Recent content by merc

I got a mostly working passthrough of the gpu by following instructions here and there, but I'd like a little more info to properly understand what's required. And AMD's ridiculously convoluted driver stack certainly isn't helping. Need some definitive answers on these questions please. 1)...

And one more update for posterity, as it turns out, the reason fstab failed to mount nfs all along, was because it was waiting for ipv6 to timeout, since I had ipv6 set to dhcp in the CT network settings. That caused a lot of problems with systemctl not starting enabled services on reboot...

Answering my own question, as I just found a workaround. Simply added a typical mount nfs share line to the CT's fstab and then created a root cron job @reboot mount -a This gets the job done, but I'd still like to know what's the deal with autofs.

Trying to automount an NFS share to a CT that's being shared from another VM, both on the same pve host. The share isn't mounted to the pve host because it's only available when the other VM sharing it is running and I don't need it on the host regardless. So, I got an ubuntu 20.04 CT...

Will do, thanks.

Need to create a VM that runs from its bootable iso. No storage required. Can't find any option to not create a hard drive for it. Something I'm not seeing?

No, I hadn't started an x server. Good point, wasn't aware of that requirement, so that explains it. As for what I need from glxinfo, it's troubleshooting amd driver's abysmal performance in lxc containers. ffmpeg doing simple h264 vaapi transcodes at literally 1.8 fps. But I doubt that's got...

glxinfo can't give any useful info because DISPLAY env variable is not set on proxmox host, even when logged in to actual console with a monitor and keyboard attached. Forcing export $DISPLAY=:0 or whatever else doesn't help either. Any ideas why is that?