Commit Graph

84 Commits (8a4ef1b836d9e66ce3ebfed9602b17b86e08bd5a)

Author SHA1 Message Date
zhenshan.cao ac4f3997ce
enhance: Reconstructing Compaction to possess persistence capability (#33265)
issue #33586

Signed-off-by: zhenshan.cao <zhenshan.cao@zilliz.com>
2024-06-05 10:17:50 +08:00
cai.zhang 77637180fa
enhance: Periodically synchronize segments to datanode watcher (#33420)
issue: #32809

---------

Signed-off-by: Cai Zhang <cai.zhang@zilliz.com>
2024-05-30 13:37:44 +08:00
yihao.dai 7730b910b9
enhance: Decouple compaction from shard (#33138)
Decouple compaction from shard, remove dependencies on shards (e.g.
SyncSegments, injection).

issue: https://github.com/milvus-io/milvus/issues/32809

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2024-05-24 09:07:41 +08:00
yihao.dai 017fd7bc25
enhance: Select L2 segments in L0Compaction as well (#32991)
/kind improvement

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2024-05-21 16:13:39 +08:00
yihao.dai 32560263fa
enhance: Query slot for compaction task (#32881)
Query slot of compaction in datanode, and transfer the control logic for
limiting compaction tasks from datacoord to the datanode.

issue: https://github.com/milvus-io/milvus/issues/32809

---------

Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
2024-05-17 18:19:38 +08:00
congqixia 8cf2cf5c94
enhance: Add `go-deadlock` as unittest only dependency (#33063)
See also #33062

This PR:

- Add `lock.RWMutex` & `lock.Mutex` alias to switch implementation based
  on build flags
- When build flags has `test` in it, use `go-deadlock` to detect
  possible deadlocks
- Replace all `sync.RWMutex` & `sync.Mutex` in datacoord pkg

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2024-05-15 16:33:34 +08:00
congqixia cedb33ceec
enhance: Improve datacoord segment filtering with collection (#32831)
See also #32165

This PR modify the `SelectSegments` interface to utilizing collection id
information when selecting segment with provided collection

---------

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2024-05-08 21:37:29 +08:00
XuanYang-cn aad3ed3835
fix: [cherry-pick]Skip changing meta if nodeID not match with channel (#31672)
See also: #31648
pr: #31665, #31694

---------

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2024-04-10 15:09:18 +08:00
aoiasd 0c153a5820
enhance: Rename update segment operator (#31121)
Signed-off-by: aoiasd <zhicheng.yue@zilliz.com>
2024-03-20 17:53:14 +08:00
XuanYang-cn 0066c016b6
enhance: Skip submit empty l0 tasks in DC (#31280)
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2024-03-19 10:13:14 +08:00
Xiaofan 8c43c5b6cb
fix: get compaction failure when datanode is actually alive (#31353)
didn't mark the compact as failure if it's simply an rpc error when
GetCompactionPlansResults
see #31352

Signed-off-by: xiaofanluan <xiaofan.luan@zilliz.com>
2024-03-18 10:01:36 -07:00
congqixia 196f0c1e1d
fix: Skip invalid compaction plan (#31045)
See also #31044

---------

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2024-03-06 21:36:59 +08:00
XuanYang-cn def72947c7
fix: Trigger l0 compaction when l0 views don't change (#30729)
Trigger l0 compaction when l0 views don't change

So that leftover l0 segments would be compacted in the end.

1. Refresh LevelZero plans in comactionPlanHandler, remove the meta
dependency
of compaction trigger v2
2. Add ForceTrigger method for CompactionView interface
3. rename mu to taskGuard
4. Add a new TriggerTypeLevelZeroViewIDLE
5. Add an idleTicker for compaction view manager

See also: #30098, #30556

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

---------

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2024-03-05 16:37:00 +08:00
XuanYang-cn 2867f50fcc
fix: Clear DN unkown compaction tasks (#30850)
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>
2024-03-01 11:31:00 +08:00
wayblink b74264881c
enhance: Refine compaction interfaces to support major compaction (#30632)
Refine compaction interfaces in datacoord, support compaction result
with more than one segment. Prepare for major compaction.

related: #30633

Signed-off-by: wayblink <anyang.wang@zilliz.com>
2024-02-19 20:52:50 +08:00
XuanYang-cn fb5e09d94d
fix: call injectDone after compaction failed (#30277)
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>
2024-02-01 14:25:04 +08:00
congqixia 0c7a96b48d
enhance: Make compaction log has traceID (#30338)
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>
2024-01-30 10:09:03 +08:00
chyezh f2985d8454
fix: compact operation on datacoord meta should preform as a transcation (#29775)
issue: #29691

Signed-off-by: chyezh <chyezh@outlook.com>
2024-01-26 16:59:00 +08:00
XuanYang-cn 0b6beb7e0f
fix: Fill in info in CompactionSegmentBinlogs (#30279)
After #28873, PartitionID and CollectionID should be filled in
CompactionSegmentBinlog so that DataNode can compose
the correct logPath. However There're some places left forgotten to fill
in the information, causing Datanode downloading `xxx/0/0/xxxx/xxxx`
binlogs during compaction

