wei liu dc18d2aa8a

fix: Use leaderID for LeaderAction stale check in scheduler (#46785) ...

issue: #46737

PR #46440 refactored checkStale to use action.Node() for all action
types, which breaks LeaderAction stale checking. For LeaderAction,
Node() returns the worker node where segment resides, but the task is
executed on the leader node (delegator).

When syncing segments from RO worker nodes to a RW delegator, using
action.Node() incorrectly marks the task as stale because the worker is
RO, even though the leader is RW and the task should proceed.

This fix:
- Uses leaderID instead of Node() for LeaderAction stale checking
- Adds detailed comments explaining the distinction
- Adds unit tests covering the RO worker with RW leader scenario

Signed-off-by: Wei Liu <wei.liu@zilliz.com>

2026-01-06 18:27:24 +08:00

..

action.go

feat: query coord support segment reopen when manifest path changes (#46394)

2025-12-17 22:15:16 +08:00

executor.go

fix: remove the streamingnode checking when loading segment (#45859)

2025-11-28 10:09:08 +08:00

mock_scheduler.go

fix: mockery too unavailable after upgrade golang version (#41481)

2025-04-24 10:46:43 +08:00

OWNERS

Add OWNERS files for querycoordv2 sub pkgs (#23489)

2023-04-18 15:52:30 +08:00

scheduler.go

fix: Use leaderID for LeaderAction stale check in scheduler (#46785)

2026-01-06 18:27:24 +08:00

task_test.go

fix: Use leaderID for LeaderAction stale check in scheduler (#46785)

2026-01-06 18:27:24 +08:00

task.go

feat: support dropping index without releasing collection (#42941)

2025-09-02 16:17:52 +08:00

utils_test.go

fix: LoadSegment failed for dup field mmap.enabel props (#44465)

2025-09-22 14:40:06 +08:00

utils.go

feat: query coord support segment reopen when manifest path changes (#46394)

2025-12-17 22:15:16 +08:00