dnl $OpenBSD: install,v 1.27 2002/10/03 00:19:49 miod Exp $ OpenBSDInstallPrelude There are several ways to install OpenBSD onto a disk. The easiest way in terms of preliminary setup is to use the OpenBSD installation dnl XXX change back if more than one floppy again. dnl cdrom, or an installation floppy. cdrom, or the installation floppy. If your machine is hooked up in a network and you can find a server to arrange for a diskless setup, which is a convenient way to install on a machine whose disk does not currently hold a usable operating system. This is difficult to get set up correctly the first time, but easy to use afterwards. (see ``Installing using a diskless setup'' below). It is also possible to install OpenBSD "manually" from a running Solaris system, using the system tools, as well as gunzip; see ``Installing from Solaris'' below. Booting from the Installation Media: Prior to attempting an installation, you should make sure that everything of value on the target system has been backed up. While installing OpenBSD does not necessarily wipe out all the partitions on the hard disk, errors during the install process can have unforeseen consequences and you will probably render the system unbootable if you start, but do not complete the installation. Having the installation media for the prior installation, be it a Solaris or OpenBSD CD-ROM or OpenBSD install diskettes is good insurance if you want to be able to "go back" for some reason. After taking care of all that, bring your system down gracefully using the shutdown(8) and/or halt(8) commands. This will get you to the monitor prompt. Booting from Floppy Disk installation media: ok boot floppy bsd This will cause the kernel contained in the floppy to be booted. Please note that PCI systems are not able to boot from floppy; also, Ultra 1, 1E and 2 systems might need a firmware update to be able to boot from floppy; refer to the ``Updating your firmware'' section earlier in this document for details. Booting From CD-ROM installation media: dnl No args! not a typo ok boot cdrom If the boot is successful, you will get a loader version message, executable sizes and then the Kernel copyright and device probe messages. Boot failure modes are typically a lot of CD-ROM drive activity, but no messages or complaints about magic numbers, checksums or formats. Booting from disk: Boot the miniroot by typing the appropriate command at the PROM: ok boot disk:b bsd If you've loaded the miniroot onto some other disk than the default drive 0, modify the boot specifier accordingly, keeping in mind the partition naming a=0, b=1... ok boot disk1:b bsd # example - scsi target 1 or # second ide drive During the boot, the kernel might not be able to determine by itself which device it has been booted from. In this case, it will ask you for the root and swap devices. Answer the device and partition you booted from as the root partition (likely sd0b or wd0b); you might want to rely on the boot messages to find the correct device name (see below). When the kernel asks your for the swap device, just validate the suggested device name; since the installation procedure does not use swap, it doesn't matter. Installing using a diskless setup: First, you must setup a diskless client configuration on a server. If you are using a OpenBSD system as the boot-server, have a look at the diskless(8) manual page for guidelines on how to proceed with this. If the server runs another operating system, you'll have to consult documentation that came with it (on SunOS systems, add_client(8) and the Sun System/Networks administrators guide constitute a good start; on Solaris systems, share(1M) is a good starting point as well). Boot your workstation from the server by entering the appropriate `boot' command at the monitor prompt: ok boot net bsd.rd Installing using the Floppy, CD-ROM, miniroot or netboot procedure: OpenBSDInstallPart2 Boot your machine from the installation media as described above. It will take a while to load the kernel especially from a floppy or slow network connection, most likely more than a minute. If some action doesn't eventually happen, or the spinning cursor has stopped and nothing further has happened, either your boot media is bad, your diskless setup isn't correct, or you may have a hardware or configuration problem. OpenBSDBootMsgs While booting, you will probably see several warnings. You may be warned that the kernel can't figure out what device it booted from. Do not be alarmed, this is completely normal. This warning occurs because while OpenBSD/MACHINE can boot from the floppy