Commit Graph

21664 Commits (update_knowhere_commit_0461ddf7766bc05e1fee635d134d498c2493a1a7)

Author SHA1 Message Date
github-actions[bot] 16193e1aa0 Update knowhere commit 2025-01-16 15:06:25 +00:00
Spade A 0461ddf776
fix: phrase match does not support offset input (#39338)
fix: #39337

Signed-off-by: SpadeA-Tang <tangchenjie1210@gmail.com>
2025-01-16 22:05:01 +08:00
XuanYang-cn 74b4369c5b
fix: Record active collections for l0Policy (#39217)
By recording the active collection lists, The l0 compaction trigger
of view change and idle won't influence each other.

Also this pr replaces the L0View cache with real L0 segments' change.
Save some memory and make L0 compaction triggers more accurate.

See also: #39187

Signed-off-by: yangxuan <xuan.yang@zilliz.com>

---------

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2025-01-16 20:19:02 +08:00
sre-ci-robot 78568906b9 Update all contributors
Signed-off-by: sre-ci-robot <sre-ci-robot@zilliz.com>
2025-01-16 12:04:29 +00:00
Duc Le Tu 2a962ad1ec
fix: apply custom producer config for kafkaHealthCheck (#39283)
issue: https://github.com/milvus-io/milvus/issues/39287
KafkaHealthCheck init without ProducerExtraConfig. This PR fix that.

---------

Signed-off-by: DLT1412 <tuduc93@gmail.com>
Co-authored-by: DucLT <duc.le1@be.com.vn>
2025-01-16 19:31:02 +08:00
foxspy 172051b050
fix: remove num_build_thread param (#39340)
issue: #39090 
The num_build_thread parameter will limit the number of build omps. This
parameter will override the effect of buildIndexThreadPoolRatio.
Removing this parameter will have no actual effect. This parameter is
actually only used in the growing index, where it will be explicitly
set.

Signed-off-by: xianliang.li <xianliang.li@zilliz.com>
2025-01-16 19:21:01 +08:00
Gao 75d7978a18
enhance: pass partition key scalar info if enable for vector mem index (#39123)
issue: #34332

---------

Signed-off-by: chasingegg <chao.gao@zilliz.com>
2025-01-16 14:33:03 +08:00
congqixia 1f6fd54146
fix: [skip e2e] Make test case assigner result deterministic (#39317)
Related to #39296

The case initialized with {100:8 ,101: 16}. After first assignment, the
slots become {100:8, 101:8} and the following result is not stable.

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2025-01-16 14:25:02 +08:00
nico 93ce4b1c97
test: modify milvus client rbac test cases (#39079)
Signed-off-by: nico <cheng.yuan@zilliz.com>
2025-01-16 11:39:01 +08:00
Spade A 8c4ba70a4c
fix: enable to build index with single segment (#39233)
fix https://github.com/milvus-io/milvus/issues/39232

---------

Signed-off-by: SpadeA-Tang <tangchenjie1210@gmail.com>
2025-01-16 11:01:06 +08:00
congqixia bca2a62b78
enhance: Handle PutOrRef collection schema failure error (#39310)
Related to previous pr #39279

When NewCollection returns nil, the error shall be returned and handled
by caller

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2025-01-16 10:13:06 +08:00
congqixia 57e5652f1a
enhance: Log error instead of panicking if load lock wait timeout (#39308)
Related to #39205
Previous PR #39206

This PR change wait timeout behavior to log error and return to avoid
making other collection read failure in only some collections have
deadlock

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2025-01-16 02:31:02 +08:00
yihao.dai a5a83a0904
fix: Fix consume blocked due to too many consumers (#38455)
This PR limits the maximum number of consumers per pchannel to 10 for
each QueryNode and DataNode.

issue: https://github.com/milvus-io/milvus/issues/37630

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2025-01-15 21:37:01 +08:00
yihao.dai 657550cf06
fix: Fix slow dist handle and slow observe (#38566)
1. Provide partition&channel level indexing in the collection target.
2. Make `SegmentAction` not wait for distribution.
3. Remove scheduler and target manager mutex.
4. Optimize logging to reduce CPU overhead.

issue: https://github.com/milvus-io/milvus/issues/37630

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2025-01-15 20:17:00 +08:00
presburger 38881bf591
enhance: prevent multiple query nodes from causing excessive occupancy of a single node, leading to GPU memory overflow (#39276) (#38617)
issue: #39276

Signed-off-by: yusheng.ma <yusheng.ma@zilliz.com>
2025-01-15 20:15:01 +08:00
yihao.dai 0df2c75b77
fix: Fix rootcoord meta mutex contention (#38799)
RootCoord meta uses copy-on-write, allowing the removal of unnecessary
copies.

