• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Implementation of Linux terminal network

Sivar

Member
My school's CS lab is about 2/3 thin clients (HP Entria II's) which are used to run Digital Unix, Linux W\ GNOME, and probably a few other Unices.
The setup needs updating, and I was wondering if I could get some recommendations from others with more experience in this than myself. Note that I haven't studied the network much yet, so will not have all that much info about the network, but hopefully enough for now.

First, it would be great to run a window manager or desktop environment that looks decent, such as GNOME or KDE, but these are horribly bloated. Can GNOME and KDE share memory efficiently when multiple users are logged in? For example, if a single instance of KDE requires 64MB of RAM, is it reasonable to expect another identical instance to use, paerhaps, another 8MB or would it use nearly as much as the first instance?
If both of these would be poor choices, what would be a good WM to use? Windowmaker? Somethign similar to Windows would be best for obvious reasons, so more exotic ones like Fluxbox would probably not be a great idea.
Let's assume KDE3 is used and that at any given time, there can be as many as 30 users logged into the server, writing and compiling small C, C++, and Java programs, browsing the web, etc. What would be a reasonable server for this setup? A PC would be preferred, but a more exotic architecture (i.e. Sun) is not out of the question. Would a dual AthlonMP 2000+ with 2GB of memory and two Atlas 10K-3 drives in RAID1 be sufficient? I don't know... I have little experience using Unix GUIs other than on the local machine.
Should this setup use LTSP? (Linux Terminal Server Project) Is LTSP production quality?

Additionally, can anyone recommend a good eight port KVM switch for the servers?

The school will likely consult several people, preferably ones that have more of a clue than myself 🙂, before anything is done, but any suggstions/tips/warnings/etc. would be helpful. This topic seemed slightly more appropriate here than the OS forum. If it is not, well, oops.
Thanks ahead of time.
 
This belongs in the OS forum. However, I will answer this the best I can.

1. Forget KDE. It ain't gonna happen. If you want a windows-like UI, try FVWM95. It's very, very Windows-like and it's slim and efficient. I don't have the URL handy at the moment, so check google.

2. Dual Athlon MP 2000+ with 2 GB will be fine if you don't run anything insane like 30 instances of Maya or VMware. Mozilla should be OK, but make a little swap space available. GCC shouldn't be a problem.

3. Storage space isn't all that important, unless there will be a lot of data stored in the process of the courses. RAID1 is nice but not mandatory. If losing any of the data on the server would be catastrophic, then it would be mandatory. 😉

4. You probably don't need to run anything but SSHD (or telnetd if you insist on your server getting rooted on a daily basis) on the server for remote access purposes. That would be all there is to it if you want text-based terminals, but since you want graphical sessions, I dunno. I know you'll need X servers running on the terminals, and something related to XDM on the server. Since I have never done anything like it, I can offer no advice.

All I can say other than that is, make sure you have a ton of bandwidth coming out of that server. 30 terminals running hundreds of apps via remote X will require immense ammounts of bandwidth. 10 Mbps on the clients and 1 Gbps on the server should provide a smooth experience.
 
Back
Top