Main Site | Join Robin Hood Coop | Projects | Events | Blog | Media | Forums | Mailing List | Twitter | Facebook

Why should Windows Server 2016 GPU Driver crashes --> whole server becomes unresponsive

I am new to Blockchain. Recently I am facing an issue, that when the GPU driver on my Windows Server 2016 crashes (AMD Blockchain driver), the server becomes unresponsive and all apps (VPN, apps, …) become no longer accessible.

And even when I hard reboot it (by pressing the power button for several seconds and rebooting it) it does not boot up correctly. The only way I found that fixes it, is to move the server back into my office, attach a screen to it and restart. When I do that it restarts without issues, no error message during startup and no message in event viewer. When I then move it back to its place (without a screen attached) it boots normally.

So to me, it looks like the GPU driver crashes (it has a single RX580 inside for mining) and only seems to recover when you boot with a screen attached…

My question is, how do I prevent the miner / GPU driver crashing to crash the whole server. This is a huge pain at the moment since it happens every few days and I can’t fix it until I am home to do the procedure described above (with lots of services running on the server that are down until I fix it).

Also, I have been eyeballing with switching to Debian as an OS on the server and use VM’s for my Windows-only stuff. How does Debian react to GPU driver crashes?

Thanks & Regards
Camillelola