[ https://issues.apache.org/jira/browse/MAPREDUCE-4376?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13402499#comment-13402499 ]
Kihwal Lee commented on MAPREDUCE-4376: --------------------------------------- There is a check for null to handle transitions from UNASSIGNED state, but the check doesn't work anymore because assignedRequest.get() throws NPE after the following change from MAPREDUCE-3921. {noformat} ContainerId get(TaskAttemptId tId) { if (tId.getTaskId().getTaskType().equals(TaskType.MAP)) { - return maps.get(tId); + return maps.get(tId).getId(); } else { - return reduces.get(tId); + return reduces.get(tId).getId(); } } {noformat} Jason has also suggested we put a time limit in these jobs so that they don't hang even if something goes wrong. > TestClusterMRNotification times out > ----------------------------------- > > Key: MAPREDUCE-4376 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-4376 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2, test > Affects Versions: 2.0.1-alpha > Reporter: Jason Lowe > Assignee: Kihwal Lee > > The TestClusterMRNotification test is often timing out. git bisect tests > narrowed it down to MAPREDUCE-3921, as the test consistently passes before > that change and times out most of the time after picking up that change. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira