This occurred after a hardware migration from an old bare-metal server w/ a Fibre Channel SAN to an ESX 4.1 host using iSCSI. I have other NW6.5 VM's running on this host and others w/o performance issues.

Server runs GW8.02 HP3: POA, MTA, WebAccess, GWIA. Server also supports mobiles via a BES and Datasync (SOAP).

During off-peak hours, the server settles down, but util is at 100% for most of the workday. The busiest thread is typically the POA. Handler threads are set to 15. Active users averages near 300. These settings worked well on the old hardware.

Complaints are streaming in from users. At this point I'm considering moving the database from the iSCSI SAN to the local SATA drives on the host to eliminate the possibility of disk latency. I don't want to keep it there long-term however, so I'm hoping to get some other suggestions.