-
Bug
-
Resolution: Done
-
Major
-
10.1.0.Final
ClusterPublisherManagerImpl assumes every segment has a primary owner, but that's not true for scattered caches:
11:02:00,693 TRACE (jgroups-5,Test-NodeA-43570:[]) [ClusterPublisherManagerImpl] Received suspect exception for id Test-NodeA-43570#8950 from node Test-NodeB-52654 when requesting segments {1} 11:02:00,693 TRACE (jgroups-5,Test-NodeA-43570:[]) [InnerPublisherSubscription] Completing subscription InnerPublisher-1432292347{requestId=Test-NodeA-43570#8950, topologyId=7} 11:02:00,693 TRACE (jgroups-5,Test-NodeA-43570:[]) [ClusterPublisherManagerImpl] Segments {1} not completed - retrying 11:02:00,693 TRACE (jgroups-5,Test-NodeA-43570:[]) [ClusterPublisherManagerImpl] Targets determined to be {null={1}} on topology 8 11:02:00,693 TRACE (jgroups-5,Test-NodeA-43570:[]) [InnerPublisherSubscription] Subscribed to Test-NodeA-43570#8950 via 1 11:02:00,694 TRACE (jgroups-5,Test-NodeA-43570:[]) [ClusterPublisherManagerImpl] Request: Test-NodeA-43570#8950 is initiating publisher request with batch size 53 from null in segments {1}
JGroupsTransport.invokeCommand(null, ...) throws a NullPointerException, but it's not logged and the publisher doesn't retry.
11:02:10,809 ERROR (testng-Test:[]) [TestSuiteProgress] Test failed: org.infinispan.scattered.stream.ScatteredStreamIteratorTest.verifyNodeLeavesAfterSendingBackSomeData[SCATTERED_SYNC, tx=false] java.util.concurrent.TimeoutException: null at java.util.concurrent.FutureTask.get(FutureTask.java:204) ~[?:?] at org.infinispan.stream.DistributedStreamIteratorTest.verifyNodeLeavesAfterSendingBackSomeData(DistributedStreamIteratorTest.java:158) ~[test-classes/:?]
- is caused by
-
ISPN-13399 Scattered cache prefetch publisher blocks
- Closed
- is related to
-
ISPN-13163 Scattered cache key iteration is slow
- Closed
- relates to
-
ISPN-13134 ArithmeticException: / by zero in ClusterPublisherManagerImpl
- Closed