BOOTP configuration mechanism Matthew Dillon dillon@backplane.com BOOTP kernels automatically configure the machine's IP address, netmask, optional NFS based swap, and NFS based root mount. The NFS server will typically export a shared read-only /, /usr, and /var to any number of workstations. The shared read-only root is typically either the server's own root or, if you are more security concious, a contrived root. The key issue with starting up a BOOTP kernel is that you typically want to export read-only NFS partitions from the server, yet still be able to customize each workstation ( or not ). /etc/rc.diskless1 is responsible for doing core mounts and for retargeting /conf/ME ( part of the read-only root NFS mount ) to /conf/$IP_OF_CLIENT. /etc/rc.conf.local and /etc/rc.local, along with other machine-specific configuration files, are typically softlinks to /conf/ME/. In the BOOTP workstation /conf/$IP/rc.conf.local, you must typically turn *OFF* most of the system option defaults in /etc/rc.conf as well as do additional custom configuration of your environment The /usr/src/share/examples/diskless directory contains a typical X session / sshd based workstation configuration. The directories involved are HT.DISKLESS/ and 192.157.86.12/. Essentially, the $IP/ directory ( which rc.diskless looks for in /conf/$IP/ ) contains all the junk. The HT.DISKLESS directory exists to hold common elements of your custom configuration so you do not have to repeat those elements for each workstation. The example /conf structure included here shows how to create a working sshd setup ( so you can sshd into the diskless workstation ), retarget xdm's pid and error files to R+W directories if /usr is mounted read-only, and retarget syslogd and other programs. This example is not designed to run out of the box and some modifications are required. >> NOTE << HT.DISKLESS/ttys contains the typical configuration required to bring X up at boot time. Essentially, it runs xdm in the foreground with the appropriate arguments rather then a getty on ttyv0. You must run xdm on ttyv0 in order to prevent xdm racing with getty on a virtual terminal. Such a race can cause your keyboard to be directed away from the X session, essentially making the session unusable. Typically you should start with a clean slate by tar-copying this example directory to /conf and then hack on it in /conf rather then in /usr/share/examples/diskless. BOOTP CLIENT SETUP Here is a typical kernel configuration. If you have only one ethernet interface you do not need to wire BOOTP to a specific interface name. BOOTP requires NFS and NFS_ROOT, and our boot scripts require MFS. If your /tmp is *not* a softlink to /var/tmp, the scripts also require NULLFS # BootP # options BOOTP # Use BOOTP to obtain IP address/hostname options BOOTP_NFSROOT # NFS mount root filesystem using BOOTP info options "BOOTP_NFSV3" # Use NFS v3 to NFS mount rootoptions options BOOTP_COMPAT # Workaround for broken bootp daemons. #options "BOOTP_WIRED_TO=de0" options MFS # Memory File System options NFS # Network Filesystem options NFS_ROOT # Nfs can be root options NULLFS # nullfs to map /var/tmp to /tmp BOOTP SERVER SETUP The BOOTP server must be running on the same logical LAN as the the BOOTP client(s). You need to setup two things: (1) You need to NFS-export /, /usr, and /var. (2) You need to run a BOOTP server. DHCPD can do this. NFS Export: Here is an example "/etc/exports" file. / -ro -maproot=root: -network 192.157.86.0 -mask 255.255.255.192 /usr -ro -maproot=root: -network 192.157.86.0 -mask 255.255.255.192 /var -ro -maproot=root: -network 192.157.86.0 -mask 255.255.255.192 In order to be an NFS server, the server must run portmap, mountd, nfsd, and rpc.statd. The standard NFS server options in /etc/rc.conf will work ( you should put your overrides in /etc/rc.conf.local on the server and not edit the distribution /etc/rc.conf, though ). BOOTP Server: This configuration file "/etc/dhcpd.conf" example is for the '/usr/ports/net/isc-dhcp' dhcpd port. subnet 192.157.86.0 netmask 255.255.255.192 { # range if you want to run the core dhcpd service of # dynamic IP assignment, but it is not used with BOOTP # workstations range 192.157.86.32 192.157.86.62; # misc configuration. # option routers 192.157.86.2; option domain-name-servers 192.157.86.2; server-name "apollo.fubar.com"; option subnet-mask 255.255.255.192; option domain-name-servers 192.157.86.2; option domain-name "fubar.com"; option broadcast-address 192.157.86.63; option routers 192.157.86.2; } host test1 { hardware ethernet 00:a0:c9:d3:38:25; fixed-address 192.157.86.11; option root-path "192.157.86.2:/"; option option-128 "192.157.86.2:/images/swap"; } host test2 { # hardware ethernet 00:e0:29:1d:16:09; hardware ethernet 00:10:5a:a8:94:0e; fixed-address 192.157.86.12; option root-path "192.157.86.2:/"; option option-128 "192.157.86.2:/images/swap"; } SWAP. This example includes options to automatically BOOTP configure NFS swap on each workstation. In order to use this capabilities you need to NFS-export a swap directory READ+WRITE to the workstations. You must then create a swap directory for each workstation you wish to assign swap to. In this example I created a dummy user 'lander' and did an NFS export of /images/swap enforcing a UID of 'lander' for all accesses. apollo:/usr/ports/net# ls -la /images/swap total 491786 drwxr-xr-x 2 root wheel 512 Dec 28 07:00 . drwxr-xr-x 8 root wheel 512 Jan 20 10:54 .. -rw-r--r-- 1 lander wheel 33554432 Dec 23 14:35 swap.192.157.86.11 -rw-r--r-- 1 lander wheel 335544320 Jan 24 16:55 swap.192.157.86.12 -rw-r--r-- 1 lander wheel 134217728 Jan 21 17:19 swap.192.157.86.6 A swap file is best created with dd: # create a 32MB swap file for a BOOTP workstation dd if=/dev/zero of=swap.IPADDRESS bs=1m count=32 It is generally a good idea to give your workstations some swap space, but not a requirement if they have a lot of memory.