", key), zap.String("state", watchInfo.State.String()),
zap.String("vChanName", vChanName))
// flow graph will leak if not release, causing new datanode failed to subscribe
node.tryToReleaseFlowgraph(vChanName)
return nil
}
log.Info("handle put event successfully", zap.String("key", key), zap.String("state", watchInfo.State.String()))
log.Info("handle put event success", zap.String("key", key),
zap.String("state", watchInfo.State.String()), zap.String("vChanName", vChanName))
return nil
}
@ -444,8 +443,8 @@ func (node *DataNode) handleDeleteEvent(vChanName string) {
// tryToReleaseFlowgraph tries to release a flowgraph
func (node *DataNode) tryToReleaseFlowgraph(vChanName string) {
log.Info("try to release flowgraph", zap.String("vChanName", vChanName))
node.flowgraphManager.release(vChanName)
log.Info("try to release flowgraph success", zap.String("vChanName", vChanName))
}
// BackGroundGC runs in background to release datanode resources
@ -455,7 +454,6 @@ func (node *DataNode) BackGroundGC(vChannelCh <-chan string) {
for {
select {
case vchanName := <-vChannelCh:
log.Info("GC flowgraph", zap.String("vChanName", vchanName))
node.tryToReleaseFlowgraph(vchanName)
case <-node.ctx.Done():
log.Warn("DataNode context done, exiting background GC")

View File

@ -128,12 +128,8 @@ func (e *channelEventManager) retryHandlePutEvent(event event) {
return
}
}
err = e.handlePutEvent(event.info, event.version)
if err != nil {
log.Warn("failed to handle put event", zap.String("vChanName", event.vChanName), zap.Error(err))
// no need to retry here,
} else {
if err == nil {
log.Info("handle put event successfully", zap.String("vChanName", event.vChanName))
return
}

View File

@ -36,12 +36,6 @@ func newFlowgraphManager() *flowgraphManager {
}
func (fm *flowgraphManager) addAndStart(dn *DataNode, vchan *datapb.VchannelInfo) error {
log.Info("received Vchannel Info",
zap.String("vChannelName", vchan.GetChannelName()),
zap.Int("Unflushed Segment Number", len(vchan.GetUnflushedSegmentIds())),
zap.Int("Flushed Segment Number", len(vchan.GetFlushedSegmentIds())),
)
if _, ok := fm.flowgraphs.Load(vchan.GetChannelName()); ok {
log.Warn("try to add an existed DataSyncService", zap.String("vChannelName", vchan.GetChannelName()))
return nil
@ -61,11 +55,7 @@ func (fm *flowgraphManager) addAndStart(dn *DataNode, vchan *datapb.VchannelInfo
log.Warn("new data sync service fail", zap.String("vChannelName", vchan.GetChannelName()), zap.Error(err))
return err
}
log.Info("successfully created dataSyncService", zap.String("vChannelName", vchan.GetChannelName()))
dataSyncService.start()
log.Info("successfully started dataSyncService", zap.String("vChannelName", vchan.GetChannelName()))
fm.flowgraphs.Store(vchan.GetChannelName(), dataSyncService)
metrics.DataNodeNumFlowGraphs.WithLabelValues(fmt.Sprint(Params.DataNodeCfg.GetNodeID())).Inc()
@ -73,14 +63,11 @@ func (fm *flowgraphManager) addAndStart(dn *DataNode, vchan *datapb.VchannelInfo
}
func (fm *flowgraphManager) release(vchanName string) {
log.Info("release flowgraph resources begin", zap.String("vChannelName", vchanName))
if fg, loaded := fm.flowgraphs.LoadAndDelete(vchanName); loaded {
fg.(*dataSyncService).close()
metrics.DataNodeNumFlowGraphs.WithLabelValues(fmt.Sprint(Params.DataNodeCfg.GetNodeID())).Dec()
}
rateCol.removeFlowGraphChannel(vchanName)
log.Info("release flowgraph resources end", zap.String("Vchannel", vchanName))
}
func (fm *flowgraphManager) getFlushCh(segID UniqueID) (chan<- flushMsg, error) {

