[
https://issues.apache.org/jira/browse/FLINK-4152?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15382258#comment-15382258
]
ASF GitHub Bot commented on FLINK-4152:
---------------------------------------
Github user tillrohrmann commented on the issue:
https://github.com/apache/flink/pull/2257
It is true that the duplicate RegisterTaskManager messages fix could have
been a separate PR. Since it surfaces as a symptom of the registration problem,
I thought about fixing it in the same PR. The separation should be clear
(TaskManager vs. RM changes).
The reason why I had to change more code than this single line is that the
RM shouldn't interfere with the TaskManager registration to be on the safe
side. As far as I've understood, the RM should tell you about a dying TM but it
should not forbid the registration of a TM at the JM. I think that this
conservative behaviour is more robust wrt the old behaviour and helps us to not
overlook other corner cases.
Furthermore, some of the changes were necessary to make the YARN RM
component more easily testable.
> TaskManager registration exponential backoff doesn't work
> ---------------------------------------------------------
>
> Key: FLINK-4152
> URL: https://issues.apache.org/jira/browse/FLINK-4152
> Project: Flink
> Issue Type: Bug
> Components: Distributed Coordination, TaskManager, YARN Client
> Reporter: Robert Metzger
> Assignee: Till Rohrmann
> Attachments: logs.tgz
>
>
> While testing Flink 1.1 I've found that the TaskManagers are logging many
> messages when registering at the JobManager.
> This is the log file:
> https://gist.github.com/rmetzger/0cebe0419cdef4507b1e8a42e33ef294
> Its logging more than 3000 messages in less than a minute. I don't think that
> this is the expected behavior.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
