[mrtg] Re: Too many hosts in single mrtg config files

[tony bourke]

Hi Gaurav,

MRTG does have limitations on how many hosts it can handle, but it's
usually in the several hundreds to thousands.  The amount of targets you
are looking to collect data for look very reasonable for an average
collector machine.

To answer your questions:

1)  LAN is fine.  Shouldn't be a problem at all
2)  Downed devices aren't much of an issue.  There is a timeout typically,
but nothing that would cause too much of a problem.
3)  Not too much of an issue either.  Spliting the config into various
files as suggested below would further help that situation, but putting
slow WAN links on a seperate file, minimizes any potential impact timeouts
would have (and that isn't even too much of an issue)

Here are a few tips that might help as well:

1:  Use a dedicated (or mostly dedicated) machine to do MRTG collection.
MRTG is disk I/O intensive at a high number of devices, and the less it's
doing other than MRTG the better.

2:  Split the machines up into different config files.  A config file for
HP, another for network, another for Solaris.  This is a real benefit for
dual processor boxes, since you've got semi-multi-threadedness out of it.

3: Spread the MRTG runs out amoung the various slots in the 5 minute
interval.  Run some 1, 6, 11, others 2, 8, 12, etc.

4: Use a SCSI system if possible.  Since there is quite a bit of disk I/O,
SCSI offloads alot of the processing so you don't hang an IDE bus, saving
processor power and running through MRTG config files quicker.

Hope that helps!

Tony

On Sun, 26 Aug 2001, Gaurav Sabharwal wrote:

>
> I have been monitoring my Cisco routers using MRTG (8 in number). Starting
> from tomorrow onwards, I will be working on monitoring traffic, CPU, memory
> utilization for 21 odd Solaris servers. In the next couple of weeks, we plan
> to have 10 HP servers, which will be monitored too. Two Cisco PIX and a few
> other network devices. Not a huge list but I would like to know what are the
> downside of using a single configuration script for all the devices. Let's
> assume the number of devices being 50 and all being on a 10/100 Mbps LAN.
> Please correct me if my understanding is incorrect.
>
> 1) All on LAN. Should not be a problem to have MRTG fetch/gather the
> information within 5 minutes interval.
> 2) Problem if devices are down. (Is that a problem?)
> 3) Problem if devices are monitored over the WAN and on a slow link.
>
> What are the other solutions that people are using for the same? Would like
> to hear back from the community.
>
> Thanks and Regards,
> - Gaurav
>
>
> --
> 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
>

-- 
-------------- -- ---- ---- --- - - - -  -  -- -  -  -  -   -     -
Tony Bourke				tony at vegan.net



--
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