[infinispan-issues] [JBoss JIRA] (ISPN-2688) BaseDistributionInterceptor and TxDistributionInterceptor do not detect properly if the key needs to be fetched remotely

[Mircea Markus (JIRA)]

     [ https://issues.jboss.org/browse/ISPN-2688?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Mircea Markus updated ISPN-2688:
--------------------------------

    Fix Version/s: 5.2.1

    
> BaseDistributionInterceptor and TxDistributionInterceptor do not detect properly if the key needs to be fetched remotely
> ------------------------------------------------------------------------------------------------------------------------
>
>                 Key: ISPN-2688
>                 URL: https://issues.jboss.org/browse/ISPN-2688
>             Project: Infinispan
>          Issue Type: Bug
>          Components: State transfer
>    Affects Versions: 5.2.0.Beta6
>            Reporter: Dan Berindei
>            Assignee: Mircea Markus
>            Priority: Critical
>             Fix For: 5.2.1, 5.3.0.Final
>
>         Attachments: stlot.log.gz
>
>
> BaseDistributionInterceptor.shouldFetchFromRemote() and TxDistributionInterceptor.remoteGetAndStoreInL1() can mistakenly decide not to fetch remotely because they check the presence of the key in data container. The key may be there _now_ but it was not there before the local execution if state transfer was in progress for this key. So it should be re-fetched rather than use the null result.
> This makes StateTransferLargeObjectTest.testForFailure fail randomly.
> The failure appears because the state transfer has not finished, yet the distribution interceptor doesn't go remotely for the key:
> {noformat}
> 14:06:45,872 TRACE (asyncTransportThread-1,NodeA:___defaultcache) [StateTransferManagerImpl] Installing new cache topology CacheTopology{id=7, currentCH=DefaultConsistentHash{numSegments=60, numOwners=3, members=[NodeA-52814, NodeB-62397, NodeC-63995], owners={0: 0 1 2, 1: 0 1 2, 2: 0 1 2, 3: 0 1 2, 4: 0 1 2, 5: 0 1 2, 6: 0 1 2, 7: 0 1 2, 8: 0 1 2, 9: 0 1 2, 10: 0 1 2, 11: 0 1 2, 12: 0 2, 13: 0 2, 14: 0 2, 15: 0 1, 16: 0 1, 17: 0 1, 18: 0 1, 19: 0 1, 20: 2 0, 21: 2 0, 22: 2 0, 23: 2 0, 24: 2 0, 25: 2 0, 26: 2 0, 27: 2 0, 28: 2 0, 29: 2 0, 30: 1 0 2, 31: 1 0 2, 32: 1 0 2, 33: 1 2, 34: 1 0, 35: 1 2, 36: 1 0, 37: 1 2, 38: 1 0, 39: 1 2, 40: 1 0, 41: 1 2, 42: 1 0, 43: 1 2, 44: 1 0, 45: 1 0, 46: 1 0, 47: 1 0, 48: 1 0, 49: 1 2, 50: 2 1, 51: 2 1, 52: 2 1, 53: 2 1, 54: 2 1, 55: 2 1, 56: 2 1, 57: 2 1, 58: 2 0, 59: 2 0}, pendingCH=DefaultConsistentHash{numSegments=60, numOwners=3, members=[NodeA-52814, NodeB-62397, NodeC-63995], owners={0: 0 1 2, 1: 0 1 2, 2: 0 1 2, 3: 0 1 2, 4: 0 1 2, 5: 0 1 2, 6: 0 1 2, 7: 0 1 2, 8: 0 1 2, 9: 0 1 2, 10: 0 1 2, 11: 0 1 2, 12: 0 2 1, 13: 0 2 1, 14: 0 2 1, 15: 0 1 2, 16: 0 1 2, 17: 0 1 2, 18: 0 1 2, 19: 0 1 2, 20: 2 0 1, 21: 2 0 1, 22: 2 0 1, 23: 2 0 1, 24: 2 0 1, 25: 2 0 1, 26: 2 0 1, 27: 2 0 1, 28: 2 0 1, 29: 2 0 1, 30: 1 0 2, 31: 1 0 2, 32: 1 0 2, 33: 1 2 0, 34: 1 0 2, 35: 1 2 0, 36: 1 0 2, 37: 1 2 0, 38: 1 0 2, 39: 1 2 0, 40: 1 0 2, 41: 1 2 0, 42: 1 0 2, 43: 1 2 0, 44: 1 0 2, 45: 1 0 2, 46: 1 0 2, 47: 1 0 2, 48: 1 0 2, 49: 1 2 0, 50: 2 1 0, 51: 2 1 0, 52: 2 1 0, 53: 2 1 0, 54: 2 1 0, 55: 2 1 0, 56: 2 1 0, 57: 2 1 0, 58: 2 0 1, 59: 2 0 1}} on cache ___defaultcache
> 14:06:46,287 TRACE (OOB-9,ISPN,NodeA-52814:___defaultcache) [StateConsumerImpl] Received keys [0, 2, 10, 94, 103, 117, 187, 189, 288, 305, 307, 376, 481, 487, 502, 729, 771, 994] for segment 50 of cache ___defaultcache from node NodeB-62397
> 14:06:46,338 INFO  (testng-StateTransferLargeObjectTest:) [StateTransferLargeObjectTest] ----Running a get on 10
> 14:06:46,351 TRACE (OOB-9,ISPN,NodeA-52814:___defaultcache ___defaultcache) [InvocationContextInterceptor] Invoked with command PutKeyValueCommand{key=10, value=org.infinispan.statetransfer.BigObject at 6ed3126d, flags=[CACHE_MODE_LOCAL, SKIP_REMOTE_LOOKUP, PUT_FOR_STATE_TRANSFER, SKIP_SHARED_CACHE_STORE, SKIP_OWNERSHIP_CHECK, IGNORE_RETURN_VALUES, SKIP_XSITE_BACKUP], putIfAbsent=false, lifespanMillis=-1, maxIdleTimeMillis=-1, successful=true} and InvocationContext [org.infinispan.context.impl.LocalTxInvocationContext at 2c6dd013]
> 14:06:46,358 TRACE (testng-StateTransferLargeObjectTest:___defaultcache) [GetKeyValueCommand] Entry not found
> 14:06:46,358 TRACE (testng-StateTransferLargeObjectTest:___defaultcache) [BaseDistributionInterceptor] Not doing a remote get for key 10 since entry is mapped to current node (NodeA-52814), or is in L1.  Owners are [NodeC-63995, NodeB-62397, NodeA-52814]
> 14:06:46,359 ERROR (testng-StateTransferLargeObjectTest:) [UnitTestTestNGListener] Test testForFailure(org.infinispan.statetransfer.StateTransferLargeObjectTest) failed.
> java.lang.AssertionError: expected object to not be null
> 	at org.testng.Assert.fail(Assert.java:89)
> 	at org.testng.Assert.assertNotNull(Assert.java:399)
> 	at org.testng.Assert.assertNotNull(Assert.java:384)
> 	at org.infinispan.statetransfer.StateTransferLargeObjectTest.assertValue(StateTransferLargeObjectTest.java:145)
> 	at org.infinispan.statetransfer.StateTransferLargeObjectTest.testForFailure(StateTransferLargeObjectTest.java:115)
> {noformat}

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira