Slashdot Mirror
Ask Slashdot: Linux Diskless Clients?
Graymalkin asks:
"I was wondering if there is a Linux program (or many
programs) that would let me have a single server on a LAN
and then a bunch of diskless clients. I'd like to make the
clients have a motherboard, memory, a processor, and a NIC
and get all their info from the server. If possible I'd
also like to have all the networked stuff loaded into the
server's memory so there's as few disk read/writes as
possible. Can I do this?"
Ideally:
.....etc)
- Dataless clients. Indicating they have a small
hard drive for swap space.DO NOT SWAP OVER
NETWORK. Trust me, it is STUPID. NFS mount everything else from your server(s).
- Booting using kernels with nfs-root capability,
this implies either arp/bootp request... OR..
- NICs with bootp roms
- RAM filesystem?? ramdisk?? Unsure how big these
can be... and if they be worth the hassle.
NFS can be tunable....
- NFS performance... how many clients per nfs server, and what type of connection (10baseX or
100BaseX or
.
.
.
List goes on...
enjoy
Read the file /usr/src/linux/Documentation/nfsroot.txt
It is a part of the linux kernel sources.
The simplest way to boot linux diskless is to use
a local hard disk or floppy.
I did this once on a 10BaseT network. It works fine but if you don't want to wait a second for every 'ls' I suggest you use 100 Mbit NICs. If your network holds 10+ machines a 10BaseT infrastructure is not recommended. Consult the mini HOWTOs at LDP for more info.
A friend of mine cranked out a bunch of thin clients with the root filesystem on CD-ROM and a floppy which holds client access and config data (useful since you'd hafta re-burn the cd to change even a simple ip-address). You just burn a root filesystem onto the CD. Way cheeper that a hard disk (~$30 US for 8x reader) and it worked great for us. Not sure if this method is referenced by any howtos.
> The simplest way to boot linux diskless is to use :-)
> a local hard disk or floppy.
That's not what I call "diskless"
There are any number of reasons to go with diskless workstations (assuming the network infrastructure can support them). The big ones for me are security and centralized management. Obviously, software upgrades become vastly easier. With no local files to modify, an admins can concentrate their time securing a single (or small number of) resource. If an individual workstation is comprimised, it can simply be power cycled and the server thouroughly investigated. I've worked with a few web server farms that are built this way, and it's reasonably effective.
P.S. - the "negligable" cost (say $100 for a 4GB IDE) adds up pretty quick when there are 100 or a 1000 machines.
If you don't need a lot of power, those cute little DEC Multias can do this pretty easily. I've never actually run one like this, but from my experience with them (I installed RedHat5.2 on it without a floppy drive, doing everything over the network), it shouldn't be too bad...
Dan Moore
pergamon@68k.org
This is what I used to make my two diskless
/
clients which I use at home to reduce noise.
I do not boast thousand installations but
this at least works.
Claus-Justus Heine had an NFS swap:
http://www.math1.rwth-aachen.de/~heine/nfs-swap
Gero Kuhlmann made a standalone bootstrapper:
http://www.han.de/~gero/netboot/index.html
--P
We are doing that exact thing at work, and let me tell you, it is BEAUTIFUL!! :)
:)
:)
/. pass)
Specs:
AMD K6/2 300 up to the 450
128 megs of RAM
ATI Rage video card
3Com 509 ethernet cards with an EEPROM we burn ourselves that handles the boot described below
I think we are building them for like $400 or so (minus the monitor).
We use BOOTP to get an IP address for each system, then we can TFTP the kernel from our server. From there on out it is really straight forward.
These systems are really quick (even over a 10BaseT network - switched network though). We are using the kdm (kde version of xdm) for users to log on, and it takes ~75 seconds from powerup until the kdm is up. Of course, these systems don't need to go down at all. But if the user decides to switch off the power, since there is no hard drive, we don't have to worry about file corruption.
