milvus/internal/datacoord/import_scheduler_test.go
yihao.dai 4e264003bf
enhance: Ensure ImportV2 waits for the index to be built and refine some logic (#31629)
Feature Introduced:
1. Ensure ImportV2 waits for the index to be built

Enhancements Introduced:
1. Utilization of local time for timeout ts instead of allocating ts
from rootcoord.
3. Enhanced input file length check for binlog import.
4. Removal of duplicated manager in datanode.
5. Renaming of executor to scheduler in datanode.
6. Utilization of a thread pool in the scheduler in datanode.

issue: https://github.com/milvus-io/milvus/issues/28521

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2024-04-01 20:09:13 +08:00

276 lines
8.2 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 datacoord
import (
"context"
"math"
"testing"
"github.com/samber/lo"
"github.com/stretchr/testify/mock"
"github.com/stretchr/testify/suite"
"github.com/milvus-io/milvus-proto/go-api/v2/schemapb"
"github.com/milvus-io/milvus/internal/metastore/mocks"
"github.com/milvus-io/milvus/internal/proto/datapb"
)
type ImportSchedulerSuite struct {
suite.Suite
collectionID int64
catalog *mocks.DataCoordCatalog
alloc *NMockAllocator
cluster *MockCluster
meta *meta
imeta ImportMeta
scheduler *importScheduler
}
func (s *ImportSchedulerSuite) SetupTest() {
var err error
s.collectionID = 1
s.catalog = mocks.NewDataCoordCatalog(s.T())
s.catalog.EXPECT().ListImportJobs().Return(nil, nil)
s.catalog.EXPECT().ListPreImportTasks().Return(nil, nil)
s.catalog.EXPECT().ListImportTasks().Return(nil, nil)
s.catalog.EXPECT().ListSegments(mock.Anything).Return(nil, nil)
s.catalog.EXPECT().ListChannelCheckpoint(mock.Anything).Return(nil, nil)
s.catalog.EXPECT().ListIndexes(mock.Anything).Return(nil, nil)
s.catalog.EXPECT().ListSegmentIndexes(mock.Anything).Return(nil, nil)
s.cluster = NewMockCluster(s.T())
s.alloc = NewNMockAllocator(s.T())
s.meta, err = newMeta(context.TODO(), s.catalog, nil)
s.NoError(err)
s.meta.AddCollection(&collectionInfo{
ID: s.collectionID,
Schema: newTestSchema(),
})
s.imeta, err = NewImportMeta(s.catalog)
s.NoError(err)
buildIndexCh := make(chan UniqueID, 1024)
s.scheduler = NewImportScheduler(s.meta, s.cluster, s.alloc, s.imeta, buildIndexCh).(*importScheduler)
}
func (s *ImportSchedulerSuite) TestProcessPreImport() {
s.catalog.EXPECT().SaveImportJob(mock.Anything).Return(nil)
s.catalog.EXPECT().SavePreImportTask(mock.Anything).Return(nil)
var task ImportTask = &preImportTask{
PreImportTask: &datapb.PreImportTask{
JobID: 0,
TaskID: 1,
CollectionID: s.collectionID,
State: datapb.ImportTaskStateV2_Pending,
},
}
err := s.imeta.AddTask(task)
s.NoError(err)
var job ImportJob = &importJob{
ImportJob: &datapb.ImportJob{
JobID: 0,
CollectionID: s.collectionID,
TimeoutTs: math.MaxUint64,
Schema: &schemapb.CollectionSchema{},
},
}
err = s.imeta.AddJob(job)
s.NoError(err)
// pending -> inProgress
const nodeID = 10
s.cluster.EXPECT().QueryImport(mock.Anything, mock.Anything).Return(&datapb.QueryImportResponse{
Slots: 1,
}, nil)
s.cluster.EXPECT().PreImport(mock.Anything, mock.Anything).Return(nil)
s.cluster.EXPECT().GetSessions().Return([]*Session{
{
info: &NodeInfo{
NodeID: nodeID,
},
},
})
s.scheduler.process()
task = s.imeta.GetTask(task.GetTaskID())
s.Equal(datapb.ImportTaskStateV2_InProgress, task.GetState())
s.Equal(int64(nodeID), task.GetNodeID())
// inProgress -> completed
s.cluster.EXPECT().QueryPreImport(mock.Anything, mock.Anything).Return(&datapb.QueryPreImportResponse{
State: datapb.ImportTaskStateV2_Completed,
}, nil)
s.scheduler.process()
task = s.imeta.GetTask(task.GetTaskID())
s.Equal(datapb.ImportTaskStateV2_Completed, task.GetState())
// drop import task
s.cluster.EXPECT().DropImport(mock.Anything, mock.Anything).Return(nil)
s.scheduler.process()
task = s.imeta.GetTask(task.GetTaskID())
s.Equal(int64(NullNodeID), task.GetNodeID())
}
func (s *ImportSchedulerSuite) TestProcessImport() {
