nv-l
[Top] [All Lists]

Re: mlm discovery process

To: nv-l@lists.tivoli.com
Subject: Re: mlm discovery process
From: Chris Cowan <chris.cowan@2ND-WAVE.COM>
Date: Wed, 20 Oct 1999 11:43:08 -0500
"Frank W. Hansen" wrote:
>
> Thanks for the information.  Chris, help me with my understanding of the
> netmon daemon and the  -g -Gmlmseed -z switches which assume the discovery
> role for that mlm domain.  If you were to provide the MLM with multiple
> subnets, through an automated collection modification, would these options
> not force discovery of subnets outside the initial MLM segment?  I'm going
> to try it, I'll keep you posted as to the results.
>

I'm almost 100% certain that the discovery can not be distributed
outside of a subnets that the MLM is directly connected to.  There's
simply no place for the MLM's engine to store all the information it
would need to do this properly.  I didn't check the SNMP polling list
(another -a flag on netmon) but that may give you an indication.

If you want to test it however, look at the netmon entry in the NV
Administrators reference.   You can clear the database, bring up netmon
and nvcold, preload the mlmDomain collection rule with nvUtil, then
start NV (so that the rule is already in place).   If you turn netmon
tracing on, your question will be answered.


The only deviation to all of this is the attended MLM (which is only
available on NV for NT).   In this case, you can choose through the use
of environment variables, whether the MLM code or the Netview Manager is
doing the discovery, status, and SNMP data collection.    This is
describe in an html doc on the NT distribution in the \usr\ov
directory.

In this case however, I believe you end up with redundant
discovery/config polling.

Attachment: chris.cowan.vcf
Description: Card for Chris Cowan

<Prev in Thread] Current Thread [Next in Thread>

Archive operated by Skills 1st Ltd

See also: The NetView Web