[lago-devel] lago command taking 100% CPU

[Doron Fediuck]

Looking at [1] you can look for the relevant thread and
gdb into it to see what's it about


[1]
https://stackoverflow.com/questions/23838498/how-to-diagnose-a-python-process-chewing-cpu-in-linux

On 28 July 2017 at 11:19, Yaniv Kaul <ykaul at redhat.com> wrote:

> When running o-s-t, and specifically when running add_secondary_storag,
> lago command is taking 100% CPU. I'm not sure where and why, and trying to
> narrow down the relevant threads did not yield a lot - it's not the Glance
> related functions, for example.
>
> I did try to make some changes to testlib.py, in an effort to reduce what
> I thought was a tight loop, but it made no effect (nevertheless, I think
> those are good changes):
> [ykaul at ykaul lago-ost-plugin]$ diff -u ovirtlago/testlib.py
> /usr/lib/python2.7/site-packages/ovirtlago/testlib.py
> --- ovirtlago/testlib.py 2017-07-28 11:15:31.506848944 +0300
> +++ /usr/lib/python2.7/site-packages/ovirtlago/testlib.py 2017-07-28
> 09:29:19.403446285 +0300
> @@ -209,7 +209,7 @@
>      return any(True for cls in cls_list if isinstance(obj, cls))
>
>
> -def assert_equals_within(func, value, timeout, allowed_exceptions=None):
> +def assert_equals_within(func, value, timeout, allowed_exceptions=None,
> initial_wait=15):
>      allowed_exceptions = allowed_exceptions or []
>      with utils.EggTimer(timeout) as timer:
>          while not timer.elapsed():
> @@ -219,12 +219,17 @@
>                      return
>              except Exception as exc:
>                  if _instance_of_any(exc, allowed_exceptions):
> +                    time.sleep(3)
>                      continue
>
>                  LOGGER.exception("Unhandled exception in %s", func)
>                  raise
>
> -            time.sleep(3)
> +            if initial_wait == 0:
> +                time.sleep(3)
> +            else:
> +                time.sleep(initial_wait)
> +                initial_wait = 0
>      try:
>          raise AssertionError(
>              '%s != %s after %s seconds' % (res, value, timeout)
>
>
>
> Any ideas how to debug this? I'm on F26, with latest Lago.
> Y.
>
>
> _______________________________________________
> lago-devel mailing list
> lago-devel at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/lago-devel
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/lago-devel/attachments/20170730/5ab280dd/attachment.html>