Commit Graph

795 Commits

Author SHA1 Message Date
congqixia
1d76565894
Add metrics for garbage collection (#27303)
Also fix second metrics usage in compaction

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-09-22 18:47:25 +08:00
SimFG
26f06dd732
Format the code (#27275)
Signed-off-by: SimFG <bang.fu@zilliz.com>
2023-09-21 09:45:27 +08:00
XuanYang-cn
916528f51a
Add initCtx in getDataSyncService (#27199)
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>
2023-09-20 16:03:23 +08:00
yah01
b4f86ea55e
Construct all success status with merr (#27226)
Signed-off-by: yah01 <yah2er0ne@outlook.com>
2023-09-20 10:57:23 +08:00
yah01
338848fbc3
Remove unused code and dup imports (#27228)
Signed-off-by: yah01 <yah2er0ne@outlook.com>
2023-09-20 10:55:23 +08:00
Xiaofan
6635398a6d
Fix Bin log concurrency by adding a pool (#27189)
Signed-off-by: xiaofan-luan <xiaofan.luan@zilliz.com>
2023-09-19 18:05:22 +08:00
XuanYang-cn
09505ea78e
Move etcd watch related code into eventmanager (#27192)
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-09-19 15:27:38 +08:00
XuanYang-cn
10116b85ac
Release before remove when releasing flowgraphs (#27191)
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>
2023-09-19 11:53:22 +08:00
yihao.dai
4b2802033d
Fix datanode panic due to concurrent compaction and delete processing (#27167)
Co-authored-by: xiaofan-luan <xiaofan.luan@zilliz.com>
2023-09-18 19:35:29 +08:00
yah01
a6b98740b7
Replace all status with only error string (#27125)
Signed-off-by: yah01 <yah2er0ne@outlook.com>
2023-09-18 14:05:28 +08:00
yihao.dai
c162c6a4c8
Increase FlushChannel retry times (#27140)
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2023-09-16 00:41:20 +08:00
yah01
168e82ee10
Fix panic while handling with the nil status (#27040)
Signed-off-by: yah01 <yah2er0ne@outlook.com>
2023-09-15 10:09:21 +08:00
yihao.dai
dd2cb1d44a
Flush by flush channels (#26859)
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2023-09-12 21:07:19 +08:00
yah01
00c65fa0d7
Refine QueryNode errors (#27013)
Signed-off-by: yah01 <yah2er0ne@outlook.com>
2023-09-12 16:07:18 +08:00
Xu Tong
9166011c4a
Add float16 vector (#25852)
Signed-off-by: Writer-X <1256866856@qq.com>
2023-09-08 10:03:16 +08:00
XuanYang-cn
7f1ae35e72
Add timeout in dispatcher, AsConsumer and Seek (#26686)
See also: #25309

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-09-08 09:51:17 +08:00
yiwangdr
337edc321b
tikv integration (#26246)
Signed-off-by: yiwangdr <yiwangdr@gmail.com>
2023-09-07 07:25:14 +08:00
bjzhjing
548c82eca5
Refactor storage.MergeInsertData() to optimize the merging process (#26839)
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>
2023-09-05 21:41:48 +08:00
congqixia
fe7f7ea237
Add cpp-build and setenv for querynode mockery regen command (#26843)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-09-05 14:13:49 +08:00
Enwei Jiao
fb0705df1b
Decouple basetable and componentparam (#26725)
Signed-off-by: Enwei Jiao <enwei.jiao@zilliz.com>
2023-09-05 10:31:48 +08:00
XuanYang-cn
b2e7cbdf4b
Remove TimeTravel in compactor (#26785)
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-09-04 17:41:48 +08:00
yihao.dai
4340cbfba2
Merge syncCPLagTooBehind policy into syncPeriodically policy (#26713)
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2023-09-04 14:37:47 +08:00
yihao.dai
7624c2b949
Improve rated log in insertBufferNode (#26788)
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2023-09-04 09:59:10 +08:00
yah01
3349db4aa7
Refine errors to remove changes breaking design (#26521)
Signed-off-by: yah01 <yah2er0ne@outlook.com>
2023-09-04 09:57:09 +08:00
congqixia
e8f1b1736e
Remove log.Error(err.error())-style log (#26783)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-09-01 13:09:01 +08:00
XuanYang-cn
8d54509e54
Fix CompactionLatency metrics (#26747)
- Refine compactor logs

