-
Type: Bug
-
Resolution: Done
-
Priority: Major - P3
-
None
-
Affects Version/s: 3.2.10
-
Component/s: Replication
-
Labels:None
-
Server Triage
-
ALL
-
- Create a MongoDB replicaset
- Do some write requests to the replicaset
- Wait a long time ( bigger than 30 seconds )
- Do some write requests again, and the secondary nodes may re-evaluate sync source
I got this warning in the log file:
"Fri Jun 29 22:00:02.585 I REPL [ReplicationExecutor] re-evaluating sync source because our current sync source's most recent OpTime is (term: 2, timestamp: Jun 29 21:50:07:1) which is more than 30s behind member xxx.xxx.xxx.xxx:7005 whose most recent OpTime is (term: 2, timestamp: Jun 29 22:00:01:1)"
The write majority request is hung up and timeout when the two secondaries are both doing "re-evaluating sync source"
The current sync source is primary node, so I think its "OpTime" should not behind the secondary nodes.