How to debug high system load?

Bryn M. Reeves bmr at redhat.com
Thu Jul 12 14:34:56 UTC 2012


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 07/12/2012 03:29 PM, Suvayu Ali wrote:
> Hi Bryn,
> 
> On Thu, Jul 12, 2012 at 03:17:01PM +0100, Bryn M. Reeves wrote:
>> 
>> That would normally also be reflected to some extent in CPU usage
>> and process activity as shown in top (unless the problem was
>> buffering them from a really slow network file system or
>> something).
>> 
> 
> I do have AFS (authenticated with Kerberos) on my laptop, although
> I'm not using it at the moment. There are no background jobs that
> should access it either. Only time AFS is used is when I explicitly
> run some jobs (through python scripts) for my research.
> 

Hi Suvayu,

It's possible that you have processes spending brief periods in
D-state - short enough that they are hard to spot in top with the
default sample period but long enough to count toward the load average
(it's possible there's some other explanation though but that's what
I'd try to rule out first).

You could use tools like blktrace, iotop and latencytop to try to
investigate further or if you're willing to install systemtap you
could use the "sleepingBeauties.stp" script to look for processes
spending >10ms in this state (and dump their stacks when they do).

Regards,
Bryn.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk/+4JAACgkQ6YSQoMYUY95XTACfWX2MwOa2GkWGujMC37NUVaOP
GDoAnjrWVT1pafqf7Qv/Yf1xs4tvta32
=FzSI
-----END PGP SIGNATURE-----


More information about the users mailing list