[
https://issues.apache.org/jira/browse/MESOS-3009?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16172446#comment-16172446
]
Jie Yu commented on MESOS-3009:
-------------------------------
I think a better approach is to create per container cgroup under systemd
hierarchy (/sys/fs/cgroup/systemd/mesos) and move the pid there. That'll
prevent the pid migration as well because essentially, systemd looks for pid in
the systemd hierarchy, and won't try to migrate the pid if the pid is in a
unmanaged cgroup under /sys/fs/cgroup/systemd/mesos. I think that's what Docker
does with their cgroupfs driver.
I did some experiments on my box. I manually created `/mesos` cgroup under
systemd hierarchy and assign a pid to it. Then, I put the same pid into a
random cpu cgroup. Then, I reload the systemd daemon and restarted a random
service as described above, the pid is not migrated.
> Reproduce systemd cgroup behavior
> ----------------------------------
>
> Key: MESOS-3009
> URL: https://issues.apache.org/jira/browse/MESOS-3009
> Project: Mesos
> Issue Type: Task
> Reporter: Artem Harutyunyan
> Assignee: Joris Van Remoortere
> Labels: mesosphere
>
> It has been noticed before that systemd reorganizes cgroup hierarchy created
> by mesos slave. Because of this mesos is no longer able to find the cgroup,
> and there is also a chance of undoing the isolation that mesos slave puts in
> place.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
