If you don't have a lot invested in the current discovery, you would
be best off clearing the databases and rediscovering (smitty nv6000...
control... Restart Automatic Map Generation) while preserving the
definitions of collections and APM settings.
If you need to save the current cut/paste of routers and networks,
I would define a Collection that identifies those things you do not
want to discover. You won't be able to do this by IP Address, since
nodes don't have addresses, but you could probably do one that
includes non-snmp nodes, or snmp nodes with the Microsoft OID
(~ 22.214.171.124.4.1\.311\.). Then look at what is in the Collection, and select
and delete them from all submaps. Remember to do this on all
maps if you have more than the 'default' map. If you don't, the nodes
are still managed somewhere, and you will get events even if they
are not in the map you are looking at. Check 'File..Open Map' to
look for other maps.
During the delete, make sure netmon is down. After the delete events
show up in the events display (may take a while), close the map and
run 'ovtopofix'. Use either -a or -A depending on how many maps
you have - see the man page. Then, with the new seedfile in place,
restart netmon and the map.
This should not be an ongoing administrative task, once they are
gone AND prevented from rediscovery. The ongoing issue does,
of course, apply to changes to network devices. In general, you
need to do a delete and rediscover of a device that has changed
a lot. The addition of new interfaces is handled by normal configuration
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
"Les Dickert" <lesdickert@HOTMAIL.COM>@tkg.com on 08/11/2000 10:34:27 AM
Please respond to IBM NetView Discussion <firstname.lastname@example.org>
Sent by: email@example.com
Subject: [NV-L] Re-Discovery under NetView 5.1 on AIX
I have come upon a NetView 5.1 under AIX that
has been set up and running for about nine months.
When it was initially set up, the seed file contained
entries that basically told netmon to discover
everything, which it did. Now we want to eliminate
all of the workstations from the monitoring, and
I have added entries to the seed file to tell netmon
to ignore these (by IP ranges). None of these workstations
appear on any submaps that have been created.
So the question is, how do I force a re-discovery with
the intent of purging these workstation objects from
all of the appropriate databases.
So far I have:
1. Run netmon -y to tell netmon to reload the seed
file and confirmed from netmon.trace that it did
reload the seed file.
2. Done an ovstop and ovstart several times.
3. Confirmed that under the Topology/Status Polling
Conf, that everything is turned on.
However, it appears that netmon is still remembering about
the workstations because we are still getting Interface up/down
and Node up/down events from netmon about these workstations.
I seems like this would be an ongoing administrative thing to
have to do, ie force a purging of old IP addresses from the
databases, not only in this case, but in cases where IP addresses
are changed by the network wizards.
NV-L List information and Archives: http://www.tkg.com/nv-l