You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Marco Massenzio (JIRA)" <ji...@apache.org> on 2015/10/06 01:45:27 UTC
[jira] [Commented] (MESOS-3587) Framework failover when framework
is 'active' does not trigger allocation.
[ https://issues.apache.org/jira/browse/MESOS-3587?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14944267#comment-14944267 ]
Marco Massenzio commented on MESOS-3587:
----------------------------------------
Would love to learn more about this area of the code; if you ([~bmahler]) are willing to shepherd this, I'd be quite keen to take this on.
> Framework failover when framework is 'active' does not trigger allocation.
> --------------------------------------------------------------------------
>
> Key: MESOS-3587
> URL: https://issues.apache.org/jira/browse/MESOS-3587
> Project: Mesos
> Issue Type: Bug
> Components: allocation, master
> Reporter: Benjamin Mahler
> Assignee: Marco Massenzio
> Priority: Minor
> Labels: mesosphere
>
> FWICT, this is just a consequence of some technical debt in the master code. When an active framework fails over, we do not go through the deactivation->activation code paths, and so:
> (1) The framework's filters in the allocator remain after the failover.
> (2) The failed over framework does not receive an immediate allocation (it has to wait for the next allocation interval).
> If the framework had disconnected first, then the failover goes through the deactivation->activation code paths.
> This also means that some tests take longer to run than necessary.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)