March 9, 20251 yr For the second time in a few days, my Unraid server's root filesystem filled to 100%, causing SMB shares and WebUI to stop working. Each time, a reboot fixes the issue completely, leaving the root filesystem nearly empty (only 6% used). Version: Unraid 6.12.14 USB drive: 57.3GB with 7.8GB allocated for root Found Jellyfin process holding deleted memfd file showing as 2TB in lsof | grep deleted After reboot: Root filesystem drops from 100% to 6% usage (457MB/7.8GB) All services resume normal operation Questions: What might be causing this rapid filling of root partition that disappears after reboot? Can that deleted file held open by Jellyfin (memfd:doublemapper) consume all available space on the root filesystem through inodes or other filesystem resources? What's the best way to monitor/prevent this?
March 9, 20251 yr Community Expert Most likely you have a mapping from a container, for example, going to an invalid mount point, and that then goes to RAM, check all mappings, they must be all mapped to a valid /mnt/user/<share> or /mnt/<disk path>/<share> path, also note that Linux is case-sensitive.
March 9, 20251 yr Author 3 hours ago, JorgeB said: Most likely you have a mapping from a container, for example, going to an invalid mount point, and that then goes to RAM, check all mappings, they must be all mapped to a valid /mnt/user/<share> or /mnt/<disk path>/<share> path, also note that Linux is case-sensitive. I think is that cache directory that is mounted in the root fs the problem: (root) ➜ ~ docker inspect -f '{{ .Name }}{{ printf "\n" }}{{ range .Mounts }}{{ printf " Source: %s\n Destination: %s\n Type: %s\n\n" .Source .Destination .Type }}{{ end }}' Jellyfin /Jellyfin Source: /var/lib/docker/volumes/0007636a411b5ffb97a471ea286855683ad3e014065ce8f57a8db5d4af52da9b/_data Destination: /cache Type: volume Source: /mnt/user/appdata/Jellyfin Destination: /config Type: bind Source: /mnt/user/UNNAS/OLD HDD 6TB/Film Destination: /data/movies Type: bind Source: /mnt/user/UNNAS/OLD HDD 6TB/musica/Plex Music/music Destination: /data/music Type: bind Source: /mnt/user/UNNAS/OLD HDD 6TB/musica/Plex Music/soundtrack Destination: /data/soundtracks Type: bind Source: /mnt/user/UNNAS/OLD HDD 6TB/Serie TV Destination: /data/tvshows Type: bind I added a cache path in the jellyfin container config, now is like this: (root) ➜ ~ docker inspect -f '{{ .Name }}{{ printf "\n" }}{{ range .Mounts }}{{ printf " Source: %s\n Destination: %s\n Type: %s\n\n" .Source .Destination .Type }}{{ end }}' Jellyfin /Jellyfin Source: /mnt/user/UNNAS/OLD HDD 6TB/Serie TV Destination: /data/tvshows Type: bind Source: /mnt/user/UNNAS/OLD HDD 6TB/Film Destination: /data/movies Type: bind Source: /mnt/user/UNNAS/OLD HDD 6TB/musica/Plex Music/soundtrack Destination: /data/soundtracks Type: bind Source: /mnt/user/UNNAS/OLD HDD 6TB/musica/Plex Music/music Destination: /data/music Type: bind Source: /mnt/user/appdata/Jellyfin/cache Destination: /cache Type: bind Source: /mnt/user/appdata/Jellyfin Destination: /config Type: bind This is ok? We should tell the mantainer of jellyfin/jellyfin:latest that it should require the mapping of the cache dir?
March 9, 20251 yr Community Expert 2 minutes ago, UnraidUser71374 said: We should tell the mantainer of jellyfin/jellyfin:latest that it should require the mapping of the cache dir? No Attach Diagnostics to your NEXT post in this thread.
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.