I noticed my home servers SSD running out of space and it ended up being my Jellyfin Docker container which wasn’t clearing the directory for transcodes in /var/lib/jellyfin/transcodes correctly.
I simply created a new directory on my media hard drive and bind mounted the above mentioned directory to it. Now Jellyfin got over 1 TB of free space to theoretically clutter. To prevent this I simply created a cronjob to delete old files in case Jellyfin isn’t.
/usr/bin/find /path/to/transcodes -mtime +1 -delete
Easy!
I transcode to ramdisk.
How is this done? If you don’t mind sharing 🤗
tmpfs is the filesystem you are looking for. You can mount it like any other filesystem in /etc/fstab.
tmpfs /path/to/transcode/dir tmpfs defaults 0 0
Can just point it to /dev/shm as a transcoding folder, for a quick and dirty way.
Otherwise you’d mount a tmpfs disk.
I have like a dozen people using my Jellyfin and sometimes 3-4 people watch something at the same time which results in a lot of transcoding data. At the moment my transcoding directory (which is cleaned every 24 hours) is almost 8 GB big. I don’t have the RAM to do this.
Starting with 10.9 you can enable segment deletion so files are cleaned up while still transcoding.
Im so looking forward to this. When i tried to use tmpfs / ramdisk, the transcoding would simply stop because there was no space left.
Version 10.9 is not even released, right?
Nope, release target is mid-April currently.
I will have to try that once my ram upgrade gets here.
You can restrict the size of the ramdisk so you do not end up killing processes. A large amount of ram is not mandatory.
Good to know. Well I have 16G now that should give me plenty to spare.