[ https://issues.apache.org/jira/browse/YARN-4613?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15110286#comment-15110286 ]
Takashi Ohnishi commented on YARN-4613: --------------------------------------- I think this fails by too early checking before the node state really changed. {code:title=TestClientRMService.java} // Now make the node unhealthy. node.nodeHeartbeat(false); // Call again nodeReports = client.getClusterNodes(request).getNodeReports(); Assert.assertEquals("Unhealthy nodes should not show up by default", 0, nodeReports.size()); {code} How about adding a status check of NodeManager? I will attach a patch. > TestClientRMService#testGetClusterNodes fails occasionally > ---------------------------------------------------------- > > Key: YARN-4613 > URL: https://issues.apache.org/jira/browse/YARN-4613 > Project: Hadoop YARN > Issue Type: Sub-task > Components: test > Environment: jenkins > Reporter: Jason Lowe > > TestClientRMService has been failing occasionally on precommits. Sample > failure: > {noformat} > Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=768m; > support was removed in 8.0 > Running org.apache.hadoop.yarn.server.resourcemanager.TestClientRMService > Tests run: 19, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 15.077 sec > <<< FAILURE! - in > org.apache.hadoop.yarn.server.resourcemanager.TestClientRMService > testGetClusterNodes(org.apache.hadoop.yarn.server.resourcemanager.TestClientRMService) > Time elapsed: 0.551 sec <<< FAILURE! > java.lang.AssertionError: Unhealthy nodes should not show up by default > expected:<0> but was:<1> > at org.junit.Assert.fail(Assert.java:88) > at org.junit.Assert.failNotEquals(Assert.java:743) > at org.junit.Assert.assertEquals(Assert.java:118) > at org.junit.Assert.assertEquals(Assert.java:555) > at > org.apache.hadoop.yarn.server.resourcemanager.TestClientRMService.testGetClusterNodes(TestClientRMService.java:215) > {noformat} -- This message was sent by Atlassian JIRA (v6.3.4#6332)