Hi, I hope this is a correct category for this post.
I have been entrusted with a SUSE Enterprise server, with proprietary computation software configured on it (without any sensible documentation). It has been configured a really long time ago and no one knows what’s really going on there, so I can’t reinstall it until I find out.
The problem is… Right before it got to me, someone bought a Tesla K40c (without video outputs), put it inside and slapped proprietary Nvidia drivers on it. Since then, DE stopped working due to (?) probably incorrect configurations. I’m no expert in Linux, have some experience playing with Arch and such, but not enterprise distributions etc. The DE is crucial for that proprietary software I mentioned.
So, let’s get to details. Originally, Gnome crashed on startup with an error “oops something went wrong”. My first guess was that Xorg is now trying to run on that Tesla, which doesn’t have video outputs. So I made a dd backup, generated a new Xorg.conf and edited a few lines to what I think would be a working configuration. I uploaded it here for you to take a look:
When I try to run gnome-session, the error I mentioned before still happens. Xorg log is also in the pastebin.
I noticed that there’s a VNC server running on it and it starts with “startx”. But I can’t connect to it (it says something about password not being configured, even though I made sure it is in a config file in correct location). Maybe that is because of Gnome not running properly?
I tried reinstalling the Nvidia drivers (newer version) and configuring everything again, but no luck.
Was I on a right track thinking it’s crashing because it is trying to run from the Tesla and not from Matrox, or could the problem be somewhere else? I found Nvidia linux drivers always acting weird on my personal computers, this Tesla seems to be a bigger beast than that.
Thanks in advance!