The server is backed up nightly, therefore everyone in the office is backed up nightly as well! VERY nice. VERY fast. VERY fun!!
Currenly we have 45 of these scattered throughout the office. Of course we had to get a nice size server to handle all these clients, but it has saved us so much time in maintainance and these things are sooooooo reliable!!
I love it (if you couldn't tell)!
Nathan Ehresman
nathan at usa.om.org
(forgot my
To get my personal opinion about diskless workstations, replace the first "S" in "diSkless" and replace it with a "C".
I strongly recommend you use surplus hard drives in your workstations as boot devices; if nothing else, it will reduce the load on your network, since your servers won't need to be throwing megabyte kernels out over and over. It's also handy to have a few megs of local swap space for those big tasks.
>Speaking of booting off of EPROM's, does anyone
>know if a flourescent style black-light puts out
>enough UV to erase an EPROM?
Try it and see but I think not.
IIRC ultra violet tubes for PCB making produce long wavelength UV. The tubes for erasing EEPROM
devices produce short wavelength UV. Short wavelength Uv is also used for germicidal purposes and apparantly causes skin cancer.
The tubes in blacklights and counterfeit banknote detectors produce purple light and a little bit of weak UV.
(any optical/physics people want to comment?)
About 6 weeks on a window ledge will do, depending on season and local climate.
Take a look at the Linux Router Project. It runs off of a floppy generally, but there are a bunch of people using disk on a chip flash systems for booting.
http://www.linuxrouter.org
And for some neat boxen...
http://www.calibri.net
These are generally really small boxes with disk on a chip for the main storage. Add some memory, a monitor and a 10bt wire and you're good to go.
We use a etherminal from Igel, which is a diskless linux based pc using a 4 or 8 mb flash diskonchip with linux on it. it has 10/100 mb ethernet and you can use the 2 serial and parallel port for connecting terminal or printer and it will act as a terminalserver.
You can find more information on http://www.igel.de/indexe.htm
We are planning to replace all our terminals with this little linux device.
O yeah it also has citrix for who really want to connect to NT servers.
It also has several terminal emulations including full screen 80/132 column.
Hans Harder, Brinkman BV, The Netherland
Email harder@brinkman.nl
Why? Because AFS is better than NFS. More secure, more flexible and faster.
Then there's Coda aswell.
Noise:
Strange nobody mentioned this, but with a decent
CPU (that doesn't need any fan) you can build up
a system that makes absolutly no noise.
Size:
Diskless stations can be made significantly smaller(in size similar to some X-terminals) since they don't need any room for harddiscs or CD-ROM drives, big power supplies etc.
> I've heard that having a swap space makes linux run 10-20% faster, even if that space is never used.
I'm going to try this with NetBSD...
Other (cheaper) eraser ideas:
- Leave the EPROM exposed in the sun for a couple of weeks...
- Expose the EPROM to a rapidly flashing strobe light for a while (less than a day more than a minute... I haven't tried this one myself).
Yes, there is a little known package by
Jacques Gelinas called netaudio. It
provides a wrapper shared library that
catches accesses to sound devices and diverts
them to the diskless PC with the sound card.
You can find the link at www.slug.org.au/etherboot
>* NFS Network file system /home, /usr, /opt between server and clients and should have / and /var separate for each.
/etc/fstab.linux4 is /etc/fstab on computer linux4). ok, it is a bit messy, and you have to prevent writes on shared files, but it would be easier to set up than 20 separate / directories. (and save a lot of space).
> This is the base system for sharing files. You can share
while this is certainly true, i'm looking for a more elegant way of doing the same thing. basically i wont to moun root and that's it. a few years ago i heard about an nfs(server) extension, that allows to use different files for different clients (eg
anyone who knows about this extension. is it still possible, are there other ways? i would like to stay close to a standard distribution (eg redhat, debian or suse).
check http://www.solucorp.qc.ca/xterminals/
Hi,
:) /dev and /var. We symbolic linked many config files from /etc into /var so that / could be mounted read-only. We symbolic linked /tmp to /var/tmp so that it too is unique. /var/spool/mail read-write to all of the clients. /tmp -- but watch out for security concerns!) /etc directory for the clients. That's a real kludge but it works. :) It takes about 75 seconds for a network computer to boot completely (KDE login screen with X running)
/dev and /var directories for a new network computer.