issue: https://github.com/milvus-io/milvus/issues/37630

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2025-01-15 20:11:08 +08:00
congqixia 82bdf9a6a8
fix: Add index param duplication check (#39289)
Related to #39288

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2025-01-15 19:23:01 +08:00
congqixia eb63334312
enhance: Add try-catch and return CStatus for NewCollection (#39279)
Related to #28795

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2025-01-15 19:17:01 +08:00
wei liu d2834a1812
enhance: Add logs for check health failed (#39208)
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
2025-01-15 17:31:00 +08:00
qixuan 27a99f6b9d
test:Add search Iterator v2 test case for milvus client (#39120)
issue: #37548

Signed-off-by: qixuan <673771573@qq.com>
2025-01-15 17:10:59 +08:00
Zhen Ye 7addeea70c
fix: wrong streaming mockery package name (#39260)
issue: #39095

Signed-off-by: chyezh <chyezh@outlook.com>
2025-01-15 15:51:00 +08:00
yihao.dai ce41778fe6
enhance: Optimize GetLocalDiskSize and segment loader mutex (#38599)
1. Make the segment loader lock protect only the resource.
2. Optimize GetDiskUsage to avoid excessive overhead.

issue: https://github.com/milvus-io/milvus/issues/37630

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2025-01-15 15:45:01 +08:00
Ted Xu e501025bba
enhance: simplify compaction tasks to reduce their memory overhead (#39121)
See #39080

---------

Signed-off-by: Ted Xu <ted.xu@zilliz.com>
2025-01-15 14:51:00 +08:00
yihao.dai f5234c3c11
enhance: Remove mutex from SegmentManger in querynode (#39051)
Remove mutex from SegmentManger in querynode to prevent mutex
contention.

issue: https://github.com/milvus-io/milvus/issues/37630

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2025-01-15 12:45:00 +08:00
jaime e8f76cd2d9
fix: unstable ut in leader_vew_manager.go file (#39161)
issue: #38672

Signed-off-by: jaime <yun.zhang@zilliz.com>
2025-01-15 12:26:59 +08:00
cai.zhang 6d45dd5666
fix: Add scalar index engine version for compatibility (#39204)
issue: #39203

---------

Signed-off-by: Cai Zhang <cai.zhang@zilliz.com>
2025-01-15 12:25:00 +08:00
cai.zhang 3a6408b237
fix: Record a map to avoid repeatedly traversing the CompactionFrom (#38925)
issue: #38811

Signed-off-by: Cai Zhang <cai.zhang@zilliz.com>
2025-01-15 10:02:58 +08:00
Chun Han ed31a5a4bf
enhance: fix inconsistenty of alias and db for query iterator(#39045) (#39216)
related: #39045

Signed-off-by: MrPresent-Han <chun.han@gmail.com>
Co-authored-by: MrPresent-Han <chun.han@gmail.com>
2025-01-15 09:48:59 +08:00
sre-ci-robot e1f5cb7427 Update all contributors
Signed-off-by: sre-ci-robot <sre-ci-robot@zilliz.com>
2025-01-15 00:01:22 +00:00
sre-ci-robot 55dcac375c
[automated] Update Knowhere Commit (#39263)
Update Knowhere Commit
Signed-off-by: sre-ci-robot sre-ci-robot@users.noreply.github.com

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2025-01-15 02:52:59 +08:00
yihao.dai ec2e77b5d7
enhance: Reduce memory usage of BF in DataNode and QueryNode (#38129)
1. DataNode: Skip generating BF during the insert phase (BF will be
regenerated during the sync phase).
2. QueryNode: Skip generating or maintaining BF for growing segments;
deletion checks will be handled in the segcore.

issue: https://github.com/milvus-io/milvus/issues/37630

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2025-01-15 01:59:01 +08:00
yihao.dai 272d95ad79
enhance: Reduce mutex contention in datacoord meta (#38219)
1. Using secondary index to avoid retrieving all segments at
`GetSegmentsChanPart`.
2. Perform batch SetAllocations to reduce the number of times the meta
lock is acquired.

issue: https://github.com/milvus-io/milvus/issues/37630

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2025-01-15 01:15:02 +08:00
Cai Yudong 5bf1b2b929
feat: Support Int8Vector in go (#38990)
Issue: #38666

Signed-off-by: Cai Yudong <yudong.cai@zilliz.com>
2025-01-14 20:43:06 +08:00
congqixia d89768f9e0
enhance: Unify LoadStateLock RLock & PinIf (#39206)
Related to #39205

This PR merge `RLock` & `PinIfNotReleased` into `PinIf` function
preventing segment being released before any Read operation finished.

