add sparse float vector support to different milvus components,
including proxy, data node to receive and write sparse float vectors to
binlog, query node to handle search requests, index node to build index
for sparse float column, etc.
https://github.com/milvus-io/milvus/issues/29419
---------
Signed-off-by: Buqian Zheng <zhengbuqian@gmail.com>
1. The Import APIs now provide detailed progress information for each
imported file, including details such as file name, file size, progress,
and more.
2. The APIs now return the collection name and the completion time.
3. Other modifications include changing jobID to jobId and other similar
adjustments.
issue: https://github.com/milvus-io/milvus/issues/28521
---------
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
See also #31143
This PR add short cut for datanoe metacache `WithSegmentIDs` filter,
which could just fetch segment from map with provided segmentIDs. Also
add benchmark for new implementation vs old one.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
This PR includes the following adjustments:
1. To prevent channelCP update task backlog, only one task with the same
vchannel is retained in the updater. Additionally, the lastUpdateTime is
refreshed after the flowgraph submits the update task, rather than in
the callBack function.
2. Batch updates of multiple vchannel checkpoints are performed in the
UpdateChannelCheckpoint RPC (default batch size is 128). Additionally,
the lock for channelCPs in DataCoord meta has been switched from key
lock to global lock.
3. The concurrency of UpdateChannelCheckpoint RPCs in the datanode has
been reduced from 1000 to 10.
issue: https://github.com/milvus-io/milvus/issues/30004
---------
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
Co-authored-by: jaime <yun.zhang@zilliz.com>
Co-authored-by: congqixia <congqi.xia@zilliz.com>
In the cache of the timeTickSender, retain only the latest stats instead
of storing stats for every time tick.
issue: https://github.com/milvus-io/milvus/issues/30967
---------
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
This PR introduces novel managerial roles for importv2:
1. ImportMeta: To manage all the import tasks;
2. ImportScheduler: To process tasks and modify their states;
3. ImportChecker: To ascertain the completion of all tasks and instigate
relevant operations.
issue: https://github.com/milvus-io/milvus/issues/28521
---------
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
If DC restarted, those unkonwn compaction tasks
will never get call back in DN, so that the segments in the compaction
task will be locked, unable to sync and compaction again, blocking cp
advance and compaction executing.
See also: #30137
---------
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
1. add coordinator graceful stop timeout to 5s
2. change the order of datacoord component while stop
3. change querynode grace stop timeout to 900s, and we should
potentially change this to 600s when graceful stop is smooth
issue: #30310
also see pr: #30306
---------
Signed-off-by: chyezh <chyezh@outlook.com>
See also #30571
When `compactionExecutor` stops one compaction task, the `stop` method
will case `injectDone` called.
However in `executeTask` when `compact` method returns error, it shall
also invoke `injectDone` as well. That the reason `Unlock of unlocked
RWMutex` panicking happened.
This PR add sync.Once to make sure that `injectDone` is called only
once. We did not remove any of the `injectDone` since removal any of
those invocation may cause logic problem.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675#30469
For a sync task, the segment could be compacted during sync task. In
previous implementation, this sync task will hold only the old segment
id as KeyLock, in which case compaction on compacted to segment may run
in parallel with delta sync of this sync task.
This PR introduces sync target segment verification logic. It shall
check target segment lock it's holding beforing actually syncing logic.
If this check failed, sync task shall return`errTargetSegementNotMatch`
error and make manager re-fetch the current target segment id.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
This PR changes the following to speed up L0 compaction and
prevent OOM:
1. Lower deltabuf limit to 16MB by default, so that each L0 segment
would be 4X smaller than before.
2. Add BatchProcess, use it if memory is sufficient
3. Iterator will Deserialize when called HasNext to avoid massive memory
peek
4. Add tracing in spiltDelta
See also: #30191
---------
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
This pr decoups importing segment from flush process by:
1. Exclude the importing segment from the flush policy, this approch
avoids notifying the datanode to flush the importing segment, which may
not exist.
2. When RootCoord call Flush, DataCoord directly set the importing
segment state to `Flushed`.
issue: https://github.com/milvus-io/milvus/issues/30359
---------
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
syncMgr.Block() will lock the segment when executing compaction.
Previous implementation was unable to Unblock thoese segments when
compaction failed. If next compaction of the same segments arrives,
it'll stuck forever and block all later compation tasks.
This PR makes sure compaction executor would Unblock these segments
after a failure compaction.
