[ https://issues.apache.org/jira/browse/MAPREDUCE-2933?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101132#comment-13101132 ]
Vinod Kumar Vavilapalli commented on MAPREDUCE-2933: ---------------------------------------------------- Looked at the latest patch. Just a couple more of comments. - _LeafQueue.assignReservedContainer()_: Need to pass an 'AbnormalContainerStatus' to completedContainer(). Passing null will eventual cause a NPE in container state-machine. - _SchedulerApp.completedContainer()_: the event parameter isn't used anymore. Can remove it here and all the way up in the call trace. > Change allocate call to return ContainerStatus for completed containers > rather than Container > ---------------------------------------------------------------------------------------------- > > Key: MAPREDUCE-2933 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-2933 > Project: Hadoop Map/Reduce > Issue Type: Sub-task > Components: applicationmaster, mrv2, nodemanager, resourcemanager > Reporter: Arun C Murthy > Assignee: Arun C Murthy > Priority: Blocker > Fix For: 0.23.0 > > Attachments: MAPREDUCE-2933.patch, MAPREDUCE-2933.patch, > MAPREDUCE-2933.patch > > > Change allocate call to return ContainerStatus for completed containers > rather than Container, we should do this all the way from the NodeManager too. -- This message is automatically generated by JIRA. For more information on JIRA, see: http://www.atlassian.com/software/jira