LTSP Unofficial FAQ! 8D
check out the IceWM Recipe, windows 98/eb dualboot, user suicide, shared phoenix and setnumlock scripts,
the pre ltsp checklist, my EPROM experience, xpde, IE 6 on Linux related links
1.- Why did you write this faq?
2.- syslogd: syslogd: unable to write to remote file handle 192.168.0.1:514
3.- What can I do with mozilla?
4.- Can you just set all the dhcpd.conf options globally?
5.- How may I start to troubleshoot X?
6.- Fatal server error:
XDMCP fatal error: Manager unwilling Host unwilling
7.- I have installed RH 8/MDK 9 and LTSP 3 four times with no results
8.- TFTP error 1 (File not found) Unable to load file?
9.- What goes in /etc/exports??
10.- GDM is not listening to my requests, how can I fux this?
11.- Is it possible to create a boot disk, that would work with most systems?
12.- Should I create /var/opt/ltsp/swapfiles directory?
13.- Can I run an X terminal client from a regualar Linux installation?
14.- Thin clients with IP's outside the range declared
with host blabla in dhcpd.conf do not get assigned a hostname. It's
just blank. What can we do to get every tc assigned a hostname,
regardless of static address or not?
15.- How can I do fail-over and load balancing on different kind of servers?
16.- How can I send a message to my X users?
17.- Screensavers take too much bandwidth, is there an image slideshow replacement
18.- Can Linux handle more than one network card to be
connected to the same switch for added performance?
19.- My GUI takes several tens of seconds to be displayed on the xterminal
20.- How may I setup my old video graphics card?
21.- How may I choose between different hosts from a terminal?
22.- How may I use PXE network cards?
23.- How may I log printing activity?
24.- Where is my display manager?
25.- I have 2 ntework cards, how do I tell dhcpd where to look?
26.- Is there an easy way of simulating the operation of
various terminals within the server for testing purposes?
27.- Is dhcp server working?
28.- How may I tell the related services to start when turning the server on?
29.- Can I boot from a compact flash IDE disk?
30.- XFS is running, but are you sure it is listening on tcp port 7100?
31.- too many open files in the system
32.- Hoy may I add the LTSP client option to my current computer?
33.- Can I boot a terminal from a CDROM?
34.- Can I boot a terminal from a floppy disk?
35.- I have little ram on my hard disks, how can I make them work ok?
36.- How can I see a workstations X sratup messages?
37.- How to fix 'could not load font 'fixed', and the terminal keeps re-trying to start X?
38.- How avoid xsession-errors filling the hard drive?
39.- Is there a text based graphing program for Linux load?
40.- How can I make sure my users cannot se other users files?
40.- How can I delete core.* files?
start - 1.- I had no place to save this data, I gues the web is the best!
Some data is on the official docs, some data I could not find anywhere so this
page will even help me just in case 8)
start - 2.- syslogd: syslogd: unable to write to remote file handle 192.168.0.1:514
This can be fixed by telling your syslogd on your server to accept
remote connections. You didn't mention which distro you are running,
but try finding the startup script, usually /etc/rc.d/init.d/syslog
Find the line that starts syslogd, and add a '-r' to it.
start - 3.- What can I do with mozilla?
Script one
<html>
<head>
<script language="javascript">
function
WinOpen(){
window.open("http://skynet.frankfurt-airport.de","","toolbar=yes,scrollbars=yes");
timeoutID=setTimeout("WinClose()", 1000);
}
function WinClose(){
window.close();
}
</script>
</head>
<body onLoad="WinOpen()">
</body>
</html>
Script 2
rm -d "/home/Internet/.mozilla/default/yui33i76.slt/lock" -f
mozilla "http://10.11.115.12/kiosk.html"
start - 4.- Can you just set all the dhcpd.conf options globally?
I edit dhcpd.conf with the following line:
shared-network WORKSTATIONS {
subnet 192.168.0.0 netmask 255.255.255.0 {
range dynamic-bootp 192.168.0.10 192.168.0.100;
if substring (option vendor-class-identifier, 0, 9) = "PXEClient" {
filename "/lts/pxelinux.0"; }
else if substring (option vendor-class-identifier, 0, 9) =
"Etherboot" {
filename "/lts/vmlinuz-2.4.19-ltsp-1"; }
}
}
After doing this all my clients are working without problems - without
editing /ec/hosts lts.conf or anything else.
