milvus-io
/
milvus
mirror of https://github.com/milvus-io/milvus.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
milvus/internal/datacoord
History
XuanYang-cn ca7ec23198
enhance: Use partitionID when delete by partitionKey (#38231) 
When delete by partition_key, Milvus will generates L0 segments
globally. During L0 Compaction, those L0 segments will touch all
partitions collection wise. Due to the false-positive rate of segment
bloomfilters, L0 compactions will append false deltalogs to completed
irrelevant partitions, which causes *partition deletion amplification.

This PR uses partition_key to set targeted partitionID when producing
deleteMsgs into MsgStreams. This'll narrow down L0 segments scope to
partition level, and remove the false-positive influence
collection-wise.

However, due to DeleteMsg structure, we can only label one partition to
one deleteMsg, so this enhancement fails if user wants to delete over 2
partition_keys in one deletion.

See also: #34665

Signed-off-by: yangxuan <xuan.yang@zilliz.com>
 		2024-12-20 11:18:46 +08:00
..
allocator fix: fix `make generate-mockery` panic with go1.22 (#36830) 2024-10-17 12:11:31 +08:00
broker fix: fix `make generate-mockery` panic with go1.22 (#36830) 2024-10-17 12:11:31 +08:00
session fix: SyncSegments rpc always failed (#38578) 2024-12-19 15:40:45 +08:00
OWNERS [skip ci]Update OWNERS files (#11898) 2021-11-16 15:41:11 +08:00
README.md [skip ci]Change etcd to lowercase (#9983) 2021-10-15 18:58:37 +08:00
analyze_meta.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
analyze_meta_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
build_index_policy.go Format the code (#27275) 2023-09-21 09:45:27 +08:00
channel.go fix: ChannelManager concurret Release and Watch bug (#38590) 2024-12-19 22:50:47 +08:00
channel_manager.go fix: ChannelManager concurret Release and Watch bug (#38590) 2024-12-19 22:50:47 +08:00
channel_manager_factory.go fix: drop collection failed if enable streaming service (#37444) 2024-11-07 10:26:26 +08:00
channel_manager_test.go fix: ChannelManager concurret Release and Watch bug (#38590) 2024-12-19 22:50:47 +08:00
channel_store.go fix: ChannelManager concurret Release and Watch bug (#38590) 2024-12-19 22:50:47 +08:00
channel_store_test.go fix: ChannelManager concurret Release and Watch bug (#38590) 2024-12-19 22:50:47 +08:00
cluster.go enhance: refine exists log print with ctx (#38080) 2024-12-14 22:36:44 +08:00
cluster_test.go enhance: Move datanode/indexnode manager to session pkg (#35634) 2024-08-22 16:02:56 +08:00
compaction.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
compaction_l0_view.go enhance: Add deltaRowCount in l0 compaction (#33997) 2024-06-20 17:46:01 +08:00
compaction_l0_view_test.go enhance: Add deltaRowCount in l0 compaction (#33997) 2024-06-20 17:46:01 +08:00
compaction_policy_clustering.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
compaction_policy_clustering_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
compaction_policy_l0.go enhance: tidy compaction logs (#37595) 2024-11-12 21:18:29 +08:00
compaction_policy_l0_test.go enhance: Separate allocator pkg in datacoord (#35622) 2024-08-22 10:06:56 +08:00
compaction_policy_single.go enhance: tidy compaction logs (#37595) 2024-11-12 21:18:29 +08:00
compaction_policy_single_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
compaction_queue.go fix: Refine compactionTask to avoid data race (#36936) 2024-10-24 09:55:28 +08:00
compaction_queue_test.go fix: Refine compactionTask to avoid data race (#36936) 2024-10-24 09:55:28 +08:00
compaction_task.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
compaction_task_clustering.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
compaction_task_clustering_test.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
compaction_task_l0.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
compaction_task_l0_test.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
compaction_task_meta.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
compaction_task_meta_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
compaction_task_mix.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
compaction_task_mix_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
compaction_test.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
