A good place to start is to try to find out whether the slave was OOM
killed, what do you get from:
$ grep -i kill /var/log/messages
$ grep -i kill /var/log/kern.log
Then, you should examine the existing cgroups setup on your machine. Was
there already a cgroups hierarchy mounted? Were controls a
Ben,
Maybe I was wrong, but the last line of the slave log was "Killed", and the
slave did exist, so I thought the slave was oom killed. If by design the
slave should not be oom killed, I can probably test more to see what
happened.
Li
On Thu, Aug 22, 2013 at 7:02 PM, Benjamin Mahler
wrote:
>
Given how long the tests took on your machine, the coordinator tests are
likely timeouts. I believe you can ignore these safely. Unfortunately
0.12.0 does not disable the cgroups tests when not run as the superuser so
you can ignore those as well (running the tests with sudo should turn them
green)
Hi Li,
Why do you think the slave was OOM killed? Is there something that pointed
you to that conclusion? All I see is the slave launched an executor, and
the executor was killed by framework a few seconds after the task was
launched.
Also, what version are you running?
Ben
On Thu, Aug 22, 201
Hello guys,
I am implementing a mesos executor and see this behavior when I enabled
cgroups isolation. It seems the slave got oom killed. I didn't expect the
slave to be oom killed in any circumstance, am I wrong?
Here are the slave log:
I0822 21:22:09.168122 15557 cgroups_isolation_module.cpp:4
Build mesos without perftools (else the compile fails):
-bash-3.2$ pwd
/data/test/mesos-cluster/mesos-0.12.0
-bash-3.2$ cd build/
-bash-3.2$ ../configure
--prefix=/data/test/mesos-cluster/mesos-0.12.0/installation
--disable-perftools
Run tests (all the Cgroups fail, but I'm not sudo so I think th
> (1) One thing particular I found unexpected is that the executors are
> shutdown if the scheduler is shutdown. Is there a way to keep executors/tasks
> running when the scheduler is down? I would imagine when the scheduler comes
> back, it could reestablish the state somehow and keep going w
Hi guys,
I am trying to implement a mesos framework. Here are some questions I have:
(1) One thing particular I found unexpected is that the executors are
shutdown if the scheduler is shutdown. Is there a way to keep
executors/tasks running when the scheduler is down? I would imagine when
the sch
thats right. all the driver methods are synchronized.
On Thu, Aug 22, 2013 at 7:04 AM, Li Jin wrote:
> Vinod,
>
> Thanks for the explanation. Just to clarify, by thread safe, do you also
> mean multiple threads can call driver.sendStatusUpdate (and other
> functions)?
>
> Li
>
>
> On Thu, Aug 2
Vinod,
Thanks for the explanation. Just to clarify, by thread safe, do you also
mean multiple threads can call driver.sendStatusUpdate (and other
functions)?
Li
On Thu, Aug 22, 2013 at 12:34 AM, Vinod Kone wrote:
> (1) Is executor per-host instance or per-task instance?
>
> Typically you run
10 matches
Mail list logo
