milvus/internal/flushcommon/syncmgr/storage_serializer.go
congqixia f94b04e642
feat: [2.6] integrate Loon FFI for manifest-based segment loading and index building (#46076)
Cherry-pick from master
pr: #45061 #45488 #45803 #46017 #44991 #45132 #45723 #45726 #45798
#45897 #45918 #44998

This feature integrates the Storage V2 (Loon) FFI interface as a unified
storage layer for segment loading and index building in Milvus. It
enables
manifest-based data access, replacing the traditional binlog-based
approach
with a more efficient columnar storage format.

Key changes:

### Segment Self-Managed Loading Architecture
- Move segment loading orchestration from Go layer to C++ segcore
- Add NewSegmentWithLoadInfo() API for passing load info during segment
creation
- Implement SetLoadInfo() and Load() methods in SegmentInterface
- Support parallel loading of indexed and non-indexed fields
- Enable both sealed and growing segments to self-manage loading

### Storage V2 FFI Integration
- Integrate milvus-storage library's FFI interface for packed columnar
data
- Add manifest path support throughout the data path (SegmentInfo,
LoadInfo)
- Implement ManifestReader for generating manifests from binlogs
- Support zero-copy data exchange using Arrow C Data Interface
- Add ToCStorageConfig() for Go-to-C storage config conversion

### Manifest-Based Index Building
- Extend FileManagerContext to carry loon_ffi_properties
- Implement GetFieldDatasFromManifest() using Arrow C Stream interface
- Support manifest-based reading in DiskFileManagerImpl and
MemFileManagerImpl
- Add fallback to traditional segment insert files when manifest
unavailable

### Compaction Pipeline Updates
- Include manifest path in all compaction task builders (clustering, L0,
mix)
- Update BulkPackWriterV2 to return manifest path
- Propagate manifest metadata through compaction pipeline

### Configuration & Protocol
- Add common.storageV2.useLoonFFI config option (default: false)
- Add manifest_path field to SegmentLoadInfo and related proto messages
- Add manifest field to compaction segment messages

### Bug Fixes
- Fix mmap settings not applied during segment load (key typo fix)
- Populate index info after segment loading to prevent redundant load
tasks
- Fix memory corruption by removing premature transaction handle
destruction

Related issues: #44956, #45060, #39173

## Individual Cherry-Picked Commits

1. **e1c923b5cc** - fix: apply mmap settings correctly during segment
load (#46017)
2. **63b912370b** - enhance: use milvus-storage internal C++ Reader API
for Loon FFI (#45897)
3. **bfc192faa5** - enhance: Resolve issues integrating loon FFI
(#45918)
4. **fb18564631** - enhance: support manifest-based index building with
Loon FFI reader (#45726)
5. **b9ec2392b9** - enhance: integrate StorageV2 FFI interface for
manifest-based segment loading (#45798)
6. **66db3c32e6** - enhance: integrate Storage V2 FFI interface for
unified storage access (#45723)
7. **ae789273ac** - fix: populate index info after segment loading to
prevent redundant load tasks (#45803)
8. **49688b0be2** - enhance: Move segment loading logic from Go layer to
segcore for self-managed loading (#45488)
9. **5b2df88bac** - enhance: [StorageV2] Integrate FFI interface for
packed reader (#45132)
10. **91ff5706ac** - enhance: [StorageV2] add manifest path support for
FFI integration (#44991)
11. **2192bb4a85** - enhance: add NewSegmentWithLoadInfo API to support
segment self-managed loading (#45061)
12. **4296b01da0** - enhance: update delta log serialization APIs to
integrate storage V2 (#44998)

## Technical Details

### Architecture Changes
- **Before**: Go layer orchestrated segment loading, making multiple CGO
calls
- **After**: Segments autonomously manage loading in C++ layer with
single entry point

### Storage Access Pattern
- **Before**: Read individual binlog files through Go storage layer
- **After**: Read manifest file that references packed columnar data via
FFI

### Benefits
- Reduced cross-language call overhead
- Better resource management at C++ level
- Improved I/O performance through batched streaming reads
- Cleaner separation of concerns between Go and C++ layers
- Foundation for proactive schema evolution handling

---------

Signed-off-by: Ted Xu <ted.xu@zilliz.com>
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Co-authored-by: Ted Xu <ted.xu@zilliz.com>
2025-12-04 17:09:12 +08:00

