Any methods to verify a Superdome server is well-configured for Windows 2008?

[Dennis-Lee]

I have checked system management homepage and run Offline Diagnostics Environment but no errors found.
We have 2 Superdome servers(clustered), 32 CPUs, 256GB RAM
1. Both of our superdomes ran well 3 months ago. I failover and failback many times at that time. I regularly backup windows image using breaking mirror(hardware RAID).
2. Someday(that said, 2 months ago), SD04 restarted unexpectedly and I found some resources were missing(half of HBAs, half of NICs). HP was on site and replaced cell board 0 and the problem is fixed.
3. Some other day(that said, 1 month ago), I failover from SD03 to SD04 and found SD04 ran strangely slow(CPU utilization, page-ins, IO queues, network utilization(Teamed,40Gb/s) are low but SAP dialog response time are 3 times higher than another node: SD03 and I found it is 5~10 times slow running any frontend programs, ex: command prompt programs, notepad). I thought it would be something wrong in the OS because HP and I couldnâ t find anything wrong in hardware event log. I used the â last known good OS imageâ (2010/1/14) to restart SD04 and I thought the problem is fixed(the users do not allow me to shutdown system from time to time for testing).
I am not challenging there is a hardware issue because I don't have any evidence. It's too strange because if the slow situation is caused by software, it should be fixed when I used the last known good OS, right? Please help, any opinions will be appreciated.