Is there nevertheless any reason to edit /etc/hosts or something else?
start - 5.- How may I start to troubleshoot X?
To debug, specify:
RUNLEVEL = 3
in lts.conf
and start X by running:
/tmp/start_ws > /tmp/x.log &
and then read its output with
cat /tmp/x.log | less
start - 6.- Fatal server error:
XDMCP fatal error: Manager unwilling Host unwilling
This is definately an XDMCP problem. I know how to fix
this in XDM and KDM (/etc/X11/xdm/Xaccess file), but I'm
not sure how to fix it in GDM. Maybe restarting GDM would
do the trick.
Also, Try this:
netstat -anp | grep ":177 "
Just to make sure that gdm is listening for remote connections.
start - 7.- I have installed RH 8/MDK 9 and LTSP 3 four times with no results
You have a typical display manager misconfiguration problem.
First, you need to determine which display manager you are using.
Either xdm, gdm or kdm.
Try: ps -elf | grep dm
and see which is running.
Then, try this:
netstat -anp | grep ":177 "
to see if anything is listening on port 177 (xdmcp). If not,
then your display manager isn't configured to listen to remote
requests.
You'll need to configure the dm to listen. If you are using
gdm, edit your /etc/X11/gdm/gdm.conf file, and find the section
that starts with '[xdmcp]'. The line below that should
say 'enabled=true' or 'enabled=yes'. If it doesn't say that,
then you need to fix it so it does.
if you are using kdm, then you need to edit /etc/kde/kdm/kdmrc
and look for a similar section with a similar answer.
If you are using xdm, then you need to edit /etc/X11/xdm/xdm-config
and find the line that says 'DisplayManager.requestPort: 0'
and make sure the line is commented out.
Once you change one of the above config files, you'll need to
restart the display manager.
For you, I recommend just rebooting the server, it's the easiest
way.
If, on the other hand, you find that there isn't a display manager
running, then you need to check your server's current runlevel.
Do this by typing:
runlevel
It should return something like:
N 5
The 2nd field indicates the current runlevel.
If yours says '3', then you need to edit your /etc/inittab file
and change the 'initdefault' line to 5 and reboot.
That should help.
start - 8.- TFTP error 1 (File not found) Unable to load file?
This file:
/etc/xinetd.d/tftp
should contain something similar to this text:
# default: off
# description: The tftp server serves files using the trivial file transfer \
# protocol. The tftp protocol is often used to boot diskless \
# workstations, download configuration files to network-aware printers, \
# and to start the installation process for some operating systems.
service tftp
{
disable = no
socket_type = dgram
protocol = udp
wait = yes
user = root
server = /usr/sbin/in.tftpd
server_args = -s /tftpboot
}
But I dont understand, why on Mandrake 9 it had this
server_args = -s /var/tftpboot
So remember, this
server_args = -s /tftpboot
is the correct one! 8)
and consecuently, in dhcpd.con, you will use something similar to:
filename "/lts/vmlinuz-2.4.9-ltsp-5";
and the servers real directory is like this:
/tftpboot/lts/vmlinuz-2.4.9-ltsp-5
This is really a big mess 8/
start - 9.- waht about those exports?
This goes in /etc/exports (depending on your net config of course)
/opt/ltsp/i386 192.168.0.0/255.255.255.0(ro,no_root_squash)
/var/opt/ltsp/swapfiles 192.168.0.0/255.255.255.0(rw,no_root_squash)
start - 10.- GDM is not listening to my requests, how can I fux this?
In this file
/etc/sysconfig/desktop
change from this
DISPLAYMANAGER=kdm
to this
DISPLAYMANAGER=gdm
and restart X with this command to take effect of the new configuration:
telinit 3; sleep 15; telinit 5 &
start - 11.- Is it possible to create a boot disk, that would work with most systems?
The Windows 2000 Remote OS Installation Service (RIS) includes a Remote Boot Floppy
Generator utility (RBFG.EXE) that creates boot floppy that supports over 25
different PCI network adapters. The RIS Boot Floppy can be used with LTSP as
long as your LTSP server is configured to support PXE clients.
3Com's free MBA on Disk offers similar functionality as long as you are using
a 3Com network adapter. http://www.3com.com/products/en_US/result.jsp?selected=5&sort=fsize&sku=3CMBA-D
Regards,
Conrad Lawes
start - 12.- Should I create /var/opt/ltsp/swapfiles directory?
