Purple screen on Windows Server 2008R2 Start-Up

[cyberjohnny]

Hi,

We have a single Intel S5500BC server (dual XEON E5520 CPU's, 32GB ECC, Adaptec RAID with 2 x RAID sets)

Originally the server was running ESXi4 when we added a new VM (Windows Server 2008 R2 Standard x64). Whenever we start this VM it causes ESX to crash (see attached purple screen). The VM sometimes starts and runs for a minute or 2, but then the purple screen.

We originally put this down to some sore of hardware incompatibility between our system and ESX4. We later upgraded to ESX4.1 (fresh install on new USB drive) hoping this would sole the problem, but it didn't. Last night we upgraded to ESXi 5.0, and the problem is still there. Although the error is a little different. It indicates a hardware problem, however this server has been running several Windows Server 2003 x86 and Windows XP Pro x86 VM's for around 2 years, and has 'never' had a problem.

We have tested the server with different Server 2008 R2 x64 images, and they all produce the same result - purple screen 1 to 2 minutes after booting.

We are currently using a loan ESX server to run our 2 x Server 2008 R2 x64 VM's, but would really like to get this rectified.

As this server is in use 7 days a week for 15+ hours a day, running in-depth hardware testing on it will be difficult, and our loan ESX is not powerful enough to run all of our VM's. So anyone that can decipher the purple screen or point me in the right direction, the advice would be much appreciated.

Cheers,

John.

[admin]

it's clear that it's a hardware error. vmkernel log file might have helped to understand the issue a right way to start diagnosing this issue it's to run hardware diagnostics tests on this machine, a support request with the hardware vendor might help.

[golddiggie]

I would do as the purple screen states and contact the hardware vendor to see if there's an update (bios or firmware) that needs to be applied to fix this. You could also open up a support case with VMware to help isolate the issue. Since the screen shows a lot of CPU messages, that could be a good place to start your investigation.

Since you've only been running 32 bit servers until now, that makes me really think it's a processor issue. Or it could be something simple as a bios setting on the host. During your next window of opportunity, make sure that the bios is set properly (check with the hardware vendor and/or VMware for that.

[cyberjohnny]

Thanks for your suggestion guys. We have the latest Intel BIOS and have been through every conceivable setting to attempt to eliminate the problem. A processor problem does seem likely though given it only happens with the 64bit OS.

[satya1]

cyberjohnny wrote:

Thanks for your suggestion guys. We have the latest Intel BIOS and have been through every conceivable setting to attempt to eliminate the problem. A processor problem does seem likely though given it only happens with the 64bit OS.

Hi you mean to say ESXi version was 32 bit ?

Yours,

Satya

[cyberjohnny]

Hi Satya1,

No, the ESXi is 64bit - I wasn't aware that there even was 32bit version? The problem is that the Windows Server 2003 (x86) / 32bit servers run without a hitch. ESXi purple screens minutes after firing up and Windows Server 2008 R2 / 64bit VM. So we are assuming the way the 64bit OS/VM access' the hardware is causing this issue?

We have scheduled down-time the weekend after next to run hardware testing. Hopefully this will find something.

Cheers,