Toady001

Trying to pass my Quadro GPU to the VM machine. Followed the steps in the OP but after rebooting my frigate docker no longer starts which also has the quadro passed. Is this intended? i.e. can you only pass through to either docker or VM and not both? GUI also freezes after passing and starting up the Windows VM with the Quadro gpu.

Hi guys, Honestly I don't know where this should be but it seems to be the most fitting. There appears to be no external forum / discord for the acestream app either. So the container is failing to start due top permissions. 2023-08-30 10:35:26,735|MainThread|acestream|failed to init app: PermissionError(13, 'Permission denied') Traceback (most recent call last): File "core/Core.pyx", line 402, in File "core/src/CoreApp.pyx", line 1390, in File "core/src/CoreApp.pyx", line 1377, in File "ACEStream/Core/Session.py", line 139, in __init__ File "/usr/local/lib/python3.8/os.py", line 223, in makedirs mkdir(name, mode) PermissionError: [Errno 13] Permission denied: '/srv/ace/.ACEStream/collected_torrent_files' That in itself is easy to understand and looking at the dockerfile there is a user being added called "ace" which I feel may be the culprit. FROM python:3.8 SHELL ["/bin/bash", "-o", "pipefail", "-c"] ENV BASE_URL="https://download.acestream.media/linux" ENV ACE_VERSION="3.1.75rc4_ubuntu_18.04_x86_64_py3.8" RUN useradd --shell /bin/bash --home-dir /srv/ace --create-home ace USER ace WORKDIR /srv/ace RUN curl --progress-bar $BASE_URL/acestream_$ACE_VERSION.tar.gz | tar xzf -;\ pip install --no-cache-dir --upgrade --requirement requirements.txt CMD mkdir /dev/shm/.ACEStream; \ ln -s /dev/shm/.ACEStream .ACEStream; \ ./start-engine \ --client-console \ --live-cache-type memory The thing is though, if the user is added surely they have the right privs inside the container, unless permissions are extending to outside the container. Perhaps anyone else that has successfuly run this container that can give some assistance?

Solved. Did a restart, and the server no longer picked up the USB stick. Not sure if that is related, but got everything booted up again and managed to log in.

For some reason the GUI was reporting an incorrect login details. Stupidly I carried on thinking I was typing it wrong and now I am locked out. Is there a timeout period for the GUI or is there a way via cli to unblock the account (somehow my ssh session is still open) Thanks

Thanks, busy restoring previous version via GUI. So I take it the fix before updating is backup data on the drive, format to xfs (same as other drives) and then update?

After updating to 6.12. One of my drives is no longer mountable. All other drives in the array are fine which is running on XFS system. The cherry on the top is that I now noticed my flash backup didn't download correctly so restoring is no longer possible now. Attached diagnbostics. toadcreek-diagnostics-20230619-1015.zip

Just to add onto the above. Have issues with one drive with btrfs fs. Jun 19 09:46:08 ToadCreek emhttpd: Mounting disks... Jun 19 09:46:08 ToadCreek emhttpd: mounting /mnt/disk1 Jun 19 09:46:08 ToadCreek emhttpd: shcmd (38): mkdir -p /mnt/disk1 Jun 19 09:46:08 ToadCreek emhttpd: /sbin/btrfs filesystem show /dev/md1p1 2>&1 Jun 19 09:46:08 ToadCreek emhttpd: ERROR: cannot scan /dev/sde1: Input/output error Jun 19 09:46:08 ToadCreek emhttpd: Label: none uuid: ece161a0-e7e9-4078-9161-9b73bbc1597a Jun 19 09:46:08 ToadCreek emhttpd: Total devices 1 FS bytes used 1.64TiB Jun 19 09:46:08 ToadCreek emhttpd: devid 1 size 3.64TiB used 1.71TiB path /dev/md1p1 Jun 19 09:46:08 ToadCreek emhttpd: /mnt/disk1: no btrfs Jun 19 09:46:08 ToadCreek emhttpd: /mnt/disk1 mount error: Unsupported or no file system Jun 19 09:46:08 ToadCreek emhttpd: shcmd (39): rmdir /mnt/disk1 Deleted the file in /sbin/btrfs with no positive result. Question though, if i had to reformat the drive, will the parity rebuild it? Don;t want to lose any data.

Update: Potential Problem Found So since the above, I have disabled hardware transcoding for plex and server is still fine after 4 days. The GPU is still transcoding frigate so seems unrelated to GPU and more related to the plex container / gpu passthrough / hardware transcoding that is causing the server to bug out.

Thanks for your feedback. I'm using plex for transcoding. I did notice last night that running htop there's a plex eae service that was maxing some cores. I believe this service has something to do with plex's skip intro function. I find it strange how transcoding has anything to do with CPU... it's supposed to be handled by the GPU but I'll disable hardware transcoding and see if it becomes more stable.

Hi guys, I have been putting this off for quiet some time to try and troubleshoot this myself but my unraid server keep crashing every few days, sometimes even within 24hrs. From what I was able to see in the GUI, is a couple of CPU threads hit 100% usage. Not always the same threads but this does happen before a freeze. The GUI becomes completely unresponsive. The only way to get it to reboot is to SSH in and force reboot with reboot -f. A graceful shutdown does not work. It started after installing nvme cache drive and a quadro p1000 for frigate and plex transcoding. I can't see any errors relating to the GPU. I did read that changing cache settings for docker and vm's could cause this if the files are split between the cache and array so i stopped docker and vm's and ran the mover. Hasn't fixed the issue. OS Version: 6.11.5 System: Dell R510 Dual Xeon® CPU X5675 / 32GB RAM Unraid Log Report.html toadcreek-diagnostics-20230605-1210.zip