SBS 2011, Hardware failure / HD Moves to identical chassis / CPU Maxed
Experts,
Our server that hosts SBS 2011 experienced a hardware failure and has been moved to a 100% identical server chassis. The server is the exact same model (IBM x336) - Same exact specs.
The RAID 1 HDs were moved into the new server and everything went as expected. I needed to reactivate the OS, as it detected a hardware change, and everything is up and running.
Of course, I didn't expect to see what I'm experiencing now with it. It appears that 3 services are in a knock-down drag-out battle to the death on CPU usage.
w3 - IIS
SQL
Exchange UM.
Ever since the change, (Approx 30 hours ago) - CPU usage on the SBS box has been 100% from a mix of those services. Of course, this makes performance of the server itself absolutely horrible.
Nothing was out of the ordinary before the failure. (And on note - the failure was attempting to upgrade RAM in the old chassis). I shut the old server down, and tried to mount the memory, where the old server coul no longer POST after that point (Ebay purchase). - So move everything over to a 100% identical chassis, and it fires up no problem.
Would anybody happen to know why the CPU usage would be as horrible as it is on this server?
Our server that hosts SBS 2011 experienced a hardware failure and has been moved to a 100% identical server chassis. The server is the exact same model (IBM x336) - Same exact specs.
The RAID 1 HDs were moved into the new server and everything went as expected. I needed to reactivate the OS, as it detected a hardware change, and everything is up and running.
Of course, I didn't expect to see what I'm experiencing now with it. It appears that 3 services are in a knock-down drag-out battle to the death on CPU usage.
w3 - IIS
SQL
Exchange UM.
Ever since the change, (Approx 30 hours ago) - CPU usage on the SBS box has been 100% from a mix of those services. Of course, this makes performance of the server itself absolutely horrible.
Nothing was out of the ordinary before the failure. (And on note - the failure was attempting to upgrade RAM in the old chassis). I shut the old server down, and tried to mount the memory, where the old server coul no longer POST after that point (Ebay purchase). - So move everything over to a 100% identical chassis, and it fires up no problem.
Would anybody happen to know why the CPU usage would be as horrible as it is on this server?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Problem was CPU Cache processing. Changing to "Write-Back" solved this issue.
ASKER
Turns out the problem was actually the BIOS config. I was shooting this pretty in depth and just couldn't figure it out, until I decided to try and fire up the old box.
I was able to get it running with a single stick of RAM, and I started poking around in the BIOS. One significant item stood out, in that the old server was configured for "Write Back" on the CPU cache, where the new one was set on "Write Through".
I figured I might as well change the new server to be exactly how the old one was, even though there shouldn't be much difference speed wise between the two options.
I was completely wrong. As soon as I booted the new server with "Write Back" selected, - I'm not joking in the slightest - the 100% CPU pegged problem went away. Currently, the server is idling at 2-3% CPU and hasn't hit above 35-40 on full load.
It shouldn't have made that much of a difference, but it most certainly did on that CPU Caching option.
Thanks for everyone chipping in to assist me with this problem.