See also: #30213

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2024-01-26 10:51:02 +08:00
congqixia 6a73860815
enhance: Add open telemetry tracing for compaction (#30168)
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>
2024-01-23 10:37:00 +08:00
smellthemoon e52ce370b6
enhance:don't store logPath in meta to reduce memory (#28873)
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>
2024-01-18 22:06:31 +08:00
aoiasd cb18f18c1d
fix: compacted segment status was flushing instead flushed and L0 segment trigger gc slowly (#29587)
relate: https://github.com/milvus-io/milvus/issues/29492

Signed-off-by: aoiasd <zhicheng.yue@zilliz.com>
2024-01-09 10:52:49 +08:00
XuanYang-cn f1b6ccf305
enhance: compaction use ChannelManager interface (#29530)
Rewrite compaction_test.go

See also: #29447

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2024-01-02 18:08:49 +08:00
XuanYang-cn 4b406e5973
enhance: Add CompactionTaskNum metrics (#29518)
See also: #27606

---------

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-12-28 15:46:55 +08:00
XuanYang-cn ae180d1628
enhance: Change ChannelManager to interface (#29300)
Rewrite cluster test
issue: #28854

---------

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-12-25 19:24:46 +08:00
SimFG 67ab0e424b
fix: Clean the compaction plan info to avoid the object leak (#29365)
issue: #29296

Signed-off-by: SimFG <bang.fu@zilliz.com>
2023-12-22 12:00:43 +08:00
XuanYang-cn 2c62bcbbf6
fix: Fix some left comments (#29027)
See also: #28814

---------

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-12-13 18:48:39 +08:00
aoiasd 3c32ba2407
enhance: pack datacoord Cluster and SessionManager with interface and mock them (#28869)
relate: https://github.com/milvus-io/milvus/issues/28861
https://github.com/milvus-io/milvus/issues/28854

---------

Signed-off-by: aoiasd <zhicheng.yue@zilliz.com>
2023-12-11 17:52:37 +08:00
XuanYang-cn 5bac7f7897
fix: Fix L0 compaction in datacoord (#28814)
See also: #27606

---------

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-12-05 18:44:37 +08:00
congqixia 3f9da2b8a0
fix: Check whether compaction plan exist before update (#28826)
Fix: #28825

Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-11-29 18:54:27 +08:00
XuanYang-cn 321c5c32e3
fix: Separate schedule and check results loop (#28692)
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>
2023-11-29 10:50:29 +08:00
XuanYang-cn 9b371067d2
feat: Add Compaction views and triggers (#27906)
- Add Compaction l0 views
- Add Compaction scheduler
- Add Compaction triggerv2
- Add Compaction view manager

See also: #27606

---------

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-11-23 17:30:25 +08:00
XuanYang-cn 0be22ccacd
Fix compaction unable to notify datanode (#28409)
See also: #28214, #28368

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-11-14 16:38:18 +08:00
XuanYang-cn 40d5c902b6
Enable getting multiple segments in plan result (#28350)
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>
2023-11-14 15:56:19 +08:00
congqixia c41df18b6d
Add compaction meta in `SyncSegmentsRequest` (#28199)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-11-07 10:06:17 +08:00
XuanYang-cn 22c089894e
Add scheduler (#27938)
See also: #27606

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-11-07 03:18:18 +08:00
yah01 90e2c63d9e
Fix getting incorrect CPU num (#28146)
Signed-off-by: yah01 <yang.cen@zilliz.com>
2023-11-06 06:02:16 +08:00
SimFG 26f06dd732
Format the code (#27275)
Signed-off-by: SimFG <bang.fu@zilliz.com>
2023-09-21 09:45:27 +08:00
cai.zhang d7cd1f2a6d
Fix bug for get index state with compaction segment (#26822)
Signed-off-by: cai.zhang <cai.zhang@zilliz.com>
2023-09-04 17:29:48 +08:00
XuanYang-cn 9598a8b236
Fix timeout task never release queue (#26593)
See also: #26413, #26566

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-08-30 11:12:27 +08:00
XuanYang-cn 35f3e263cd
Fix lock usage in DataCoord compaction (#26032)
See also: #26009

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
2023-08-01 08:55:04 +08:00
congqixia 37636b2f4d
Add compaction worker max parallel config item (#25834)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-07-24 10:17:00 +08:00
congqixia 8d343bf75a
Make compaction rpc timeout and parallel maxium configurable (#25672)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-07-18 14:25:20 +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
xige-16 33c2012675
Add more metrics (#25081)
Signed-off-by: xige-16 <xi.ge@zilliz.com>
2023-06-26 17:52:44 +08:00
Xiaofan f57fe6d70b
Add compaction log (#24976)
Signed-off-by: xiaofan-luan <xiaofan.luan@zilliz.com>
2023-06-19 14:18:41 +08:00
congqixia 41af0a98fa
Use go-api/v2 for milvus-proto (#24770)
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
2023-06-09 01:28:37 +08:00
MrPresent-Han b517bc9e6a
refine balance mechanism including:(#23454) (#23763) (#23791)
1. balance granuity to replica to avoid influence unrelated replicas
2. avoid balance back and forth

Signed-off-by: MrPresent-Han <jamesharden11122@gmail.com>
2023-05-04 12:22:40 +08:00
jaime c9d0c157ec
Move some modules from internal to public package (#22572)
Signed-off-by: jaime <yun.zhang@zilliz.com>
2023-04-06 19:14:32 +08:00
SimFG 4a90490a67
Fix the `segment not found` error (#22772)
Signed-off-by: SimFG <bang.fu@zilliz.com>
2023-03-17 17:27:56 +08:00