View File

@ -202,8 +202,8 @@ func newReplica(ctx context.Context, rc types.RootCoord, cm storage.ChunkManager
// try to cache latest schema
_, err := replica.getCollectionSchema(collID, 0)
if err != nil {
log.Warn("failed to get schema when create replica", zap.Int64("collID", collID), zap.Error(err))
return nil, err
err2 := fmt.Errorf("replica get schema for collection %d failed:%w", collID, err)
return nil, err2
}
return replica, nil
@ -758,17 +758,12 @@ func (replica *SegmentReplica) getCollectionID() UniqueID {
// If you want the latest collection schema, ts should be 0.
func (replica *SegmentReplica) getCollectionSchema(collID UniqueID, ts Timestamp) (*schemapb.CollectionSchema, error) {
if !replica.validCollection(collID) {
log.Warn("Mismatch collection for the replica",
zap.Int64("Want", replica.collectionID),
zap.Int64("Actual", collID),
)
return nil, fmt.Errorf("not supported collection %v", collID)
return nil, fmt.Errorf("mismatch collection, want %d, actual %d", replica.collectionID, collID)
}
if replica.collSchema == nil {
sch, err := replica.metaService.getCollectionSchema(context.Background(), collID, ts)
if err != nil {
log.Error("Grpc error", zap.Error(err))
return nil, err
}
replica.collSchema = sch

View File

@ -120,17 +120,12 @@ func (fsw *flushedSegmentWatcher) enqueueInternalTask(segmentID UniqueID) {
defer fsw.internalNotifyFunc()
fsw.internalTaskMutex.Lock()
defer fsw.internalTaskMutex.Unlock()
log.Ctx(fsw.ctx).Info("flushedSegmentWatcher enqueueInternalTask", zap.Int64("segmentID", segmentID))
if _, ok := fsw.internalTasks[segmentID]; !ok {
fsw.internalTasks[segmentID] = &internalTask{
state: indexTaskPrepare,
segmentInfo: nil,
}
}
log.Ctx(fsw.ctx).Info("flushedSegmentWatcher already have the task success", zap.Int64("segmentID", segmentID))
}
func (fsw *flushedSegmentWatcher) internalScheduler() {
@ -157,8 +152,6 @@ func (fsw *flushedSegmentWatcher) internalRun() {
fsw.internalTaskMutex.RLock()
segmentIDs := make([]UniqueID, 0, len(fsw.internalTasks))
if len(fsw.internalTasks) > 0 {
log.Ctx(fsw.ctx).Debug("IndexCoord flushedSegmentWatcher schedule internal tasks",
zap.Int("internal task num", len(fsw.internalTasks)))
for segID := range fsw.internalTasks {
segmentIDs = append(segmentIDs, segID)
}

View File

@ -674,7 +674,6 @@ func (i *IndexCoord) DropIndex(ctx context.Context, req *indexpb.DropIndexReques
// GetIndexInfos gets the index file paths from IndexCoord.
func (i *IndexCoord) GetIndexInfos(ctx context.Context, req *indexpb.GetIndexInfoRequest) (*indexpb.GetIndexInfoResponse, error) {
log.Debug("IndexCoord GetIndexFilePaths", zap.String("indexName", req.IndexName), zap.Int64s("segmentIDs", req.SegmentIDs))
if !i.isHealthy() {
log.Warn(msgIndexCoordIsUnhealthy(i.serverID))
return &indexpb.GetIndexInfoResponse{
@ -719,15 +718,11 @@ func (i *IndexCoord) GetIndexInfos(ctx context.Context, req *indexpb.GetIndexInf
}
}
log.Info("IndexCoord GetIndexFilePaths ", zap.Int("segIDs num", len(req.SegmentIDs)),
zap.Int("file path num", len(ret.SegmentInfo)))
return ret, nil
}
// DescribeIndex describe the index info of the collection.
func (i *IndexCoord) DescribeIndex(ctx context.Context, req *indexpb.DescribeIndexRequest) (*indexpb.DescribeIndexResponse, error) {
log.Info("IndexCoord receive DescribeIndex", zap.Int64("collID", req.CollectionID))
if !i.isHealthy() {
log.Warn(msgIndexCoordIsUnhealthy(i.serverID))
return &indexpb.DescribeIndexResponse{