compaction_trigger.go enhance: refine exists log print with ctx (#38080) 2024-12-14 22:36:44 +08:00
compaction_trigger_test.go enhance: improve mix compaction performance by removing max segment limitations (#38344) 2024-12-11 20:38:42 +08:00
compaction_trigger_v2.go enhance: refine exists log print with ctx (#38080) 2024-12-14 22:36:44 +08:00
compaction_trigger_v2_test.go enhance: Separate allocator pkg in datacoord (#35622) 2024-08-22 10:06:56 +08:00
compaction_view.go enhance: Add deltaRowCount in l0 compaction (#33997) 2024-06-20 17:46:01 +08:00
const.go enhance: pass partition key scalar info if enabled when build vector index (#29931) 2024-01-24 00:04:55 +08:00
errors.go Refine DataCoord status (#27262) 2023-09-26 17:15:27 +08:00
errors_test.go Format the code (#27275) 2023-09-21 09:45:27 +08:00
garbage_collector.go enhance: refine exists log print with ctx (#38080) 2024-12-14 22:36:44 +08:00
garbage_collector_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
go_channel_singleton.go fix: Fix improper use of offset in HybridSearch (#36244) 2024-09-13 22:05:15 +08:00
handler.go fix: Support get segments from current segments view (#38512) 2024-12-18 18:00:54 +08:00
handler_test.go fix: Support get segments from current segments view (#38512) 2024-12-18 18:00:54 +08:00
import_checker.go fix: Use the correct RootPath when decompressing binlog in stats task (#38341) 2024-12-11 16:16:42 +08:00
import_checker_test.go fix: Use the correct RootPath when decompressing binlog in stats task (#38341) 2024-12-11 16:16:42 +08:00
import_job.go fix: Revert "enhance: Support db for bulkinsert (#37012)" (#37420) 2024-11-07 17:02:25 +08:00
import_meta.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
import_meta_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
import_scheduler.go enhance: refine exists log print with ctx (#38080) 2024-12-14 22:36:44 +08:00
import_scheduler_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
import_task.go enhance: Use internal json(sonic) to replace std json lib (#37708) 2024-11-18 10:46:31 +08:00
import_util.go fix: Use the correct RootPath when decompressing binlog in stats task (#38341) 2024-12-11 16:16:42 +08:00
import_util_test.go fix: Use the correct RootPath when decompressing binlog in stats task (#38341) 2024-12-11 16:16:42 +08:00
index_engine_version_manager.go enhance: Add `go-deadlock` as unittest only dependency (#33063) 2024-05-15 16:33:34 +08:00
index_engine_version_manager_test.go Optimize the codec code of session (#27360) 2023-10-01 10:33:30 +08:00
index_meta.go enhance: refine exists log print with ctx (#38080) 2024-12-14 22:36:44 +08:00
index_meta_test.go enhance: Add ctx param to KV operation interfaces (#38154) 2024-12-05 15:16:41 +08:00
index_service.go enhance: refine exists log print with ctx (#38080) 2024-12-14 22:36:44 +08:00
index_service_test.go enhance: alterindex & altercollection supports altering properties (#37437) 2024-12-11 10:20:42 +08:00
job_manager.go fix: Use the correct RootPath when decompressing binlog in stats task (#38341) 2024-12-11 16:16:42 +08:00
job_manager_test.go feat: add segment/channel/task/slow query render (#37561) 2024-11-12 17:44:29 +08:00
knapsack.go enhance: improve mix compaction performance by removing max segment limitations (#38344) 2024-12-11 20:38:42 +08:00
knapsack_test.go enhance: try compact small segments first if they may compose a full segment (#37709) 2024-12-02 16:12:38 +08:00
meta.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
meta_test.go enhance: Add ctx param to KV operation interfaces (#38154) 2024-12-05 15:16:41 +08:00
meta_util.go Make VChannelInfo read compatible with 2.0.2 (#17995) 2022-07-07 21:58:22 +08:00
metrics_info.go enhance: add disk stats within system metrics (#38033) 2024-12-06 16:32:41 +08:00
metrics_info_test.go enhance: add list index and segment index retrieval API for WebUI (#37861) 2024-11-22 16:58:34 +08:00
mock_channel_store.go fix: ChannelManager concurret Release and Watch bug (#38590) 2024-12-19 22:50:47 +08:00
mock_channelmanager.go feat: add segment,pipeline, replica and resourcegroup api for WebUI (#37344) 2024-11-07 11:52:25 +08:00
mock_cluster.go fix: fix `make generate-mockery` panic with go1.22 (#36830) 2024-10-17 12:11:31 +08:00
mock_compaction_meta.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