s.catalog.EXPECT().SaveImportJob(mock.Anything).Return(nil)
s.catalog.EXPECT().SaveImportTask(mock.Anything).Return(nil)
var task ImportTask = &importTask{
ImportTaskV2: &datapb.ImportTaskV2{
JobID: 0,
TaskID: 1,
CollectionID: s.collectionID,
State: datapb.ImportTaskStateV2_Pending,
FileStats: []*datapb.ImportFileStats{
{
HashedStats: map[string]*datapb.PartitionImportStats{
"channel1": {
PartitionRows: map[int64]int64{
int64(2): 100,
},
PartitionDataSize: map[int64]int64{
int64(2): 100,
},
},
},
},
},
},
}
err := s.imeta.AddTask(task)
s.NoError(err)
var job ImportJob = &importJob{
ImportJob: &datapb.ImportJob{
JobID: 0,
CollectionID: s.collectionID,
PartitionIDs: []int64{2},
Vchannels: []string{"channel1"},
Schema: &schemapb.CollectionSchema{},
TimeoutTs: math.MaxUint64,
},
}
err = s.imeta.AddJob(job)
s.NoError(err)
// pending -> inProgress
const nodeID = 10
s.alloc.EXPECT().allocN(mock.Anything).Return(100, 200, nil)
s.alloc.EXPECT().allocTimestamp(mock.Anything).Return(300, nil)
s.cluster.EXPECT().QueryImport(mock.Anything, mock.Anything).Return(&datapb.QueryImportResponse{
Slots: 1,
}, nil)
s.cluster.EXPECT().ImportV2(mock.Anything, mock.Anything).Return(nil)
s.cluster.EXPECT().GetSessions().Return([]*Session{
{
info: &NodeInfo{
NodeID: nodeID,
},
},
})
s.scheduler.process()
task = s.imeta.GetTask(task.GetTaskID())
s.Equal(datapb.ImportTaskStateV2_InProgress, task.GetState())
s.Equal(int64(nodeID), task.GetNodeID())
// inProgress -> completed
s.cluster.ExpectedCalls = lo.Filter(s.cluster.ExpectedCalls, func(call *mock.Call, _ int) bool {
return call.Method != "QueryImport"
})
s.cluster.EXPECT().QueryImport(mock.Anything, mock.Anything).Return(&datapb.QueryImportResponse{
State: datapb.ImportTaskStateV2_Completed,
}, nil)
s.scheduler.process()
task = s.imeta.GetTask(task.GetTaskID())
s.Equal(datapb.ImportTaskStateV2_Completed, task.GetState())
// drop import task
s.cluster.EXPECT().DropImport(mock.Anything, mock.Anything).Return(nil)
s.scheduler.process()
task = s.imeta.GetTask(task.GetTaskID())
s.Equal(int64(NullNodeID), task.GetNodeID())
}
func (s *ImportSchedulerSuite) TestProcessFailed() {
s.catalog.EXPECT().SaveImportJob(mock.Anything).Return(nil)
s.catalog.EXPECT().SaveImportTask(mock.Anything).Return(nil)
var task ImportTask = &importTask{
ImportTaskV2: &datapb.ImportTaskV2{
JobID: 0,
TaskID: 1,
CollectionID: s.collectionID,
NodeID: 6,
SegmentIDs: []int64{2, 3},
State: datapb.ImportTaskStateV2_Failed,
},
}
err := s.imeta.AddTask(task)
s.NoError(err)
var job ImportJob = &importJob{
ImportJob: &datapb.ImportJob{
JobID: 0,
CollectionID: s.collectionID,
PartitionIDs: []int64{2},
Vchannels: []string{"channel1"},
Schema: &schemapb.CollectionSchema{},
TimeoutTs: math.MaxUint64,
},
}
err = s.imeta.AddJob(job)
s.NoError(err)
s.catalog.EXPECT().AddSegment(mock.Anything, mock.Anything).Return(nil)
s.cluster.EXPECT().QueryImport(mock.Anything, mock.Anything).Return(&datapb.QueryImportResponse{
Slots: 1,
}, nil)
s.cluster.EXPECT().GetSessions().Return([]*Session{
{
info: &NodeInfo{
NodeID: 6,
},
},
})
for _, id := range task.(*importTask).GetSegmentIDs() {
segment := &SegmentInfo{
SegmentInfo: &datapb.SegmentInfo{ID: id, IsImporting: true},
}
err = s.meta.AddSegment(context.Background(), segment)
s.NoError(err)
}
for _, id := range task.(*importTask).GetSegmentIDs() {
segment := s.meta.GetSegment(id)
s.NotNil(segment)
}
s.cluster.EXPECT().DropImport(mock.Anything, mock.Anything).Return(nil)
s.catalog.EXPECT().DropSegment(mock.Anything, mock.Anything).Return(nil)
s.scheduler.process()
for _, id := range task.(*importTask).GetSegmentIDs() {
segment := s.meta.GetSegment(id)
s.Nil(segment)
}
task = s.imeta.GetTask(task.GetTaskID())
s.Equal(datapb.ImportTaskStateV2_Failed, task.GetState())
s.Equal(0, len(task.(*importTask).GetSegmentIDs()))
s.Equal(int64(NullNodeID), task.GetNodeID())
}
func TestImportScheduler(t *testing.T) {
suite.Run(t, new(ImportSchedulerSuite))
}