> I don't really understand your question and I suspect it has a lot to do with
> the fact that many people are confused by the LTSP name. Some people are
> aware of Terminal Services in the Windows/Citrix world. For those that
> aren't, Windows Terminal server runs a login session on the server and sends
> the screen's output information back to a remote client machine. This is
> similar to remote control software such as PCAnywhere. However, it differs
> from remote control programs in that the server can host multiple different
> sessions, hundreds if the hardware is up to it, at one time. Whereas remote
> control programs can only support a single session at one time.
>
> Those familiar with X11 in the Unix world know that X11 though technically
> very different, offers very similar functionality in the sense that it can
> "export the display" to a remote client, which X11 confusingly calls the
> server. In any case a program is executed on a particular machine and the
> screen output from that program is displayed on a remote client, as is the
> case with Windows Terminal server.
>
> Linux Terminal Server is completely different. The purpose of this software is
> to remotely boot the operating system itself. That is, a client machine that
> has no operating system, such as a diskless workstation, uses a Boot-ROM on
> the clients network card to connect to the LTSP server and download a copy of
> the operating system to the local machine's memory. The local client machine
> then executes its programs locally on the client machine just like a regular
> PC with a disk drive would. But, because there may be no hard drive in the
> machine, when the client is turned off, it loses everything in the sense of
> an operating system. When it is turned on again it again connects to the LTSP
> server and downloads the OS.
>
> This type of arrangement was once common in the Windows world too. Where,
> diskless workstations connected to a central server and downloaded a copy of
> DOS into memory. It then ran a network copy of Windows 3.1. The practice was
> discontinued when hard drives became cheaper and Windows 95 came out. Things
> like the registry made diskless workstations very difficult to do.
>
> Now, to answer your question. I do not believe that it is possible to download
> the vmlinuz image from the LTSP server and "boot" it from within Windows.
> Once Windows boots PXE support is bypassed, it only runs right after the POST
> when the PC is first booted. I also do not see the point of trying this. If
> you wish to have Linux access from within Windows simple get an X11 client
> and use the exported display, as you said in your post.
>
> It is however possible, to have a Windows workstation with a PXE network card
> in it download and boot Linux from an LTSP server. You simply need to choose
> the PXE network boot option when you first power on the PC. This would give
> the effect of having a dual boot Windows/Linux machine even though only
> Windows would be stored on the local hard drive. Choosing a hard disk boot by
> bypassing the PXE boot process would thus boot Windows.
>
Actually you gave me one of the better explanations as far as the difference
between the MS Terminal Server world and the Linux Terminal Service/XServer
world. This makes it much more clearer to me now and I think I might have found
a possible solution that would work but I would like to know if anyone has tried
this in their environment. I found a free XServer that is java based called
WeirdX (http://www.jcraft.com/wiredx/index.html) and am curious as to anyone's
experiences with this XServer so far? Does it perform well?
Unfortunately I'm still kinda green on all this stuff although you think I would
have caught on by now but for some reason it escapes me. Once I figure it out
for myself I'll be better so for now I'm just reading a lot of stuff but any
offers of advice are always greatly appreciated.
I also know that I can use VNC but from my understanding you must spawn a VNC
process and each VNCSession is given a port number. What we are trying to do in
the lab is to redirect from a website to a full blown session that a user gets a
desktop and everything. With VNC I would have to first spawn the process and
then direct the browser to the particular port which I'm sure we can do with a
script but I was looking for other options.
Again, I thank you for a great response as well as all those that have assisted
(one of the reasons I love LUG's :)
Pace
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 18:46:48 EDT