[mrtg] Re: Solaris SNMP

Neil Calvert ncalvert at cabletron.com
Wed Sep 1 18:05:11 MEST 1999


Useful info, especially if I was running MRTG on Linux or Solaris! :)

Unfortunately, I'm running MRTG on an NT box (crowd goes 'boo!') and using it to
remotely query devices. I don't see any way for a perl script under NT to go out
and grab vmstat info from a UNIX box.. but I may be wrong! Any thoughts on how
this may be accomplished? (Perhaps a cron script on the unix station that
creates a text file, which the NT box remotely reads.. probably would need
Solaris NFS for NT with a drive on the NT box mapped onto the sun station
however..) Does this sound semi-feasible..?

-----Original Message-----
From: Kyle Maxwell <kyle.maxwell at telops.gte.com>
To: mrtg-bounce at list.ee.ethz.ch <mrtg-bounce at list.ee.ethz.ch>;
mrtg at list.ee.ethz.ch <mrtg at list.ee.ethz.ch>
Date: Wednesday, September 01, 1999 9:30 AM
Subject: [mrtg] Re: Solaris SNMP


>To track a specific process, you should be able to do something like
>
>ps -A | grep <appname> | awk 'print $1'
>
>to get the process ID (that awk portion may be a little off). You can put in
other options to ps to get other data you want about that process.
>
>For general CPU system/idle time, I have a tiny script that essentially does
>
>vmstat 1 2 | sed '$!D' | awk -f statget.awk
>
>where statget.awk prints out the fields I'm interested in. There's probably
more "elegant" ways to get this done, but as a quick and dirty hack, it did
well.
>
>Obviously, all this totally goes around SNMP, but it's quick, eh?
>
>--
>Kyle Maxwell
>Systems Engr (Firewall Security)
>GTE Enterprise Information Protection Systems
>---------- Original Text ----------
>
>From: "Neil Calvert" <ncalvert at cabletron.com>, on 8/31/99 4:16 PM:
>
>Hi all.
>
>I'm looking into the possibility of using MRTG against a Solaris SNMP agent
(the basic Solstice agent) and I was wondering if people are doing this already,
and if so, what applications they've found it useful for? Currently I can see
mrtg being able to graph SNMP stats, interface traffic, and so on, but I'd like
to get more OS specific information like total CPU time vs. Idle time. The Sun
mib lets you see cpu time as an instanced mib where every process on the system
is listed. The problem with this of course is that it will be dynamically
changing as processes are starting and stopping. Is there any way to 'alias' an
executable name to a process, so you ask it to graph say 'application A CPU
time' and it goes off and finds out what 'application A's processID is to get
the mib value?
>
>
>I'm guessing this is probably a script hack but if there are any sun mibs out
there that may help I'd be grateful to hear of them.
>
>Neil
>
>
>--
>* To unsubscribe from the mrtg mailing list, send a message with the
>  subject: unsubscribe to mrtg-request at list.ee.ethz.ch
>* The mailing list archive is at http://www.ee.ethz.ch/~slist/mrtg
>

--
* To unsubscribe from the mrtg mailing list, send a message with the
  subject: unsubscribe to mrtg-request at list.ee.ethz.ch
* The mailing list archive is at http://www.ee.ethz.ch/~slist/mrtg


More information about the mrtg mailing list