Trello Board -> Emby

[jaredallard 0]

I've recently started running a media server for a bunch of friends of mine, and well, it's pretty hard to keep track of different media and requests that they have. So, I started a pretty detailed Trello board to track all of the requests, and keep the sources / MAL (MyAnimeList, yes it's all Anime). That way I was able to track media and organize what's processing (converted to device friendly formats), what's deployed and have a log for that.

 

 

Trello Board

 

 

 

 

 

Anyways, enough back story, the past two days I've been working on a quick software stack that utilizes Docker and DigitalOcean. It's not done, but it can:

 

• Track Trello card movements, i.e Requests -> Ready
• Spin up Digital Ocean servers for processing and tear them down after usage ($$$ saving)(this is almost done)
• Deploy Media to a Emby server via sftp
• Scaleable converting w/ redis-backed job queue and docker containers

 

Github: https://github.com/jaredallard/media-stack

 

Reason I'm posting this here is in case anyone else is looking for software to make re-encoding media using external services a lot cheaper (I can take this from $120/month to $60-70 by not having a server always on). Disk space is still an issue, but oh well, that's what you get for hosting U/HD media in the cloud. Also, if anyone has anything they'd like to see from a platform like this, let me know! Ideally I'll have Emby running in docker connected to the stack (but I have a production environment I have to reduce downtime on, so that'll be slower)

Edited January 28, 2018 by jaredallard

[Luke 42077]

Very cool, thanks !

[jaredallard 0]

In case anyone was looking into this, it no longer spins up Digital Ocean droplets as we got some on-prem hardware that's good enough to handle converting jobs -- might finish that feature in the future as it still routes through the deploy service.

 

I expanded this to support Kubernetes and to have a helm chart, so after setting up a cluster with different roles like:

 

  triton.state=stable (public facing machines, get's Emby & events microservice)

  triton.role=ops (kubernetes machine, other ops stuff, doesn't effect the stack I think)

  triton.role=bigdata (schedule converter instances, for high CPU/etc machines)

  triton.role=storage (originally for whatever machine had emby on it, but that's now NFS, but schedules emby/twilight pod onto it)

 

You can just helm install --name media media-stack and it'll work! Hopefully someone finds this useful for their Emby needs, but the docs are a bit lacking so feel free to PM me if you need help deploying it and I'll help in any way I can (even if that's just installing Kubernetes on some nodes)

Edited February 25, 2018 by jaredallard