Quick Migration between processor architectures

[pufferdude]

I'm replacing an old Intel-based vSphere host with a AMD-based host. When I tried a test Quick Migration of one of the less-important VMs, the process was successful all the way to the point of restarting the migrated VM on the destination host, at which point vCenter showed an error on the new VM saying that it couldn't resume/power on because the processor architecture was different (or something to that effect.)

It then gave me the option of "fixing the problem and try again" or "discarding the saved state" of the VM so it could be restarted and recognize the new hardware. By the time I went to click "discard", Veeam had decided the power-on was taking too long and stopped the job and reverted the VM back to the source host.

ALL of that makes sense to me and would be expected when trying to migrate a VM in this way between architectures. My question is: SHOULD I do Quick Migration again and choose to discard the "saved state" (and if so, would I lose data?), or am I better off just forgetting quick migrate and instead un-registering the VM on source host, copy VM files, and re-register on destination host, since the VM has to be completely rebooted, anyway?

It seems like I don't gain much from moving my VMs via Quick Migration in this scenario, but wanted to ask the experts.

[Andreas Neufert]

Hello, thanks for reporting this.
We should be able to detect this scenario and shutdown the VM instead of putting it into resume. Please open a support ticket and share the number here.

Support can give you as well a workaround based on a reg key that forces us to use shutdown instead of resume when memory is bigger than x. If you set a really low value there, you will not experience the issue.

As I said workaround. It would be better to check why we can not detect the situation correctly.

[pufferdude]

Thanks. I opened case #05059069 for this.

[pufferdude]

Just to close this thread in case someone comes upon it... the case didn't go anywhere and I had them close it. It finally (today) got assigned to a Tier 2 person, but they wanted lots of logs and processor versions and such and I just don't have the time or ability to get all that info to them, since I long moved past this original issue (I just ended up cold vmotioning the VMs) and don't even have the old hosts anymore. The Tier 2 tech seemed to imply that Veeam "should have" recognized the differing architecture and refused to live migrate, thus all the data they want to look at... but I just don't have the time to get that for them. Sorry!