`mesos ps` takes an eternity

[stevenschlansker]

We have a cluster with about 20 tasks running, and running mesos ps takes almost 2 minutes.

[steven@host:~]% time mesos ps
(wait 2 minutes)
   TIME   STATE     RSS     CPU   %MEM           COMMAND          USER                                              ID          
<output>

mesos ps  1.85s user 0.19s system 1% cpu 1:58.50 total

[grampelberg]

Yeah .... I've looked into this a couple times and it is usually a mesos specific thing. At the moment, requests to /monitor/statistics.json has to go hit cgroups for every task running on the box. There's an open issue to make that whole thing quite a bit faster.

If you're looking for specific tasks/apps, you can filter ps the same way all the other commands filter. I know it isn't the best solution, but it works for now.

[stevenschlansker]

Thanks for the explanation. Do you have a link to the other issue handy? I'd like to subscribe to it :)