Hi, We have finished a yarn application and deploy it to hadoop 2.6.0 cluster. But if one machine in cluster is down. Our application will hang on NMClientAsyncImpl.stop() (util we kill it 5 hour later). The last log is:
953851 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a14k17307.em21.tbsite.net:8041 953852 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a24a21449.em21.tbsite.net:8041 953853 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : e92e09611.em21.tbsite.net:8041 953854 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a24a21477.em21.tbsite.net:8041 953855 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : e92e09590.em21.tbsite.net:8041 953856 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a14k17314.em21.tbsite.net:8041 953857 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a24a21474.em21.tbsite.net:8041 953858 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : e92e09574.em21.tbsite.net:8041 953859 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : e92e09579.em21.tbsite.net:8041 953860 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a14k17332.em21.tbsite.net:8041 953861 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a14k17300.em21.tbsite.net:8041 953862 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : e93g17455.em21.tbsite.net:8041 953863 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : e92e09595.em21.tbsite.net:8041 953864 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a14k17309.em21.tbsite.net:8041 953865 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a14k17294.em21.tbsite.net:8041 953866 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a14k17316.em21.tbsite.net:8041 953867 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : e92e09592.em21.tbsite.net:8041 953868 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a24a21472.em21.tbsite.net:8041 953869 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : a14k17337.em21.tbsite.net:8041 953870 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : e92e09580.em21.tbsite.net:8041 953871 17/09/27 10:46:52 INFO impl.ContainerManagementProtocolProxy: Opening proxy : e92e09619.em21.tbsite.net:8041 Is any way to set a timeout for NMClientAsyncImpl when call it stop? we init it by : NMClientAsync nmClient = NMClientAsync.createNMClientAsync( nmCallbackHandler); and then : yarnConfig = new YarnConfiguration(); nmClient .init(yarnConfig); nmClient .start(); our code not modify nmClient.