[ https://issues.apache.org/jira/browse/YARN-910?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13760943#comment-13760943 ]
Vinod Kumar Vavilapalli commented on YARN-910: ---------------------------------------------- This seems logically wrong, though it is fine in the current state of the code. {code} + //if event type is KILL_CONTAINER it means the CONTAINER_INIT was never + // sent for the event, thus no need to send the CONTAINER_STOP + if (event.getType() != ContainerEventType.KILL_CONTAINER) { + container.dispatcher.getEventHandler().handle(new AuxServicesEvent + (AuxServicesEventType.CONTAINER_STOP, container)); + } {code} You should also look if currentState is NEW in addition to the event being KILL_CONTAINER, as that is the transition that you explicitly wanted to avoid. Or better yet, just send it to the aux-service all the time for it to handle containers' stop whether or not they are inited properly? That should okay. > Allow auxiliary services to listen for container starts and completions > ----------------------------------------------------------------------- > > Key: YARN-910 > URL: https://issues.apache.org/jira/browse/YARN-910 > Project: Hadoop YARN > Issue Type: Improvement > Components: nodemanager > Affects Versions: 2.1.0-beta > Reporter: Sandy Ryza > Assignee: Alejandro Abdelnur > Attachments: YARN-910.patch, YARN-910.patch, YARN-910.patch > > > Making container start and completion events available to auxiliary services > would allow them to be resource-aware. The auxiliary service would be able > to notify a co-located service that is opportunistically using free capacity > of allocation changes. -- 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