We are just finishing a rollout of 10.2.2 (upgrade from ZDM 7). Have
about 1000 workstations, and 2 primary servers.

Both primaries are running Server 2003 R2. One is a Poweredge 1950, 2
dual-core CPUs with 4 GB, the other is an ESX 4.0 VM with 2 virtual CPUs
and 4 GB.

The default closest server rule points to the physical server, with the
VM second on the list. CPU usage, network utilization on both servers
is very low. Memory load ("commit charge") is 3.2 GB on the physical
server and 2.3 GB on the VM.

During test and pilot, we were very happy with the response time in ZCC.
Almost every mouse click gave a neary instant response. But the more
workstations that are registered, the slower it has gotten. Now it's
like wading through molasses.

The strange part is that the response is equally slow on both primary
servers. I am assuming (since there seems to be no way to monitor),
that the physical server has the bulk of the client connections. The
secondary shouldn't be very busy, but ZCC is still just as slow.

We don't have any complaints about the client side. Login, refresh, and
launching of apps is just about as fast now as it was during the pilot.

The database is Oracle 10g on a separate "big iron" linux server. The
dba tells me he cant see any significant latency or anything.

I am tempted to install a third primary server just for ZCC, but thought
I would try here first. I'm actually a little doubtful it will help anyway.

Any advice appreciated. Regards,
Phillip E. Thomas