We’ve recently had some performance problems on one of our mail servers. Server availability was dropping down to 0 for minutes at a time.
We did some analysis using Task Manager, Perfmon, and the analysis tools on our SAN, and we think the bottleneck was disk read on the data drive - data is on a RAID 5 disk array on an EMC SAN. It’s reasonably well spec’ed, and has been running for 18 months or so without any problems.
We’re fairly sure we’ve eliminated the SAN itself - it’s not reporting any failing disks, or any problems that we can see. Nor is the fibre the problem, which leaves us with the Domino server. None of the server tasks were using any more than 1-2% CPU, but clearly something was maxing out the disk. According to Task Manager, nserver.exe was the process where I/O Read bytes was increasing the most.
Is there any way to track down which user or which database is responsible for high disk I/O ? Any debugging tools that we can use?