---------

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2025-01-14 18:38:59 +08:00
aoiasd da07993082
enhance: report auth log to hook when grpc authenticate failed (#38943)
Signed-off-by: aoiasd <zhicheng.yue@zilliz.com>
2025-01-14 17:47:04 +08:00
junjiejiangjjj 097d167e96
doc: Update tools info (#39244)
Signed-off-by: junjie.jiang <junjie.jiang@zilliz.com>
2025-01-14 16:53:00 +08:00
Zhen Ye fd84ed817c
enhance: add broadcast operation for msgstream (#39040)
issue: #38399

- make broadcast service available for msgstream by reusing the
architecture streaming service

---------

Signed-off-by: chyezh <chyezh@outlook.com>
2025-01-14 15:14:59 +08:00
congqixia da1b786ef8
enhance: Utilize "find0" in segment.find_first (#39229)
Related to #39003

Previous PR #39004 has to clone & flip bitset due to bitset does not
support find0 operator. #39176 added this feature so clone & flip could
be removed now.

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2025-01-14 14:14:58 +08:00
Zhen Ye 3e788f0fbd
enhance: record memory size (uncompressed) item for index (#38770)
issue: #38715

- Current milvus use a serialized index size(compressed) for estimate
resource for loading.
- Add a new field `MemSize` (before compressing) for index to estimate
resource.

---------

Signed-off-by: chyezh <chyezh@outlook.com>
2025-01-14 10:33:06 +08:00
Buqian Zheng 5e38f01e5b
enhance: update knowhere version (#39212)
Signed-off-by: Buqian Zheng <zhengbuqian@gmail.com>
2025-01-14 10:21:05 +08:00
Alexander Guzhva 3447ff7310
enhance: [bitset] extend op_find() to be able to search both 0 and 1 (#39176)
issue: #39124 

`bitset::find_first()` and `bitset::find_next()` now accept one more
parameter, which allows to search for `0` bit instead of `1` bit

Signed-off-by: Alexandr Guzhva <alexanderguzhva@gmail.com>
2025-01-14 09:50:58 +08:00
yihao.dai 702347bbfd
fix: Fix incorrect memory estimation for small segments (#38813)
Skip estimation index memory logic for segments without index file.

issue: https://github.com/milvus-io/milvus/issues/37630

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2025-01-13 21:49:12 +08:00
sre-ci-robot cc8e6ebd5e
[automated] Bump milvus version to v2.5.3 (#39215)
Bump milvus version to v2.5.3
Signed-off-by: sre-ci-robot sre-ci-robot@users.noreply.github.com

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2025-01-13 18:37:03 +08:00
congqixia b3daae8e40
fix: Make KeyLock obj pool size unlimited (#39182)
Related to #39181

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2025-01-13 18:04:58 +08:00
Bingyi Sun a00ba861a4
fix: Fix in filter search result is empty if pk type is varchar (#39106)
https://github.com/milvus-io/milvus/issues/39107

Signed-off-by: sunby <sunbingyi1992@gmail.com>
2025-01-13 16:14:58 +08:00
smellthemoon accc9e7fbf
fix: fail to get empty index num rows (#39155)
#39125

Signed-off-by: lixinguo <xinguo.li@zilliz.com>
Co-authored-by: lixinguo <xinguo.li@zilliz.com>
2025-01-13 16:04:58 +08:00
sthuang 5c5948cb70
fix: rbac custom group privilege level check (#39164)
related: https://github.com/milvus-io/milvus/issues/39086

Signed-off-by: shaoting-huang <shaoting.huang@zilliz.com>
2025-01-13 11:24:59 +08:00
Zhen Ye 5f94954bb4
fix: data race when accessing field_ when retrieving (#39151)
issue: #39148

Signed-off-by: chyezh <chyezh@outlook.com>
2025-01-13 11:23:04 +08:00
wei liu cc5d59392a
fix: channel unbalance during stopping balance progress (#38971)
issue: #38970
cause the stopping balance channel still use the row_count_based policy,
which may causes channel unbalance in multi-collection case.

This PR impl a score based stopping balance channel policy.

Signed-off-by: Wei Liu <wei.liu@zilliz.com>
2025-01-13 11:21:06 +08:00
Buqian Zheng 640a49ffb6
fix: fix chunk cache madvise when sparse raw data is mmaped (#39145)
instead of marking as not supported,
`ChunkedSparseFloatColumn::DataByteSize` can simply use the impl of
super class.

issue: https://github.com/milvus-io/milvus/issues/39158

Signed-off-by: Buqian Zheng <zhengbuqian@gmail.com>
2025-01-13 10:34:57 +08:00