Create this directory if it doesnt exist already
/var/opt/ltsp/swapfiles
start - 13.- Can I run an X terminal client from a regualar Linux installation?
x :0 -query ip.of.terminal.server
for example
x :0 -query 192.168.0.254
acording to tls example configutarions <8)
or you could make it start straight into X terminal mode with the last line in the inittab being set as follows...
x:5:respawn:/usr/X11R6/bin/X -query ip.of.terminal.server
for example
x:5:respawn:/usr/X11R6/bin/X -query 192.168.0.254
acording to tls example configutarions <8)
start - 14.- Thin clients with IP's outside the range declared
with host blabla in dhcpd.conf do not get assigned a hostname. It's just
blank. What can we do to get every tc assigned a hostname, regardless of
static address or not?
We are running ISC's V3.0pl1 dhcpd, and the following is my
dhcpd.conf section that returns a hostname to the client:
group ltsp-pxe-ws {
use-host-decl-names on;
option log-servers 192.168.6.19;
next-server 192.168.6.19;
filename "pxelinux.0";
option vendor-encapsulated-options 09:0f:80:00:0c:4e:65:74:77:6f:72:6b:20:62:6f:6f:74:0a:07:00:50:72:6f:6d:70:74:06:01:02:08:03:80:00:00:47:04:80:00:00:00:ff;
host ws001z6 {
hardware ethernet 00:08:c7:fc:27:58;
fixed-address ws001z6;
option host-name "ws001z6";
}
}
My dhcpd is running on a different box from the tftpd server, which
is on a different box than any of my nfs mounts, which is a different
box than my named server. For some reason (since I'm not a dhcpd guru)
if I don't include the 'option host-name', then the client is assigned
the same name as my default 'option host-name'. That is, for a
definition:
host ws001z6{
hardware ethernet 00:08:c7:fc:27:58;
fixed-address ws001z6;
}
"ws001z6" is not passed in my scenario. Is it because I've included
an option host-name that has global scope?
Anyways, perhaps you have a similar setup.
start - 15.- How can I do fail-over and load balancing on different kind of servers?
Using dhcp for fail-over and load balancing really isn't very
complicated... and it offers load balancing options that can be
explicitly defined as Hans said. The only difference between a primary
and secondary server are a few additional lines in dhcpd.conf.
This is the primary:
failover peer "ltsp" {
primary;
address 172.16.0.1;
port 519;
peer address 172.16.0.2;
peer port 520;
mclt 3600;
max-response-delay 30;
max-unacked-updates 10;
load balance max seconds 3;
# split 128;
hba ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:
00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00;
}
This is the secondary:
failover peer "ltsp" {
secondary;
address 172.16.0.2;
port 520;
peer address 172.16.0.1;
peer port 519;
max-response-delay 30;
max-unacked-updates 10;
load balance max seconds 3;
}
That's it!
If you use the hba option instead of split, you can proportion the
server responses based on their speed and memory. This allows the mix
of
a P-III 500 with a P-4 1.6G where each will be optimally utilized. The
faster machine, in this case, can handle roughly 3 times the number of
clients... and can be configured to do so using the hba bit map.
Doing this allows the slower (and cheaper) server to float on the
network as an active backup without the risk of it becoming overloaded
during normal use by "unlucky" clients. In the event of a primary
failure, things will slow down for everyone... but they won't stop.
There are a lot of ways to approach this and the system you described
would have the advantage of handling more then two servers while dhcp
load balancing is currently limited to two.
-Tom
start - 16.- How can I send a message to my X users?
Hans, Delz,
the question is not exactly moot - there *are* instances when
sending a message to all the logged in users is very valuable. For old
style tty terminals there is a "wall" command that does exactly what
Delz
wants, for X terminals we need xmessage, the way Hans points out.
Here is what I use (I call it "wallx"):
#!/bin/bash
users=$(who | awk '{print $1}')
for user in $users
do
display=$(who | grep ^"$user" | awk '{print $2}')
#uncomment the following line and comment out next if access_control
#su $user 2>/dev/null -c "xmessage -center -timeout 30 -display
$display $user, please log out now"&
xmessage -center -timeout 30 -display $display $user, please log
out now 2>/dev/null &
done
#end sript
notice that i use direct xmessage, because i have
disable_access_control=y
in lts.conf, so concievably, i could get away with just one run of
"who"
good luck, julius
start - 17.- Screensavers take too much bandwidth, is there an image slideshow replacement?