I'm helping set up a network which currently has about 40 Linux-based network computers which boot from Flash ROM on the network card. They have no floppy nor hard disk. Here are a few suggestions:
1. We use Etherboot (http://www.slug.org.au/etherboot) for our network boot ROM, using a 3c905b-txnm network card.
2. Use a healthy amount of RAM to prevent swapping and encourage the client machine to cache as much of its filesystem as possible. We used 128MB. Be aware that even without a swap partition the machine still _can_ swap -- it will just drop pages in demand-paged executables and shared objects, and read them back from the executable when needed. You don't want this to happen on a 10mbps network
3. The main filesystems that have to be unique for each machine are
4. Depending on your email setup, you may want to share the server's
5. We use KDE as the GUI. There are some quirks in KDE that we had to work around, mostly dealing with a user being logged in on more than one machine (some pid files had to be moved to directories in
6. Locking over NFS appears to be broken in Linux 2.2. Thus we couldn't use YP. Instead, we force password changes on the server and regularly cp the password file and shadow file to the exported
7. We used RedHat and had to modify rc.sysinit extensively and re-order much of the boot-time scripting. I've heard that other distributions like Slackware are much easier to set up for network computing than RedHat. It was quite a chore....
8. In our setup, our mostly-10mbps switched network works just fine. We're using 100mbps capable cards in the event we rewire our building
9. Be sure to disable unneeded processes for the client systems! You probably won't need apache, samba, etc. on the clients....
But -- once it is set up, network computing is very low-maintenance. And, adding new machines is just a matter of configuring the NIC, installing the boot ROM, and adding an entry for bootpd. We have a script that auto-creates the
They have: http://www.siemens.de/pc/business/scovery/index.ht m
Does anyone know if it's possible for the diskless clients to have sound through X applications run at the server?
The workstations will need a sound card, of course, but is X able to distribute the sound?
--
--
The differences between the two projects include 1) size of the images, 2) number of supported NICs, and 3) available features. Nonetheless, they aim for compatibility between each other and it is often worthwhile trying both and finding out which one works better for your particular configuration. (As a co-author of Etherboot, I am somewhat biased).
Rather than asking your questions on Slashdot you probably stand a better changes getting answers to your technical questions, if you subscribed to the combined mailing list for the two projects. There also is an archive of all the messages ever posted on this list.
Both Etherboot and Netboot allow to load the Boot Prom image from floppy disk while you are still testing your environment. Once everything works, you can decide to burn an EPROM and eliminate the need for the floppy disk. Of course, with modern NIC cards things are easier, because some of them already include a FlashPROM.
Another very promising project is NILO. It has originally been started by one of the Etherboot authors and it is currently being written from scratch under commercial sponsorship (the entire project will is GPL'd). See the home page for all the details.
Eventually NILO will probably obsolete Etherboot and Netboot for everything but very special purposes. It currently is under active development and while we would appreciate more alpha testers it isn't really very useable for real-world applications. There will be announcements on Freshmeat about the progress of this project.
check out the network block device included with the kernel. You can use it as swap for diskless machines.
-paul.
I use Friend/Foe + mod-point modifiers as a karma/reputation system.
There are BIOS'es that you can tell to boot off the net. If you get you're BIOS so far to boot off the net you'd simply use tftp protocol or bootp protocol to have a kernel loaded (typically residing in /tftpboot..)
Indeed BTW diskless clients have often disks, for swap
I changed it to int rd_size = 131072; so I could have a 128M ramdisk.
