Cannot restore VMs from Azure to AWS

[pmichelli]

Case : 06172714

We are able to backup successfully 2 VMs from Azure. One is Ubuntu 20.04, the other is Windows Server 2022
We have a copy job from Azure repo to local repo

I have tried to restore with and without the helper VM. It always fails (each OS) at the importing part. Here are my error messages. I don't know what to do and really need to get this done ASAP as we are cancelling the Azure tenant in a month.

Ubuntu :
7/13/2023 11:08:36 AM Error Importing VM Error: Failed to import machine to Amazon EC2: ClientError: Unsupported kernel version 5.15.0-1041-azure

Windows :
7/13/2023 11:44:17 AM Error Restore job failed Error: Failed to import machine to Amazon EC2: The t2.medium instance type does not support an AMI with a boot mode of UEFI. Only instance types built on the Nitro System support UEFI. Specify an instance type that supports UEFI, and try again.

[HannesK]

Hello,
as far as I see, the error messages have the solution

The Ubuntu machine has a kernel that is too new. So you need to downgrade to a supported kernel. Here are the requirements for import to AWS

The Windows machine has UEFI boot and a compatible configuration is shown in the error message. On the other hand, UEFI is supported in general for import (except a few AWS regions which should not be relevant looking on the country where you come from). This link might help to find a compatible EC2 instance type.

Best regards,
Hannes

[pmichelli]

Hi Hannes

Thank you for replying. I guess with all the time spent I overlooked the very basic things. I was able to recover the Windows 2022 VM by selecting a T3 tier from AWS.

The issue seems to be with the Ubuntu 20.04. According to the matrix on AWS, they support kernel 5.4, which was released in 2019. It is not feasible to downgrade the kernel to one so old. It would break things and my security team would want to hang me I suppose this isn't a Veeam problem, rather the constraints AWS puts on us.

The odd part is we run Ubuntu 20 and 22 in AWS and are always patching them to the latest kernels. It seems the limitation is on the importing of a VM.

Have you run across this as well? It would be really nice if VBR told us this before it starts the restore vs after it already deploys a helper VM and imports the data.

Is my only option really to try and downgrade the kernel to something very old?

[HannesK]

yes, I have seen import issues... that's why I know that AWS documentation

Is my only option really to try and downgrade the kernel to something very old?

depending on what the application is, I would just spin up a new VM in AWS and copy the data with rsync or whatever directly instead of investing time in the import process.

[pmichelli]

Thank you as always HannesK. That is exactly the path we've chosen. New VM and rsync. Much less headaches