Hi there
My system looks like this:
Host: Sun X4150 (2x Intel Xeon x5460 quad-core processors, 32GB RAM, 4 disks in RAID5 configuration), running Windows Server 2003 Enterprise (fully patched), with latest build of VMWare Server 2.0
VM's: three running Solaris 10
five running Server 2003.
At random intervals (usually when one of the VM's is doing something major like rebooting, deploying a large web application, etc but sometimes for no obvious reason) the host machine grinds almost to a halt, dragging the VM's with it. At these times, CPU and Memory useage are running fairly low, but Disk IO is extremely high - queue of between 200 and 300 operations (compared to a normal queue length of under 6).
I've followed all common sense ideas to try reduce the load on the disk, such as uninstalling antivirus tools, avoiding disk-intensive (or for that matter, ANY) operations on the host, confirming that write-caching is enabled, and so on but there has been no noticeable change.
Any other suggestions for ideas which I may have missed?