[GEODE-9944] NPE could occur if HARegion is being created again but not fully initialized

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.15.0
• Fix Version/s: 1.15.0
• Component/s: client queues
• Labels: GeodeOperationAPI pull-request-available unreleased
• Assignee: Eric Shu
• Reporter: Eric Shu
• Created: 11/Jan/22 22:06
• Updated: 22/Jun/22 20:45
• Resolved: 13/Jan/22 17:59

Description

The stack trace for the NPE is:

fatal 2022/01/08 12:45:33.175 PST bridgegemfire7_host1_25026 <Pooled Message Processor 7> tid=0x148] Uncaught exception processing QueueSynchronizationProcessor$QueueSynchronizationMessage@340d586b processorId=4029 sender=rs-FullRegression15142100a3i3large-hydra-client-18(bridgegemfire6_host1_25009:25009)<ec><v5>:41006
java.lang.NullPointerException
        at org.apache.geode.internal.cache.ha.QueueSynchronizationProcessor$QueueSynchronizationMessage.getDispatchedEvents(QueueSynchronizationProcessor.java:160)
        at org.apache.geode.internal.cache.ha.QueueSynchronizationProcessor$QueueSynchronizationMessage.process(QueueSynchronizationProcessor.java:127)
        at org.apache.geode.distributed.internal.DistributionMessage.scheduleAction(DistributionMessage.java:376)
        at org.apache.geode.distributed.internal.DistributionMessage$1.run(DistributionMessage.java:441)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at org.apache.geode.distributed.internal.ClusterOperationExecutors.runUntilShutdown(ClusterOperationExecutors.java:444)
        at org.apache.geode.distributed.internal.ClusterOperationExecutors.doProcessingThread(ClusterOperationExecutors.java:391)
        at org.apache.geode.logging.internal.executors.LoggingThreadFactory.lambda$newThread$0(LoggingThreadFactory.java:120)
        at java.lang.Thread.run(Thread.java:748)

This could occur when a client is not able to re-auth in time, and the server removes the HARegionQueue to the client. When the client is able to re-auth later, the new HARegionQueue is created again. There is a race that when the server process the above message, the HARegionQueue is not fully initialized and cause this NPE.

Issue Links

• links to: GitHub Pull Request #7259