You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Benjamin Hindman (JIRA)" <ji...@apache.org> on 2014/08/01 02:17:38 UTC
[jira] [Commented] (MESOS-1656) Do not remove docker container
until gc process runs
[ https://issues.apache.org/jira/browse/MESOS-1656?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14081724#comment-14081724 ]
Benjamin Hindman commented on MESOS-1656:
-----------------------------------------
Yes, great idea [~jaybuff]. This is going to be a bit harder as currently a containerizer is not part of the GC cleanup process. We'll first want to introduce a new method on containerizer that signifies "GC/cleanup" of a container and then have the DockerContainerizer do the final remove then.
> Do not remove docker container until gc process runs
> ----------------------------------------------------
>
> Key: MESOS-1656
> URL: https://issues.apache.org/jira/browse/MESOS-1656
> Project: Mesos
> Issue Type: Improvement
> Components: containerization
> Reporter: Jay Buffington
>
> The current docker containerizer implementation that is up for review at https://reviews.apache.org/r/23771/ does a {{docker rm}} as soon as a task fails. This makes debugging difficult, if not impossible. MESOS-1652 will aid with this, but it does not address the use case of diagnosing a failure which involves inspecting the state of a container.
> [~tnachen] and I discussed this on IRC, see http://wilderness.apache.org/channels/?f=mesos/2014-07-31#1406842111
--
This message was sent by Atlassian JIRA
(v6.2#6252)