The program isn't really a screen saver, it is just a graphic program
called xloadimage. You should find it in most distros. To run it as a
screen saver, replace the programs: section of your xscreensaver config
file with something like:
default-n: /usr/bin/X11/xloadimage -fullscreen \
-onroot -border black \
/home/kiosk/images/kimage1.jpg \n\
default-n: /usr/bin/X11/xloadimage -fullscreen \
-onroot -border black \
/home/kiosk/images/kimage2.jpg \n
You will find the per-user xscreensaver config file in the user's
home directory as .xscreensaver. There is also a global config file in
your X11/app-defaults directory named XScreensaver. You need to eliminate
all the other programs from both of these
files so that they do not rotate along with your 'slide show.'
Set the screensaver to mode: random and the cycle: to something like
30 seconds
Pete
start - 18.- Can Linux handle more than one network card to be connected to the same switch for added performance?
Oh yes there si a solution and I personally love showing it off to
all Windoze centric buddys cus' it is a pain if even possible to do with
windows. Channel bonding( sometimes called
agregated/trunking/ramping/etc,etc) is Ideal for this particular problem, best of all teh linux
kernel can do it with out any addons and standard nics. The one real issue
is wether or not your switch will handle it. It does take a better than
bargain basement model to does this. If your switch can, then check in
the /usr/src/linux/Documentation/networking/bonding.txt for the details
to set it up.
A second more painful option, but should work on most setups is to
dual-home the server and use two ips for the server. In dhcp is there is
an option to "load-balance" between 2(exactly 2) servers
(http://theseus.sourceforge.net/projects/ets/supplemental.html#DHCP_Load_Balancing_Optimization
for details). Treat teh tow server ips as seperate servers to give a
false appearance of two servers and balance the clients between the two
nics (this only good on a switch) or possibly smarter if you expect
heavy internet use is to install a dedicated nic for the clients and a
second one for the rest of the network(again this only works on switches),
this probably a good idea any tine you have an LTSP server that is
expecting lots of traffic from things other than the clients such as
internet,NFS mounted dir. and the like.
Evan
start - 19.- My GUI takes several tens of seconds to be displayed on the xterminal.
If you are getting things working at all, then it is likely
NOT an ipmasq problem.
My guess is a DNS or /etc/hosts problem.
Make sure your workstation is listed in the /etc/hosts file,
or make sure you've got it setup properly in your dns tables,
along with the proper reverse. There are several services that
want to map the IP address back to a hostname.
start - 20.- How may I setup my old video graphics card?
All VGA monitors, no matter how old, can support ANY color depth.
Any video hardware out there should run with XFree86 4.2.0 either with a native
driver for the card, or by using the "vesa" driver, "fbdev" or with the "vga"
driver as a last resort.
3.3.6 was dropped in RHL 8.0 because it was a huge maintenance problem and
caused a lot of user confusion and distribution integration confusion, as well
as major problem whenever secuirty issues were found - since nobody upstream is
interested in 3.3.6 anymore.
start - 21.- How may I choose between different hosts from a terminal?
Open this file
/opt/ltsp/i386/etc/rc.local
The part to modify is originally something like this:
if [ -w /proc/progress ]; then
echo "/usr/X11R6/bin/${XBINARY} ${ACC_CTRL} -query ${XDM_SERVER} >/dev/tty3 2>&1" >/tmp/start_ws
else
echo "/usr/X11R6/bin/${XBINARY} ${ACC_CTRL} -query ${XDM_SERVER}" >/tmp/start_ws
fi
you have to change from 'query' to 'indirect', so it will look like this:
if [ -w /proc/progress ]; then
echo "/usr/X11R6/bin/${XBINARY} ${ACC_CTRL} -indirect ${XDM_SERVER} >/dev/tty3 2>&1" >/tmp/start_ws
else
echo "/usr/X11R6/bin/${XBINARY} ${ACC_CTRL} -indirect ${XDM_SERVER}" >/tmp/start_ws
fi
start - 22.- How may I use PXE network cards?
