Linux Proxies backing up Windows 11 VMs with vTPM

[pmichelli]

I have posted in the past that when I was running Ubuntu 22 proxies they would start throwing errors and fail jobs after approximately 2 months. I would need to remove them as proxies and Linux machines, then re-add them back to make them start working again.

Recently as a test I decided to drop one proxy at each site from Ubuntu 22 and replace it with Windows 11. I have jobs that use 2 proxies for load balancing and redundancy.

Each time I had this proxy failure on Ubuntu I noticed it was always proxy 2 at each site. After really digging into this, it seems for almost every single run, proxy 2 would be the one that would get to backup my Windows 11 machines with vTPM modules (I did not have these when I was running Veeam 11 and had no problems with Ubuntu proxies either). It would always fail over to NBD when trying to backup Windows 11 which I now understand is due to the fact they are encrypted. Veeam says we should use proxies that are encrypted to backup encrypted VMs.

Now that proxy 2 is Windows 11 as well, the Windows 11 VMs all seem to hit this proxy during the job run and are using Hotadd, which is what I want.

Now here in the interesting part : I have not had any failures like I did when Ubuntu 22 was acting as proxy 2.

Looking at the Veeam logs it seems the errors were always related to NBD failures on proxy 2 with Ubuntu. I never had a problem with Proxy 1.

So to the devs that may read this forum. Perhaps you can investigate this behaviour. I think there is a problem when an Ubuntu (or maybe other Linux) proxies are trying to backup Encrypted VMs or VMs with vTPM.

Just wanted to let you know. I am not in a position to go back to a second Ubuntu proxy as support was not very helpful and always claimed they did not know Linux well (why offer it as an option when the teams can't support it properly). I could possibly spin up a second "dev" instance of Veeam 12 and test this theory to hopefully generate logs that may help solve this once and for all. If you feel that would be helpful, let me know and I can try and find some time to do this.

Here are the old SRs. Perhaps the logs are still available and may help shed some light

Case #06200095
Case #06043966

[Andreas Neufert]

The 2 tickets are closed I think already.
One ticket had the issue that we could not connect by TCP902 to one of the ESXi hosts.
Please check again that there is no communication issue (Check firewalls, hardware and on the OS).
Even for HotAdd processing we need to connect to this port on the ESXi host to perform some of the operations.
The second ticket was about a communication issue between our VBR server and one of the proxies. The case comments listed that the reinstall of our transport components solved the issue.

It looks like that you have further issues and I highly suggest to open again a support case and potentially escalate the case (there is a button "talk to a manager" in the support system). It would be really hard to analyze what is happening here through the forum.

As well please check if you implemented all needed pre requirements: https://helpcenter.veeam.com/docs/backu ... ml?ver=120