milvus/internal/datanode/flow_graph_delete_node.go
XuanYang-cn e081eb287d
Refine flowgraph nodes for readability (#19864)
- delete compaction tasks from executor while dropping

See also: #19780

Signed-off-by: yangxuan <xuan.yang@zilliz.com>

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2022-10-20 16:39:29 +08:00

299 lines
9.2 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 datanode
import (
"context"
"fmt"
"reflect"
"sync"
"github.com/opentracing/opentracing-go"
"go.uber.org/zap"
"github.com/milvus-io/milvus-proto/go-api/schemapb"
"github.com/milvus-io/milvus/internal/common"
"github.com/milvus-io/milvus/internal/log"
"github.com/milvus-io/milvus/internal/metrics"
"github.com/milvus-io/milvus/internal/mq/msgstream"
"github.com/milvus-io/milvus/internal/storage"
"github.com/milvus-io/milvus/internal/util/retry"
"github.com/milvus-io/milvus/internal/util/trace"
)
// DeleteNode is to process delete msg, flush delete info into storage.
type deleteNode struct {
BaseNode
ctx context.Context
channelName string
delBuf sync.Map // map[segmentID]*DelDataBuf
channel Channel
idAllocator allocatorInterface
flushManager flushManager
clearSignal chan<- string
}
func (dn *deleteNode) Name() string {
return "deleteNode-" + dn.channelName
}
func (dn *deleteNode) Close() {
log.Info("Flowgraph Delete Node closing")
}
func (dn *deleteNode) showDelBuf(segIDs []UniqueID, ts Timestamp) {
for _, segID := range segIDs {
if v, ok := dn.delBuf.Load(segID); ok {
delDataBuf, _ := v.(*DelDataBuf)
log.Debug("delta buffer status",
zap.Uint64("timestamp", ts),
zap.Int64("segment ID", segID),
zap.Int64("entries", delDataBuf.GetEntriesNum()),
zap.String("vChannel", dn.channelName))
}
}
}
// Operate implementing flowgraph.Node, performs delete data process
func (dn *deleteNode) Operate(in []Msg) []Msg {
if in == nil {
log.Debug("type assertion failed for flowGraphMsg because it's nil")
return []Msg{}
}
if len(in) != 1 {
log.Error("Invalid operate message input in deleteNode", zap.Int("input length", len(in)))
return []Msg{}
}
fgMsg, ok := in[0].(*flowGraphMsg)
if !ok {
log.Warn("type assertion failed for flowGraphMsg", zap.String("name", reflect.TypeOf(in[0]).Name()))
return []Msg{}
}
var spans []opentracing.Span
for _, msg := range fgMsg.deleteMessages {
sp, ctx := trace.StartSpanFromContext(msg.TraceCtx())
spans = append(spans, sp)
msg.SetTraceCtx(ctx)
}
// update compacted segment before operation
if len(fgMsg.deleteMessages) > 0 || len(fgMsg.segmentsToFlush) > 0 {
dn.updateCompactedSegments()
}
// process delete messages
var segIDs []UniqueID
for i, msg := range fgMsg.deleteMessages {
traceID, _, _ := trace.InfoFromSpan(spans[i])
log.Info("Buffer delete request in DataNode", zap.String("traceID", traceID))
tmpSegIDs, err := dn.bufferDeleteMsg(msg, fgMsg.timeRange)
if err != nil {
// error occurs only when deleteMsg is misaligned, should not happen
err = fmt.Errorf("buffer delete msg failed, err = %s", err)
log.Error(err.Error())
panic(err)
}
segIDs = append(segIDs, tmpSegIDs...)
}
// display changed segment's status in dn.delBuf of a certain ts
if len(fgMsg.deleteMessages) != 0 {
dn.showDelBuf(segIDs, fgMsg.timeRange.timestampMax)
}
// process flush messages
if len(fgMsg.segmentsToFlush) > 0 {
log.Debug("DeleteNode receives flush message",
zap.Int64s("segIDs", fgMsg.segmentsToFlush),
zap.String("vChannelName", dn.channelName))
for _, segmentToFlush := range fgMsg.segmentsToFlush {
buf, ok := dn.delBuf.Load(segmentToFlush)
if !ok {
// no related delta data to flush, send empty buf to complete flush life-cycle
dn.flushManager.flushDelData(nil, segmentToFlush, fgMsg.endPositions[0])
} else {
err := retry.Do(dn.ctx, func() error {
return dn.flushManager.flushDelData(buf.(*DelDataBuf), segmentToFlush, fgMsg.endPositions[0])
}, getFlowGraphRetryOpt())
if err != nil {
err = fmt.Errorf("failed to flush delete data, err = %s", err)
log.Error(err.Error())
panic(err)
}
// remove delete buf
dn.delBuf.Delete(segmentToFlush)
}
}
}
// process drop collection message, delete node shall notify flush manager all data are cleared and send signal to DataSyncService cleaner
if fgMsg.dropCollection {
dn.flushManager.notifyAllFlushed()
log.Debug("DeleteNode notifies BackgroundGC to release vchannel", zap.String("vChannelName", dn.channelName))
