[
https://issues.apache.org/jira/browse/FLINK-5810?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15985113#comment-15985113
]
ASF GitHub Bot commented on FLINK-5810:
---------------------------------------
Github user tillrohrmann commented on the issue:
https://github.com/apache/flink/pull/3394
Thanks for the thorough review @StephanEwen. I'll will pull in your commit
with the improvements.
Concerning the other points you've raised:
- I think you're right. The timeouts should be configurable. I would like
to do as a follow up because the PR is already big enough.
- You're also right that the timeout logic is not properly working. I agree
that the (2) approach is probably less brittle. Additionally we would get rid
of the timeout identifier generation.
- I think it is fair to assume that the number of slots don't change over
the lifetime of a `TaskManager`. I will change the implementation accordingly.
> Harden SlotManager
> ------------------
>
> Key: FLINK-5810
> URL: https://issues.apache.org/jira/browse/FLINK-5810
> Project: Flink
> Issue Type: Sub-task
> Components: Distributed Coordination
> Affects Versions: 1.3.0
> Reporter: Till Rohrmann
> Assignee: Till Rohrmann
>
> Harden the {{SlotManager}} logic to better cope with lost messages.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
