Re: [Ltsp-discuss] Understanding LTSP

[Michael P. <ma...@gm...>]

On 30/10/13 10:44, A.J. Maurin wrote:
> After looking through forums, documentation, manuals, wikipedia articles
> and more, my understanding of LTSP is thus:
>
> * thin client boots from wake-on-lan, loads pxe from uefi
> * thin client loads pxelinux, then initializes network from dhcp
> * after downloading initrd via tftp, mounts nbd and nfs drives
> * initializes the x window system via ssh forwarding and ltsp window manager
> * finishes loading and stabilizes
>
> one thing remains unclear, and no amount of googling and manual perusal
> has really enlightened me.
>
> How does GLX forwarding work? i understand that the x server uses glx to
> access opengl-based 3D and 2D acceleration, and since the x server is
> running on the thin client, the rendering relies on the thin client's
> graphics card.
>
> I don't mind this, since it's a lot easier to build a bunch of machines
> with beefy graphics cards and high-quality NICs than it is to load up on
> RAM, CPU cores, and very high quality motherboards, etc. etc.
>
> If anyone knows exactly how GLX works with the X forwarding in LTSP, I'd
> appreciate it.
>
> If anyone has tried using virtualbox or vmware over LTSP, please let me
> know how it went! Thank you in advance!
>

I don't know about the GLX question, but I have used virtualbox on LTSP 
machines. Virtualbox worked pretty good over the network, we mainly used 
it for one application for accountants. I have since moved over to using 
KVM instead of Virtualbox for all these images as it has a lot of 
advantages such as easier maintenance, all images accessible from any 
thin client, can move sessions around, faster, open source, more control 
over hardware, leave sessions running when thin client disconnects, etc.

from
Mick