[Inexistence]

Multiple XBMCs using XEN server OR independant

Advice needed please.

Current situation: I’m currently running XBMC on one PC that is shared to 4 locations in my home using a HDMI 4 X 4 matrix.

Goal: Have each location use its own dedicated XBMC setup.

Other info: I have an equipment rack in an isolated "server room" in the basement. It has the XBMC machine, matrix switch, NAS, etc. Video is sent to each location using cat6 HDMI extenders. I want to keep all equipment in this rack. Each system will share a database and have access to the same data.

XBMC community, please tell me your opinion.

Option 1: Build a Xen server and have 4 virtual XBMC machines.

I do not know if anyone has had any success with this. This could also be used to run other services such as asterisks and such. Xen has the ability to directly access hardware. The Xen server would have 4 video cards with HDMI output.

Option 2: Build 4 1U or 2U rack mount Mini –ITX systems for each location to be housed in the equipment rack. Each system can then boot using PXE so no hard drives are required.

Question: What way would you go and why? Thank you.

[GortWillSaveUs]

I wish I could come close to going your way.
Sounds like a pretty awesome setup. An isolated "server room". Sweet!

While I haven't worked directly with XEN, I do work with VMWare.
Is there any way for you to test out Option 1 to see if it works? That option seems to be the way top go,..if it works.

[kuchar]

You won't be able to use hardware accelerated video decoding inside multiple Xen domains. You can dedicate graphic card to one VM, but then it cannot be used by any other.

[Inexistence]

From my understanding Xen can do hardware acceleration. Plus each system will have its own dedicated video card. Here is some information I have found in another forum. ----> http://communities.vmware.com/message/2030744

------------- From the Xen FAQ:

"

Xen 4.0.0 is the first version to support VGA graphics adapter passthrough to Xen HVM (fully virtualized) guests. This means you can give HVM guest full and direct control of the graphics adapter, making it possible to have high performance full 3D and video acceleration in a virtual machine.

Xen VGA passthrough requires IOMMU (Intel VT-d) support from the motherboard chipset, from the motherboard BIOS and from Xen. Please see the VTdHowTo wiki page for more information about Xen VT-d support and supported chipsets with IOMMU.

Xen 4.0.0 supports VGA passthrough of the primary graphics adapter (the one that gets initialized and is used when you power-on the computer) only. VGA passthrough of secondary graphics adapters currently requires additional extra patches to Xen.

Initial Xen VGA graphics passthru support was committed to xen-unstable development repository in September 2009, and more support for Intel IGD passthru was added in February 2010." --------------------

I will need to give this a try and see how far I can get.

Does anyone have XBMC running in such an enviroment?

[Bstrdsmkr]

Haven't tried it, but theoretically at least if you have one video card per monitor, it should work. A good quad core CPU will be needed for running addons, but If you've got the hardware, I don't see a technical reason it shouldn't work.

Sounds pretty sweet!

[AHiredGoon]

Appologies if this is a bit late.

I reccommend (and use myself) a multiseat setup for XBMC.

This is extremely similar to your concept of xen, except multiseat is only one system. Each seat has its own user for the xbmc files (autologin TTY's, XDM etc). I use MySQL to manage my library.

pros:
if using VDPAU, your horsepower reqs are quite low (depending on addons as Bstrdsmkr mentioned)
extra hardware savings can buy a better system.
Adding a new client can be just buying a 20 dollar card and remote

Potential Issues:
Windows: no clue if windows can do this, sorry
OpenGL: Linux currently can use ONE amd, nvidia or mesa opengl library (required to run XBMC), neither at the same time. So your cards must all support the same.
Remotes: different remotes or addressing is required to keep all buttons distinct.

My Setup: I actually run a multiseat with 2 nvidia cards on my Dom0 (Nvidia vgapassthrough is preliminary right now so having multiple domu would require identical cards due to memory bar passing). I use the binary driver and get VDPAU on both seats. Its a bit tricky on the Dom0 side of things but for a regular system its not (ubuntu site has some good guides).


I will note that if you used AMD for cards you could likely do the multiple domu thing pretty easy (plus mix the clients up with windows and linux). AMD passthrough is simple and works great (7950 in my setup).

[tedstevens]

Appologies if this is a bit late.

I reccommend (and use myself) a multiseat setup for XBMC.

This is extremely similar to your concept of xen, except multiseat is only one system. Each seat has its own user for the xbmc files (autologin TTY's, XDM etc). I use MySQL to manage my library.

pros:
if using VDPAU, your horsepower reqs are quite low (depending on addons as Bstrdsmkr mentioned)
extra hardware savings can buy a better system.
Adding a new client can be just buying a 20 dollar card and remote

Potential Issues:
Windows: no clue if windows can do this, sorry
OpenGL: Linux currently can use ONE amd, nvidia or mesa opengl library (required to run XBMC), neither at the same time. So your cards must all support the same.
Remotes: different remotes or addressing is required to keep all buttons distinct.

My Setup: I actually run a multiseat with 2 nvidia cards on my Dom0 (Nvidia vgapassthrough is preliminary right now so having multiple domu would require identical cards due to memory bar passing). I use the binary driver and get VDPAU on both seats. Its a bit tricky on the Dom0 side of things but for a regular system its not (ubuntu site has some good guides).


I will note that if you used AMD for cards you could likely do the multiple domu thing pretty easy (plus mix the clients up with windows and linux). AMD passthrough is simple and works great (7950 in my setup).

What sort of resource savings are you getting with this setup vs just running two VMs, each with their own card passed through?