-
Type: Bug
-
Resolution: Duplicate
-
Priority: Major - P3
-
None
-
Affects Version/s: 3.2.5
-
Component/s: Replication
-
None
-
Replication
-
Fully Compatible
-
ALL
-
-
Repl 15 (06/03/16), Repl 16 (06/24/16)
-
(copied to CRM)
MongoDB server always complain "Fatal assertion 18750" after abnormal mongod termination.
Sometimes mongod server is hang and no response, at that time mongod server almost not response (SERVER-23778). So I can't terminate mongod normally.
So I kill mongod with SIGKILL. Or sometimes linux kill mongod because excessive memory usage. Anyway always mongod complain this assertion error and stop replicating during startup after mongod abnormal shutdown.
2016-04-20T22:51:04.192+0900 I REPL [ReplicationExecutor] This node is test02-mongo3:27017 in the config 2016-04-20T22:51:04.193+0900 I REPL [ReplicationExecutor] Member test02-mongo1:27017 is now in state SECONDARY 2016-04-20T22:51:04.193+0900 I REPL [ReplicationExecutor] Member test02-mongo2:27017 is now in state PRIMARY 2016-04-20T22:51:05.183+0900 I REPL [ReplicationExecutor] syncing from: test02-mongo2:27017 2016-04-20T22:51:05.184+0900 I REPL [SyncSourceFeedback] setting syncSourceFeedback to test02-mongo2:27017 2016-04-20T22:51:05.185+0900 I ASIO [NetworkInterfaceASIO-BGSync-0] Successfully connected to test02-mongo2:27017 2016-04-20T22:51:05.187+0900 I REPL [rsBackgroundSync] Starting rollback due to OplogStartMissing: our last op time fetched: (term: 15, timestamp: Apr 20 10:20:00:7). source's GTE: (term: 16, timestamp: Apr 20 10:20:00:7) hashes: (-157997914050892151/-8410528485294631657) 2016-04-20T22:51:05.187+0900 I - [rsBackgroundSync] Fatal assertion 18750 UnrecoverableRollbackError: need to rollback, but in inconsistent state. minvalid: (term: 16, timestamp: Apr 20 10:20:01:2) > our last optime: (term: 15, timestamp: Apr 20 10:20:00:7) 2016-04-20T22:51:05.187+0900 I - [rsBackgroundSync] ***aborting after fassert() failure
Is this normal behavior or there might be something wrong on my configuration.
And is there any fast way to recover without full re-sync ?
I found a few bug-report from JIRA, but they say it's already fixed.
So I report this issue again.
Thanks.
- duplicates
-
SERVER-7275 node can't roll back if behind minValid
- Closed
-
SERVER-7200 use oplog as op buffer on secondaries
- Closed