Then I added an /etc/rc.d/rc.ramdisk:
# Install the ramdisk module. /dev/ram0 131072 /dev/ram0 /data/mysql/var/RD /data/mysql/var/Archive/Archive.* /data/mysql/var/RD /data/mysql/var/Econ/Econ.* /data/mysql/var/RD /data/mysql/var/RD
insmod rd
# format the ramdisk
dd if=/dev/zero of=/dev/ram0 bs=1k count=128k
# Make an e2fs filesystem
mke2fs -c -v -m0
# mount the filesystem
mount
# Copy the data into the ramdisk
cp -a
cp -a
chown -R mysql:db
It'll work if you're more patient than Job. An eraser don't set you back too much. With a timer, it'll be worth every penny. Keep in mind that you can "cook" an EPROM if you expose it too much.
I have a CNet (CN600Eplus) ethernet card, and a boot rom for it, but the rom apparently uses the RPL protocol. Anyone know where I can get a RPL daemon for Linux, or of a place that has some information on the RPL protocol?
Would it be practical to use diskless workstations as Beowulf cluster nodes? Say, dual P3-450's (or hacked Celerons), 256MB of RAM (128meg DIMMs are under $100 apiece!), 100Mbit Ethernet (dumb hubs are cheap), booting off floppy, and not much else. Would this work for raytracing, SETI@Home, decryption, and similar tasks? You could build dual P3 nodes for not much more than a grand, and dual Celeron nodes for substantially less. Abit's upcoming dual Socket 370 board would be perfect for this.
Diskless nodes is a path many take when they build Beowulf clusters -- makes it easier to replace nodes &/or expand the cluster.
Christopher A. Bohn
cb
Oooh! What does this button do!?
Well, at my lab most of us use our PCs, Linux boxes, Macs etc chiefly as VNC terminals to our real work running somewhere on the network....
But we may be biased in that direction!
It was cool just to see how they did it. The machine itself was a bit slow, and I wanted a little more (better local browser, possibly NIS, graphical login?, etc). They have a new model out thats bigger faster better, but I haven't seen it yet.
http://www.csis.gvsu.edu/matt/projects/wopr/diskle ss.html
Diskless Beowulf
IMHO yer gonna need a local page & swap file, and I'd be inclined to load fonts locally as well.
Probably the first thing to go check is the assorted howto's about setting this up. Howto.linuxberg.com is a good place to go. As far as specific hardware, look for network cards with the boot rom, or have a floppy drive and one disk to get things going on boot.
-----
I don't think that the network cards that support boot proms are expensive. I got some NE2000 compatible PCI 10baseT cards that support PROMS for US$20. I'm not sure if it supports 32k though. The only expensive part would be the writer and eraser.
Scuttlemonkey is a troll
Just as a footnote, your setup used to be extremely common in the late-80s, early-90s with DOS-client LANs. Even when the clients had hard disks (which weren't cheap), applications were run off the network because it made administration simpler. There usually was a boot menu that launched from the Novell login scripts that insulated the users from the DOS prompt.
What killed this setup was Windows 3.1. Usually swapping over the network was extremely slow, and all of your Windows clients had to have exactly the same hardware for it to work (at least for the stuff it handled like video and mouse). Windows also did not like it's system directory marked read-only.
--
Business. Numbers. Money. People. Computer World.
Instead of expensive network cards with boot proms, you could also get EPROM floppy emulator cards like this or this and then copy a working boot floppy image onto it. You can find these and more by asking Alta Vista for: solid near state and emulate and floppy and EPROM
- "History shows again and again how nature points out the folly of men" -- Blue Oyster Cult, 'Godzilla'
Because I'm tired of writing it all over again, I put it on http://acs.lavsa.com/sskracic/diskless/.
This describes one real-world setup and also sheds some light on motivation. In my experience, deploying diskless (Linux) clients led to significant decrease in hardware and software cost, and (most important!) greatly reduced admin time.
