2025-01-07, 09:12 PM
(This post was last modified: 2025-01-07, 09:19 PM by alexauran. Edited 2 times in total.)
I've tried running as micro (my user) with the following results:
https://pastebin.com/EF2LW3qk
I'm running proxmox on a poweredge r740xd with jellyfin in a debian VM, with a truenas VM acting as a NAS holding my media on 10 HDDs via proxmox passthrough to the truenas VM. This works great; my LAN pcs think the truenas VM is an appliance, and my VMs on the same machine connect to it via cifs, but with near local disk access speeds. I get my storage with parity on Raidz2 with all of my VMs and LAN sharing it nicely. All of the VMs run from an 800Gb enterprise SSD which I don't want to clog up with low priority metadata. I originally tried to put my data dir on the truenas, since it's local physically, but I assume the small delay of the truenas disk controller/manager made plex and jellyfin upset, because they were crawling. My next thought was to consolidate all of my large metadata folders from programs like plex and jellyfin to one location, mount that folder via fstab to a large HDD (visible in the VM via a passthrough in proxmox) that is separate from the truenas. The HDD is empty so there are no capacity issues, and if I need to rebuild the jellyfin/media VM then the metadata is on a separate drive which I can backup, I spin up the new VM, pass the drive through to the new VM, and then point JELLYFIN_DATA_DIR at it. That way I don't have to do all the setup again and generate all the thumbnails, etc. I'm new to linux so this may be dumb or inefficient. This setup seemingly works with plex (/home/micro/.databases/plexmediaserver) in the same debian vm (haven't rebuilt the vm to fully test it yet.) Now, it's just become my white whale for learning linux better. If this is a bad idea, I'm all ears.
https://pastebin.com/EF2LW3qk
I'm running proxmox on a poweredge r740xd with jellyfin in a debian VM, with a truenas VM acting as a NAS holding my media on 10 HDDs via proxmox passthrough to the truenas VM. This works great; my LAN pcs think the truenas VM is an appliance, and my VMs on the same machine connect to it via cifs, but with near local disk access speeds. I get my storage with parity on Raidz2 with all of my VMs and LAN sharing it nicely. All of the VMs run from an 800Gb enterprise SSD which I don't want to clog up with low priority metadata. I originally tried to put my data dir on the truenas, since it's local physically, but I assume the small delay of the truenas disk controller/manager made plex and jellyfin upset, because they were crawling. My next thought was to consolidate all of my large metadata folders from programs like plex and jellyfin to one location, mount that folder via fstab to a large HDD (visible in the VM via a passthrough in proxmox) that is separate from the truenas. The HDD is empty so there are no capacity issues, and if I need to rebuild the jellyfin/media VM then the metadata is on a separate drive which I can backup, I spin up the new VM, pass the drive through to the new VM, and then point JELLYFIN_DATA_DIR at it. That way I don't have to do all the setup again and generate all the thumbnails, etc. I'm new to linux so this may be dumb or inefficient. This setup seemingly works with plex (/home/micro/.databases/plexmediaserver) in the same debian vm (haven't rebuilt the vm to fully test it yet.) Now, it's just become my white whale for learning linux better. If this is a bad idea, I'm all ears.