[
https://issues.apache.org/jira/browse/MESOS-856?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13837141#comment-13837141
]
Ian Downes commented on MESOS-856:
----------------------------------
We've observed freeze-thaw loops continuing indefinitely when the processes
contained in the cgroup were in non-interruptible D state but also for
processes that are supposedly in runnable R state but cannot be killed with
kill -9. In this scenario the freezer may actually be freezing but we still
cannot kill the process(es) and will loop.
We should consider how to handle/report this scenario.
> TasksKiller may run forever because the cgroup cannot be frozen.
> ----------------------------------------------------------------
>
> Key: MESOS-856
> URL: https://issues.apache.org/jira/browse/MESOS-856
> Project: Mesos
> Issue Type: Bug
> Reporter: Jie Yu
> Assignee: Ian Downes
> Fix For: 0.17.0
>
>
> Possible fix: rely on process kill loop instead of the cgroup freezer.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