Speaking of booting off of EPROM's, does anyone know if a flourescent style black-light puts out enough UV to erase an EPROM? I built a simple programmer, but I don't want to buy an eraser, but I do own a black-light. Any help would be appreciated.
"All those tubes and wires and careful notes!"
This is already supported in the kernel. You just need to export the server filesystems with NFS, and set up the clients to boot with NFS as the root filesystem.
What you want is basically to just set up the server with the software for the clients exported over NFS, and a bootable area for the clients. IE, you'll set up a directory tree on the server which will work as a root directory for your clients (with some separate writable space for each client). You'll then probably end up making a boot disk for your clients (easier than getting your NICs to use a boot rom), with a linux kernel on it set up for booting with NFS as the root directory.
Except for the little trivialities, like:
this is a complete no-brainer. Nothing to it.
A few tips:
A couple other things:
If you have enough memory its not a problem. Enough memory depends on what you're going to run - if they're just basically going to be X terms, 64 should be heaps. They make very fast xterms. Even as workstations they will probably run ok, but you may need more memory.
Fonts over the network is not really a problem either, either nfs or a font server is fine. I've had pools of terminals netbooting and running off a single UE2 server with few performance problems. I mean generally X11 applications dont use that many fonts anyway - and you're not likely to be running gimp are you?
Of course, that may depend on what you consider reasonable performance
The actual setup for net-booting linux isn't too hard either, there are HOWTO's or other doco on setting up bootroms and so forth to boot a kernel over the network.
__// `Thinking is an exercise to which all too few brains
_
\\/ are accustomed' - First Lensman
The FreeBSD Handbook (http://www.freebsd.org/handbook/diskless.html)
has a quick guide to diskless operation.
You can run netboot from an eprom or a diskette to bootstrap the kernel from the net too.
We (http://www.ing.unipi.it) actually run Diskless FreeBSD with per host configuration.
More info (not updated!) about our local setup: http://www.iet.unipi.it/~luigi/diskless981113/
P.S. We run Debian GNU/Linux and RedHat too, no OS war intended.
Also have a look at Ken's presentation on this page for info on setting up a discless PC as a X terminal.
University of California Deptarment of Mathematics
/etc.)
has 38 diskless workstations. Currently they all use a kernel disk which uses nfsroot to mount file systems off a central server. Basicly all of our Grad Students and Visiting Research Profs use them.
Looking for a way to get bootroms for the Intel EEPro cards -- unfortunatly the first 18 systems we bought have the non-boot-prom-socketed NetGeat 310TXs so we can't use proms on those..
This is basicly a home-brew setup, always looking for better ways to do it. One thing thats neet about this setup is that we export most of the filesystems read-only (had to hack some of the mount utils to keep from fiddling with files in
Wheeeee...
Thank you for the very informative post! Have you tried this setup over a DSL VPN or any such link?
Thanks!
Eric
Thank you for the very informative post! Have you tried this setup over a DSL VPN or any such link?
Thanks!
Eric
Thank you for the very informative post! Have you tried this setup over a DSL VPN or any such link?
Thanks!
Eric
could look up the netboot and etherboot projects
(http://google.com will lead you there)
I've been using software from Sun for about three years now called AutoClient, that allows not diskless workstations, but cached OS workstations. /usr.
Most of the advice you've gotten saying not to go totally diskless is right; no matter how much memory a Unix system has, it really need a swap space and swapping over the network is unbearable.
Autoclient uses a cached filesystem for root and usr, which means that when it queries the server for files, it stores them in a filesystem that only has structure while the system is runing. This does mean longer boot phases, but my systems are rarely down. It is much nicer for administration, since I can change all the workstation root directories with a foreach loop, and then execute a command to tell the clients to update their root directories. Also, I only have to apply patches to the server, and patches get applied to all the root directories of the clients. They all share a common
I've been debating setting up something similar for Linux with a friend, though we'll probably try it with sparc systems first, since it's considerably easier to boot a sparc from the net than a pc.