Simply download the .lzpxe from rom-o-matic what works with your NIC
(e.g. the rtl8139 if you have a NIC with this chip)
Then put something like this in your dhcpd.conf and restart the daemon:
host ws004 {
hardware ethernet 00:30:00:03:B5:49;
fixed-address 192.168.0.4;
if substring (option vendor-class-identifier, 0, 9) = "PXEClient"
{ filename "/data/tftpboot/lts/eb-5.0.7-rtl8139.lzpxe"; }
else if substring (option vendor-class-identifier, 0, 9) = "Etherboot"
{ filename "/data/tftpboot/lts/vmlinuz-2.4.18-ltsp-1";
}
}
start - 23.- How may I log printing activity?
You can increase the logging levels in the cups config file:
/etc/cups/cupsd.conf
With: LogLevel debug
LogLevel can be set to:
debug2
debug
info
warn
error
none
start - 24.- Where is my display manager?
Figuring out which display manager is pretty easy.
On the server, as root, do this:
netstat -anp | grep ":177 "
It should display the PID/program
As for customizing the background, that is entirely dependent
upon which display manager is running.
For XDM, look at the Xsetup scripts in /etc/X11/xdm
For GDM, look for the gdm.conf file.
For KDM, look for the kdmrc file
Hope that helps,
Jim McQuillan
start - 25.- I have 2 ntework cards, how do I tell dhcpd where to look?
With two cards, you need to tell dhcpd which interface to run on. So,
you need to edit:
/etc/rc.d/init.d/dhcpd
and add/edit a line:
INTERFACES="ethX"
where ethX is the interface you want dhcpd to run on.
After changed, restart dhcpd:
/etc/rc.d/init.d/dhcpd restart
cheers,
-rob
start - 26.- Is there an easy way of simulating the operation of
various terminals within the server for testing purposes?
Using vmware and a floppy bootable image for a pcnetfastiii nic.
- file->wizard create a new config for a new vm.
- setting->configuration-editor set floppy to the (file)
eb-5.0.x-pcnetfastpciii.lzdsk
- start vmware with this config
- you may need to tell dhcpd about new mac-addresses (and restart it)
- if you want to run more than one client,
- copy the vmware configuration into a new directory,
add a line like the following to the config file (.vxm or .cfg):
ethernet0.address = 00:50:56:XX:YY:ZZ
(as you may expect each vm needs a unique mac-address
see the text below from vmware for values for XX:YY:ZZ
(in short: 00 < XX < 3H, 00 < YY|ZZ < FF))
again you may need to config/restart dhcpd
start vmware with this config
grtx,
Johan
-------------- begin from vmware documentation -----------------------
Changing the MAC Address of a Virtual Machine
When a virtual machine is powered on, VMware Workstation automatically
assigns each of its virtual network adapters a MAC address. MAC stands
for media access control. A MAC address is the unique address assigned
to each physical network device.
The software guarantees that virtual machines are assigned unique MAC
addresses within a given host system. However, the software does not
guarantee that a given virtual machine is assigned the same MAC address
every time it is powered on. In addition, VMware Workstation does its
best, but cannot guarantee, to automatically assign unique MAC
addresses
for virtual machines running on multiple host systems.
If you want to guarantee that the same MAC address is assigned to a
given virtual machine every time, or if you want to guarantee a unique
MAC address for each virtual machine within a networked environment,
you
can assign the address manually instead of allowing VMware Workstation
to assign it automatically.
To manually assign the same, unique MAC address to any virtual machine,
use a text editor to add the following line to its configuration file
(the .vmx file on a Windows host or .cfg file on a Linux host):
ethernet0.address = 00:50:56:XX:YY:ZZ
where XX must be a valid hex number between 00h and 3Fh, and YY and ZZ
must be valid hex numbers between 00h and FFh. Because VMware
Workstation virtual machines do not support arbitrary MAC addresses,
the
above format must be used.
So long as you choose XX:YY:ZZ so it is unique among your hard-coded
addresses (where XX is a valid hex number between 00h and 3Fh, and YY
and ZZ are valid hex numbers between 00h and FFh), conflicts between
the
automatically assigned MAC addresses and the manually assigned ones
should never occur.
start - 27.- Is dhcp server working?
type these commands to find out:
ps fauxww | grep dhcp
netstat -an --program | grep :67
start - 28.- How may I tell the related services to start when turning the server on?
chkconfig --level 345 dhcpd on
chkconfig --level 345 nfs on
chkconfig --level 345 tftpd on
chkconfig --level 345 portmap on
or use ntsysv to set tftpd, dhcpd, portmap and nfs on system startup 8)
If you believe you have already set them up and only want to make sure they
are on, type this commands:
chkconfig --list tftp
chkconfig --list nfs
chkconfig --list dhcpd
chkconfig --list portmap
start - 29.- Can I boot from a compact flash IDE disk?
