Details
-
Bug
-
Resolution: Incomplete
-
Major
-
5.0.0
-
Centos 7
-
Untriaged
-
Unknown
Description
1. Create a cluster with 3 nodes cluster with a bucket in it. Have a large number of documents in the bucket. Have background mutations in progress throughout the test
2. Enable autofailover with 5 sec as timeout
3. On any of the non orchestrator node, restart the couchbase server. Wait for the server to be back up.
We expect the failover to be initated if the node is not back up within 5 sec or bucket is not warmed up within 5 sec. But the failover is initiated after ~15 sec consistently.
http://qa.sc.couchbase.com/job/autofailover-restart-server/10/consoleFull has the tests where the tests took longer than 5 sec to initiate the timeout.
CBCollectinfo and test logs for failed tests can be found here at /data/workspace/autofailover-restart-server/logs/testrunner-17-Mar-05_17-28-17/test_<test no> at 172.23.105.89 (user:root, password:couchbase) machine. Test logs can be found