Recent linux kernels offer the possibility to boot a linux box entirely from network, by loading its kernel and root filesystem from a server. In that case, the client may use several ways to get the first instructions it has to execute when booting: home made eproms, special network cards implementing the RARP, BOOTP or DHCP protocols, cdroms, or bootloaders loaded from a boot floppy or a local hard drive.
Logilab sponsored this HOWTO. Check their website for new versions of this document. I also thank the etherboot, netboot, plume and linux terminal server project developers and webmasters, who made it really possible to boot a Linux worstation over a network.
Very special thanks go to Ken Yap, member of the etherboot project, whose comments greatly helped to improve the quality of this document.
I also thank Jerome Warnier, main developer of the plume project, Pierre Mondié, Kyle Bateman, Peter T. Breuer, Charles Howes, and Thomas Marteau for their comments and contributions.
Sometimes, buying a diskless linux computer will be cheaper than building! Checkout the list of commercial sites given in appendix, which are selling diskless linux network cards and diskless computers. These companies do mass production of linux diskless computers selling millions of units and thereby reducing the cost per unit.
Diskless computers will become more and more popular in the next years. They will be very successful because of the availability of very high-speed network cards at very low prices. Today 100 Megabit per second (12.5 MB per sec transfer rate) network cards are common and in about 1 to 2 years 1000 MBit (125 MB per sec transfer rate) network cards will become very cheap and will be the standard.
In near future, monitor manufacturers will place the CPU, NIC, RAM right inside the monitor to form a diskless computer. This eliminates the diskless computer box and saves space. The monitor will have outlet for mouse, keyboard, network RJ45 and power supply.
The following are benefits of using diskless computers:
Total cost of ownership is very low in case of diskless computers. Total cost of ownership is cost of initial purchasing + cost of maintenance. The cost of maintenance is usually 3 to 5 times the cost of initial computer purchase and this cost is recurring year after year. In case of diskless computers, the cost of maintenance is completely eliminated.
All the backups are centralized at one single main server.
No need of UPS battery, air-conditioning, dust proof environment for diskless clients, only server needs UPS battery, A/C and dust proof environment.
A better protection from virus attack - Some computer virus cannot attack diskless computers as they do not have any hard disk. This kind of virus cannot do any damage to diskless computers. Only one single server box needs to be protected against virus attack. This saves millions of dollars for the company by avoiding installation of vaccines and cleaning the hard disks.
Servers can have large powerful/high performance hard disks, can optimize the usage of disk space via sharing by many diskless computer users. Fault tolerance of hard disk failure is possible by using RAID on main server.
On some installations: sharing of central server RAM memory by many diskless computer users. For example, if many users are running a web browser remotely on a server, then there will be only one copy of this web browser in its RAM.
Very few system administrators required to maintain central server.
Zero administration at diskless client side. Diskless computers are absolutely maintenance free and troublefree.
Long life of diskless clients.
Eliminates install/upgrade of hardware, software on diskless client side.
Eliminates cost of cdrom, floppy, tape drive, modem, UPS battery, printer parallel ports, serial ports etc...
Can operate in places like factory floor where a hard disk might be too fragile.
The most important thing in order to boot from network is to have an equipment which enables the stations to execute a bootloader, which will get the kernel on the server and launch it. Another solution is to use a device which will load a local kernel, which will mount the root filesystem on the server. There are several solutions: home made eproms containing the first instructions to execute when booting the station, boot with BOOTP/DHCP capable network adapters, or a local floppy, a tiny hard drive, or a cdrom to load the kernel. Note that some vendors also sell network booting capable stations: for instance, some Sun stations implement the BOOTP protocol.
Other hardware requirements depend on the configuration you plan to use: on some sites, every application run by the stations is executed remotely on the server, this implies that a very high-performance server is required, but only light stations are required: depending on what they will have to do, 80486 CPUs with 16 MB of RAM may be enough. On the other side, if application programs are really executed locally on the stations, the requirements for the stations depend completely on these applications. In that case, only a small server is required. A 80486 CPU with 32 MB of RAM will be sufficient for a small number of stations, but more memory will be necessary in very large installations with hundreds or thousands of machines. Note the server's CPU does not really matter for such an installation.
Linux kernel version 2.0 or above sources are required. All tools required to build a linux kernel are also necessary (see the linux kernel documentation for more information on this).
A BOOTP daemon (a DHCP daemon may also do fine, but I won't explain how to configure this), a NFS daemon (if you want to mount the root filesystem on a remote server), are also required. We will also need a TFTP daemon if you plan to load the kernel remotely. At last, the mknbi utility provided with the etherboot distribution, and, if you use LanWorks EPROMs, like those included in the 3c905 3com ethernet adapter, you will also need the imggen utility, available at http://www.ltsp.org/contrib/.
This documentation has been written for experimented system administrators, who are already aware of linux fundamentals, like the use of grep, sed, and awk, basic shell programming, the init process and the boot scripts, kernel compilation, and NFS server configuration. Experience of kernel arguments passing should also help. Information on these subjects can be found respectively in the grep, sed, awk, and bash man/info pages, in the Bootdisk-HOWTO, the From-PowerUp-To-Bash-Prompt-HOWTO, the Kernel-HOWTO, the BootPrompt-HOWTO, the bootparam man page, the rdev man page, the NFS-HOWTO, and the exports manual page.
There are many sources of information on network booting, but, and this is why I wrote this HOWTO, none describes all the existing ways of booting over a network, and much of them are specific to a way of operating. The most useful to me has been the documentation provided by the linux terminal server project, although I did not use the packages they recommend, and I have chosen to describe here how to proceed without these packages, because they configure things so that every application program is executed remotely on a server. Useful information can also be found on the etherboot project's homepage.
At last, you can also find useful but succinct information in the kernel's source tree, in /usr/src/linux/Documentation, assuming your kernel source tree resides in /usr/src/linux.