dn.clearSignal <- dn.channelName
}
for _, sp := range spans {
sp.Finish()
}
return in
}
// update delBuf for compacted segments
func (dn *deleteNode) updateCompactedSegments() {
compactedTo2From := dn.channel.listCompactedSegmentIDs()
for compactedTo, compactedFrom := range compactedTo2From {
// if the compactedTo segment has 0 numRows, remove all segments related
if !dn.channel.hasSegment(compactedTo, true) {
for _, segID := range compactedFrom {
dn.delBuf.Delete(segID)
}
dn.channel.removeSegments(compactedFrom...)
continue
}
var compactToDelBuff *DelDataBuf
delBuf, loaded := dn.delBuf.Load(compactedTo)
if !loaded {
compactToDelBuff = newDelDataBuf()
} else {
compactToDelBuff = delBuf.(*DelDataBuf)
}
for _, segID := range compactedFrom {
if value, loaded := dn.delBuf.LoadAndDelete(segID); loaded {
compactToDelBuff.updateFromBuf(value.(*DelDataBuf))
// only store delBuf if EntriesNum > 0
if compactToDelBuff.EntriesNum > 0 {
dn.delBuf.Store(compactedTo, compactToDelBuff)
}
}
}
log.Debug("update delBuf for compacted segments",
zap.Int64("compactedTo segmentID", compactedTo),
zap.Int64s("compactedFrom segmentIDs", compactedFrom),
)
dn.channel.removeSegments(compactedFrom...)
}
}
func (dn *deleteNode) bufferDeleteMsg(msg *msgstream.DeleteMsg, tr TimeRange) ([]UniqueID, error) {
log.Debug("bufferDeleteMsg", zap.Any("primary keys", msg.PrimaryKeys), zap.String("vChannelName", dn.channelName))
primaryKeys := storage.ParseIDs2PrimaryKeys(msg.PrimaryKeys)
segIDToPks, segIDToTss := dn.filterSegmentByPK(msg.PartitionID, primaryKeys, msg.Timestamps)
segIDs := make([]UniqueID, 0, len(segIDToPks))
for segID, pks := range segIDToPks {
segIDs = append(segIDs, segID)
rows := len(pks)
tss, ok := segIDToTss[segID]
if !ok || rows != len(tss) {
return nil, fmt.Errorf("primary keys and timestamp's element num mis-match, segmentID = %d", segID)
}
var delDataBuf *DelDataBuf
value, ok := dn.delBuf.Load(segID)
if ok {
delDataBuf = value.(*DelDataBuf)
} else {
delDataBuf = newDelDataBuf()
}
delData := delDataBuf.delData
for i := 0; i < rows; i++ {
delData.Pks = append(delData.Pks, pks[i])
delData.Tss = append(delData.Tss, tss[i])
// this log will influence the data node performance as it may be too many,
// only use it when we focus on delete operations
//log.Debug("delete",
// zap.Any("primary key", pks[i]),
// zap.Uint64("ts", tss[i]),
// zap.Int64("segmentID", segID),
// zap.String("vChannelName", dn.channelName))
}
// store
delDataBuf.updateSize(int64(rows))
metrics.DataNodeConsumeMsgRowsCount.WithLabelValues(fmt.Sprint(Params.DataNodeCfg.GetNodeID()), metrics.DeleteLabel).Add(float64(rows))
delDataBuf.updateTimeRange(tr)
dn.delBuf.Store(segID, delDataBuf)
}
return segIDs, nil
}
// filterSegmentByPK returns the bloom filter check result.
// If the key may exists in the segment, returns it in map.
// If the key not exists in the segment, the segment is filter out.
func (dn *deleteNode) filterSegmentByPK(partID UniqueID, pks []primaryKey, tss []Timestamp) (map[UniqueID][]primaryKey, map[UniqueID][]uint64) {
segID2Pks := make(map[UniqueID][]primaryKey)
segID2Tss := make(map[UniqueID][]uint64)
buf := make([]byte, 8)
segments := dn.channel.filterSegments(partID)
for index, pk := range pks {
for _, segment := range segments {
segmentID := segment.segmentID
exist := false
switch pk.Type() {
case schemapb.DataType_Int64:
int64Pk := pk.(*int64PrimaryKey)
common.Endian.PutUint64(buf, uint64(int64Pk.Value))
exist = segment.pkStat.pkFilter.Test(buf)
case schemapb.DataType_VarChar:
varCharPk := pk.(*varCharPrimaryKey)
exist = segment.pkStat.pkFilter.TestString(varCharPk.Value)
default:
//TODO::
}
if exist {
segID2Pks[segmentID] = append(segID2Pks[segmentID], pk)
segID2Tss[segmentID] = append(segID2Tss[segmentID], tss[index])
}
}
}
return segID2Pks, segID2Tss
}
func newDeleteNode(ctx context.Context, fm flushManager, sig chan<- string, config *nodeConfig) (*deleteNode, error) {
baseNode := BaseNode{}
baseNode.SetMaxQueueLength(config.maxQueueLength)
baseNode.SetMaxParallelism(config.maxParallelism)
return &deleteNode{
ctx: ctx,
BaseNode: baseNode,
delBuf: sync.Map{},
channel: config.channel,
idAllocator: config.allocator,
channelName: config.vChannelName,
flushManager: fm,
clearSignal: sig,
}, nil
}