[ https://issues.apache.org/jira/browse/MAPREDUCE-4273?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13492713#comment-13492713 ]
Yu Gao commented on MAPREDUCE-4273: ----------------------------------- org.apache.hadoop.mapred.TestClusterMRNotification fails as well without this patch. It looks like a job state transition problem, and has been tracked by MAPREDUCE-4774 > Make CombineFileInputFormat split result JDK independent > -------------------------------------------------------- > > Key: MAPREDUCE-4273 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-4273 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: client > Affects Versions: 1.0.3 > Reporter: Luke Lu > Assignee: Yu Gao > Attachments: mapreduce-4273-branch-1.patch, > mapreduce-4273-branch-2.patch, mapreduce-4273.patch > > > The split result of CombineFileInputFormat depends on the iteration order of > nodeToBlocks and rackToBlocks hash maps, which makes the result HashMap > implementation hence JDK dependent. > This is manifested as TestCombineFileInputFormat failures on alternative JDKs. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira