milvus/internal/querynode/validate.go
zhenshan.cao 08a4dff41d
Add logic for search merging and a simple task scheduler for read tasks (#17022)
Signed-off-by: zhenshan.cao <zhenshan.cao@zilliz.com>

Co-authored-by: cai.zhang <cai.zhang@zilliz.com>
Co-authored-by: bigsheeper <yihao.dai@zilliz.com>

Co-authored-by: cai.zhang <cai.zhang@zilliz.com>
Co-authored-by: bigsheeper <yihao.dai@zilliz.com>
2022-05-23 16:41:58 +08:00

138 lines
4.6 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 querynode
import (
"errors"
"fmt"
"go.uber.org/zap"
"github.com/milvus-io/milvus/internal/log"
)
func validateOnHistoricalReplica(replica ReplicaInterface, collectionID UniqueID, partitionIDs []UniqueID, segmentIDs []UniqueID) ([]UniqueID, []UniqueID, error) {
var err error
var searchPartIDs []UniqueID
// no partition id specified, get all partition ids in collection
if len(partitionIDs) == 0 {
searchPartIDs, err = replica.getPartitionIDs(collectionID)
if err != nil {
return searchPartIDs, segmentIDs, err
}
} else {
for _, id := range partitionIDs {
_, err = replica.getPartitionByID(id)
if err != nil {
return searchPartIDs, segmentIDs, err
}
searchPartIDs = append(searchPartIDs, id)
}
}
log.Debug("read target partitions", zap.Int64("collectionID", collectionID), zap.Int64s("partitionIDs", searchPartIDs))
col, err2 := replica.getCollectionByID(collectionID)
if err2 != nil {
return searchPartIDs, segmentIDs, err2
}
// all partitions have been released
if len(searchPartIDs) == 0 && col.getLoadType() == loadTypePartition {
return searchPartIDs, segmentIDs, errors.New("partitions have been released , collectionID = " +
fmt.Sprintln(collectionID) + "target partitionIDs = " + fmt.Sprintln(searchPartIDs))
}
if len(searchPartIDs) == 0 && col.getLoadType() == loadTypeCollection {
return searchPartIDs, segmentIDs, nil
}
var newSegmentIDs []UniqueID
if len(segmentIDs) == 0 {
for _, partID := range searchPartIDs {
segIDs, err2 := replica.getSegmentIDs(partID)
if err2 != nil {
return searchPartIDs, newSegmentIDs, err
}
newSegmentIDs = append(segmentIDs, segIDs...)
}
} else {
newSegmentIDs = segmentIDs
for _, segmentID := range newSegmentIDs {
var segment *Segment
if segment, err = replica.getSegmentByID(segmentID); err != nil {
return searchPartIDs, newSegmentIDs, err
}
if !inList(searchPartIDs, segment.partitionID) {
err = fmt.Errorf("segment %d belongs to partition %d, which is not in %v", segmentID, segment.partitionID, searchPartIDs)
return searchPartIDs, newSegmentIDs, err
}
}
}
return searchPartIDs, newSegmentIDs, nil
}
func validateOnStreamReplica(replica ReplicaInterface, collectionID UniqueID, partitionIDs []UniqueID, vChannel Channel) ([]UniqueID, []UniqueID, error) {
var err error
var searchPartIDs []UniqueID
var segmentIDs []UniqueID
// no partition id specified, get all partition ids in collection
if len(partitionIDs) == 0 {
searchPartIDs, err = replica.getPartitionIDs(collectionID)
if err != nil {
return searchPartIDs, segmentIDs, err
}
} else {
for _, id := range partitionIDs {
_, err = replica.getPartitionByID(id)
if err != nil {
return searchPartIDs, segmentIDs, err
}
searchPartIDs = append(searchPartIDs, id)
}
}
log.Debug("read target partitions", zap.Int64("collectionID", collectionID), zap.Int64s("partitionIDs", searchPartIDs))
col, err2 := replica.getCollectionByID(collectionID)
if err2 != nil {
return searchPartIDs, segmentIDs, err2
}
// all partitions have been released
if len(searchPartIDs) == 0 && col.getLoadType() == loadTypePartition {
return searchPartIDs, segmentIDs, errors.New("partitions have been released , collectionID = " +
fmt.Sprintln(collectionID) + "target partitionIDs = " + fmt.Sprintln(searchPartIDs))
}
if len(searchPartIDs) == 0 && col.getLoadType() == loadTypeCollection {
return searchPartIDs, segmentIDs, nil
}
for _, partID := range searchPartIDs {
segIDs, err2 := replica.getSegmentIDsByVChannel(partID, vChannel)
log.Debug("get segmentIDs by vChannel",
zap.Any("collectionID", collectionID),
zap.Any("vChannel", vChannel),
zap.Any("partitionID", partID),
zap.Any("segmentIDs", segIDs))
if err2 != nil {
return searchPartIDs, segmentIDs, err2
}
segmentIDs = append(segmentIDs, segIDs...)
}
return searchPartIDs, segmentIDs, nil
}