Details
-
Bug
-
Resolution: Fixed
-
Critical
-
5.5.0, 5.5.1, 5.5.2, 5.5.3
-
Untriaged
-
Unknown
Description
[Cloned from MB-33750 to track fixing that issue in 5.5.4]
We have seen a couple of cases in which a rogue lease acquirer process is left running on a node that used to be the orchestrator. This can happen via a race condition that's exposed if a supervisor that is shutting down normally, is for some reason brutally shutdown leaving an orphan child process.
The rogue lease acquirer interferes with the operation of the orchestrator causing, for instance, rebalance to fail.
Attachments
Issue Links
- relates to
-
MB-33750 Ensure that when a new node picks up orchetratorship, the lease_acquirer process on the old orchestrator is shutdown
- Closed
For Gerrit Dashboard: MB-33853 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
107989,2 | MB-33853 Don't use async:perform for leader_lease_acquire_worker. | vulcan | ns_server | Status: MERGED | +2 | +1 |
107997,1 | Merge remote-tracking branch 'couchbase/vulcan' into merge | alice | ns_server | Status: MERGED | +2 | +1 |
107999,3 | MB-33853 Adjust master advertised version to 5.5.4. | vulcan | ns_server | Status: MERGED | +2 | +1 |
108039,1 | Merge remote-tracking branch 'couchbase/vulcan' into merge | alice | ns_server | Status: MERGED | +2 | +1 |
108465,1 | Merge remote-tracking branch 'couchbase/alice' into merge | master | ns_server | Status: MERGED | +2 | +1 |