See also: #26743

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-08-31 17:35:03 +08:00
smellthemoon
6069a7d42b
Log out memory in MB (#26262)
Signed-off-by: lixinguo <xinguo.li@zilliz.com>
Co-authored-by: lixinguo <xinguo.li@zilliz.com>
2023-08-30 17:13:00 +08:00
congqixia
cc584551fb
Remove unused hash value setting in DDNode (#26575)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-08-28 11:30:26 +08:00
SimFG
b9bc6681ae
Use atomic.bool for the isWatchFailed param in the tickler (#26558)
Signed-off-by: SimFG <bang.fu@zilliz.com>
2023-08-23 13:22:27 +08:00
XuanYang-cn
27691e843c
Fix channel checkpoint stuck (#26534)
See also: #23621

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-08-22 18:10:22 +08:00
XuanYang-cn
08fd28b30b
Only do gracefully stop when DN Stop (#26399)
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-08-21 11:16:20 +08:00
SimFG
3be4ac4022
Fix datanode/datacoord continuous restart (#26470)
Signed-off-by: SimFG <bang.fu@zilliz.com>
2023-08-20 21:20:24 +08:00
Enwei Jiao
533f0ddf6d
Add amazonlinux image, prepare to remove Centos7 image, remove openblas image (#26368)
Signed-off-by: Enwei Jiao <enwei.jiao@zilliz.com>
2023-08-18 18:32:19 +08:00
congqixia
fbb5d32cb6
Make write binlog in parallel (#26325)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-08-17 15:44:18 +08:00
congqixia
f371ec0316
Add data sync service close log for watch stuck debugging (#26332)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-08-16 10:57:32 +08:00
Enwei Jiao
78bc688d16
Remove QueryMsgStream in MqFactory interface (#26374)
Signed-off-by: Enwei Jiao <enwei.jiao@zilliz.com>
2023-08-16 10:29:33 +08:00
congqixia
f8bcf60e55
Adjust some confusing Warning log to INFO (#26356)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-08-15 17:49:33 +08:00
wayblink
21eeb37ce9
Save binlog timestampFrom, timestampTo meta when compact (#26203)
Signed-off-by: wayblink <anyang.wang@zilliz.com>
2023-08-08 21:17:21 +08:00
congqixia
b9850ce5c0
Fix copylocks linter errors (#26217)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-08-08 20:41:15 +08:00
MrPresent-Han
b84f5c560a
refine log out for datanode(#25763) (#26195)
Signed-off-by: MrPresent-Han <chun.han@zilliz.com>
2023-08-08 16:49:08 +08:00
MrPresent-Han
fb933fe64d
fix sync memory policy lose effect when inserting too many partitions(#25763) (#26154)
Signed-off-by: MrPresent-Han <chun.han@zilliz.com>
2023-08-07 09:59:07 +08:00
XuanYang-cn
830f06783a
Reduce log volumes of DC&DN (#26060)
See also: #26057

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-08-02 17:13:11 +08:00
XuanYang-cn
84253f255e
Fix datanode graceful stop panic (#25932)
See also: #25925

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-07-28 10:11:08 +08:00
groot
a6808e6484
Fix bulkinsert row count error (#25869)
Signed-off-by: yhmo <yihua.mo@zilliz.com>
2023-07-25 11:29:00 +08:00
cai.zhang
779d677eb2
Remove lock for compaction sync segments (#25711)
Signed-off-by: cai.zhang <cai.zhang@zilliz.com>
2023-07-24 17:25:00 +08:00
jaime
8e5d6193f9
Add a timeout config for bulkinsert request (#25789)
Signed-off-by: jaime <yun.zhang@zilliz.com>
2023-07-24 14:23:00 +08:00
congqixia
3c503afe7c
Use typeutil.ConcurrentMap instead of sync.Map (#25846)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-07-24 10:23:01 +08:00
congqixia
31b40b3ccc
Use typeutil.ConcurrentMap instead of sync.Map in fg manager (#25835)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-07-24 10:21:07 +08:00
Enwei Jiao
66fdc71479
Refactor logs in DataCoord & DataNode (#25574)
Signed-off-by: Enwei Jiao <enwei.jiao@zilliz.com>
2023-07-14 15:56:31 +08:00
yiwangdr
b9189b9f41
Organize mocks from types.go (#25466)
Signed-off-by: yiwangdr <yiwangdr@gmail.com>
2023-07-14 10:12:31 +08:00