Apart form that, this PR also refines some logs and clean some codes of
compaction, compactor:
1. Log segment count instead of segmentIDs to avoid logging too many
segments
2. Flush RPC returns L1 segments only, skip L0 and L2
3. CompactionType is checked in `Compaction`, no need to check again
inside compactor
4. Use ligter method to replace `getSegmentMeta`
5. Log information for L0 compaction when encounters an error
See also: #30213
---------
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
See also #27606
Previously l0 linear compaction will scan all target segment id from
metacache for each line of delta entry, which is not needed since
compaction target segments shall be all immutable.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
This PR introduces novel importv2 roles for datanode:
1. Executor: To execute tasks, a import task will be divided into the
following steps: read data -> hash data -> sync data;
2. Manager: To manage all the tasks;
issue: https://github.com/milvus-io/milvus/issues/28521
---------
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
See also #27675
This PR adds back MemoryHighSyncPolicy implementation. Also change
MinSegmentSize & CheckInterval to configurable param item.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675
BloomFilterSet.current shall be reset after RollStats, otherwise it will
keep tracking whole segment data causing the false positive ratio larger
than expected.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #30167
After support open telemetry tracing, we want to have traceID as well,
this PR adds util functions to set traceID with span & propagate traceID
between different context.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27606
`MultiRead` actually download file in sequence, which may lead to large
time consumption during l0 compaction download phase.
This PR make l0 compactor download deltalogs in parallel utilizing conc
package & io pool.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #30111
Segments could be "Flushed" only by `FlushSegments` grpc call from
datacoord by design. There are two possible reason to cause one segment
got flushed multiple times.
- Segment is in flushing state during multiple epoch in flowgraph
- Segment is flushed by flushTs & Flush segments
So this pr fix:
- Remove state change logic form FlushTs policy
- Change Flush segment into three stage way: Sealed->Flushing->Flushed
preventing multiple Flushed=true operations.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Resolves#30167
This PR add tracing for all compaction from the task start in datacoord
and execution procedures in datanode.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
This PR discontinuing the subscription to the mq and, instead, employing
the channel checkpoint as the DML and starting position for the import
segments.
issue: https://github.com/milvus-io/milvus/issues/30106
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
See also: #30121#27675
This PR changes the delete buffering logic:
- Write buffer shall buffer insert first
- Then the delete messages shall be evaluated
- Whether PK matches previous Bloom filter, which ts is always smaller
- Whether PK matches insert data which has smaller timestamp
- Then the segment bloom filter is updates by the newly buffered pk rows
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
This PR also fixes bugs in l0 compactor where
l0 results would never be removed from datanode
See also: #30099
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
don't store logPath in meta to reduce memory, when service get
segmentinfo, generate logpath from logid.
#28885
Signed-off-by: lixinguo <xinguo.li@zilliz.com>
Co-authored-by: lixinguo <xinguo.li@zilliz.com>
See also: #29657
Datanode Compactor use estimated row number from schema to decide when
to sync the batch of data when executing compaction. This est value
could go way from actual size when the schema contains variable field(
say VarChar, JSON, etc.)
This PR make compactor able to check the actual buffer data size and
make it possible to sync when buffer is actually beyong max binglog
size.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also: #27675
The bloom filter set initialized new BF with fixed configured `n`. This
value is always larger than the actual batch size and causes generated
BF using more memory.
This PR make write buffer to initialize BF with estimated batch size
from schema & configuration value.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
fix: #29757
In previous code, `ColumnBasedInsertMsgToInsertData` adds empty field if
the insertMsg parameter does not have the column schema defined. This
may lead to unexpected behavior of caller functions.
This PR:
- Add column missing check
- Add column length check
- Generate BlobInfo for ColumnBasedInsertMsgToInsertData result
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675
`Allocator.Alloc` and `Allocator.AllocOne` might be invoked multiple
times if there were multiple blobs set in one sync task.
This PR add pre-fetch logic for all blobs and cache logIDs in sync task
so that at most only one call of the allocator is needed.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675
Fix logic problem introduced by #29413, which is serializer tries to
merge statslog list while level segments do not have statslog. This
shall result returning error. `writeBufferBase` ignores this error but
it shall only ignore `ErrSegmentNotFound`.
This PR add logic checking segment level before execution of merging
statslog list. And add error type check for getSyncTask failure.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675
For now, Level zero segments shall always be synced as `Flushed` ones.
