dlaffin1012 1 Posted January 13, 2020 Posted January 13, 2020 (edited) Hello, I have recently changed my server/network infrastructure. I went from a single openmediavault server to replicated proxmox servers. Openmediavault now is in a VM. So I can put Emby back into a docker inside openmediavault again and just up the resources of the VM or I can run it on a standalone box such as a fedora server no gui, or fedora workstation, or ubuntu, etc. I do have a quadro card to pass through to emby. Is there a benfit of each. It seems easy enough to do either option but I am curious to see which would either perform better or if its just better to consolidate openmediavault and emby since openmediavault is where it will pull its info from. I will have a main server and backup server so each server will have its own storage array, openmediavault and emby setup. Curious to hear your replies. Thank You Edited January 13, 2020 by dlaffin1012
Luke 42077 Posted January 13, 2020 Posted January 13, 2020 Hi, the benefit of using a native package over docker is that hardware accelerated transcoding will be easier to setup, if you wish to use that. Aside from that it should be pretty much the same.
dlaffin1012 1 Posted January 13, 2020 Author Posted January 13, 2020 Is it significantly harder to use hardware accelerated transcoding when using docker?
Solution Luke 42077 Posted January 13, 2020 Solution Posted January 13, 2020 Well the main difference is that we only consider it experimental. It can work, but will likely require setup on your part, and at this point we don't yet have all of the answers of what you might have to do. If you're using Nvidia though, then this might help: https://emby.media/community/index.php?/topic/76937-docker-hwa-nvidia-instructions/ 1
dlaffin1012 1 Posted January 13, 2020 Author Posted January 13, 2020 So in other words, with some decent knowledge of this, its just better and easier to run it on its own VM if you use hardware accelerated transcoding?
dlaffin1012 1 Posted January 13, 2020 Author Posted January 13, 2020 I may just still run a separate VM. Makes me more comfortable knowing that I don't have docker making a third layer of virtualization. Is adding the media from a separate vm a little more complex? I assume since it would run through SMB rather than directly within the same OS?
Luke 42077 Posted January 13, 2020 Posted January 13, 2020 It could be, yes, but as long as it's shared over the network you should be fine.
dlaffin1012 1 Posted January 13, 2020 Author Posted January 13, 2020 Technically it would have less performance wouldn't it? You add the latency of the network. Dang I make my own decisions harder by asking questions rather than just making a decision.
Volfan6415 19 Posted January 14, 2020 Posted January 14, 2020 I run emby in its own Ubuntu VM. All of the files are stored on a separate VM that runs windows home server 2011. . Both vms are on the same esxi host and the drives are added via SMB. Prior to visualizing my servers i ran emby on a dedicated box running WHS 2011. I have not noticed any performance differences. I am using ESXi for the VM layer. With esxi if both vms are on the same host and in the same lan (or vlan) network communication is done at hardware speed and is not limited by the speed of the network. There may be some loss due to SMB inefficiencies in linux but I have not seen or noticed them. Granted i have not done any specific testing of this. I have a collection of roughly 1000 movies and a similar number of TV shows. The files vary from under 1 gb up to 154 gb for some uncompressed blu-rays. Within my lan of course the various players have direct access to the files via network paths (setup in emby) so locally it makes no difference if EMBY and the files are on the same server. The only time the files have to go "through the emby server" are when they have to be transcoded to stream out over the network. I have not had any performance issues with streaming any of the movies over the internet.
dlaffin1012 1 Posted January 14, 2020 Author Posted January 14, 2020 I really appreciate that input. Makes me more comfortable with splitting it. I currently installed emby on a fedora server install and My media in a separate vm under openmediavault with a netapp card passed through that connects to a separate array via qsfp. I figure making it easier for access to the gpu would be best.
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now