Yesterday morning, a number of users complained about Outlook/OWA being slow. After taking a look around the server (which was not very responsive), I noticed the HD LED on the front was on solid -- and has been since then. Nothing has been changed on the server. A reboot didn't resolve the problem. Exchange is the only application on this single, dedicated server (no Front End).
The Exchange Database and Log files are D: drive partition on a RAID 5 volume (3x150GB) in a Dell PowerEdge SC1420 with a Dell CERC SATA 1.5/6ch RAID Controller, 2GB of RAM (3069MB Recommended Initial and Maximum Page File Size on C: drive).
According to the RAID Storage Manager, all of the drives are reporting "Optimal" and none have SMART errors. The Read Cache has been enabled but the Write Cache never has been (since I've been here over the past 2 years). I'm considering turning that on, but not if I don't have to to help prevent file corruptions if there's a hard shutdown.
I ran Perfmon and can easily see the high Disk Write value but not sure why. RPC requests are relatively low. Screenshot attached.
Stopping all of the Exchange Services significantly reduces the disk utilization and the server is very responsive (via RDP). The utilization skyrockets again when they're turned back on.
Before yesterday, Backup Exec was backing up Exchange at ~1,500MB/sec. When it kicked in last night, it was running at about 64MB/sec before I stopped it.
Running an Analysis in Disk Defragmenter shows 43% Total Fragmentation and 87% File Fragmentation. I understand that doesn't necessarily cause problems like this.
I managed to run ESEUTIL /D on the Public Folder store last night. It took 7 hours for a 2.1GB EDB and STM file. The PRIV1.EDB is about 44GB and the STM is about 22GB. I also understand we're reaching the hard 75GB limitation so we're planning on a new server early next year.
Hopefully that's enough information for you. If not, please let me know what else you might need. Thanks!!