[ https://issues.apache.org/jira/browse/YARN-5829?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15939381#comment-15939381 ]
ASF GitHub Bot commented on YARN-5829: -------------------------------------- Github user kambatla commented on a diff in the pull request: https://github.com/apache/hadoop/pull/201#discussion_r107799584 --- Diff: hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSSchedulerNode.java --- @@ -131,10 +203,58 @@ void addContainersForPreemption(Collection<RMContainer> containers) { /** * Remove container from the set of containers marked for preemption. + * Reserve the preempted resources for the app that requested + * the preemption. * * @param container container to remove */ void removeContainerForPreemption(RMContainer container) { containersForPreemption.remove(container); } + + /** + * The Scheduler has allocated containers on this node to the given + * application. + * @param rmContainer Allocated container + * @param launchedOnNode True if the container has been launched + */ + @Override + protected synchronized void allocateContainer(RMContainer rmContainer, + boolean launchedOnNode) { + super.allocateContainer(rmContainer, launchedOnNode); + Resource allocated = rmContainer.getAllocatedResource(); + if (!Resources.isNone(allocated)) { + for (FSAppAttempt app: resourcesPreemptedPerApp.keySet()) { --- End diff -- Given this is a map, why aren't we just doing a map.get()? > FS preemption should reserve a node before considering containers on it for > preemption > -------------------------------------------------------------------------------------- > > Key: YARN-5829 > URL: https://issues.apache.org/jira/browse/YARN-5829 > Project: Hadoop YARN > Issue Type: Sub-task > Components: fairscheduler > Reporter: Karthik Kambatla > Assignee: Miklos Szegedi > > FS preemption evaluates nodes for preemption, and subsequently preempts > identified containers. If this node is not reserved for a specific > application, any other application could be allocated resources on this node. > Reserving the node for the starved application before preempting containers > would help avoid this. -- This message was sent by Atlassian JIRA (v6.3.15#6346) --------------------------------------------------------------------- To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org