Hi Slava, Thank you very much for your answer.
Juan From: Вячеслав Коптилин [mailto:slava.kopti...@gmail.com] Sent: Monday, July 10, 2017 12:21 PM To: user@ignite.apache.org Subject: Re: TcpDiscoverySpi worker thread failed with assertion error Hi Juan, U.currentTimeMillis(), which is based on System.currentTimeMillis(), is not monotonic, so there is no guarantee that returned values will always be greater than all previous calls. please see the following ticket: https://issues.apache.org/jira/browse/IGNITE-5562 Thanks, Slava. 2017-07-10 12:34 GMT+03:00 Juan Barani <juan.bar...@unitedplanet.com<mailto:juan.bar...@unitedplanet.com>>: Hi everyone, We are testing a new application running Ignite 2.0 with 3 nodes running on Azure platform. We left the cluster running a couple of days without load, and randomly one of the Nodes stopped with the following stacktrace: INFO 2017-07-07T23:07:28,475 - org.apache.ignite.internal.IgniteKernal%IntrexxDataGrid[grid-timeout-worker-#15%IntrexxDataGrid%] FreeList [name=IntrexxDataGrid, buckets=256, dataPages=3, reusePages=0] INFO 2017-07-07T23:07:28,475 - org.apache.ignite.internal.IgniteKernal%IntrexxDataGrid[grid-timeout-worker-#15%IntrexxDataGrid%] FreeList [name=IntrexxDataGrid, buckets=256, dataPages=9, reusePages=0] ERROR 2017-07-07T23:08:10,632 - org.apache.ignite.spi.discovery.tcp.TcpDiscoverySpi[tcp-disco-msg-worker-#2%IntrexxDataGrid%] TcpDiscoverSpi's message worker thread failed abnormally. Stopping the node in order to prevent cluster wide instability. java.lang.AssertionError: -21 at org.apache.ignite.spi.discovery.tcp.internal.TcpDiscoveryStatistics.onMessageSent(TcpDiscoveryStatistics.java:317) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.sendMessageAcrossRing(ServerImpl.java:2980) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.processMetricsUpdateMessage(ServerImpl.java:5001) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.processMessage(ServerImpl.java:2562) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.processMessage(ServerImpl.java:2364) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$MessageWorkerAdapter.body(ServerImpl.java:6460) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.body(ServerImpl.java:2450) [ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.IgniteSpiThread.run(IgniteSpiThread.java:62) [ignite-core-2.0.0.jar:2.0.0] ERROR 2017-07-07T23:08:10,661 - org.apache.ignite.spi.discovery.tcp.TcpDiscoverySpi[tcp-disco-msg-worker-#2%IntrexxDataGrid%] Runtime error caught during grid runnable execution: IgniteSpiThread [name=tcp-disco-msg-worker-#2%IntrexxDataGrid%] java.lang.AssertionError: -21 at org.apache.ignite.spi.discovery.tcp.internal.TcpDiscoveryStatistics.onMessageSent(TcpDiscoveryStatistics.java:317) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.sendMessageAcrossRing(ServerImpl.java:2980) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.processMetricsUpdateMessage(ServerImpl.java:5001) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.processMessage(ServerImpl.java:2562) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.processMessage(ServerImpl.java:2364) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$MessageWorkerAdapter.body(ServerImpl.java:6460) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.discovery.tcp.ServerImpl$RingMessageWorker.body(ServerImpl.java:2450) ~[ignite-core-2.0.0.jar:2.0.0] at org.apache.ignite.spi.IgniteSpiThread.run(IgniteSpiThread.java:62) I researched online, and it seems that the thread fails when there is a long gc pause (>10 s), or not so much memory. In all those cases it comes also an InterruptedException. However, that does not seem to be the case, since we do have enough memory, and no load whatsoever. What strikes me the most is the assertion error: TcpDiscoverSpi's message worker thread failed abnormally. Stopping the node in order to prevent cluster wide instability. java.lang.AssertionError: -21 This is an assertion in method onMessageSent: assert time >= 0 : time; which in turns comes from (ServerImpl.java:2980): spi.stats.onMessageSent(msg, tstamp0 - tstamp); tstamp0 and tsamp are in turn values returned from U.currentTimeMillis(); which if I understand correctly should return an ever increasing number (in 10 seconds interval). Hence I cannot figure out why this asserted value is negative. Can someone give us a hint what we should do? Thanks! Juan