191 lines
5.5 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 syncmgr
import (
"context"
"strconv"
"github.com/samber/lo"
"go.uber.org/zap"
"github.com/milvus-io/milvus-proto/go-api/v2/schemapb"
"github.com/milvus-io/milvus/internal/flushcommon/metacache"
"github.com/milvus-io/milvus/internal/storage"
"github.com/milvus-io/milvus/pkg/v2/log"
"github.com/milvus-io/milvus/pkg/v2/proto/etcdpb"
"github.com/milvus-io/milvus/pkg/v2/util/merr"
)
type storageV1Serializer struct {
schema *schemapb.CollectionSchema
pkField *schemapb.FieldSchema
inCodec *storage.InsertCodec
metacache metacache.MetaCache
}
func NewStorageSerializer(metacache metacache.MetaCache, schema *schemapb.CollectionSchema) (*storageV1Serializer, error) {
pkField := lo.FindOrElse(schema.GetFields(), nil, func(field *schemapb.FieldSchema) bool { return field.GetIsPrimaryKey() })
if pkField == nil {
return nil, merr.WrapErrServiceInternal("cannot find pk field")
}
meta := &etcdpb.CollectionMeta{
ID: metacache.Collection(),
Schema: schema,
}
inCodec := storage.NewInsertCodecWithSchema(meta)
return &storageV1Serializer{
schema: schema,
pkField: pkField,
inCodec: inCodec,
metacache: metacache,
}, nil
}
func (s *storageV1Serializer) serializeBinlog(ctx context.Context, pack *SyncPack) (map[int64]*storage.Blob, error) {
if len(pack.insertData) == 0 {
return make(map[int64]*storage.Blob), nil
}
log := log.Ctx(ctx)
blobs, err := s.inCodec.Serialize(pack.partitionID, pack.segmentID, pack.insertData...)
if err != nil {
return nil, err
}
result := make(map[int64]*storage.Blob)
for _, blob := range blobs {
fieldID, err := strconv.ParseInt(blob.GetKey(), 10, 64)
if err != nil {
log.Error("serialize buffer failed ... cannot parse string to fieldID ..", zap.Error(err))
return nil, err
}
result[fieldID] = blob
}
return result, nil
}
func (s *storageV1Serializer) serializeBM25Stats(pack *SyncPack) (map[int64]*storage.Blob, error) {
if len(pack.bm25Stats) == 0 {
return make(map[int64]*storage.Blob), nil
}
blobs := make(map[int64]*storage.Blob)
for fieldID, stats := range pack.bm25Stats {
bytes, err := stats.Serialize()
if err != nil {
return nil, err
}
blobs[fieldID] = &storage.Blob{
Value: bytes,
MemorySize: int64(len(bytes)),
RowNum: stats.NumRow(),
}
}
return blobs, nil
}
func (s *storageV1Serializer) serializeStatslog(pack *SyncPack) (*storage.PrimaryKeyStats, *storage.Blob, error) {
if len(pack.insertData) == 0 {
return nil, nil, nil
}
var rowNum int64
var pkFieldData []storage.FieldData
for _, chunk := range pack.insertData {
chunkPKData := chunk.Data[s.pkField.GetFieldID()]
pkFieldData = append(pkFieldData, chunkPKData)
rowNum += int64(chunkPKData.RowNum())
}
stats, err := storage.NewPrimaryKeyStats(s.pkField.GetFieldID(), int64(s.pkField.GetDataType()), rowNum)
if err != nil {
return nil, nil, err
}
for _, chunkPkData := range pkFieldData {
stats.UpdateByMsgs(chunkPkData)
}
blob, err := s.inCodec.SerializePkStats(stats, pack.batchRows)
if err != nil {
return nil, nil, err
}
return stats, blob, nil
}
func (s *storageV1Serializer) serializeMergedPkStats(pack *SyncPack) (*storage.Blob, error) {
segment, ok := s.metacache.GetSegmentByID(pack.segmentID)
if !ok {
return nil, merr.WrapErrSegmentNotFound(pack.segmentID)
}
// Allow to flush empty segment to make streaming service easier to implement rollback transaction.
stats := lo.Map(segment.GetHistory(), func(pks *storage.PkStatistics, _ int) *storage.PrimaryKeyStats {
return &storage.PrimaryKeyStats{
FieldID: s.pkField.GetFieldID(),
MaxPk: pks.MaxPK,
MinPk: pks.MinPK,
BFType: pks.PkFilter.Type(),
BF: pks.PkFilter,
PkType: int64(s.pkField.GetDataType()),
}
})
if len(stats) == 0 {
return nil, nil
}
return s.inCodec.SerializePkStatsList(stats, segment.NumOfRows())
}
func (s *storageV1Serializer) serializeMergedBM25Stats(pack *SyncPack) (map[int64]*storage.Blob, error) {
segment, ok := s.metacache.GetSegmentByID(pack.segmentID)
if !ok {
return nil, merr.WrapErrSegmentNotFound(pack.segmentID)
}
stats := segment.GetBM25Stats()
// Allow to flush empty segment to make streaming service easier to implement rollback transaction.
if stats == nil {
return nil, nil
}
fieldBytes, numRow, err := stats.Serialize()
if err != nil {
return nil, err
}
blobs := make(map[int64]*storage.Blob)
for fieldID, bytes := range fieldBytes {
blobs[fieldID] = &storage.Blob{
Value: bytes,
MemorySize: int64(len(bytes)),
RowNum: numRow[fieldID],
}
}
return blobs, nil
}
func hasBM25Function(schema *schemapb.CollectionSchema) bool {
for _, function := range schema.GetFunctions() {
if function.GetType() == schemapb.FunctionType_BM25 {
return true
}
}
return false
}