[
https://issues.apache.org/jira/browse/MESOS-1630?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14085645#comment-14085645
]
Bernd Mathiske commented on MESOS-1630:
---------------------------------------
[[email protected]] Let's first make sure I understand what is asked here. My
reading of this issue description is as follows.
1. A framework runs to completion or is otherwise removed and therefore
regarded as "completed".
2. No more tasks of it are running.
3. A completely unrelated framework starts up and is assigned the exact same
framework ID.
4. It starts new tasks.
5. Master::reconcile then asks the slave to shut them down instead of letting
them run.
The fix is to "forget" that a framework with the reused ID is "among the
completed".
If that is all correct and that is the accurate fix, then I should be able to
program this very quickly, indeed. The test may take longer, though. Let me get
back to you tomorrow on that.
> Remove framework from completedFrameworks if framework re-registers.
> --------------------------------------------------------------------
>
> Key: MESOS-1630
> URL: https://issues.apache.org/jira/browse/MESOS-1630
> Project: Mesos
> Issue Type: Bug
> Components: master
> Affects Versions: 0.14.0, 0.14.1, 0.14.2, 0.17.0, 0.16.0, 0.15.0, 0.18.0,
> 0.18.1, 0.18.2, 0.19.0, 0.19.1
> Reporter: Benjamin Hindman
> Assignee: Bernd Mathiske
> Priority: Critical
>
> If a framework gets removed, for example, because it unregisters with the
> master (i.e., due to MESOS-1550), but then the same framework ID is reused
> when a framework re-registers (which we currently allow) then we should
> remove the framework from Master::completedFrameworks otherwise when a slave
> re-registers then in Master::reconcile we'll notice that the slave is
> runnings tasks from a completed framework and tell the slave to shutdown that
> framework, thus shutting down all of the tasks.
> This should be easily fixed by removing the framework from
> completedFrameworks when a framework re-registers with the same ID as a
> completed framework.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
