[nv-l] NetView for Windows2k Performance

["Denis PIAROTAS" <denis.piarotas@RTE-FRANCE.COM>]

Hello all,

We work in a Win2k environnement with
a Netview v7.1.4 FP1. It's a dedicated one obviously. The DB is the Windows'
one (Access format).

This deals with about 4000 objects,
sooner or later 5000 (ie 400 to 500 nodes), and we expected to optimise
the server and the application. 

Problem : As a WAN VPN supervision link
comes to down, we have often one or two hours to resynchronize the server
and have an available server for the admins. Fisrt we suspect the DBs size
(not exactely the case, snmpcollect is 260 Mo and tess_db 30 Mo).

So we began to set the parameters to
the better way the can be by non-specialist NetView engineers : ovtopofix
-a,  calculation of ovwdb size cache, log check (all of \usr\ov\log\)
and of netmon.seed + location.conf files.

One important thing is we don't want
to reset the DB files at first sight, or radically resinstall all the maps
and objects : it's a too large and long operation for us for many reasons.

So our solution is to explore the parameters
ways : log file study (there's maybe things to do when many traps are generated
in nv.log), the daemons parameters, the cache in the DB (2 Mo by default
with Access, it seems shorter to me by the way), ...

As anyone experienced such a situation,
or some ideas about the elements I list here ?

Any help is welcomed and appreciated.

Regards,

Denis