I have following configuration: 1. active/passive cluster, Win2000, SQL2000 2. 4 CPUs in each node, 4GB RAM in each node 3. Disk arrays configured for optimal performance (different arrays for diff things, diff RAID levels etc.) I have few periods during the day when CPU utilization is stuck between 30 and 50% without a single peak on any side. That can last for more than hour. All CPUs are behaving the same way. Response on the client side is very slow. The only process in that period of time that is using CPUs extensively is sqlservr.exe. I've checked deadlocks - nothing. I was running Profiler to check slow queries - nothing. Other perfmon counters seams to be OK (one does give me concern actually - buffer cache hit ratio is actually over 100, around 100.150 or similar?!). Without any performed action, this behaviour stops as sudden as it started and utilization on all CPUs is back on expected 0-100%! Any ideas? Appreciate your help!