mock_compaction_plan_context.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
mock_handler.go fix: Support get segments from current segments view (#38512) 2024-12-18 18:00:54 +08:00
mock_index_engine_version_manager.go fix: fix `make generate-mockery` panic with go1.22 (#36830) 2024-10-17 12:11:31 +08:00
mock_job_manager.go fix: Use the correct RootPath when decompressing binlog in stats task (#38341) 2024-12-11 16:16:42 +08:00
mock_segment_manager.go enhance: Reduce segmentManager lock granularity (#37836) 2024-12-17 14:12:52 +08:00
mock_subcluster.go fix: fix `make generate-mockery` panic with go1.22 (#36830) 2024-10-17 12:11:31 +08:00
mock_test.go fix: revert optimize CPU usage for CheckHealth requests (#35589) (#38555) 2024-12-19 00:38:45 +08:00
mock_trigger_manager.go fix: fix `make generate-mockery` panic with go1.22 (#36830) 2024-10-17 12:11:31 +08:00
partition_stats_meta.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
partition_stats_meta_test.go enhance: clean compaction task in compactionHandler (#38170) 2024-12-19 12:36:47 +08:00
policy.go fix: drop collection failed if enable streaming service (#37444) 2024-11-07 10:26:26 +08:00
policy_test.go fix: Fill nil schema so that Milvus can watch channel for those upgraded from 2.2 to 2.4 #35695 (#35694) 2024-08-27 10:36:59 +08:00
segment_allocation_policy.go fix: Fix large growing segment (#37388) 2024-11-07 17:12:25 +08:00
segment_allocation_policy_test.go enhance: Change the fixed value to a ratio for clustering segment size (#35076) 2024-08-01 22:04:14 +08:00
segment_info.go enhance: Remove unnecessary clone in SetState (#37697) 2024-11-15 15:06:30 +08:00
segment_info_test.go fix: Return all compactTo segments after support split (#36361) 2024-09-20 14:11:11 +08:00
segment_manager.go enhance: Reduce segmentManager lock granularity (#37836) 2024-12-17 14:12:52 +08:00
segment_manager_test.go enhance: Reduce segmentManager lock granularity (#37836) 2024-12-17 14:12:52 +08:00
segment_operator.go enhance: Update stats task to optional (#35947) 2024-09-12 20:37:08 +08:00
segment_operator_test.go fix: Prevent clone when selecting segments from meta (#30928) 2024-03-01 13:59:02 +08:00
server.go fix: revert optimize CPU usage for CheckHealth requests (#35589) (#38555) 2024-12-19 00:38:45 +08:00
server_test.go fix: revert optimize CPU usage for CheckHealth requests (#35589) (#38555) 2024-12-19 00:38:45 +08:00
services.go enhance: Use partitionID when delete by partitionKey (#38231) 2024-12-20 11:18:46 +08:00
services_test.go enhance: Reduce segmentManager lock granularity (#37836) 2024-12-17 14:12:52 +08:00
stats_task_meta.go feat: add segment,pipeline, replica and resourcegroup api for WebUI (#37344) 2024-11-07 11:52:25 +08:00
stats_task_meta_test.go fix: Record the nodeID before assigning tasks (#36371) 2024-09-28 17:21:15 +08:00
sync_segments_scheduler.go fix: Support get segments from current segments view (#38512) 2024-12-18 18:00:54 +08:00
sync_segments_scheduler_test.go fix: datacoord stuck at stopping progress (#36852) 2024-10-17 12:13:37 +08:00
task_analyze.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
task_index.go fix:remove unnecessary error logs (#38245) 2024-12-08 17:20:40 +08:00
task_scheduler.go enhance: Refine task scheduler logs (#38334) 2024-12-11 15:00:44 +08:00
task_scheduler_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
task_stats.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
task_stats_test.go enhance: refine the datacoord meta related interfaces (#37957) 2024-11-26 19:46:34 +08:00
types.go fix: Record the nodeID before assigning tasks (#36371) 2024-09-28 17:21:15 +08:00
util.go fix: revert optimize CPU usage for CheckHealth requests (#35589) (#38555) 2024-12-19 00:38:45 +08:00
util_test.go enhance: Separate allocator pkg in datacoord (#35622) 2024-08-22 10:06:56 +08:00

README.md

Data Coordinator

Data cooridnator(datacoord for short) is the component to organize DataNodes and segments allocations.

Dependency

  • KV store: a kv store has all the meta info datacoord needs to operate. (etcd)
  • Message stream: a message stream to communicate statistics information with data nodes. (Pulsar)
  • Root Coordinator: timestamp, id and meta source.
  • Data Node(s): could be an instance or a cluster, actual worker group handles data modification operations.