New server questions

[rayfield]

Longtime Linux user (mostly Ubuntu) but never set up a headless server - while I'm comfortable with managing my laptops via SSH I kinda feel like I'd just as soon plug in a monitor and keyboard to set it up and then manage it remotely using a dummy HDMI plug and Remmina - is that possible/a good idea? Any caveats or suggestions?

On the other hand, if I need to be concerned about system overhead with the GUI (which would probably be straight Gnome as I've never been terribly fond of xfce) and/or it's super simple to manage Jellyfin only via CLl, I don't mind doing it that way. Guess I could also use Portainer via CLI tho for now I'm comfortable with the browser interface.

Also, since ISTR from the docs that installation via apt seems to be preferred for Debian, is that the way to go?(Can't really remember but I think I opted for Docker after native installation didn't go smoothly the first time so I opted for Docker but it's been a couple years.)

[tmsrxzar]

there is no gui component to the jellyfin server so no benefit in a desktop therefor no need for a remote desktop
(outside of the web gui which can be managed from Any browser, not desktop specific)

personally, i use portainer+docker+jellyfin server because i already use docker for other containers for various things and it works perfect
i would recommend this for portability, any time you want to rebuild or relocate the server install it is exceptionally simple

if this is a dedicated machine for jellyfin server; it could be just as easy to go bare metal and then no need for the extras of docker or portainer, pretty sure this is still fairly portable as long as you setup your mount paths correctly

[TheDreadPirate]

If you intend to have all your libraries and users migrate to your new server, you would need to continue to use docker.

If you don't care about that, we have auto-install scripts for debian and ubuntu. Unless memory usage is a concern a GUI is fine, but all the important admin tasks are done on the command line.

[Venson]

Longtime Linux user (mostly Ubuntu) but never set up a headless server - while I'm comfortable with managing my laptops via SSH I kinda feel like I'd just as soon plug in a monitor and keyboard to set it up and then manage it remotely using a dummy HDMI plug and Remmina - is that possible/a good idea? Any caveats or suggestions?

On the other hand, if I need to be concerned about system overhead with the GUI (which would probably be straight Gnome as I've never been terribly fond of xfce) and/or it's super simple to manage Jellyfin only via CLl, I don't mind doing it that way. Guess I could also use Portainer via CLI tho for now I'm comfortable with the browser interface.

Also, since ISTR from the docs that installation via apt seems to be preferred for Debian, is that the way to go?(Can't really remember but I think I opted for Docker after native installation didn't go smoothly the first time so I opted for Docker but it's been a couple years.)

I agree. Managing just JF is an file operation task only, give or take a few systemd commands.

I would see it a bit as a learning curve, if you are already familiar with linux (ubuntu) you should have your fair share of terminal work already done and doing that over SSH is really no big difference. There is a bit of overhead running a GUI version of linux, however its negligible in the grand theme of things, a bit more ram some CPU cycles and so on. 

Really there is no _prefered_ version of linux for jellyfin. If we support it in our doku it should work the same. Keep ubuntu if you are familiar with it, it will work just fine. 

Docker is a bit of a different beast. If you are new to it, i would not recommend using JF as a start, as it has its own set of "challenges" especially when it comes to file permissions and HWA transcoding. But if you try to get it running, and succeed you pretty much will know 60% of what docker will do for you. Add an reverse proxy and its around 80%. But prepare for some documentation to read. If you want to keep everything as is with docker, the migration should be more or less easy esp. when you have a compose file. Just ensure all paths are updated to be the same in the container as before and run "docker compose up -d". Done. Thats the big benefit of docker.

[rayfield]

Thanks for the helpful responses - just looked up how to move Portainer containers and it's indeed awfully simple, so should be set. I maybe used to understand that the web interface was just that, but after a certain age, the brain decides some facts are disposable.