Veeam backup for Proxmox, numa option for the worker

[nimda]

I have in use 3 different Proxmox servers (Fujitsu Primergy) on the current kernel 6.8.12-5-pve and lvm-thin storage. All three servers have a current bios and processor microcode.

Two of them are able to create the worker VM by CLI successfully. But one shows "rcu_sched self-detected stall on CPU" on its console view, when more than one cpu is selected.

I contacted Proxmox and it turns out that the worker VM should have enabled "numa" for the cpu. They showed me where I can temporarely add "numa=1" as default setting and then the worker run like a charm.

Therefore I would suggest Veeam to add a numa option to the options for the worker VM.

[rovshan.pashayev]

Hello,

Your request has been registered. Thank you!

[rovshan.pashayev]

Also,

Adding "numa=1" to the settings is temporary, as you mentioned. After redeployment of the worker, you will need to do it again.
Please also open a support case and share the case number here. The support team will instruct you on how to deploy workers with this setting enabled.