[mrtg] Re: Problems with mrtg returning 0

[Steve Shipway]

> So its not possible to say show total DHCP requests over the 
> 5minutes, but not ever increasing, such that in 5 minute 
> invervals there were 3, 5, 1 then
> 4 requests. Is it not possible to show these on the graph as 
> they are (e.g not dividing by time) and without including the 
> last result (like the 'Gauge' option).

Why not use 'Options[]: perminute' (without the gauge option) to get total
DHCP requests per minute?  If this is too small, the use 'Factor[]: 5' to
get the per-5min total (youd need to relabel your axis of course).  You'd
also need to be using RRDTool as the numbers would be decimals - and youd
also need 'Options[]: integer' to make it round off to an integer.

For this sort of thing, here we use perminute and graph the average
requests/min or perhour if this is too small.

Alternatively, if you are more interested in total requests per
day/week/month then use the routers2 frontend, it will calculate these for
you.

> Would it only be possible to do this with an external script?

Yes.

> Another question is what happens when a counter is reset upon 
> reboot, i rebooted just now and alot of graphs spiked to 
> 32,000, is this because the counter got reset to 0 and MRTG 
> thought the counter had overflowed and reset? Does it not 
> check when the device was last booted and decided from that?

In that order, yes, and no.  You should set MaxBytes and/or AbSMax to let
MRTG more accurately spot these spikes.

> And have you had any experience with external scripts getting 
> data from WMI, its hellish slow, i just timed it and it took 
> 45 seconds to run mrtg once, and thats only 3 WMI scripts.
 
I've never use WMI, only PerfMon (via the pNSclient agent) and it works fast
enough.

HTH

Steve


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