As the title says, I just upgraded to 10.9.9 and now I can’t connect to the server. I have it installed on a docker container on a headless ubuntu machine. This is the first time I’ve updated after installing and successfully using the server for several months. I followed the instructions in the first response here: link
Originally, the update/install seemed to work just fine. I opened up a client on a browser and it successfully connected and prompted the install wizard. So I went through it and re-pointed to all of my media. However, at the very last step when clicking submit, it got stuck. I saw online that this happened to others and what worked for them was to just try again.
So, after some time, I restarted the server and restarted the above process, except now I’m unable to connect at all.
When trying to connect from my Nvidia Shield, it tells me to complete the setup on a browser. When connecting on a browser, it asks me to select a server, lists the server, and then fails to connect to it with the error “We’re unable to connect to the selected server right now. Please ensure it is running and try again”. Connecting on my android phone and connecting in incognito prompts the same error.
Also, I’m unable to see any new logs now. ~/jellyfin/config/log has 3 log files from the last 3 days. But the latest one seems to have stopped around the time of the upgrade and shows no errors. I’ve updated the logging.json file as per the instructions here and this does not help: debug-logging
Does anyone have any advice here?
UPDATE: Issue seems to be solved. Tailscale seemed to be causing some problem on the client-side. Disabling and reattempting seemed to work. The library is still being scanned, so I haven’t validated playback. But I can at least navigate my content now.
UPDATE 2: Playback is validated
What’s the output of docker log?
And what image are you using for your container?
jellyfin/jellyfin:latest
Do yourself a favor and do not use the latest tag. Always use the tag for the explicit version you want. Makes things a lot more stable.
You also mentioned that jellyfin prompted you with the installation wizard after the update. It seems like you fixed your problem but for the next time: this is not expected behaviour. A short downtime right after the new container started is normal but afterwards the server should run normally with the same configuration as before.
These are the last logs before they stopped showing up:
spoiler
To clarify I mean specifically the output of the “docker logs jellyfin” command, not the log file you were looking at.
If jellyfin isn’t starting properly it may not make it to its own logs, but the docker output will likely catch more.
Oh! I wasn’t aware of these logs. Yep, I’m seeing an error repeating many times here: “Unhandled exception. System.UnauthorizedAccessException: Access to the path ‘/config/config/migrations.xml’ is denied.”
Yeah there’s your answer. Check what user jellyfin is running as and fix the permissions on your files.
So, I’m connected through SSH as the admin user. Lets call it salsa. When calling docker run, I passed the --user 1000:1000, which matches salsa. This actually results in the browser not even finding jellyfin. Returning some 4xx (probably 401).
However, originally I was running docker run without the --user field, which yields the result I mentioned before. Where it shows Jellyfin, lists the server, but then fails to connect.
From this, docker logs show no errors:
spoiler
EDIT: Also, with this, ~/jellyfin/config/logs is still showing no new logs.
I use docker compose but I also set the uid/gid to 1000.
Do you have a local path mounted into /config of the container? You might just need to chown -r 1000:1000 your config dir
Actually, I managed to fix the issue. Tbh, idk why/how this fixed it and perhaps I should’ve mentioned this as well, but I have tailscale set up on my server to allow remote connections. I was connecting to the server through tailscale even though I’m in the local network. This should work just fine and has in the past, but perhaps it doesn’t when first signing into a fresh install. I disabled tailscale on my client and retried and that worked. I was able to reattempt the install wizard, completed it, and now have a working server once again.
I haven’t yet validated playback as my library is being rescanned now, but I’m at least able to navigate. I’m also unsure why I was able to attempt the wizard before just for it to get stuck…
Thanks for the help! Your swift responses have been appreciated and I learned a good bit :)
Just in case, I tried fixing the permissions anyway with
sudo chown user:salsa /config/config/migration.xml
. However, this gives another error:chown: invalid user: 'user:salsa'
. Even though it is a valid user and shows up when checkinggetent passwd | grep 'salsa'
Your syntax is wrong. It’s user:group, as in salsa:salsa. Not literally “user” but the user/group ownership you want to set.