The fundamental idea is as follows: the PC will get its IP address from the boot server via the bootp protocol, using 0.0.0.0 as the initial IP address and its kernel via the tftp protocol. (-- Booting across segments (via router) not a simple question, so either put both the server and the diskless boxes on the same lan segment or configure an UDP helper address in your router to the address of the server. Refer to your router product manual for further info.--)
For this follow the steps below.
Get the nfsboot package (the package is available from your favourite linux mirror site in the /pub/Linux/system/Linux-boot directory). It contains a booteprom image for the wd8013 card which can be directly burned in.
There are alternative ways to prepare the PC:
These images contain a bootp and tftp client. You need to prepare a linux kernel too, which contains the nfs-root option.
It can be found in package bootpd-2.4.tar.gz (which can be found on your favourite linux mirror site in the /pub/Linux/system/Network/boot.net directory). Get the package, compile and install it. If your other Unix box happens to be a Slackware Linux then you may skip this step for the standard distributions contain a bootpd. The daemon can be run either directly by issuing command
______________________________________________________________________
bootpd -s
______________________________________________________________________
· /etc/inetd.conf to remove the hashmark from the start of these
lines:
______________________________________________________________________
# tftp dgram udp wait root /usr/sbin/in.tftpd tftpd /export
# bootps dgram udp wait root /usr/sbin/in.bootpd bootpd
______________________________________________________________________
· insert or uncomment the following two lines in /etc/services:
______________________________________________________________________
bootps 67/tcp # BOOTP server
tftp 69/udp # TFTP server
______________________________________________________________________
· restart inetd by
______________________________________________________________________
kill -HUP <process id of inetd>.
______________________________________________________________________
First of all, bootpd have a config file called bootptab which usually resides in /etc. You must modify it by inserting the IP addresses of your gateway, dns server, and the ethernet address(es) of your diskless machine(s). An example /etc/bootptab:
______________________________________________________________________
global.prof:\
:sm=255.255.255.0:\
:ds=192.168.1.5:\
:gw=192.168.1.19:\
:ht=ethernet:\
:bf=linux:
machine1:hd=/export/root/machine1:tc=global.prof:ha=0000c0863d7a:ip=192.168.1.140:
machine2:hd=/export/root/machine2:tc=global.prof:ha=0800110244e1:ip=192.168.1.141:
machine3:hd=/export/root/machine3:tc=global.prof:ha=0800110244de:ip=192.168.1.142:
______________________________________________________________________
global.prof is a general template for host entries, where
After this, every machine must have a line:
TFTP (Trivial File Transfer Protocol) is a file transfer protocol, such as ftp, but it's much simpler to help coding it in EPROMs. TFTP can be used in two ways:
Almost every Unix implementation contains tfpt server, probably you don't need to install your own one.
Install tftpd, make sure it's active in /etc/inetd.conf, typical line
tftp dgram udp wait root /usr/sbin/tcpd in.tftpd /tftpboot
You must compile a kernel for the DC that includes NFS support and NIC driver compiled in (not modules). Answer yes to Root file system on NFS? and BOOTP support?
After building the kernel, run mknbi-linux from the Etherboot distribution on it. Install this tagged image as /tftpboot/(bf attribute in bootptab).
This may contain packages a, ap, n, and x of the Slackware distribution. To install more is OK; however the above packages suffice for the purposes of a diskless X terminal. For the installation you need a working Linux system. Find some disk space on the remote machine and export it read-write. Mount the exported directory onto somewhere (e.g. /mnt) on the file system of the Linux box. Start Linux setup and change the root option in the setup from / to /mnt. Then setup the above packages as usual. If you want to run no more than one diskless Linux then no changes are needed. On the other hand, if you plan to use more than one diskless machine then the above setup will not work because some files and directories must be private to the machines. The problem can be bypassed by moving the /usr (it contains no private data) and then create a separate subdir for each diskless machine. For example, if /export/linux/machine1 were mounted to /mnt then the directory structure after the initial setup will look like
______________________________________________________________________
/export/linux/machine1/bin
/export/linux/machine1/sbin
/export/linux/machine1/lib
/export/linux/machine1/etc
/export/linux/machine1/var
/export/linux/machine1/usr
______________________________________________________________________
After the changes you will have
______________________________________________________________________
/export/linux/machine1/bin
/export/linux/machine1/sbin
/export/linux/machine1/lib
/export/linux/machine1/etc
/export/linux/machine1/var
/export/linux/usr
______________________________________________________________________
Now create the subdirectories for the other machines. Assume for now that your diskless machines are called machine1, machine2, machine3, etc.; then you may use the following bash script to setup the other directories
______________________________________________________________________
cd /export/linux
for x in machine2 machine3 ; do
mkdir $x; cd $x
(cd ../machine1; tar cf - *) | tar xvf -
done
______________________________________________________________________
Then do the following export:
______________________________________________________________________
# This file is /etc/export
# for remote linux X terminals by Buci
# this line is only once
/export/root/usr -access=linuxnet
# these lines once for every host
/export/root/machine1 rw=machine1,root=machine1
/export/root/machine2 rw=machine2,root=machine2
/export/root/machine3 rw=machine3,root=machine3
______________________________________________________________________
Now it is time to configure the tftp server. If you do not need secure tftp then everything is quite simple for your clients can be booted from the /export directory.