Definitely try that (please!). :)
Assuming your CF card is on /dev/hda, partition the drive with a Linux
partition (type 83) using fdisk (or your favorite Linux partition
tool),
create an ext2 (or your favorite) filesystem on the partition, create
the file structure necessary for the bootloader, copy the Etherboot
LILO
code to the device, and initialize the bootloader on the CF device.
1) Partition the CF device with a Linux partition scheme:
Type 'fdisk /dev/hda' at a command prompt. This will bring up the
paritioning tool under Linux. Delete all existing partitions on the
device (unless you want to save something). Create a single primary
partition with ID 1. Make sure the parition type is Linux (should be
83, if I remember correctly). Write out the parition table and exit
fdisk.
2) Create a filesystem on the newly created partition:
Type 'mke2fs /dev/hda1' at a command prompt. This will create an EXT2
filesystem on the first partition on your CF device. The device is now
prepared to be mounted and files placed on it. You can mount it by
typing 'mount /dev/hda1 /mnt' (or a location you would like to use; we
will assume /mnt).
3) Add the files for LILO and Etherboot for your network card:
Create the following directories on your CF device: boot, etc, tmp.
The
Etherboot files go in /boot along with specific LILO files. The
lilo.conf file goes in /etc.
To create the directories, type 'cd /mnt' at a command prompt. Type
'mkdir boot etc tmp' from that location.
Copy the parts of LILO from your current installation (if you use LILO
as your bootloader) or download and install LILO on your workstation
and
copy the relevant files. Most likely you will want the /boot/message
file and the /boot/boot.b files. Copy them from /boot to /mnt/boot.
Copy the /etc/lilo.conf file to /mnt/etc/lilo.conf. Copy the LILO
Etherboot file from wherever you downloaded it (www.rom-o-matic.net has
them) to /mnt/boot. Edit /mnt/etc/lilo.conf and change the
configuration to point to the /boot/ file instead of
the kernel image it was pointing to.
4) Initialize the bootloader on the CF device on the MasterBootRecord
of the CF device:
You will need to make sure that the /mnt/etc/lilo.conf file has
boot=/dev/hda (or whatever your CF device will be when inserted). If
it
is currently not that device, you will need to determine which BIOS
disk
it will be and "modify" it with bios=... and disk=... lines in
/mnt/etc/lilo.conf. I will leave that as an excercise in reading the
man pages for lilo and lilo.conf for you to figure out. E-mail the
list
if you need assistance.
To initialize the CF device's MBR to allow it to boot, you will need to
run the /sbin/lilo command from a command prompt. LILO needs to know
that you intend to install the bootloader on a device other than your
currently running boot drive, so you need to specify the "root"
location
for the configuration file, etc. to LILO when running the command. You
do this with the '-r' option to lilo.
Type 'lilo -v -r /mnt' to install the bootloader onto your CF device.
Reboot and enjoy Etherboot/LTSP! :)
start - 30.- XFS is running, but are you sure it is listening on tcp port 7100?
Try this:
netstat -an | grep ":7100 "
If you don't see a line showing 'LISTENING', then you
need to configure xfs to listen.
Look at the /etc/X11/fs/config file, I know on RH7.2
and 8, there is a setting for 'no-listen', that should
be commented out.
For RH7.1, you may need to modify the way XFS is
started, I don't remember for sure.
Hope that helps,
start - 31.- too many open files in the system
echo "65536" >/proc/sys/fs/file-max
start - 32.- Hoy may I add the LTSP client option to my current computer?
Dead easy. Go to the www.rom-o-matic.org and get the Etherboot ROM Image
for your NIC in LILO Bootable ROM (.lzlilo) Image format. Save it in
your /boot directory. Add it as an option into your /etc/grub.conf
file. Reboot.
