[
https://issues.apache.org/jira/browse/MESOS-9305?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16645874#comment-16645874
]
longfei commented on MESOS-9305:
--------------------------------
[~chhsia0]
I found one possible reason, just as you said.
We have two different mesos slaves deployed on one host.
!image-2018-10-11-10-40-11-326.png!
The mesos_titan_hbhl's cgroups_root is a subdir of mesos_ufo's.
!image-2018-10-11-10-45-35-369.png!
If we restart mesos_ufo's slave now. We can see that mesos_titan_hbhl is gone.
!image-2018-10-11-10-45-48-170.png!
That's because mesos_titan_hbhl is treated as an ophan to mesos_ufo and is
removed since.
[~jieyu] OK, I'll send a PR soon. Thanks.
> Create cgoup recursively when calling prepare on containers
> -----------------------------------------------------------
>
> Key: MESOS-9305
> URL: https://issues.apache.org/jira/browse/MESOS-9305
> Project: Mesos
> Issue Type: Improvement
> Affects Versions: 1.5.1, 1.6.1, 1.7.0
> Reporter: longfei
> Assignee: longfei
> Priority: Critical
> Attachments: image-2018-10-11-10-40-11-326.png,
> image-2018-10-11-10-45-35-369.png, image-2018-10-11-10-45-48-170.png
>
>
> This is my case:
> My cgroups_root of mesos-slave is some_user/mesos under /sys/fs/cgroup。
> It happens that this some_user dir may be gone for some unknown reason, in
> which case I can no longer create any cgroup and any task will fail.
> So I would like to change
>
> {code:java}
> Try<Nothing> create = cgroups::create(
> hierarchy,
> infos[containerId]->cgroup);
> {code}
> to
> {code:java}
> Try<Nothing> create = cgroups::create(
> hierarchy,
> infos[containerId]->cgroup,
> true);
> {code}
> in CgroupsIsolatorProcess::prepare in
> src/slave/containerizer/mesos/isolators/cgroups/cgroups.cpp.
> However, I'm not sure if there's any potential problem doing so. Any advice?
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
