[mrtg] Newbie question
Andrew.Hughes at aquilent.com
Thu Aug 21 19:15:26 CEST 2008
Steve, thanks very much for the response. I will pursue what you have suggested.
I did search for some error logs but found none. Where are they hiding?
Is mrtg by default going to run every 5 minutes without my intervention, or are you saying I should change the task to run every 5 minutes instead rather than the 10 I set up?
From: Steve Shipway [mailto:s.shipway at auckland.ac.nz]
Sent: Monday, August 18, 2008 6:25 PM
To: Hughes, Andy; 'mrtg at lists.oetiker.ch'
Subject: RE: Newbie question
There are two problems here.
First, you are running data collection every 10 minutes. Unless you initially set the Interval to be 10, then this will mean you have 50% of your data unknown, since you are (by default) supposed to poll every 5 minutes. As your have a default XFF of 0.5 this can result in all your data becoming 'unknown'.
Secondly, your data polling may be failing - but its hard to say as I don't know what you're monitoring. Possibly one of your monitored devices is down, or unreachable, or the community string is wrong? Check your error logs. If you are failing to poll, or the poll is taking >5min, then you will get unknowns stored. The Routers2 frontend will show this as a grey block to indicate 'data unknown' or 'device unpollable', but native MRTG (and 14all) will default to the 'unknasprev' option (see below)
Now, if you have the 'unknasprev' option set in Routers2, or you are using native MRTG (which does unknasprev by default), then 'unknown' data will be displayed as the last known good value. In your case, you may have mostly unknowns, so you'd get a single, flat, solid green area. This appears to be what's happening. If the time on the x-axis is changing, then you know your graph is being re-created, but the polling is failing and the unknasprev option is just logging the last known good value.
Check your polling. Probably, the SNMP community has changed, or the network is not working the same, or there is another problem. Run the MRTG poll from the command line to see if there are any obvious error messages. Make sure your polling is really running. Schedule it every 5 mins instead of every 10 mins so the XFF (see the RRDtool documentation) doesn't log everything as zero.
From: mrtg-bounces at lists.oetiker.ch [mailto:mrtg-bounces at lists.oetiker.ch] On Behalf Of Hughes, Andy
I installed mrtg two weeks ago to monitor traffic to/from an individual router. All seemed to be well; I watched the graphs move as the week went on etc. For some reason, the graph no longer moves at all (i.e. no peaks and valleys)--the entire thing is just solid green.
I run mrtg every 10 minutes by a scheduled task that calls a batch file in the \bin directory (Server 2003) that has a single line: perl mrtg mrtg.cfg I notice that when I was on the server initially when mrtg was running properly, the command window would briefly flash when the batch file ran. Now it stays open much, much longer before disappearing?
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the mrtg