NPEs during virtual database shutdown when one backend state is UNKNOWN
-----------------------------------------------------------------------
Key: SEQUOIA-900
URL: https://forge.continuent.org/jira/browse/SEQUOIA-900
Project: Sequoia
Type: Bug
Components: Core
Versions: Sequoia 2.10.4
Reporter: Mathieu Peltier
Assigned to: Mathieu Peltier
Fix For: Sequoia 2.10.5, Sequoia 3.0
Test case 1:
- Start RAIDB1 cluster of 2 controllers with 1 backend, load vdb and enable all
backends
- Restore first backend of controller 1 from a bad dump: backend state should
become UNKNOWN
- Shutdown controller 1 <-- NPE was thrown here
Caused by: java.lang.RuntimeException: No deadlock detection thread to stop on
backend c1b1
at
org.continuent.sequoia.controller.loadbalancer.BackendTaskQueues.terminateDeadlockDetectionThread(Back\
endTaskQueues.java:1902)
at
org.continuent.sequoia.controller.backend.DatabaseBackend.terminateDeadlockDetectionThread(DatabaseBac\
kend.java:2996)
at
org.continuent.sequoia.controller.loadbalancer.raidb1.RAIDb1.terminateThreadsAndConnections(RAIDb1.jav\
a:2049)
at
org.continuent.sequoia.controller.loadbalancer.raidb1.RAIDb1.disableBackend(RAIDb1.java:2020)
at
org.continuent.sequoia.controller.virtualdatabase.protocol.DisableBackendsAndSetCheckpoint.handleMessa\
geMultiThreaded(DisableBackendsAndSetCheckpoint.java:201)
Test case 2:
- Start RAIDB1 cluster of 2 controllers with 1 backend, load vdb and enable all
backends
- Kill first backend of controller 1
- Execute write request so that backend 1 is automatically disabled
- Restart virtual database on controller 1
- Check that state of backend 1 of controller 1 is UNKNOWN
- Shutdown controller 1 <-- NPE
Caused by: java.lang.NullPointerException
at
org.continuent.sequoia.controller.backend.DatabaseBackend.terminateDeadlockDetectionThread(DatabaseBac\
kend.java:2996)
at
org.continuent.sequoia.controller.loadbalancer.raidb1.RAIDb1.terminateThreadsAndConnections(RAIDb1.jav\
a:2049)
at
org.continuent.sequoia.controller.loadbalancer.raidb1.RAIDb1.disableBackend(RAIDb1.java:2020)
at
org.continuent.sequoia.controller.virtualdatabase.protocol.DisableBackendsAndSetCheckpoint.handleMessa\
geMultiThreaded(DisableBackendsAndSetCheckpoint.java:201)
... 2 more
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://forge.continuent.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
