I’m running a Plex server off my Nvidia Shield, which is connected to a Samba server 1TB drive that runs off my WRT1900ACS OpenWrt router––def not the most efficient setup but it works. I also run a seedbox on the other drive connected to the router. Please keep this in mind when you look at my config. There’s limited memory and processing power on the router (512MB DDR3) so everything is built with that in mind. Believe it or not, Plex streams well.
I am a little confused between rclone mount and rclone cache. Do all the options listed on the mount manpage apply to the cache manpage, From my understanding it seems the --vfs options are only applicable to non-cache mounts, but it’s not clear.
Anyway, below is my config. Questions right now:
Do all the commands I put in mount apply to cache?
I haven’t tested adding more movies to the drive and triggering a Plex rescan. Not quite sure how I’m going to do that yet, since right now my downloads are on another drive and get uploaded via a separate script, but with this I would have to manually update the library… any suggestions?
Any suggestions to my config below to free up more memory? My memory on the router goes from 30% free to 5% free (linux has great built-in memory management) but I thought with this config no memory would be getting used. Any suggestions? VSZ is 979m 159%
For those who are curious: Plex takes about 5 secs to load the file data then another 10 seconds to buffer the video. Load on the router is around .5 to .9 whenever I stream one movie. Haven’t tried multiple yet. Seeking takes about 10 seconds to jump around.
I’m thinking about transferring this process to the Nvidia SHIELD since it has 3GB memory and I want my router focused on router things. However I can’t find an rclone android app, which means I’m going to have to go hard and root it, etc. Any plans for an android app to avoid this?
You can run Plex on that stuff assuming that you don’t transcode anything. Once you hit a transcode, things are going to crawl as there isn’t much power to support that.
Thanks. While I wait for my other post to get approved (marked by the spam filters) do you know the proper way to kill the rclone daemon? Whenever I kill the process there is always a problem with fuse:
Wed Jan 2 13:06:43 2019 daemon.err uhttpd[3715]: df: /mnt/a/Movies: Socket not connected
But fusermount -uz is only for fuse, the rclone daemon will still run, right? So it sounds like I have to fusermount -uz first and then kill process?
Regarding your comment on transcoding, I see where you’re coming from, but the shield can handle transcodes pretty well. I’m not sure the drive caching will impact it or not, but yes I’ll test it.
felix@gemini:~$ ps -ef | grep rclone | grep Test
felix 14271 14133 0 12:48 pts/1 00:00:00 rclone mount gcrypt: /Test
felix@gemini:~$ fusermount -uz /Test
felix@gemini:~$ ps -ef | grep rclone | grep Test
felix@gemini:~$
as an example.
You’d get plex to scan via a few ways. You can use the integration with Sonarr/Radarr to kick off a scan. You can use something like https://github.com/l3uddz/plex_autoscan to kick off a scan.
You can run the plex CLI and scan after your process moving the files finishes.
You are right, I ran it again and it worked. Weird. Thank you!!!
Any idea on how I get plex to scan from a different system? Ideally, I’d like to run it on my transmission script torrent done, which is a script that runs after the download completes. After the download completes, I upload the file to the cloud, then I want the script to trigger a Plex scan.
It seems like Sonarr/Radarr, Plex Autoscan, and Plex CLI all are required to run off the machine hosting the PMS.
None of them are required to be on the same system. You can just point it to the proper host name/IP and use the right user name/password and it works.