If a secure tftp is used then you can either make a full /export/linux directory structure under /tftpboot (with a single real kernel and symbolic links for the other machines), or let the /export directory be the boot directory of the secure tftpd. Or, if you have a separate tftpboot directory then, similarly, you need only the original directory structure with a single kernel and symbolic links for the others. You can achieve this setup by typing the following:
______________________________________________________________________
mkdir -p /tftpboot/export/linux/machine1
cd /tftpboot/export/linux/machine1
cp /export/linux/machine1/<name of the kernel> .
______________________________________________________________________
Then type the following:
______________________________________________________________________
mkdir -p /tftpboot/export/linux/machine2
cd ../machine2
ln -s ../machine2/<name of the kernel>
______________________________________________________________________
Finally, you must insert
______________________________________________________________________
/sbin/mount nfs_server:/export/linux/usr /usr
______________________________________________________________________
as the first line of
______________________________________________________________________
/export/linux/<machinex>/etc/rc.d/rc.S
______________________________________________________________________
where <machinex> stands for machine1, machine2, etc.
A small program that runs as a BIOS extension, usually on an EPROM on the NIC. It handles the BOOTP query and TFTP loading and then transfers control to the loaded image.
It uses TCP/IP protocols but the loaded image doesn't have to be Linux. The loaded image can be anything, even DOG.
There are two free implementations of TCP/IP net loaders: Etherboot and Netboot : Etherboot uses built-in drivers while Netboot uses Packet drivers.
They can also be loaded from a floppy for testing and for temporary setups.
The DC requests to mount /tftpboot/(IP address of DC) (in 2.1 and above: /tftpboot/(name of DC in bootptab) ) as its / by NFS from server. You must export this from the server (rw, no_root_squash) because the DC wants to write on it (log files, etc).
The / must contain /sbin, /bin, /lib, /etc, /var, /tmp, /root, /dev and /proc.
/sbin, /bin, /lib/ can be a copy of an existing RH5 system. They can be shared between all DCs. But hard links only. BTW, don't link to server originals.
/etc, /var and /dev should be non-sharable copies. Customise /etc/sysconfig/network, /etc/sysconfig/network-scripts/ifcfg-eth0, /etc/fstab, /etc/conf.modules, and others. Turn off all network services you don't need. Remove all stuff you don't need from /var, e.g. RPM db, lpd files.
/root and /proc should just exist. /tmp should exist and be mode 1777.
You probably want to create /usr and /home mount points. /usr can be mounted ro.
About 10 MB per DC plus about 15 MB of shared files should be sufficient. BTW: if your DCs are quite similar, the kernel image can also be shared.
Here is an illustrative script to create the first root filesystem.
#!/bin/sh
if [ $# != 1 ]
then
echo Usage: $0 client-IP-addr
exit 1
fi
cd /
umask 022
mkdir -p /tftpboot/$1
# just make these ones
for d in home mnt proc tmp usr
do
mkdir /tftpboot/$1/$d
done
chmod 1777 /tftpboot/$1/tmp
touch /tftpboot/$1/fastboot
chattr +i /tftpboot/$1/fastboot
# copy these ones
cp -a bin lib sbin dev etc root var /tftpboot/$1
cat <<EOF
Now, in /tftpboot/$1/etc, edit
sysconfig/network
sysconfig/network-scripts/ifcfg-eth0
fstab
conf.modules
and configure
rc.d/rc3.d
EOF
Here is an illustrative script to duplicate the root filesystem
#!/bin/sh
if [ $# != 2 ]
then
echo Usage: $0 olddir newdir
exit 1
fi
cd /tftpboot
if [ ! -d $1 ]
then
echo $1 is not a directory
exit 1
fi
umask 022
mkdir -p $2
# just make these ones
for d in home mnt proc tmp usr
do
mkdir $2/$d
done
chmod 1777 $2/tmp
touch $2/fastboot
chattr +i $2/fastboot
# link these ones
for d in bin lib sbin
do
(cd $1; find $d -print | cpio -pl ../$2)
done
# copy these ones
for d in dev etc root var
do
cp -a $1/$d $2
done
cat <<EOF
Now, in /tftpboot/$2/etc, edit
sysconfig/network
sysconfig/network-scripts/ifcfg-eth0
fstab (maybe)
conf.modules (maybe)
and configure
rc.d/rc3.d
EOF
RH5 wants to fsck the root FS. I stopped this with a /fastboot. But init script wants to delete it, so I did chattr +i /fastboot
/etc/localtime is a link to TZ file in /usr/share/... I made it a copy.
Turn off /etc/rc.d/rc6.d/K97network or it will disable the network before root FS is done with.
X server wants to write into /usr/X11R6/lib/X11/xkb/compiled. I made this a link to /etc/X11/kbd/compiled
Remember your DC will keep appending to log files so have logrotate or something deal with them at regular intervals.
On the server, make sure the DC is matched by a clause in /etc/X11/xdm/Xaccess and comment out the :0 in /etc/X11/xdm/Xservers. Then make sure that xdm is run from the init scripts.
On the client, run X -query server
You will get the xdm login box and then all your X clients will run on the server.
For other applications use - you could use diskless technique for netboot routers, print servers (but should not be spooling print server), standalone apps, etc.