[collectd] High CPU usage of collectd when server is under load

[XANi]

On Wed, 28 Oct 2009 10:44:01 +0100, Florian Forster <octo at verplant.org>
wrote:
> Hi Mariusz,
> 
> On Tue, Oct 27, 2009 at 01:19:42AM +0100, XANi wrote:
> > I've installed collectd on server (amd64 debian stable) and when I
> > have high load collectd can eat like 90-100% of one CPU (and its not
> > iowait, disk is fine), what can be reason of that ?
> 
> are you absolutely sure it's not I/O? Eating up cycles of only one CPU
> is the typical behavior for I/O-intensive processes.
http://img9.imageshack.us/img9/4417/cpu0.png
http://img9.imageshack.us/img9/4273/cpu6.png
(other cpu's look similar to cpu0)
I thougth this isn't IO because there is not much iowait but maybe I am
wrong ? There was quite heavy MySQL traffic when it happened (about
~120 inserts/sec).
> > most of them are in pretty much default config, only rrdtool have
> > 
> >        RRARows 12000
> >        CacheTimeout 600
> >        CacheFlush   1200
> >        RandomTimeout 100
> > 
> > What can be reason of such high load ? Collectd works normal (barely
> > 1% load) when ther is light/no load on server
> 
> How many RRD-files do you have when under load and how often are they
> updated? What kind of hardware do you use?
284 files with standard 10 sec interval.
Hardware is Xeon E5410 at 2.33GHz, 16 GB RAM + 2x1TB SATA on Hardware RAID.


-- 
Mariusz Gronczewski (XANi) <xani666 at gmail.com>
GnuPG: 0xEA8ACE64
http://devrandom.pl

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: not available
Url : http://mailman.verplant.org/pipermail/collectd/attachments/20091028/08237933/attachment.pgp