This PR fixes when level zero segments selected by policies other than
flush ts policy will be synced as growing state.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
the array type can't be compacted, the system could continue with the
inserted segments, but these segments can be never compacted
fix#29503
---------
Signed-off-by: yah01 <yah2er0ne@outlook.com>
See also #27675
Since serialization segment buffer does not related to sync manager can
shall be done before submit into sync manager. So that the pk statistic
file could be more accurate and reduce complex logic inside sync
manager.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #29092
`FlushSegments` transfer only `Growing` segment to flushing, if the
segment is in `Sealed` state before Datanode watch channel, the state
will never got satisfied for a segment be selected to be flushed.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
related to #28427
Add a jitter in syncStatleBuffer policy so all segments won't flush at
the same time
Signed-off-by: xiaofanluan <xiaofan.luan@zilliz.com>
See also #29223
This PR make `conc.Pool` resizable by adding `Resize` method for it.
Also make newly added datanode `MaxParallelSyncMgrTasks` config
refreshable
---------
Signed-off-by: Congqi.Xia <congqi.xia@zilliz.com>
Since the sync manager is global in datanode now, the old
`maxParallelSyncTaskNum` does not fit into current implementation
anymore.
This PR add a new param item for sync mgr parallel control and enlarge
default value
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Related to #27675
The timestamp from, to field is not field for new implementation of
writebuffer & sync manager
This pr fills these field for better log information
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #28924
The compaction task generated before datanode finish `SaveBinlogPath`
grpc call contains segments which are still in Growing state DataNode
shall verify each non-levelzero segments before submit compaction task
to executor
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
- Change flowgraphManager to fgManagerImpl
- Change close to stop
- change execute to controlMemWaterLevel
- Change method name of fgManager for readability
- Add mockery for fgmanager
Issue: #28853
---------
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
Related to #28628
Compacted segment syncing counter is not set correctly in sync task and
the bf write buffer shall not use compacted segment as candidate when
buffering delta data
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
This PR:
- Separates compaction scheduler and check results loop So that slow in
check-loop doesn't influence execution.
- Cleans compaction tasks when drop a vchannel so dropped-channel's
compaction tasks won't be checked over and over again.
- Skips meta change when meta's already changed, avoid panic
- Remove not inuse injectDone(bool) parameter
See also: #28628, #28209
---------
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
Related to #28736#28748
See also #27675
Previous PR: #28646
This PR fixes `SegmentNotFound` issue when compaction happens multiple
times and the buffer of first generation segment is sync due to stale
policy
Now the `CompactSegments` API of metacache shall update the compactTo
field of segmentInfo if the compactTo segment is also compacted to keep
the bloodline clean
Also, add the `CompactedSegment` SyncPolicy to sync the compacted
segment asap to keep metacache clean
Now the `SyncPolicy` is an interface instead of a function type so that
when it selects some segments to sync, we colud log the reason and
target segment
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675
When L0 segment contains only delta data, merged statslog shall be
skiped when performing sync task
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675
Compacted segment info shall be removed after all buffer belongs to it
is sync-ed.
This PR add the cleanup function after triggerSyncTask logic:
- The buffer is stable and protected by mutex
- Cleanup fetches compacted & non-sync segment
- Remove segment info only there is no buffered maintained in manager
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
remove some unnecessary assignments, for the reason that
commonpbutil.NewMsgBase has default value.
Signed-off-by: lixinguo <xinguo.li@zilliz.com>
Co-authored-by: lixinguo <xinguo.li@zilliz.com>
See also #28660
This pr add request timeout config item for etcd kv request timeout
Sync the default timeout value to same value for etcdKV & tikv config
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #28628
Previous compaction task blocked the segment sync task and may block the
flowgraph when sync task is generated by auto sync policy This
`BlockAll` call will block forever and cause whole fg stuck
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
create goroutine only once when getOrCreateMergedTimeTickerSender
Signed-off-by: lixinguo <xinguo.li@zilliz.com>
Co-authored-by: lixinguo <xinguo.li@zilliz.com>
fix#28629
orignal unit test close channel before setting the segment id, so there
is a chance that test read segment id before setting it change unit test
behavior to wait future return now
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675, comment from #27874
This PR changes the `IsFull` logic of insert buffer
Changing the limit from rowNum to buffer size,
this shall help form better binlog file when schema has variable-length
field
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #28575
Add zero-length check for `storage.NewPrimaryKeyStats`. This function
shall return error when non-positive rowNum passed.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675
- Fix LevelZero segment cannot be flushed
- Add level option for syncTask
- Invoke `AddSegment` when new LevelZero segment is allocated
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675
Since `MetaWriter` need `*datapb.FieldBinlog` struct, sync task now
generate FieldBinlog directly
Also fix merged statslog not generated if last task has no insert
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
This PR removes all the commented code and files from PR #28320
For naming issue:
- Renaming `MinCheckpoint` to `EarliestPosition`, see #28320 comment
- Renaming `writebuffer.Mananger` to `BufferMananger`, see #27874
comment
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #27675
This PR make previously merged refactory of datanode go online
- Use write node to replace insert/delete node
- Use write buffer manager to control all buffers
- Use sync manager to control sync tasks instead of flush manager
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Compaction plan result contained one segment for one plan. For l0
compaction would write to multiple segments, this PR expand the segments
number in plan results and refactor some names for readibility.