Bob Campbell
Just read the Diskless-HOWTO. It should come with your linux-distribution; usually it's located in /usr/doc. btw I have a bootrom-burner, so if you need a custom bootrom, just mail me.
0x or or snor perron?!
Why is a swap file needed?
;)
At the moment RAM is so cheap that you can
get 256MB of RAM for the price of fairly low
end HD. I would have thought that with current systems, virtual memory is largely obsolete.
Most people seem to be considering diskless
systems as slow dumb terminals, but this doesn't
need to be the case. Todays low end systems
(Celeron or K6-2) cost very little and can
act as extremely fast network computers.
I've actually been considering building a
dual PIII diskless system...
I've also seen for sale a diskless server,
which gets it's data from a separate RAID
array.
Last year Siemens had Linux-based NCs at their CeBIT booth. They were diskless, but had a 8 or 16 MB flash memory to boot into the web AFAIK. Maybe they have some info in the web.
Check out Gero's netboot pages It features a complete system to boot any intel machine with bootp/tftp. If you don't have access to an eprom writer for your ethernet cards, it also works with a floppy. Allows you to network boot MS/DOS (also WinSucks9x) and of course Linux
Our school used to have Dell 486/25's set up as diskless clients, running Netware 3 and Windows 3.1. You logged into Netware from DOS, then it automatically loaded windows. The machines wouldn't boot if the network was down.
This increased security no end, especially once they removed the floppy drives. Also PCs set up like this are less nickable.
Anyway, I'm sure the hardware exists to do it.
I, too, have set up diskless clients, and love 'em. This guy is right on target (see original message)
You don't even need to go PIII, either. A nice cheap little AMD K6-2 (or the new K7, I suppose) would probably do just find, for considerable less money. (I have a AMD K6-2 350 that, on a good day, keeps up with a dual Intel PII-233)
Hi,
/home and /usr file system :-).
:-).
/etc/hosts in the long range.
/home, /usr, /opt between server and clients and should have / and /var separate for each.
/clients/client1. Then compile a kernel with root NFS and setup DHCP or Bootp. With only a few changes on your server you should be able to boot a diskless client. The changes would be DHCP, an entry in /etc/hosts or DNS, and NFS server with exporting /clients/client1).
:-).
/etc/exports right. And then the client will boot and you can make adjustments in the root file system.
:-).
:-).
we use a lot of diskless Linux clients here at several locations of our company and they work great.
If you have a company environment with a central database and/or file server you depend on a central server anyway so it wont bother you if the clients dont work without the server.
As for administration its the best thing to do. Its more like having ASCII terminals. You buy a new box, hook it on the network, make a few configuration entries on the server and as soon as you boot the new client it will have all the same software as the other clients, it will have an identical configuration and it will have access to all your personal files.
And if you care for a good integrated network configuration (NIS, NFS, DNS, etc.) you can really say that as Sun puts it "the network is the computer". People would be able to log in from every workstation and have the same desktop and applications everywhere. No need for complex registry things, just a shared
Often people think a diskless client could be compared to X terminals or things like Microsoft Terminal Server, but this is not true! With a diskless clients each user has its own CPU and memory. They only share the disks. In times where every office application has animated cartoon characters talking to each other you need a lot of CPU power
Another misconception (in my belief) is, that a diskless client is a "thin client", meaning a cheap client. We use our diskless clients for Java and office applications. This cant be done with an old 486. This cant even be done with good performance with a JavaStation (at least the last one Ive seen). But it can be done with a Pentium III 500 Mhz. And theyre cheap enough (at least for serious company use and more so when you look at TCO). The main benefit with diskless clients is not saving the money for a disk drive, but easy system administration and less problems for the users!
BTW, the performance is ok. Its a bit different from "normal" configurations. You might need to wait a little longer for the first load of a program after a reboot, but programs like Netscape or Applixware Office start in a couple of seconds and it will be faster the second time. After the initial loading you wont see any difference, at least if you dont have huge files as in image or audio processing.
