Disk use at 100% on terminal server, RAID5 SCSI
Posted on 2005-05-06
One of my clients is experiencing slow performance within terminal server sessions. They have a LAN with 25 workstations and 1 server. The server is acting as a all in one (domain controller, DHCP server, DNS server etc), it also runs terminal server sessions for the clients as they have a large ACT!6 database which when run over the network with ACT installed on each machine was very slow when performing queries but when running on the TS it ran like a dream (that was then).
Now I know your going to say its not wise to run TS and domain controller and all on 1 server but, when 1st installed about 4 months ago all was running fantatically quick, better than I would have ever expected. Over the past couple of weeks the client has been calling me to report that the TS sessions were running very slow and so i ran some tests, CPU and memory useage are quite low, but as soon as more than a couple of TS users log on and start using the ACT database, disk useage goes through the roof and starts running at 100% (with an occasional dip) this is causing problems.
Could any of you shed any light on what might be happening and how i can get things back to how they were when it was all 1st installed?
The server specs are:
Make: Dell PowerEdge 1600SC
CPU: Dual Xeon 2.4GHz
RAM: 1GB DDR ECC SDRAM
Drives: 3x 36GB SCSI Ultra320 (10,000rpm) at RAID5
RAID Ctrlr: PERC4 SC, 64Mb
Network is 100Mb
*All TS users are logging on using a single username, on average there are 14 TS sessions running at any one time.
Action taken so far:
-changed settings in actdiag to the best for performance (fastest)
-disabled oppertunistic locking
-re-indexed the database