Win2K3 guest boot failure after esxi host crash(es)

[z99dray]

Hi,

I have a Windows 2003 R2 guest OS running on a ESXi 3.5 host (U2 + 143129 patch level) that has started to show a blank screen after the bios messages appear, as opposed to the usual Windows logo and loading screen expected.

The HOST and VM are only two days old and the VM has the same disk image attached that it always had. The esxi host has been locking up though and the only way to get it back is to physically power off the host machine. At first I thought that perhaps the disk image had become corrupted, but I've been able to boot a Windows CDROM in the same VM to get into the recovery console and confirm that the disk image is accessible by the host.

I was also able to create a new guest, point it at the same disk image and the OS loaded. So, I believe it's something to do with the original guest or it's configuration rather than the disk image. Recreating a new guest each time this happens is less than ideal and I would like to know how to stop this happening in the future.

Can anyone suggest what the problem could be?

Thanks

[Lightbulb]

Check your /var/log/vmkernel log for signs of disk issues. It sounds as if there may be some sort of corruption of the datastore.

Is this new hardware?

Does your VMFS datastore reside on local storage?

[Lightbulb]

Sorry just noticed this is ESXi not sure how accessible any logging will be, my bad.

Still verify that there are no hardware issues. Check for BIOS and firmware updates yada yada.

[z99dray]

Hi

Thanks for you reply. I've got the latest bios for the host and there doesn't appear to be anything obvious in the logs. I don't think it's a host or datastore (which is local) issue, because the same vmware disk image will boot into the OS if I create a new guest virtual machine and point it to the image of the virtual machine that won't boot.

cheers..

[z99dray]

Hi,

Well, I've managed to resolve this just by rebooting the host! I'm not sure if booting the vm disk image from another virtual machine fixed the problem or whether a simple reboot would have done it in the first place.

Cheers.