First of all, almost everyone's object database contains a few hints and stub
objects. These are created as a result of new node discovery. netmon reads the
ARP cache of the nodes he does know and builds these hints to flesh out later.
If he cannot get to them, or your seedfile says he shouldn't, then they remain
as hints.
What is in your seedfile? If you are using entries based on wildcards or OIDs,
then a large number of hints and stubs can be normal. When you do that, these
stub objects represent addresses which netmon found that you don't want in
topology or on the map. But there has to be some record of them somewhere.
Ovtopofix will remove them, but as soon as you start new node discovery again,
they will come back. I would guess that your new position in the network gives
you visibility to a lot more nodes than it did before, which is often what
happens when you move from test to production.
Your choices to change this behavior are
(1) use a different kind of seedfile -- one which specifically lists the id's
you want to discover only plus an invalid range to limit netmon to just what is
in the file.
(2) turn off new node discovery and run ovtopofix, and don't turn it on again,
unless you create a seed file as discussed in (1)
(3) raise your ovwdb cache size accordingly and understand that these objects
in the object database are there to prevent unnecessary work by netmon after he
has put the stub in the database.
netmon's job is to discover your network. If you only want him to discover a
small portion of it, then you must accept the fact that either you will do the
work in specifying the seedfile, or netmon will do the work building stub
objects. It is a trade-off on who puts in the work.
James Shanks
Tivoli (NetView for UNIX) L3 Support
"McCarter, Bill" <mccarter_we@NNS.COM> on 08/19/99 12:13:49 PM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on NetView
<NV-L@UCSBVM.UCSB.EDU>
To: NV-L@UCSBVM.UCSB.EDU
cc: (bcc: James Shanks/Tivoli Systems)
Subject: Netview Discovery - Newby needs help
I've been running Netview5.1, Tivoli framework , TEC event console for about
6 months in a lab, but I was still connected directly to the corporate
backbone. (IP ex. 172.16.160.123). Everything seemed fine . The output of
ovobjprint -S showed 2200 objects .
I went to production this week and had to move to a new subnet (IP ex.
172.16.4.201). This required me to do a reinstall in both the netview world
and Tivoli world. My books on netview say moving to a new subnet was
unsupported and required new maps to be built. again this was not a problem
because I'm a small shop just going to production.
Now my server is running like a dog. The output of ovobjprint -S shows 22000
objects???? I'm using the same seedfile I was using in the lab, but
something is screwed up.
I ran ovtopofix -a and got 17000 of theses lines???
deleting object 125902 (INCOMPLETE NODE).
deleting object 126033 (INCOMPLETE NODE).
deleting object 126538 (INCOMPLETE NODE).
deleting object 127267 (INCOMPLETE NODE).
deleting object 127566 (INCOMPLETE NODE).
deleting object 128085 (INCOMPLETE NODE).
deleting object 123767 (HINT).
deleting object 129629 (HINT).
deleting object 129913 (HINT).
deleting object 130569 (HINT).
My ovobjprint -S now look like it did before 2200 objects. BUT, when I start
to discover new nodes the object count rises quickly to 22000 again and
ovtopofix -a gives me all those above lines.
netmon.lrf looks like this ...>
OVs_YES_START:nvsecd,ovtopmd,trapd,ovwdb:-P,-s/usr/OV/conf/nnsseedfile:OVs_W
ELL_
BEHAVED:15:
Any ideas would be greatly appreciated.......
Thank You
Bill
Bill McCarter
Enterprise Systems Management
Newport News Shipbuilding, Dept. OO5
Phone (757) 688-2602
Email mccarter_we@nnscom
|