(2024-03-18, 11:22 PM)matchaman Wrote: It seems that the new docker image is working well with minimal skipping/stuttering, but I am now having an issue where whenever I stop the docker container, it resets the jellyfin server to a like-new setup, making me reconfiugre everything. How do I fix that issue?

(2024-03-18, 11:22 PM)matchaman Wrote: Is there a fix for the Windows 11 OpenCL driver issue, or one being worked on, or would adding a graphics card help fix the issue? Or what is the best method of hosting the jellyfin server? On bare linux? Or does that require a lot of conifig with drivers as well?

(2024-03-18, 12:10 AM)TheDreadPirate Wrote: There is a known issue with Intel's Windows 11 OpenCL driver.  That is likely what caused the issue on Windows.

(2024-03-18, 11:30 PM)TheDreadPirate Wrote:

(2024-03-18, 11:22 PM)matchaman Wrote: It seems that the new docker image is working well with minimal skipping/stuttering, but I am now having an issue where whenever I stop the docker container, it resets the jellyfin server to a like-new setup, making me reconfiugre everything. How do I fix that issue?

It sounds like you didn't provide a volume for the container to write /config to.  /config is where Jellyfin writes all of its persistent data to.  You also need to provide a volume for Jellyfin to write /cache to so Jellyfin doesn't have to regenerate that after each restart.  Both should be on a NVMe volume.

(2024-03-18, 11:22 PM)matchaman Wrote: Is there a fix for the Windows 11 OpenCL driver issue, or one being worked on, or would adding a graphics card help fix the issue? Or what is the best method of hosting the jellyfin server? On bare linux? Or does that require a lot of conifig with drivers as well?

You can revert the Windows Intel driver to one of their November or December 2023 driver releases.  I believe their January 2024 driver is the first to have this issue.

The best method for hosting Jellyfin is the one you are comfortable with managing.  Each platform has its benefits and drawbacks.

(2024-03-20, 12:25 AM)matchaman Wrote:

(2024-03-18, 12:10 AM)TheDreadPirate Wrote: There is a known issue with Intel's Windows 11 OpenCL driver.  That is likely what caused the issue on Windows.

(2024-03-18, 11:30 PM)TheDreadPirate Wrote:

(2024-03-18, 11:22 PM)matchaman Wrote: It seems that the new docker image is working well with minimal skipping/stuttering, but I am now having an issue where whenever I stop the docker container, it resets the jellyfin server to a like-new setup, making me reconfiugre everything. How do I fix that issue?

It sounds like you didn't provide a volume for the container to write /config to.  /config is where Jellyfin writes all of its persistent data to.  You also need to provide a volume for Jellyfin to write /cache to so Jellyfin doesn't have to regenerate that after each restart.  Both should be on a NVMe volume.

(2024-03-18, 11:22 PM)matchaman Wrote: Is there a fix for the Windows 11 OpenCL driver issue, or one being worked on, or would adding a graphics card help fix the issue? Or what is the best method of hosting the jellyfin server? On bare linux? Or does that require a lot of conifig with drivers as well?

You can revert the Windows Intel driver to one of their November or December 2023 driver releases.  I believe their January 2024 driver is the first to have this issue.

The best method for hosting Jellyfin is the one you are comfortable with managing.  Each platform has its benefits and drawbacks.

I will have to try out the windows version as well to see if it is easier for me. What driver am I rolling back? Is it the Intel UHD Graphics driver? I tried rolling that back and it went back to 2006 which doesn't make any sense. After doing the update, it went to a 2022 driver.

After making the mounts for the /config and /cache, it stopped resetting every time I had to reboot or restart the docker container, but now I am having the stuttering issue again. My cache and config are on mirrored HDDs with 7200rpm (Seagate Ironwolf 12TB, 7200rpm, 256mb cache), not an NVMe, so I don't know if that is an issue or not. I don't have an NVMe to attach it to unfortunately, but may be able to look into a solution for that. Attached are new logs as well as a screenshot of my current transcoding settings. I only took a screenshot of settings I have changed, anything outside of the screenshot should be left as the default options. In the TrueNAS docker settings, I did confirm that I gave the container access to 3 gpu/i915s, which seems to be the max I can give before running into other issues.

(2024-03-23, 06:16 PM)TheDreadPirate Wrote: 1)  You are using Jellyfin unstable.  Use at your own risk.  Switch to 10.8.13 if you actually want to use Jellyfin day-to-day.
2)  What does your docker compose look like?

(2024-03-23, 09:07 PM)TheDreadPirate Wrote: "jellyfin/jellyfin:10.8.13-1" is the image name I use in mine.  You can also use the "latest" tag, which will auto update to the latest stable release.  That should use the docker hub website.

(2024-03-23, 09:08 PM)matchaman Wrote:

(2024-03-23, 09:07 PM)TheDreadPirate Wrote: "jellyfin/jellyfin:10.8.13-1" is the image name I use in mine.  You can also use the "latest" tag, which will auto update to the latest stable release.  That should use the docker hub website.

I might not have said it clearly in my last post, but using the tag "latest" results in an error that it can't find anything.

(2024-03-23, 09:07 PM)TheDreadPirate Wrote: "jellyfin/jellyfin:10.8.13-1" is the image name I use in mine.  You can also use the "latest" tag, which will auto update to the latest stable release.  That should use the docker hub website.