My recommendations for say 10 diskless clients with heavy workload would be (things may vary):
Server:
256 MB main memory (you cant have enough here)
Pentium III 500 MHz (the current user-space NFS needs a lot of CPU Power)
SCSI disks (with diskless clients parallelization is important)
Clients:
128 MB main memory (they shouldnt swap)
Pentium III 500 MHz (could be less depending on your applications, we use Java)
Floppy (we boot from a floppy)
NIC (buy a good one with a stable driver, as every bit will go through here; we use DEC Tulip based PCI 10/100 NICs).
Network:
It might work ok with 10MBit, but you should really go for 100MBit.
Buy a dual speed hub.
Now to the Linux setup:
First thing the client needs is a Linux kernel in his memory. You can use a boot PROM or start with a floppy which Id recommend. Just compile a kernel and dd it onto a floppy disk.
The kernel should have drivers for your favorite NICs compiled in and it should have support for root file system on NFS and Bootp support. After building the kernel you have to use rdev to actually configure the root file system.
After you start the kernel, it will ask a Bootp (DHCP) server for its IP address and root file system. This will then be mounted via NFS and the rest goes like any other Linux box.
If you want a really integrated solution, you have to set up the following subsystems:
* DNS Domain Name Server
Have entries for the server and all clients. Better than
* Mail
Use the server as SMTP and POP3 server.
* NIS/YP Network Information System
Use NIS for a common login administration.
* DHCP/Bootp
Use DHCP for configuration of diskless clients.
* Printer spooling
This is a bit tricky with apsfilter, but having network printers or printserver helps.
* NFS Network file system
This is the base system for sharing files. You can share
For start make a copy of your Linux filesystem and name it say
I did this using a SuSE Linux 5.3 but it should work with every distribution (we still use this older version, because in a company environment you just cant change your OS every other day
Just try it out step after step: first boot the client and see if you get a Bootp request at the server. Then Bootp should give the client a new IP address. Both should be fairly easy. The tricky part might be getting a root file system from the NFS server. But then, you only have to get
Good luck and best wishes!
Michael
P.S.: I made my first Unix diskless client system about 10 years ago using DECstations running DECs Ultrix because our university had ordered several DECstations but only one disk because of a money shortage
I believe, that the ability to run on diskless clients is a huge advantage for Linux compared to Windows NT. As I said, the Microsoft Terminal Server is a different concept and its worse in performance because all users share a CPU.
But on the other hand, if you push Linux so far and make heavy use of all daemons ranging from DHCP to LDAP, you will also see its limitations. Coming from different persons, different parts of the system might not work as good as possible with others. Examples include the integration of NIS. I mean, most things in Linux work, sort of, but it could be better. Does anybody know how RedHat scales in this area, as I never actually tried it.
If you feel the need to criticize me for "Linux bashing" please make your own diskless client configuration first and install it in several companies with real users and real world problems and then come back. But to calm you: a large Linux installation is better than a large NT installation any day
Checkout our website at www.dresults.com,
we are selling network cards with bootroms,
eprom burners, and complete workstations with
the bootrom and everything. You can also purchase
through Linux Systems Labs (www.lsl.com).
If you run the applications on the host, and
just run the Linux kernel, Xserver, lpd
on the workstation, you don't even need to worry
about swap.
Also, we are using Net Audio Services (NAS) to handle sound.
we are running a 50 diskless linux stations for
the undergraduate students (CS dept. Hebrew Univ.)
this is our most reliable configuration for large numbers of workstations. app. servers supporting
80+ NCD X-Terminals are far less stable.
the author of the scripts for diskless station administration is planning to post them soon on the www.mosix.cs.huji.ac.il, so check there for
updates
There is a company here in South Africa who has successfully done this for about 3000 workstations (AFAIK). They are called UCS. I don't know how much they'll tell you, but it never hurts to ask...
--- Brian the Wise Friend to Small Fury Animals Everywhere...