-
Type:
Bug
-
Resolution: Fixed
-
Priority:
Major - P3
-
Affects Version/s: None
-
Component/s: Replication
-
None
-
Fully Compatible
-
ALL
-
v5.0
-
Repl 2021-05-31
-
42
-
None
-
3
-
None
-
None
-
None
-
None
-
None
-
None
When crashAfterRollbackTruncation is false, we didn't wait for rollback to begin, so it's possible that when calling awaitNodesAgreeOnPrimary(), secondary1 has not started rollback, but already saw the new primary, and thus awaitNodesAgreeOnPrimary() will not see the network exception, but after this, rollback starts and connection closed, so the replSetFreeze() command sees the network exception. We should wait for rollback to complete before stepping it up.