Windows 10 based server BSoD when direct play is initiated from client

It sounds crazy right? I thought so too. First off, I work for the state of Texas, managing complex vSphere environments, so I am not a stranger to figuring things out on my own. Here’s whats going on.

I recently jumped over to Jellyfin from Plex, and removed all remnants of the PS software. Since installing jellyfin, as a service, I was getting random BSoDs in the event viewer. Some of my users reported that it would just be very slow sometimes and not show cover art, then start working. (Obviously rebooting in the background). So I started doing some testing of my own. I cannot repro the issue every time, but after about 3-4 days, sometimes sooner depending on actual usage, it will happen. Every time it happens, I have diagnosed it to a client requesting a direct play stream. My Shield TV will play most content directly, so I was able to get it to reboot once on its own.

I am wondering if there is any logging I can provide, turn on, or examine? Could this be an issue with caching? I am at a loss, but I do not like that my machine just reboots. I keep a user logged in for other reasons, and its annoying to find that its rebooted since I last logged in, just about every time lately.

Other than this, I love the package, and have been raving about it to my colleagues at work.

That’s really, really weird. Is there anything caught by the Event Viewer close to the actual crash? We use .NET Core, so I’d imagine if there’s some kind of crazy runtime error, it should get caught there.

I will say though, I’ve personally tested both the nightly and 10.3.7 on Windows 10 Enterprise 1903 18362.295 over the past week, and Direct Play (albeit to a local browser on the same machine) has been fine…

The only reference other than the kernel-power event was that the computer rebooted from a bugcheck. I will investigate the dump and get back to you. Thanks for your reply.