Details
-
Bug
-
Resolution: Fixed
-
Critical
-
7.1.0, 7.1.1, 7.1.2, 7.1.3
-
Untriaged
-
0
-
Yes
Description
Checkpoint Manager with P2P (MB-9982) can experience potential hang due to the handler not watching out for when checkpoint manager exits.
The stack trace of interest would be:
# 0x9c2d92 github.com/couchbase/goxdcr/pipeline_svc.respToGcCh+0x772 /home/couchbase/jenkins/workspace/couchbase-server-unix/goproj/src/github.com/couchbase/goxdcr/pipeline_svc/checkpoint_manager.go:2769
|
# 0x9c2d0f github.com/couchbase/goxdcr/pipeline_svc.(*CheckpointManager).stopTheWorldAndMergeCkpts+0x6ef /home/couchbase/jenkins/workspace/couchbase-server-unix/goproj/src/github.com/couchbase/goxdcr/pipeline_svc/checkpoint_manager.go:2762
|
# 0x9c9594 github.com/couchbase/goxdcr/pipeline_svc.(*CheckpointManager).periodicMergerImpl+0x194 /home/couchbase/jenkins/workspace/couchbase-server-unix/goproj/src/github.com/couchbase/goxdcr/pipeline_svc/checkpoint_manager.go:3487
|
As a result, over time, there will be a large number of Checkpoint Manager zombies lying around and unable to be cleaned up, leading to memory leak.