John
My /etc/grub.conf is shown here (a dual boot Win98 and RH8.0 system -
now a triple boot system with LTSP as the default option):
# grub.conf generated by anaconda
#
# Note that you do not have to rerun grub after making changes to this
# file NOTICE: You have a /boot partition. This means that
# all kernel and initrd paths are relative to /boot/, eg.
# root (hd0,1)
# kernel /vmlinuz-version ro root=/dev/hda3
# initrd /initrd-version.img
#boot=/dev/hda
default=0
timeout=10
splashimage=(hd0,1)/grub/splash.xpm.gz
title LTSP
root (hd0,1)
kernel /eb-5.0.7-3c509.lzlilo
title Red Hat Linux (2.4.18-14)
root (hd0,1)
kernel /vmlinuz-2.4.18-14 ro root=LABEL=/
initrd /initrd-2.4.18-14.img
title Microsoft Windows 98SE
rootnoverify (hd0,0)
chainloader +1
John McCreesh
start - 33.- Can I boot a terminal from a CDROM?
grab the mini-disk copy of knoppix and change the last line in the
inittab
to query your ltsp server
/usr/X11R6/bin/X -query server.domain.com
To modify the iso copy it to disk and do a loop mount on the iso.
Reburn and
enjoy...
Knoppix has darn good hardware detection so it should make it work
almost
flawless
Cliff Baeseman
start - 34.- Can I boot a terminal from a CDROM?
Sure, get your nic`s DOS executable ROM image(COM) from www.romomatic.net
and a freedos image floppy, put the floppy image on the floppy disk,
and edit autoexec.bat to run your ROM image, quite a simple task really 8)
start - 36.- How can I see a workstations X startup messages?
X is not configuring properly for your client
workstation. I've found it useful to capture the
output of X startup on the client from a runlevel 3
prompt like so:
# /tmp/start_ws > /tmp/x.out 2>&1
Then examine the file, looking for failure messages.
Chances are it doesn't like your mouse, monitor,
graphics adapter, or maybe not finding the fonts it
needs. This will simplify troubleshooting.
- Tom
start - 37.- How to fix 'could not load font 'fixed', and the terminal keeps re-trying to start X?
XFS is running, but are you sure it is listening
on tcp port 7100 ?
Try this:
netstat -an | grep ":7100 "
If you don't see a line showing 'LISTENING', then you
need to configure xfs to listen.
Look at the /etc/X11/fs/config file, I know on RH7.2, there
is a setting for 'no-listen', that should be commented out.
For RH7.1, you may need to modify the way XFS is started,
I don't remember for sure.
Hope that helps,
Jim McQuillan
[email protected]
I had a font problem for icewm (LTSP server, MDK8.2), and solved it
like
this :
in the script "/etc/rc.d/init.d/xfs", there is a line launching xfs
with
the option "-port -1". I had to modify it "-port 7100". Then I had my
font server and icewm worked.
Or disable it and launh it in /etc/rc.local : "xfs -port 7100 -daemon"
I recently had a crash on my server, and had to reinstall it (MDK9). I
did something else : copy all the fonts of the server in the ltsp
directory:
"cp -R -f /usr/X11R6/lib/X11/fonts/*
/opt/ltsp/i386/usr/X11R6/lib/X11/fonts"
It's not very smart, but it works fine, without xfs.
Maybe one of these two could help ?
Romain.
start - 38.- How avoid xsession-errors filling the hard drive?
Had another case of a ~/.xsession-errors file growing wildly and
filling
up a hard drive. Yuck.
After doing some poking around, I think I have a hack to keep this from
happening.
Comment out the top section of /etc/X11/xdm/Xsession
errfile="$HOME/.xsession-errors"
if cp /dev/null "$errfile" 2> /dev/null ; then
chmod 600 "$errfile"
exec > "$errfile" 2>&1
else
errfile=`mktemp -q /tmp/xses-$USER.XXXXXX`
if [ $? -eq 0 ]; then
exec > "$errfile" 2>&1
fi
fi
and replace it with this:
errfile="/dev/null"
exec > "$errfile" 2>&1
This will redirect all of this error output to the bit bucket.
-Eric
start - 39.- Is there a text based graphing program for Linux load?
nice.. $ sudo tload -d 1 /dev/tty9 &
Now I can anytime, and see the processor loading.
Set -d to a sane update period. Maybe as high as '-d 6' (six
minutes - 10 *'s per hour..)
thanks for that..
/steve
start - 40.- How can I make sure my users cannot se other users files?
Write a wrapper script for ls (i.e. ls -la $1 | grep $USER) then
alias ls to your wrapper script in bashrc. A user would only be able to
see their own files.
