'Seems to hang' is my cue to run vmstat and monitor (top?) to see if it is
busy or what.
If you unmanage a collection with a large number of members, expect it to
take a while.
'A while' is relative, and depends on your resources. Have you checked the
object
count lately? (ovobjprint -S) Make sure it is about 20% below the value of
ovwdb cache
size (default is 5000). I would not kill ovwdb unless I saw that the cache
setting was
in need of being increased, in which case you will have little choice.

 It should not have anything to do with your housekeeping, I wouldn't
think, except for the
case where your seedfile uses the new functions to exclude things by oid,
or name
wildcard, etc. You may have built up a large number of hints, and an
ovtopofix -a will
remove those, possibly dropping the object count back down below the cache
setting in which case you would see an improvement in performance.  They
will
start building up again after you restart netmon.  Therefore you should set
the cache
to match the highest number of objects you generally get. I have taken to
adding
address range restrictions as much as possible to keep this count down,
even when
I use the oid entries.

My standard housekeeping instructions (it does not hurt to do this weekly)
=================================================================
0.)   ovobjprint -S      (note number of objects, compare to ovwdb cache)
1.)   ovmapcount -a
2.)   ovstop netmon
3.)   ovtopofix -a    (watch the output for instructions)
                               (use -A if you have multiple maps with
 different contents)
4.)   ovstop ovtopmd
5.)   ovtopofix -Cv   (only with the 'classic' database format, not
 nvturbo)
6.)   ovstart
7.)   ovobjprint -S          ( note number of objects in database)

Cordially,

Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking



Hi all,

I have been playing around with collections lately and have run into some
small problems. When I create a collection, everything works fine. But when
I go to unmanage a collection, Netview seems to hang. I have noticed that a
ovwdb process seems to hang.When I kill it, I get my Netview back, but the
symbol names have reverted back to their original names and such. I have
not been doing much cleaning up of the databases, which I suspect is the
problem. If that is in fact the problem, what type of routine maintenance
are people running on their netview management stations?
I am running Netview 5.1 (where's my 5.1.1 cd!?!?! haha), on AIX 4.3.2 over
Framework 3.6...

Any comments would be greatly appreciated!

Matt Ashfield
mda AT unb DOT ca

PS. I may as well add my lingering question as well... Is there anyone else
out there trying with no success to measure traffic over PVC's and SVC's
within ATM interfaces?