.\" Copyright (c) 1994 Gordon W. Ross, Theo de Raadt .\" Updated by Luigi Rizzo, Robert Watson .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" 3. The name of the author may not be used to endorse or promote products .\" derived from this software without specific prior written permission. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR .\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES .\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. .\" IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, .\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT .\" NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, .\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY .\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT .\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF .\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. .\" .\" $FreeBSD: head/share/man/man8/diskless.8 224708 2011-08-08 13:13:22Z rmacklem $ .\" .Dd August 7, 2011 .Dt DISKLESS 8 .Os .Sh NAME .Nm diskless .Nd booting a system over the network .Sh DESCRIPTION The ability to boot a machine over the network is useful for .Em diskless or .Em dataless machines, or as a temporary measure while repairing or re-installing file systems on a local disk. This file provides a general description of the interactions between a client and its server when a client is booting over the network. .Sh OPERATION When booting a system over the network, there are three phases of interaction between client and server: .Bl -enum .It The stage-1 bootstrap, typically PXE built into your Ethernet card, loads a second-stage boot program. .It The second-stage boot program, typically .Xr pxeboot 8 , loads modules and the kernel, and boots the kernel. .It The kernel .Tn NFS mounts the root directory and continues from there. .El .Pp Each of these phases are described in further detail below. .Pp First, the stage-1 bootstrap loads the stage-2 boot program over the network. The stage-1 bootstrap typically uses .Tn BOOTP or .Tn DHCP to obtain the filename to load, then uses .Tn TFTP to load the file. This file is typically called .Pa pxeboot , and should be copied from .Pa /boot/pxeboot into the .Tn TFTP directory on the server, which is typically .Pa /tftpdir . .Pp The stage-2 boot program then loads additional modules and the kernel. These files may not exist on the .Tn DHCP or .Tn BOOTP server. You can use the .Ic next-server option available in .Tn DHCP configurations to specify the server holding the second stage boot files and kernel. The stage-2 program uses .Tn NFS or .Tn TFTP to obtain these files. By default, .Tn NFS is used. If you are using .Xr pxeboot 8 , you can install a version that uses .Tn TFTP by setting .Li LOADER_TFTP_SUPPORT=YES in your .Xr make.conf 5 , then recompiling and reinstalling .Xr pxeboot 8 via the command listed below. It is often necessary to use .Tn TFTP here so you can place a custom kernel in .Pa /tftpdir/ . If you use .Tn NFS and do not have a custom root file system for the .Nm client, the stage-2 boot will load your server's kernel as the kernel for the .Nm machine, which may not be what you want to have happen. .Bd -literal -offset indent cd /usr/src/sys/boot/i386 make clean; make; make install cp /boot/pxeboot /tftpdir/ .Ed .Pp In phase 3, the kernel acquires IP networking configuration in one of two ways, and then proceeds to mount the root file system and start operation. If the phase 2 loader supports passing network configuration to the kernel using the kernel environment, then the kernel will configure the network interface using that information. Otherwise, it must use .Tn DHCP or .Tn BOOTP to acquire configuration information. The boot scripts recognize a .Nm startup and perform the actions found in .Pa /etc/rc.d/resolv , .Pa /etc/rc.d/tmp , .Pa /etc/rc.d/var , and .Pa /etc/rc.initdiskless . .Sh CONFIGURATION In order to run a .Nm client, you need the following: .Bl -bullet .It An .Tn NFS server which exports a root and .Pa /usr partitions with appropriate permissions. The .Nm scripts work with read-only partitions, as long as root is exported with .Fl maproot Ns =0 so that some system files can be accessed. As an example, .Pa /etc/exports can contain the following lines: .Bd -literal -offset indent -ro -maproot=0 -alldirs /usr -ro -alldirs .Ed .Pp where .Aq ROOT is the mount point on the server of the root partition. The script .Pa /usr/share/examples/diskless/clone_root can be used to create a shared read-only root partition, but in many cases you may decide to export (again as read-only) the root directory used by the server itself. .It A .Tn BOOTP or .Tn DHCP server. .Xr bootpd 8 can be enabled by uncommenting the .Dq Li bootps line in .Pa /etc/inetd.conf . A sample .Pa /etc/bootptab can be the following: .Bd -literal -offset indent .default:\\ hn:ht=1:vm=rfc1048:\\ :sm=255.255.255.0:\\ :sa=:\\ :gw=:\\ :rp=":": :ha=0123456789ab:tc=.default .Ed .Pp where .Aq SERVER , .Aq GATEWAY and .Aq ROOT have the obvious meanings. .It A properly initialized root partition. The script .Pa /usr/share/examples/diskless/clone_root can help in creating it, using the server's root partition as a reference. If you are just starting out, you should simply use the server's own root directory, .Pa / , and not try to clone it. .Pp You often do not want to use the same .Pa rc.conf or .Pa rc.local files for the .Nm boot as you do on the server. The .Nm boot scripts provide a mechanism