[mrtg] Re: Forking MRTG

[Pavel Ruzicka]

Hello,

> > When I tried forking, it seemed like the forked mrtg's died
> > very quickly - indicating to me that the forked processes
> > were only hitting a few targets each, and then the main mrtg
> > process would continue on and hit the rest of them.
> 
> the reason for forking is to battle snmp latency, so with forking
> several snmp requests run in paralelle. Unless your targets are
> very fast and cole this can give you an almost linear speedup ...

For speedup I use this options in mrtg.cfg:
Forks: 10
SnmpOptions: retries=>2, timeout=>1

I have Linux 2.2.20 PentiumII 400MHz with 384MB Ram, MRTG 2.9.17.
In MRTG I have about 60 routers and switches and 1800 targets.

In normal situation MRTG running for about 45 seconds. But if some bigger
router goes down, 5 minutes is not enough and I have whitespaces in graphs.
Sometimes in this critical situation MRTG eats all the RAM and kernel
kills a CRON ;-)

> > > your problem with the router going offline does not get solved by
> > > this ... but it does not become worse either. What could be done
> > That would be very useful to be able to skip queries to known
> > dead hosts. Is there a global config command that does this?
> not yet :-)

It is very good idea, what I miss in mrtg.
For example, when I get 3 timeouts from one device, I skip the device
for one measuring cycle.

> tobi

Pavel Ruzicka

--
Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
Archive     http://www.ee.ethz.ch/~slist/mrtg
FAQ         http://faq.mrtg.org    Homepage     http://www.mrtg.org
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi