UnJustice

Thank you for the response. I read it earlier today, but I was saving up a post for once I did everything I could. So, here's some updates and answers to questions here (I numbered your statements for ease in figuring out what I'm referring to: 1) Yes, but after the issues began. So yes at this point and for yesterday, but no for the very first time I had issues. I tried rolling back from 6.12.3 to 6.11.1 and the docker.image disappeared and some other stuff was being weird. I probably could've fixed it, but also... I just didn't wanna mess around with that stuff. So I went back to the current 6.12.3 2) I attached the Plex stuff without the GPU redactions. 3) I tried going to the 470.xx driver today after reading this. It gave the same exact output, the same error, so I went ahead and went back to the latest driver. 4) I figured the UUID didn't matter, but, you know. I've uploaded the Plex docker stuff along with the Nvidia driver plugin info and nvidia-smi info. I'm fairly confident I copied the numbers over correctly and formatted it correctly as it has (I posted proof yesterday) worked at certain times. Also, as the screenshots show, Plex working was some sort of fluke. I'm honestly lost for even trying to explain what's going on to myself at this point. The GPU hardware is absolutely fine. I've used it, and still can (I did it yesterday and day before just to see), as a VM passthrough GPU. It works perfectly fine in a Win 11 VM for performing encoding tasks. So, it's hard for me to lean too much in the direction of bad hardware (not that you implied such, this is just my own mind here running out of ideas). Also, I used a totally different Plex docker image. Previously I was using the one from linuxserver and now I'm using the binhex plex-pass one (I do have plex-pass lifetime license, just to answer any question there). I made it completely fresh using a different directory and re-added my media and stuff. Everything is functional and fine- but it still won't load using the Nvidia runtime parameter. (This also happened to the linuxserver version too, just for clarity. It was working, but then it stopped for some reason. I think I noticed the container had stopped running after I tried to run unmanic (which didn't run).) 5) I did today after reading this. My unraid is now up running right now, running perfectly fine (minus this issue of course), with a fresh UEFI (copied contents folder) install. I changed out the USB drive because the old drive was... very old. No joke, I found it in a library about 20 years ago! So, it was time to retire it anyway, and I wanted to absolutely rule out any hardware issues on the part of the USB. I had purchased a drive almost a year ago and never felt like dealing with swapping it out...this was an excuse. I've also figured out/fixed some other issues that kept causing unclean shutdowns and then parity checks: one was an rclone mounting script that was failing to unmount. I fixed that so it unmounts now. The other, still a bit of a mystery, is a hassOS VM I run. If I don't force it to shutdown manually, it won't shutdown during the normal shutdown procedures. Anyway, I've been getting clean reboots/shutdowns since identifying and working around these issues. I mean, I don't honestly know what else to mess with here. I've tried reseating and even put the GPU in a different PCIe slot. Same result. I've tried deleting the docker image as I mentioned before and same result even when I have literally just Plex installed/running in the docker. No VMs, just the array running, the plugins, and Plex. It still gives the error. It's got to be something somewhere that I changed or a minor setting in my UEFI settings, I don't know. The only other OS/software solution I'm finding myself considering is a clean USB and see what happens. I should've tried that today, but I wasn't thinking (dinner time, you understand). Perhaps that's my next thing. Just try a completely clean USB boot and see what happens? My understanding is unraid reinstalls every run into RAM though, so, I don't know that I could've messed up something so badly in the flash drive, but it's possible, maybe. I'd be curious your opinion on even bothering to try that. I have other hardware I've considered testing out basically just for fun/science at this point. I have an AMD card (6500XT I think?) that I know for a fact works. I don't use it because it can't handle encoding though, which sucks. I also have an old 700-something series Nvidia card that was ripped out of a server or something. It also works but also doesn't support encoding. Neither can serve my overall intended purpose, but I might try them just to see what happens at this point. Especially the Nvidia card. I would use my desktop's 3080 but the setup to rip that thing out and put it back in later makes me very sad to consider trying. I also, generally, try not to play with my toys that are working. Well, I'll look for your reply and again I appreciate your time and help. tower-diagnostics-20230822-2136.zip Edit: Ok, I'm 99% sure I found the issue. It's actually completely unrelated to the nvidia driver plugin and Plex, et al. apps that would use the plugin. Short version: The Krusader docker is causing it for some reason. I understand this isn't the place to ask for support on that, so I'm not. I'm just simply reporting that it was the thing causing the runc error. Longer version: After several hours of deleting and rebuilding all my docker containers, I figured out a couple things. Perhaps this will be helpful for someone else in the future. First, once the OCI runc issue was flagging, I absolutely HAVE to restart unraid. Turning docker off, waiting, then on against does not work. Stopping the array also doesn't work. It has to be a full reboot to make the error stop popping up. Once I figured that out I quickly realized I could resolve this by process of elimination. Start with a core of containers I sort of know have no issues due to using them for over a year. I installed just the basics to get my plex server online. Traefik, the accompanying database, redis, plex itself. And that worked! Like yesterday (or rather two days ago at this point considering the current hour where I am. oh boy). Ok, so I knew I was onto something. I reinstalled unmanic. It also worked! I sat there and stopped and started the GPU-using containers repeatedly. I stopped docker service and restarted it a couple times. Everything was working fine with these few containers only. I rebooted the server and everything was still working. Great! So I kept installing from there expanding my core group of containers to more and more based on how long I had been using them and vibes basically. I did five at a time and then I'd test for failure by starting/stopping over and over on the containers and docker service itself. I'd reboot the server just to check that too trying to force something to fail and give that error. Eventually, I worked my way down to about 10 containers. Krusader was added and immediately when I stopped Plex and tried to restart it blam-o! I got the error again. I uninstalled Krusader, did a clean reboot, and all was fine. Reinstalled Krusader and blam-o! again. I'm honestly not sure where to begin on that container and making it work. I rarely use it anyway, so may just keep it uninstalled for my sanity.

Hello. I'd like to be able to use my Nvidia 1070 in do some encoding for me via unmanic or tdarr. I also plan on using it for Plex, although I don't expect much usage there. I have used this plugin in the past and it seemed to work fine. I had it uninstalled for a while and recently reinstalled it. It was working on 6.11.1 and the other day I upgraded to 6.12.3. I'm not 100% sure when the issues began, but this was the most major thing I've changed recently. This is the error message when I click apply and unraid redeploys the docker: docker: Error response from daemon: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running hook #1: error running hook: exit status 1, stdout: , stderr: Auto-detected mode as 'legacy' nvidia-container-cli: initialization error: open failed: /proc/sys/kernel/overflowuid: permission denied: unknown. Over the last week or so I've been reading through the comments here, especially the recent ones, and attempting various suggestions. I have tried completely turning off iGPU virtualization (Intel 8700K), restarted the server probably 50 times, deleted the docker.image twice and even started using the directory option. Same results. I was getting this error in Plex over and over and over and then suddenly, and I really do not know what I changed this time, the 1070 became an option in my Plex server and seems to be working! but unmanic and tdarr still do not work. I have posted the unmanic related screenshots, but just be aware tdarr is doing the same thing. Plex seems to be working (?) for now anyway. Please let me know if you need other info or screenshots. Thanks for your time helping people and for all your work as well. tower-diagnostics-20230821-1941.zip

tl;dr for future people: go here: https://your-unraid-ip:47990 in any same-network browser Did you figure it out? Those errors are telling you, in order: open the webUI to finish setup (sunshine devs should provide a link there, but, hey) the port is being is being used (by sunshine which is already running- there's a .sh script which runs at this docker's startup) the .json doesn't exist because the webUI setup hasn't been completed yet since sunshine is running, all these ports are unavailable nothing is running because sunshine is just sitting and waiting for the user to finish setup to do so go to this url in any browser on the same network: https://your-unraid-ip-here:47990 yes, it is https with a self-signed cert. i don't know why. just go with it. ignore the cert warning because it's self-signed and you know that. once you go to that url it's all very self-explanatory from there, but if you want a ton more info or questions go to the Sunshine official setup guide here: https://docs.lizardbyte.dev/projects/sunshine/en/latest/index.html

I was/am having this issue too. Here's what's causing the issue for me as far as I can tell: I have gluetun setup to provide vpn to several docker containers. When I was setting it up, whatever I was reading advised me that containers that rely on other containers (ie deluge relies on gluetun being started to run it will do absolutely nothing if gluetun isn't started, since I always want vpn on that container) have issues post-updates. Someone recommended a CA docker app called "Rebuild-DNDC" which, I'm not an expert, but I believe does what the name says... stops the container it's tied to (gluetun) and does some sort of rebuild. All I know is it works. The problem is, it seems, that when you run a mass auto update of all docker apps, something in the unraid docker gets locked into a repeated loop. Sometimes 3 times, sometimes it appears like infinite loop. The solution for me, reliably, has been: -to let all the apps update -Refresh the browser (so the update window disappears). -Click start on the Rebuild-DNDC container. You'll get an error like "container doesn't exist." It seems like unraid, when you manually try to run the Rebuild app, kills the infinite update loop, deletes the container, rebuilds the container -and then everything begins running as it should again. Possible fix I will try in the future: disabling autorun on Rebuild-DNDC and see if that fixes the looping. Hope this helps someone.