Hello, Silverblue user with stupid question here, about JellyFin server. Is running podman container better, than flatpak, what is the difference?
I tried using Jellyfin server flatpak version, it worked some time, than stopped scanning new files. I liked JellyFin, so checked their website and podman container seemed like a proper way for Silverblue. So I tried it and it works fine.
You picked the right method!
Flatpaks for server software is such an antipattern, but people keep wanting to try it lol.
I only recently started to use flatpaks, because of switching to Silverblue. Some flatpaks apps, like Syncthingy and Dropbox, worked fine, but I missed some automation and scripting, so just layered syncthing and rclone.
But Podman is mostly new thing for me, I tried docker some time before, but did not found it convenient. I now have two images in Podman, first is 1.9Gb for PhotoStructure vs 200mb appimage (it did not worked with SELinux). And second one is JellyFin 1gb vs 400mb Flatpak. And while space is not important, I just could not find some good explanations if podman containers are good for dekstop usage, like performance and memory usage, or maybe Flatpaks are better, or maybe packages are even better (but lots of app are distributed only as docker image).
When adding a volume to a podman container, you have to add a
Z flag at the end of the volume so selinux allows it. For example:
Podman containers are definitely the way to go. If you deploy them correctly you can have them auto update whenever the system boots up or the container restarts. You can also have it auto-generate a systemd file for you so you can manage them like a service.
I mean the app in AppImage did not worked. For Podman I used GUI app to create container, there was a checkbox priveleged security, I guess Z makes same?
I do not get about managing them as a service, I use
podman container start in configs. Isnt it same?
Edit: I see, it needed for auto-updates, I will do it.
No. Privileged security means it needs root level access, like to access a certain file system or to hardware, etc.
Thanks! I learned some things, my containers are now managed with systemd and they are rootless, also made convinent switches with Guillotine extension. But than I noticed my containers are not persistent and recreated each time, which is OK I guess, but why previously they were not removed when not in use?
If you have your containers deployed so they can auto-update that is normal. As long as you have your volumes configured right, each time it creates the container it will use the same volumes so all your data and config is persistent.
The images are not removed unless you do it manually. If you configured the container for auto-update it will pull the latest image before starting the container. If not, then you would need to pull it manually:
# To see all downloaded images:
# To remove an image:
podman rmi <image name or ID>
# To update an image:
podman pull <image name>