You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by John Kim <dr...@gmail.com> on 2015/07/02 07:10:35 UTC
Mesos/Marathon is not update docker image.
I have Mesos cluster consists of 9 Ubuntu 14.04 machines on cloud
environment.
ZooKeeper and Mesos-Master is running on 3 machines and Mesos-Slave is
running on 6 machines.
Also, Marathon is running on master node.
Docker registry is private resistry.
I am trying to deploy this container via Marathon:
https://mesosphere.github.io/marathon/docs/rest-api.html#post-/v2/apps
The application is successfully deployed and operated.
After a few days, I am trying to update this container via Marathon:
https://mesosphere.github.io/marathon/docs/rest-api.html#put-/v2/apps/%7Bappid%7D
Then, container is deploying, it becomes locked.
If the Zookeer, Marathon, Mesos-Master is restart, and then you can deploy
and operate normally.
A Search fo /var/log/mesos reveals:
E0526 14:05:29.145584 31390 slave.cpp:2662] Failed to update resources for
container 465b12d4-65a3-4f03-873f-3f41601b1db5 of executor
device-api.1c7a63a7-0353-11e5-8ec8-56847afe9799
running task device-api.1c7a63a7-0353-11e5-8ec8-56847afe9799 on status
update for terminal task, destroying container: Failed to determine cgroup
for the 'cpu' subsystem:
Failed to read /proc/32725/cgroup: Failed to open file
'/proc/32725/cgroup': No such file or directory
Tell us the reason or solution.
Thank you.
Re: Mesos/Marathon is not update docker image.
Posted by Timothy Chen <tn...@gmail.com>.
Hi John,
You should run your mesos-slave container with --pid=host so that it's
able to find the processes that is forked from the slave.
Thanks,
Tim
On Wed, Jul 1, 2015 at 10:10 PM, John Kim <dr...@gmail.com> wrote:
> I have Mesos cluster consists of 9 Ubuntu 14.04 machines on cloud
> environment.
>
> ZooKeeper and Mesos-Master is running on 3 machines and Mesos-Slave is
> running on 6 machines.
>
> Also, Marathon is running on master node.
>
> Docker registry is private resistry.
>
> I am trying to deploy this container via Marathon:
> https://mesosphere.github.io/marathon/docs/rest-api.html#post-/v2/apps
>
> The application is successfully deployed and operated.
>
> After a few days, I am trying to update this container via Marathon:
> https://mesosphere.github.io/marathon/docs/rest-api.html#put-/v2/apps/%7Bappid%7D
>
> Then, container is deploying, it becomes locked.
>
> If the Zookeer, Marathon, Mesos-Master is restart, and then you can deploy
> and operate normally.
>
>
> A Search fo /var/log/mesos reveals:
>
> E0526 14:05:29.145584 31390 slave.cpp:2662] Failed to update resources for
> container 465b12d4-65a3-4f03-873f-3f41601b1db5 of executor
> device-api.1c7a63a7-0353-11e5-8ec8-56847afe9799
> running task device-api.1c7a63a7-0353-11e5-8ec8-56847afe9799 on status
> update for terminal task, destroying container: Failed to determine cgroup
> for the 'cpu' subsystem:
> Failed to read /proc/32725/cgroup: Failed to open file
> '/proc/32725/cgroup': No such file or directory
>
> Tell us the reason or solution.
>
> Thank you.
Re: Mesos/Marathon is not update docker image.
Posted by Adam Bordelon <ad...@mesosphere.io>.
What version of Mesos are you using? Something similar was fixed in 0.22.1.
See https://issues.apache.org/jira/browse/MESOS-1837 and
https://issues.apache.org/jira/browse/MESOS-2583
If you're running mesos-slave inside a Docker container, there are other
complications to consider. See
https://issues.apache.org/jira/browse/MESOS-2115 for soon-to-be-released
(0.23) improvements there.
On Wed, Jul 1, 2015 at 10:10 PM, John Kim <dr...@gmail.com> wrote:
> I have Mesos cluster consists of 9 Ubuntu 14.04 machines on cloud
> environment.
>
> ZooKeeper and Mesos-Master is running on 3 machines and Mesos-Slave is
> running on 6 machines.
>
> Also, Marathon is running on master node.
>
> Docker registry is private resistry.
>
> I am trying to deploy this container via Marathon:
> https://mesosphere.github.io/marathon/docs/rest-api.html#post-/v2/apps
>
> The application is successfully deployed and operated.
>
> After a few days, I am trying to update this container via Marathon:
>
> https://mesosphere.github.io/marathon/docs/rest-api.html#put-/v2/apps/%7Bappid%7D
>
> Then, container is deploying, it becomes locked.
>
> If the Zookeer, Marathon, Mesos-Master is restart, and then you can deploy
> and operate normally.
>
>
> A Search fo /var/log/mesos reveals:
>
> E0526 14:05:29.145584 31390 slave.cpp:2662] Failed to update resources for
> container 465b12d4-65a3-4f03-873f-3f41601b1db5 of executor
> device-api.1c7a63a7-0353-11e5-8ec8-56847afe9799
> running task device-api.1c7a63a7-0353-11e5-8ec8-56847afe9799 on status
> update for terminal task, destroying container: Failed to determine cgroup
> for the 'cpu' subsystem:
> Failed to read /proc/32725/cgroup: Failed to open file
> '/proc/32725/cgroup': No such file or directory
>
> Tell us the reason or solution.
>
> Thank you.
>