Pete
start - 41.- How can I delete core.* files?
find / -type f -name "core.*" -exec rm -rf {} \;
1.- Tom Lisjac has written a nice HOWTO for setting up two identical LTSP
servers on the same subnet here:
http://theseus.sourceforge.net/projects/ets/ets-howto.html
2.- This is a good guide to set up GDM
http://www.owlriver.com/tips/gdm-setup/
3.- Embedded Terminal Server (ETS) HOWTO
http://theseus.sourceforge.net/projects/ets/ets-howto.html
4.- Calculator for video timings
http://www.hut.fi/Misc/Electronics/faq/vga2rgb/calc.html
5.- Project: Folding @ Home
http://www.extremeoverclocking.com/articles/howto/FAH_Diskless_Farm_1.html
6.- Calculator for video timings (for newbies)
http://www.hut.fi/Misc/Electronics/faq/vga2rgb/calc.html
7.- Modeline tool (for power users)
http://www.dkfz-heidelberg.de/spec/linux/modeline/
8.- Streaming Multimedia Plugin for Unix Netscape
http://fredrik.hubbe.net/plugger.html
9.- 30 second guide to making Etherboot CDs
http://www.gnuenterprise.org/~jcater/ltsp/Etherboot-On-CDR-HOWTO.txt
10.- The prime monster (This site introduced me to the terminal idea!)
http://ws9.jobnegotiator.com/index.html
11.- OT - wall content article
http://www.linuxdevices.com/articles/AT3718756592.html
12.- Mailing list archives
http://www.mail-archive.com/ltsp-discuss%40lists.sourceforge.net/
http://marc.theaimsgroup.com/?l=ltsp-discuss
13.- Uk Home Linux
http://uk.homelinux.org/
14.- LTSP, Citrix and a bunch of clients
http://www.hughes.cam.ac.uk/~pd213/ltsp-instructions.html
15.- Diskless NetBSD HOW-TO
http://www.netbsd.org/Documentation/network/netboot/
16.- Knoppix makes a great GUI installer for Debian
http://www.linuxworld.com/site-stories/2002/1104.barr.html
17.- Booting Windows 98 with Ehterboot
http://www.geocrawler.com/lists/3/SourceForge/5299/0/8052616/
18.- Statistical load balancing
http://marc.theaimsgroup.com/?l=ltsp-discuss&m=104460796613713&w=4
19.- Mozilla Kiosk
http://kiosk.mozdev.org/
20.- vnc-X-session
http://www.pcxperience.org/thinclient/software/vnc-X-session-20010321.tar.gz
21.- Transparent proxy cache
http://squid.visolve.com/white_papers/trans_caching.htm
22.- How to Install Linux PPC on Power Macintosh 6100
http://www.infernosoft.com/beowoof/installingLinux.html
23.- Garrick's Little FAQs
http://speculation.org/garrick/faqs.html
24.- XF86 for VNC
http://xf4vnc.sourceforge.net/
25.- IceWM
http://www.maol.ch/icewm/FAQ/uk/IceWM-FAQ.html
http://www.icewm.org/files/en/guide/
http://themes.freshmeat.net/browse/925/?topic_id=925
http://pinsa.escomposlinux.org/sromero/linux/icewm/icewm-2-en.html
https://tsa.dyndns.org/mirror/xvadim/index.shtml
25.- Real Player for Unix
http://scopes.real.com/real/player/unix/unix.html
26.- K12 LTSP Wiki Wiki
http://k12ltsp.org/wiki/
27.- Triple head Linux, possible in LTSP
http://digitalfreaks.org/~lavalamp/apple_cinema_23inch/
28.- LTSP goodies and i-opener files
http://home.hiwaay.net/~jeffj1/ltsp/
29.- Custom Linux Progress Patch
http://www.brainlessinc.com/linux/custom_logo.html
30.- Seul/Edu Educational Application Index
http://richtech.ca/seul/
31.- Remote X Thin Client (Over the internet of another platform)
http://www.burridge.nscc.ns.ca/Intranet/Xvnc/server_installation.htm
32.- Teacher Tool
http://yhslug.tux.org/TeacherTool.py
33.- HowTo: Build your own Linux PDC Killer!
http://network.gouldacademy.org/modules.php?name=News&file=article&sid=12
34.- Icoutils, Im running a cyber cafe and needed custom icons to replicate the windowsXP look and feel with IceWM
http://www.student.lu.se/~nbi98oli/icoutils.html