Hello!

I have a poweredge 400SC with a P4 2.4 and a gig of RAM.  I'm running Dell's CERC IDE raid controller with one raid 5 container.

The system is running the latest (as of this past month) drivers available and is routinely patched for MS-flaws.

I am using the built in gigabit ethernet running at 1000Mb/s.  The system runs OK, and when I run disk benchmarking software on the server locally, I get the kind of disk IO I'd expect.  When I run some basic load tests from other machines on the network, it performs as I'd expect.  It is running an MSDE database, and when users are using the database, it locks up very often, as if the machine is telling everything to hold off while it writes the contents of memory to disk.  This generally lasts about 10-30 seconds, during which time, the keyboard and mouse are unresponsive (although they do cache the inputs) and all server functions are stopped.  Everyone who is on the network has to sit and wait while the server swaps to disk.  It happens with enough frequency that it is a massive impediment to the work flow.

Has anyone else had similar experience?  the same software was running under windows 2000 server in the same environment on a P3-800 w/ 768Mb of ram, and it was much healthier.

Any ideas?