[
https://issues.apache.org/jira/browse/GEODE-7921?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17074657#comment-17074657
]
ASF subversion and git services commented on GEODE-7921:
--------------------------------------------------------
Commit 2ac3de788502d1a40189bed06ce86779ee47bdf8 in geode's branch
refs/heads/develop from Bruce Schuchardt
[ https://gitbox.apache.org/repos/asf?p=geode.git;h=2ac3de7 ]
GEODE-7921: NullPointerExceptions logged during auto-reconnect (#4898)
Don't deliver cache-level messages that were queued while disconnected
for auto-reconnect. During auto-reconnect there is a QuorumChecker that
receives messages on the jgroups channel and tries to establish
communications with a quorum of the old membership view. It may also
get JoinRequest messages and other membership-level messages but I
observed one case where it also queued cache-level messages when the
property disable-tcp was set to true (funnelling all comms through
jgroups).
I also added some null checks to the LatestLastAccessTimeMessage and a
small test for that.
> NullPointerExceptions logged during auto-reconnect
> --------------------------------------------------
>
> Key: GEODE-7921
> URL: https://issues.apache.org/jira/browse/GEODE-7921
> Project: Geode
> Issue Type: Bug
> Components: membership
> Reporter: Bruce J Schuchardt
> Assignee: Bruce J Schuchardt
> Priority: Major
> Fix For: 1.13.0
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> In a test with disable-tcp=true I found a huge number of NPEs logged when a
> server auto-reconnected.
> {noformat}
> [fatal 2020/03/25 15:26:57.136 PDT <Pooled Message Processor 2> tid=0x17f]
> Uncaught exception processing LatestLastAccessTimeMessage@546d3459
> processorId=0
> sender=rs-FullRegression25213428a3i32xlarge-hydra-client-22(gemfire3_host1_32677:32677)<ec><v5>:41009
> java.lang.NullPointerException
> at
> org.apache.geode.internal.cache.LatestLastAccessTimeMessage.process(LatestLastAccessTimeMessage.java:66)
> at
> org.apache.geode.distributed.internal.DistributionMessage.scheduleAction(DistributionMessage.java:376)
> at
> org.apache.geode.distributed.internal.DistributionMessage$1.run(DistributionMessage.java:440)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at
> org.apache.geode.distributed.internal.ClusterOperationExecutors.runUntilShutdown(ClusterOperationExecutors.java:442)
> at
> org.apache.geode.distributed.internal.ClusterOperationExecutors.doProcessingThread(ClusterOperationExecutors.java:389)
> at
> org.apache.geode.logging.internal.executors.LoggingThreadFactory.lambda$newThread$0(LoggingThreadFactory.java:119)
> at java.base/java.lang.Thread.run(Thread.java:834)
> {noformat}
> These came from a queue of messages built up by the QuorumChecker that were
> processed before a new cache was built. The QuorumChecker shouldn't be
> queueing cache operations. It only needs to queue membership messages.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
