[
https://issues.apache.org/jira/browse/FLINK-1154?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14214574#comment-14214574
]
ASF GitHub Bot commented on FLINK-1154:
---------------------------------------
Github user StephanEwen commented on a diff in the pull request:
https://github.com/apache/incubator-flink/pull/208#discussion_r20431011
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/instance/InstanceManager.java
---
@@ -113,6 +113,7 @@ public void shutdown() {
this.cleanupStaleMachines.cancel();
for (Instance i : this.registeredHostsById.values()) {
+ i.stopInstance();
--- End diff --
I would strongly vote to remove that. There is no need to kill a
TaskManager just because it lost its heartbeat. The TaskManager may very well
reconnect later and be available again.
> YarnTaskManagerRunner does not shut down
> ----------------------------------------
>
> Key: FLINK-1154
> URL: https://issues.apache.org/jira/browse/FLINK-1154
> Project: Flink
> Issue Type: Bug
> Components: YARN Client
> Reporter: Till Rohrmann
> Assignee: Robert Metzger
>
> The YarnClient does not shut down the YarnTaskManagerRunner when the client
> receives a stop command or is terminated. This causes a resource leak on the
> cluster.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
