SUMMARY: SunNetManager monitoring LOCAL disk space, NOT NFS or LOFS.

From: Janis Lykakis (
Date: Fri Apr 03 1998 - 07:03:25 CST

Sorry, I keep making errors when trying to send this mail.
I hope this time I got it right.

Original question:
> I really really hope this is not a RTFM, but I've read the entire
> Solstice Site/SunNet/Domain manager Admin Guide, browsed the entire
> internet, "grepped" throug 3 years of sun-managers mails, to no avail.
> What I want to do has to be easy.
> I'm running SunNetManager 2.3 on a Solaris 2.6 machine.
> We've got several Solaris and SunOS servers. I installed
> the sun provided agents on all servers.
> One of the things we want to use SNM for is monitor diskspace,
> So I used: "send Predefined" --> "event request" --> "When disk is full"
> Set the capacity attribute to 98%. BUT THE DAMN THING ALSO MONITORS
> NFS and LOFS DISKS!!! I do not want that. We have several dedicated
> dataservers with disks that fill up frequently.
> I tried to do something with the key. but the key is MountedOn.
> I do not want to define seperate requests for every partition I want
> to monitor, or do I ?? I want every new disk to automatically be
> monitored.
> Or do I need to make scripts that run from the cron, check df -kl and
> add events-requests to SNM?
> Please help me out here, I'm stuck.
> Furthermore, I tried to find and archive of "old" snm-people summaries
> or something like that, to make sure my question has not been asked hundred's
> of times, but I cannot find it. Does it excist?

This luckilly was no RTFM but it was a question that had been asked before.

Unfortunately, I did not get the answer I was hoping to get, as is turns out,
if I want to stick with SNM, there are just about two options:

1 Monitor all disks (also NFS LOFS CDROM PROC etc.), filter the output with
  a script. This is not what I'll go for.
2 Create events for every partition you want to monitor. That's what I'll do.
  I've got a script that does that for me. It add's events to the management DB.
  downside is, you have to save your DB, exit SNM, run the script and start
  snm -i, then load the DB. But it's good enough for me.

Final option, throw away SNM and go for BigBrother and/or Spong.

There are the replies I got. Thanks very much guy's for your thoughts and

M R Madhusudhan:

        I was snm 2.2.3 but found not much of use. Now i switched over to
Public domain software Bigbrother and Spong. Which are more flexible,
more support and web based. I can give the URL if u are interested.

        Spong a perl script version of BB available at.
Alternatively u can look at MRTG an SNMP based tool mainly used for graphing.
link performance and other snmp variables. I think u should be knowing this
if not and intreasted will let u know the URL.
Damir Delija:
 Usally this question arise each year.
There are few possible work arounds for this situation
One (which I've tried) is to send event to script which will
filter out NFS,CDROM, PROC and simillar things
failure is taht you can not see result of script in console ...

Other is to write down new manager which is capable of filtering
disk better then sun orginal

Also it is possible to analyze data and event logs to filter out
this real disk situations (again you must use scripts to do
such things and results can not be seen in console on map )
Robert Angelino:

Yes. I used to have the same problem with cdroms which were mounted on
machines I had the exact same request going to. My suggestion is to write
a script which filters out NFS, cdrom, anything you don't want and page
if the max disk space usage is met for the disks you want to monitor.
You will probably have to edit the request and add the script.
Eric W. Timberlake:

Did anyone ever respond to you on this? My approach was indeed an event monitor
per partition. Less than efficient if you ask me. Would love to hear about
alternative methodologies..

Just curious.

Janis Lykakis.

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:35 CDT