[ https://issues.jboss.org/browse/ISPN-5054?page=com.atlassian.jira.plugin.... ] Dan Berindei commented on ISPN-5054: ------------------------------------ The problem seems to be that the test doesn't wait for the partitions to actually enter degraded mode before starting the iteration. In replicated mode all the keys are local, so it can finish successfully between the JGroups view being installed and the cache topology being updated. {noformat} 11:23:21,567 TRACE (testng-EntryRetrieverReplPartitionHandlingTest:) [BasePartitionHandlingTest] New views installed 11:23:21,567 TRACE (testng-EntryRetrieverReplPartitionHandlingTest:) [InvocationContextInterceptor] Invoked with command org.infinispan.commands.read.EntryRetrievalCommand@318802 and InvocationContext [org.infinispan.context.impl.NonTxInvocationContext@32a81ceb] 11:23:21,568 ERROR (testng-EntryRetrieverReplPartitionHandlingTest:) [UnitTestTestNGListener] Test testRetrievalWhenPartitionIsDegrated(org.infinispan.partitionhandling.EntryRetrieverReplPartitionHandlingTest) failed. org.testng.TestException: Method EntryRetrieverDistPartitionHandlingTest.testRetrievalWhenPartitionIsDegrated()[pri:0, instance:org.infinispan.partitionhandling.EntryRetrieverReplPartitionHandlingTest@4010d494] should have thrown an exception of class org.infinispan.partitionhandling.AvailabilityException 11:23:21,568 ERROR (transport-thread-NodeA-p30559-t4:) [PreferConsistencyStrategy] ISPN000318: Cache ___defaultcache lost at least half of the stable members, assuming a network split and entering degraded mode. Current members are [NodeA-23274, NodeB-35032], lost members are [NodeC-63977, NodeD-22740], stable members are [NodeA-23274, NodeB-35032, NodeC-63977, NodeD-22740] {noformat} -- This message was sent by Atlassian JIRA (v6.3.11#6341)