-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: Replication
-
None
-
Fully Compatible
-
ALL
-
Repl 2020-10-19, Repl 2020-11-02
-
46
In 4.4 (but no other version), we kill the syncing node with a LOG_FATAL when an initial sync is happening and it is removed from the configuration. We attempt to cleanly shut down the initial syncer first. However, if this results in the initial syncer running out of attempts, we will instead get a LOG_FATAL with that error instead.
We should just LOG_FATAL with the error about being removed; there's no point in trying to cleanly shut down when we're going to kill the server anyway.
(This was uncovered when cancellation of the initial syncer was fixed by SERVER-50320)
- causes
-
SERVER-51726 Maintain status code compatibility when main executor is shut down
- Closed
- is related to
-
SERVER-50320 Fix cancellation races in initial syncer
- Closed