[
https://issues.apache.org/jira/browse/MESOS-3086?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14635490#comment-14635490
]
Joerg Schad edited comment on MESOS-3086 at 7/21/15 5:57 PM:
-------------------------------------------------------------
Yes as long as the freezer subcomponent is available all is good. just in case
that is not present we fall back to a very simple mechanism which fails if
there are still tasks present (see
https://github.com/apache/mesos/blob/0.22.1/src/linux/cgroups.cpp#L1728).
was (Author: js84):
Yes as long as the freezer subcomponent is available all is good. just in case
that is not present we fall back to a very simple mechanism which fails if
there are still tasks present (see
https://github.com/apache/mesos/blob/master/src/linux/cgroups.cpp#L1725).
> Create cgroups TasksKiller for non freeze subsystems.
> -----------------------------------------------------
>
> Key: MESOS-3086
> URL: https://issues.apache.org/jira/browse/MESOS-3086
> Project: Mesos
> Issue Type: Bug
> Reporter: Joerg Schad
> Assignee: Joerg Schad
> Labels: mesosphere
>
> We have a number of test issues when we cannot remove cgroups (in case there
> are still related tasks running) in cases where the freezer subsystem is not
> available.
> In the current code
> (https://github.com/apache/mesos/blob/0.22.1/src/linux/cgroups.cpp#L1728) we
> will fallback to a very simple mechnism of recursivly trying to remove the
> cgroups which fails if there are still tasks running.
> Therefore we need an additional (NonFreeze)TasksKiller which doesn't rely
> on the freezer subsystem.
> This problem caused issues when running 'sudo make check' during 0.23 release
> testing, where BenH provided already a better error message with
> b1a23d6a52c31b8c5c840ab01902dbe00cb1feef / https://reviews.apache.org/r/36604.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
