mirror of
https://gitee.com/milvus-io/milvus.git
synced 2025-12-06 17:18:35 +08:00
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>
186 lines
6.5 KiB
Go
186 lines
6.5 KiB
Go
package syncmgr
|
|
|
|
import (
|
|
"context"
|
|
|
|
"github.com/cockroachdb/errors"
|
|
"github.com/samber/lo"
|
|
"go.uber.org/zap"
|
|
|
|
"github.com/milvus-io/milvus-proto/go-api/v2/commonpb"
|
|
"github.com/milvus-io/milvus/internal/flushcommon/broker"
|
|
"github.com/milvus-io/milvus/internal/flushcommon/metacache"
|
|
storage "github.com/milvus-io/milvus/internal/storage"
|
|
"github.com/milvus-io/milvus/pkg/v2/log"
|
|
"github.com/milvus-io/milvus/pkg/v2/proto/datapb"
|
|
"github.com/milvus-io/milvus/pkg/v2/util/commonpbutil"
|
|
"github.com/milvus-io/milvus/pkg/v2/util/merr"
|
|
"github.com/milvus-io/milvus/pkg/v2/util/retry"
|
|
)
|
|
|
|
// MetaWriter is the interface for SyncManager to write segment sync meta.
|
|
type MetaWriter interface {
|
|
UpdateSync(context.Context, *SyncTask) error
|
|
DropChannel(context.Context, string) error
|
|
}
|
|
|
|
type brokerMetaWriter struct {
|
|
broker broker.Broker
|
|
opts []retry.Option
|
|
serverID int64
|
|
}
|
|
|
|
func BrokerMetaWriter(broker broker.Broker, serverID int64, opts ...retry.Option) MetaWriter {
|
|
return &brokerMetaWriter{
|
|
broker: broker,
|
|
serverID: serverID,
|
|
opts: opts,
|
|
}
|
|
}
|
|
|
|
func (b *brokerMetaWriter) UpdateSync(ctx context.Context, pack *SyncTask) error {
|
|
checkPoints := []*datapb.CheckPoint{}
|
|
// only current segment checkpoint info
|
|
segment, ok := pack.metacache.GetSegmentByID(pack.segmentID)
|
|
if !ok {
|
|
return merr.WrapErrSegmentNotFound(pack.segmentID)
|
|
}
|
|
|
|
insertFieldBinlogs := append(segment.Binlogs(), storage.SortFieldBinlogs(pack.insertBinlogs)...)
|
|
statsFieldBinlogs := append(segment.Statslogs(), lo.MapToSlice(pack.statsBinlogs, func(_ int64, fieldBinlog *datapb.FieldBinlog) *datapb.FieldBinlog { return fieldBinlog })...)
|
|
|
|
deltaFieldBinlogs := segment.Deltalogs()
|
|
if pack.deltaBinlog != nil && len(pack.deltaBinlog.Binlogs) > 0 {
|
|
deltaFieldBinlogs = append(deltaFieldBinlogs, pack.deltaBinlog)
|
|
}
|
|
|
|
deltaBm25StatsBinlogs := segment.Bm25logs()
|
|
if len(pack.bm25Binlogs) > 0 {
|
|
deltaBm25StatsBinlogs = append(segment.Bm25logs(), lo.MapToSlice(pack.bm25Binlogs, func(_ int64, fieldBinlog *datapb.FieldBinlog) *datapb.FieldBinlog { return fieldBinlog })...)
|
|
}
|
|
|
|
checkPoints = append(checkPoints, &datapb.CheckPoint{
|
|
SegmentID: pack.segmentID,
|
|
NumOfRows: segment.FlushedRows() + pack.batchRows,
|
|
Position: pack.checkpoint,
|
|
})
|
|
|
|
// Get not reported L1's start positions
|
|
startPos := lo.Map(pack.metacache.GetSegmentsBy(
|
|
metacache.WithSegmentState(commonpb.SegmentState_Growing, commonpb.SegmentState_Sealed, commonpb.SegmentState_Flushing),
|
|
metacache.WithLevel(datapb.SegmentLevel_L1), metacache.WithStartPosNotRecorded()),
|
|
func(info *metacache.SegmentInfo, _ int) *datapb.SegmentStartPosition {
|
|
return &datapb.SegmentStartPosition{
|
|
SegmentID: info.SegmentID(),
|
|
StartPosition: info.StartPosition(),
|
|
}
|
|
})
|
|
|
|
// L0 brings its own start position
|
|
if segment.Level() == datapb.SegmentLevel_L0 {
|
|
startPos = append(startPos, &datapb.SegmentStartPosition{SegmentID: pack.segmentID, StartPosition: pack.StartPosition()})
|
|
}
|
|
|
|
getBinlogNum := func(fBinlog *datapb.FieldBinlog) int { return len(fBinlog.GetBinlogs()) }
|
|
log.Info("SaveBinlogPath",
|
|
zap.Int64("SegmentID", pack.segmentID),
|
|
zap.Int64("CollectionID", pack.collectionID),
|
|
zap.Int64("ParitionID", pack.partitionID),
|
|
zap.Any("startPos", startPos),
|
|
zap.Any("checkPoints", checkPoints),
|
|
zap.Int("binlogNum", lo.SumBy(insertFieldBinlogs, getBinlogNum)),
|
|
zap.Int("statslogNum", lo.SumBy(statsFieldBinlogs, getBinlogNum)),
|
|
zap.Int("deltalogNum", lo.SumBy(deltaFieldBinlogs, getBinlogNum)),
|
|
zap.Int("bm25logNum", lo.SumBy(deltaBm25StatsBinlogs, getBinlogNum)),
|
|
zap.String("manifestPath", pack.manifestPath),
|
|
zap.String("vChannelName", pack.channelName),
|
|
)
|
|
|
|
req := &datapb.SaveBinlogPathsRequest{
|
|
Base: commonpbutil.NewMsgBase(
|
|
commonpbutil.WithMsgType(0),
|
|
commonpbutil.WithMsgID(0),
|
|
commonpbutil.WithSourceID(b.serverID),
|
|
),
|
|
SegmentID: pack.segmentID,
|
|
CollectionID: pack.collectionID,
|
|
PartitionID: pack.partitionID,
|
|
Field2BinlogPaths: insertFieldBinlogs,
|
|
Field2StatslogPaths: statsFieldBinlogs,
|
|
Field2Bm25LogPaths: deltaBm25StatsBinlogs,
|
|
Deltalogs: deltaFieldBinlogs,
|
|
|
|
CheckPoints: checkPoints,
|
|
|
|
StartPositions: startPos,
|
|
Flushed: pack.pack.isFlush,
|
|
Dropped: pack.pack.isDrop,
|
|
Channel: pack.channelName,
|
|
SegLevel: pack.level,
|
|
StorageVersion: segment.GetStorageVersion(),
|
|
WithFullBinlogs: true,
|
|
ManifestPath: pack.manifestPath,
|
|
}
|
|
err := retry.Handle(ctx, func() (bool, error) {
|
|
err := b.broker.SaveBinlogPaths(ctx, req)
|
|
// Segment not found during stale segment flush. Segment might get compacted already.
|
|
// Stop retry and still proceed to the end, ignoring this error.
|
|
if !pack.pack.isFlush && errors.Is(err, merr.ErrSegmentNotFound) {
|
|
log.Warn("stale segment not found, could be compacted",
|
|
zap.Int64("segmentID", pack.segmentID))
|
|
log.Warn("failed to SaveBinlogPaths",
|
|
zap.Int64("segmentID", pack.segmentID),
|
|
zap.Error(err))
|
|
return false, nil
|
|
}
|
|
// meta error, datanode handles a virtual channel does not belong here
|
|
if errors.IsAny(err, merr.ErrSegmentNotFound, merr.ErrChannelNotFound) {
|
|
log.Warn("meta error found, skip sync and start to drop virtual channel", zap.String("channel", pack.channelName))
|
|
return false, nil
|
|
}
|
|
|
|
if err != nil {
|
|
return !merr.IsCanceledOrTimeout(err), err
|
|
}
|
|
|
|
return false, nil
|
|
}, b.opts...)
|
|
if err != nil {
|
|
log.Warn("failed to SaveBinlogPaths",
|
|
zap.Int64("segmentID", pack.segmentID),
|
|
zap.Error(err))
|
|
return err
|
|
}
|
|
|
|
pack.metacache.UpdateSegments(metacache.SetStartPosRecorded(true), metacache.WithSegmentIDs(lo.Map(startPos, func(pos *datapb.SegmentStartPosition, _ int) int64 { return pos.GetSegmentID() })...))
|
|
pack.metacache.UpdateSegments(metacache.MergeSegmentAction(
|
|
metacache.UpdateBinlogs(insertFieldBinlogs),
|
|
metacache.UpdateStatslogs(statsFieldBinlogs),
|
|
metacache.UpdateDeltalogs(deltaFieldBinlogs),
|
|
metacache.UpdateBm25logs(deltaBm25StatsBinlogs),
|
|
), metacache.WithSegmentIDs(pack.segmentID))
|
|
return nil
|
|
}
|
|
|
|
func (b *brokerMetaWriter) DropChannel(ctx context.Context, channelName string) error {
|
|
err := retry.Handle(ctx, func() (bool, error) {
|
|
status, err := b.broker.DropVirtualChannel(context.Background(), &datapb.DropVirtualChannelRequest{
|
|
Base: commonpbutil.NewMsgBase(
|
|
commonpbutil.WithSourceID(b.serverID),
|
|
),
|
|
ChannelName: channelName,
|
|
})
|
|
err = merr.CheckRPCCall(status, err)
|
|
if err != nil {
|
|
return !merr.IsCanceledOrTimeout(err), err
|
|
}
|
|
return false, nil
|
|
}, b.opts...)
|
|
if err != nil {
|
|
log.Warn("failed to DropChannel",
|
|
zap.String("channel", channelName),
|
|
zap.Error(err))
|
|
}
|
|
return err
|
|
}
|