[infinispan-issues] [JBoss JIRA] (ISPN-8782) Occasional NPE on PersistentUUIDManagerImpl.mapAddresses

[Dan Berindei (JIRA)]

    [ https://issues.jboss.org/browse/ISPN-8782?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13559141#comment-13559141 ] 

Dan Berindei commented on ISPN-8782:
------------------------------------

The problem is that {{LocalTopologyManagerImpl.stop()}} removes the local node's persistent UUID from the map, and this happens before {{ClusterTopologyManagerImpl}} stops.

> Occasional NPE on PersistentUUIDManagerImpl.mapAddresses
> --------------------------------------------------------
>
>                 Key: ISPN-8782
>                 URL: https://issues.jboss.org/browse/ISPN-8782
>             Project: Infinispan
>          Issue Type: Bug
>          Components: Core
>            Reporter: Sanne Grinovero
>            Assignee: Dan Berindei
>             Fix For: 9.3.0.Final
>
>
> When running the testsuite I occasionally see stacktraces such as:
> 23:47:50,649 FATAL (transport-thread-RetryMechanismTest[null, tx=false]-NodeT-p64820-t5) [PersistentUUIDManagerImpl] Cannot find mapping for address RetryMechanismTest[null, tx=false]-NodeT-51212 java.lang.NullPointerException
> 	at org.infinispan.topology.PersistentUUIDManagerImpl.mapAddresses(PersistentUUIDManagerImpl.java:70)
> 	at org.infinispan.topology.ClusterCacheStatus.startQueuedRebalance(ClusterCacheStatus.java:824)
> 	at org.infinispan.topology.ClusterCacheStatus.queueRebalance(ClusterCacheStatus.java:125)
> 	at org.infinispan.partitionhandling.impl.PreferAvailabilityStrategy.onPartitionMerge(PreferAvailabilityStrategy.java:232)
> 	at org.infinispan.topology.ClusterCacheStatus.doMergePartitions(ClusterCacheStatus.java:597)
> 	at org.infinispan.topology.ClusterTopologyManagerImpl.lambda$recoverClusterStatus$6(ClusterTopologyManagerImpl.java:519)
> 	at org.infinispan.executors.LimitedExecutor.runTasks(LimitedExecutor.java:144)
> 	at org.infinispan.executors.LimitedExecutor.access$100(LimitedExecutor.java:33)
> 	at org.infinispan.executors.LimitedExecutor$Runner.run(LimitedExecutor.java:174)
> They seem harmless as it seems like the test results aren't affected, but it's still looking quite bad. Would be best if someone had a look.



--
This message was sent by Atlassian JIRA
(v7.5.0#75005)