[ https://issues.apache.org/jira/browse/MAPREDUCE-5355?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Chuan Liu updated MAPREDUCE-5355: --------------------------------- Attachment: MAPREDUCE-5355-branch-2.patch Attaching a patch. The cause is due to ambiguity in where the staging directory is when using default stating dir configuration, i.e. /tmp/hadoop-yarn/staging. In my environment, the test does not run on C drive. Thus the MiniMRYarn cluster will refer to D:/tmp/hadoop-yarn/staging. AM still runs on C drive and refers to C:/tmp/hadoop-yarn/staging as staging directory, and cannot find job files that should be localized to staging directory. > MiniMRYarnCluster with localFs does not work on Windows > ------------------------------------------------------- > > Key: MAPREDUCE-5355 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-5355 > Project: Hadoop Map/Reduce > Issue Type: Bug > Affects Versions: 3.0.0, 2.1.0-beta > Reporter: Chuan Liu > Assignee: Chuan Liu > Priority: Minor > Attachments: MAPREDUCE-5355-branch-2.patch > > > When MiniMRYarnCluster configured to run on localFs instead of remoteFs, i.e. > MiniDFSCluster, the job will fail on Windows. The error message looks like > the following. > {noformat} > java.io.IOException: Job status not available > {noformat} > In my testing, the following unit tests hit this exception. > * TestMRJobsWithHistoryService > * TestClusterMRNotification > * TestJobCleanup > * TestJobCounters > * TestMiniMRClientCluster > * TestJobOutputCommitter > * TestMRAppWithCombiner > * TestMROldApiJobs > * TestSpeculativeExecution -- 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