- Name refactory: - CompactionStateResult -> CompactionPlanResult -
CompactionResult -> CompactionSegment
See also: #27606
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
- Use explicit lifetime control methods: `Start` and `Stop`
- Allow control retry option
- Make sure tt sender worker exit after `Stop` return
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Passing initCtx to all IO funcs in newDataSyncService,
so when ctx.Canceled, newDataSyncService would return.
See also: #25309
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
GetAndRemove removes the fg from manager immediately,
while the flowgraph is still releasing. This PR will remove
the fg from flowgraphManager AFTER flowgraphs released.
- Add Remove for ConcurrentMap
- Move collections() into flowgraph manager
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
Benchmark Milvus with https://github.com/qdrant/vector-db-benchmark and
specify the datasets as 'deep-image-96-angular'. Meanwhile, do perf
profiling during 'upload + index' stage of vector-db-benchmark and see
the following hot spots.
39.59%--github.com/milvus-io/milvus/internal/storage.MergeInsertData
|
|--21.43%--github.com/milvus-io/milvus/internal/storage.MergeFieldData
| |
| |--17.22%--runtime.memmove
| |
| |--1.53%--asm_exc_page_fault
| ......
|
|--18.16%--runtime.memmove
|
|--1.66%--asm_exc_page_fault
......
The hot code path is in storage.MergeInsertData() which updates
buffer.buffer by creating a new 'InsertData' instance and merging both
the old buffer.buffer and addedBuffer into it. When it calls golang
runtime.memmove to move buffer.buffer which is with big size (>1M), the
hot spots appear.
To avoid the above overhead, update storage.MergeInsertData() by
appending addedBuffer to buffer.buffer, instead of moving buffer.buffer
and addedBuffer to a new 'InsertData'. This change removes the hot spots
'runtime.memmove' from perf profiling output. Additionally, the 'upload
+ index' time, which is one performance metric of vector-db-benchmark,
is reduced around 60% with this change.
Signed-off-by: Cathy Zhang <cathy.zhang@intel.com>
Support Database(#23742)
Fix db nonexists error for FlushAll (#24222)
Fix check collection limits fails (#24235)
backward compatibility with empty DB name (#24317)
Fix GetFlushAllState with DB (#24347)
Remove db from global meta cache after drop database (#24474)
Fix db name is empty for describe collection response (#24603)
Add RBAC for Database API (#24653)
Fix miss load the same name collection during recover stage (#24941)
RBAC supports Database validation (#23609)
Fix to list grant with db return empty (#23922)
Optimize PrivilegeAll permission check (#23972)
Add the default db value for the rbac request (#24307)
Signed-off-by: jaime <yun.zhang@zilliz.com>
Co-authored-by: SimFG <bang.fu@zilliz.com>
Co-authored-by: longjiquan <jiquan.long@zilliz.com>
1. balance granuity to replica to avoid influence unrelated replicas
2. avoid balance back and forth
Signed-off-by: MrPresent-Han <jamesharden11122@gmail.com>
congqixia
yihao.dai
jaime
XuanYang-cn
Buqian Zheng
Ted Xu
wayblink
chyezh
yiwangdr
smellthemoon
SimFG
Bingyi Sun
Xu Tong
MrPresent-Han
yah01
Xiaofan
groot
yah01
Jiquan Long
Filip Haltmayer
aoiasd
bjzhjing
Enwei Jiao
cai.zhang
wei liu
xige-16
PowderLi
zhenshan.cao