Boot issues and corrupt file system after replication

[GabesVirtualWorld]

Hi
I'm migrating VMs between two datacenters for a customer using replication. The majority of the migrations is going without issues, but we have 3 VMs now in which we run into some issues:
- Windows 2000 VM HW Level 7 on ESX 4.0.0 (yeah I know). When we failed over this VM, Windows wouldn't boot and report STOP error 0x00000007B. Which is probably something with VMware Tools / Disk drivers.
- Windows 2003 VM HW Level 7. When we failed over this VM, Windows kept on rebooting because of a BSOD with STOP error 0x00000024.

For the Windows 2003 VM it turned out that the ntfs was corrupt. Using a XP boot CD and running chkdsk /f /r brought the VM alive, but all NTFS permissions were gone. SQL wouldn't start anymore and some other apps had problems. After doing some fixes with resetting security I had everything running again, but since we couldn't trust this install any longer, we reverted back and will soon try again when the customer offers a new time window.

I wasn't able to fix the Windows 2000 VM. I think i first need to upgrade the VMware Tools on the VM and then give it another try.

Now, eventually I will get this fixed, but I'm surprised by the errors I'm running into. I would expect the VMs to be 100% identical on both sides. The original VMs do boot without issues, something that I wasn't sure about when the replica started to fail. I was very much afraid the original would now fail too, but these run without issues. When doing the failover we don't upgrade the HW, we don't change the settings of the VM. The procedure we use for failover:
- Shutdown original
- Resync the job
- In Veeam perform the failover
- Check the replica VM to see if it boots.

What could be wrong here?
PS: All replication jobs run without issues !!!

[Gostev]

Hi Garbie, please include support case ID for all technical issues you are posting, as requested when you click New Topic. As explained there, moderators will remove topics about errors or issues if support case ID is not included.

One idea I have is that those VMs were created outside of vSphere Client (with some P2V process etc). Such VMs may have non-standard virtual disk sizes and/or layout, which can be causing all sorts of issues. But this can really be anything, including bad target ESXi or malfunctioning area of storage. Without seeing the logs, we can be guessing forever...

Thanks!

[GabesVirtualWorld]

Ok, thank you.

[dellock6]

Gabrie,
I have too a doubt since those are all "old" VM, maybe their virtual disks are not multiple of 1KB (kylobyte)?
We have a KB (knowledge base this time, no pun intended) in regards, I know it's not about replication, but maybe VMware is creating problem with that strange size even on replicas:

http://www.veeam.com/kb1848

Luca.

[vmJoe]

This sounds like you did not get an application consistent copy of the source VM.

Did you replicate this VM whilst the server was powered down, then failover to that powered off restore point? It's been a while since I touched a W2K machine, but I don't believe that VSS support came along until W2K3. So chances are if this was replicated live then you are going to have issues when you failover to a replica.

[Gostev]

He gets BSOD at OS boot time, not application issues.

[GabesVirtualWorld]

The way I replicate the W2000 VM:
- while running replicate daily
- job has VSS disabled
- shutdown source VM
- replicate the shutdown VM
- do the failover
- boot replica which then fails with BSOD

[GabesVirtualWorld]

Since it is Win2000 it might indeed be not alligned properly. Will have to check that, but I only have very small down time window for it.

[tsightler]

Alignment or anything like that shouldn't cause any problems as Veeam doesn't care, or even know, anything about the underlying OS when it's doing a replication. It would just copy the disk as it is.

I'm guessing that perhaps it has something to do with the configured SCSI adapter, perhaps it's one so old that a newer ESXi versions emulate a slightly newer version that the old Windows 2000 driver does not even recognize. That's the normal error you see when Windows 2000 can't find the disk during boot. What SCSI adapter is being emulated on the ESX4 version and what driver version is being used in the Windows 2000 OS?