through which you can override various files in .Pa /etc (as well as other subdirectories of root). .Pp One difference that you should pay particular attention to is the value of .Va local_startup in .Pa /etc/defaults/rc.conf . A typical value for a .Nm boot is .Va mountcritremote , however your needs may be different. .Pp The scripts provide four overriding directories situated in .Pa /conf/base , .Pa /conf/default , .Pa /conf/ , and .Pa /conf/ . You should always create .Pa /conf/base/etc , which will entirely replace the server's .Pa /etc on the .Nm machine. You can clone the server's .Pa /etc here or you can create a special file which tells the .Nm boot scripts to remount the server's .Pa /etc onto .Pa /conf/base/etc . You do this by creating the file .Pa /conf/base/etc/diskless_remount containing the mount point to use as a basis of the .Nm machine's .Pa /etc . For example, the file might contain: .Pp .Dl 10.0.0.1:/etc .Pp Alternatively, if the server contains several independent roots, the file might contain: .Pp .Dl 10.0.0.1:/usr/diskless/4.7-RELEASE/etc .Pp This would work, but if you copied .Pa /usr/diskless/4.7-RELEASE to .Pa /usr/diskless/4.8-RELEASE and upgraded the installation, you would need to modify the .Pa diskless_remount files to reflect that move. To avoid that, paths in .Pa diskless_remount files beginning with .Pa / have the actual path of the client's root prepended to them so the file could instead contain: .Pp .Dl /etc .Pp The .Nm scripts create memory file systems to hold the overridden directories. Only a 2MB partition is created by default, which may not be sufficient for your purposes. To override this, you can create the file .Pa /conf/base/etc/md_size containing the size, in 512 byte sectors, of the memory disk to create for that directory. .Pp You then typically provide file-by-file overrides in the .Pa /conf/default/etc directory. At a minimum, you must provide overrides for .Pa /etc/fstab , /etc/rc.conf , and .Pa /etc/rc.local via .Pa /conf/default/etc/fstab , /conf/default/etc/rc.conf , and .Pa /conf/default/etc/rc.local . .Pp Overrides are hierarchical. You can supply network-specific defaults in the .Pa /conf/ Ns Ao Ar BROADCASTIP Ac Ns Pa /etc directory, where .Aq Ar BROADCASTIP represents the broadcast IP address of the .Nm system as given to it via .Tn BOOTP . The .Pa diskless_remount and .Pa md_size features work in any of these directories. The configuration feature works on directories other then .Pa /etc , you simply create the directory you wish to replace or override in .Pa /conf/{base,default,,}/* and work it in the same way that you work .Pa /etc . .Pp Since you normally clone the server's .Pa /etc using the .Pa /conf/base/etc/diskless_remount , you might wish to remove unneeded files from the memory file system. For example, if the server has a firewall but you do not, you might wish to remove .Pa /etc/ipfw.conf . You can do this by creating a .Pa /conf/base/ Ns Ao Ar DIRECTORY Ac Ns Pa .remove file. For example, .Pa /conf/base/etc.remove , which contains a list of relative paths that the boot scripts should remove from the memory file systems. .Pp As a minimum, you normally need to have the following in .Pa /conf/default/etc/fstab .Bd -literal -offset indent : / nfs ro 0 0 :/usr /usr nfs ro 0 0 .Ed .Pp You also need to create a customized version of .Pa /conf/default/etc/rc.conf which should contain the startup options for the .Nm client, and .Pa /conf/default/etc/rc.local which could be empty but prevents the server's own .Pa /etc/rc.local from leaking onto the .Nm system. .Pp In .Pa rc.conf , most likely you will not need to set .Va hostname and .Va ifconfig_* because these will be already set by the startup code. Finally, it might be convenient to use a .Ic case statement using .Li `hostname` as the switch variable to do machine-specific configuration in case a number of .Nm clients share the same configuration files. .It The kernel for the .Nm clients, which will be loaded using .Tn NFS or .Tn TFTP , must include support for the NFS client: .Pp .D1 Cd "options NFSCL" .D1 Cd "options NFS_ROOT" .Pp If you are using a boot mechanism that does not pass network configuration to the kernel using the kernel environment, you will also need to include the following options: .Pp .D1 Cd "options BOOTP" .D1 Cd "options BOOTP_NFSROOT" .D1 Cd "options BOOTP_COMPAT" .Pp .Em Note : the PXE environment does not require these options. .Pp The .Nm booting environment relies on memory-backed file systems to support temporary local storage in the event that the root file system is mounted read-only; as such, it is necessary to add the following to the device section of the kernel configuration: .Pp .D1 Cd "device md" .Pp If you use the firewall, remember to default to .Dq open , or your kernel will not be able to send/receive the .Tn BOOTP packets. .El .Sh SECURITY ISSUES Be warned that using unencrypted .Tn NFS to mount root and user partitions may expose information such as encryption keys. .Sh SEE ALSO .Xr ethers 5 , .Xr exports 5 , .Xr make.conf 5 , .Xr bootpd 8 , .Xr mountd 8 , .Xr nfsd 8 , .Xr pxeboot 8 , .Xr reboot 8 , .Xr tftpd 8 .Pp .Pa ports/net/etherboot .Sh BUGS This manpage is probably incomplete. .Pp .Fx sometimes requires to write onto the root partition, so the startup scripts mount MFS file systems on some locations (e.g.\& .Pa /etc and .Pa /var ) , while trying to preserve the original content. The process might not handle all cases.