drive, the kernel itself lacks a floppy driver. On some machines, you might also get this kind of warning: "No counter-timer -- using %tick at 179MHz as system clock." Unless the frequency displayed does not match your CPU speed, this is nothing to worry about. You will next be asked for your terminal type. If you are installing from a keyboard/monitor console, the default of "sun" if correct. If you are installing from a serial console you should choose the terminal type from amongst those listed. (If your terminal type is xterm, just use vt100). OpenBSDInstallPart3({:- or "wd0" for IDE drives-:}) OpenBSDInstallPart4 OpenBSDInstallPart5(sd0) OpenBSDInstallNet({:-CD-ROM, NFS, -:},nofloppy) OpenBSDFTPInstall OpenBSDHTTPInstall OpenBSDTAPEInstall OpenBSDCDROMInstall OpenBSDNFSInstall OpenBSDDISKInstall({:-"wdN" or -:},{:-only -:}) OpenBSDCommonFS OpenBSDCommonURL After completing an installation: Now try a reboot. (If needed, swap your scsi id's first). The UltraSPARC OpenFirmware will normally load the kernel from the device and filename as instructed by the ``boot-device'' and ``boot-file'' variables. If the ``boot-file'' variable is empty, the OpenBSD bootloader will look for a kernel named ``bsd'', unless a different filename has been specified in the boot command. To reset this variable to its default, empty, value, type the following: ok set-default boot-file OpenBSDCongratulations If you will be running your OpenBSD system from a serial console, you may need to edit /etc/ttys and change the terminal type, and getty method from "sun" and "suncons" to "vt100" and "std.9600" or something similar. Also when running from a serial console, you may wish to adjust the eeprom settings for input-device, output-device, screen-#columns, and screen-#rows as appropriate. In order to use 'tip' on OpenBSD/MACHINE, you'll need to edit /etc/ttys and add "local" to the end of the tty configuration line, and run 'ttyflags -a' to put your changes into effect. Installing from Solaris: You need a machine running under Solaris to install OpenBSD. You will also need at least the following pieces: - the *.tgz files you want to install (as a minimum, base{:--:}OSrev.tgz and etc{:--:}OSrev.tgz) - gunzip (GNU gzip) Solaris binary - a boot block file from a Solaris machine that matches your machine type, for ufs filesystem, such as /usr/platform/`uname -i`/lib/fs/ufs/bootblk - an OpenBSD kernel All these pieces, except the boot block and the GNU utilities are supplied in the OpenBSD/MACHINE distribution. You need to format and partition the disk using Solaris (since OpenBSD/MACHINE uses Sun compatible disk labels.) Give yourself adequate partition sizes. Here is an example layout: solaris# prtvtoc -s /dev/rdsk/c0t1d0s2 * First Sector Last * Partition Tag Flags Sector Count Sector Mount Directory 0 2 00 0 80000 79999 / 1 3 01 80000 256000 335999 2 5 00 0 4165271 4165270 3 7 00 336000 100000 435999 /tmp 4 7 00 436000 100000 535999 /var 5 7 00 536000 400000 935999 /var/tmp 6 4 00 936000 3229271 4165270 /usr Use Solaris to newfs the partitions which will have filesystems on them. (OpenBSD's filesystem format is almost identical to Solaris). solaris# newfs /dev/rdsk/c0t1d0s0 [... lots of output] Repeat for any other partition (in this example, /dev/rdsk/c0t1d0s3, /dev/rdsk/c0t1d0s4, /dev/rdsk/c0t1d0s5 and /dev/rdsk/c0t1d0s6). dnl XXX I had no time to check the -O restriction is still necessary. NOTE: If you are able to, there is a performance benefit from newfs'ing using OpenBSD. If you newfs using the OpenBSD newfs command, be sure to use the -O flag for your / partition, so that newfs will use the 4.3BSD filesystem format, rather than the new 4.4BSD filesystem format. If you forget, you will not be able to boot -- the Solaris boot blocks do not understand the extended 4.4BSD filesystem format. Mount those partitions in a tree formation, under /mnt; ie: solaris# df -k Filesystem kbytes used avail capacity Mounted on [...] /dev/dsk/c0t1d0s0 38427 0 38427 0% /mnt /dev/dsk/c0t1d0s3 48249 0 48249 0% /mnt/tmp /dev/dsk/c0t1d0s4 48249 0 48249 0% /mnt/var /dev/dsk/c0t1d0s5 193536 0 193536 0% /mnt/var/tmp /dev/dsk/c0t1d0s6 1564024 0 1564024 0% /mnt/usr Place the boot block in /mnt (your new root partition), and use the Solaris command "installboot" to make it work. The installboot man page says to do something like this: solaris# cp /usr/platform/`uname -i`/lib/fs/ufs/bootblk /mnt/bootblk solaris# sync; sync solaris# /usr/sbin/installboot /mnt/bootblk /dev/rdsk/c0t1d0s0 You can now extract the provided "*.tgz files onto your disk. solaris# ls -FC base{:--:}OSrev.tgz etc{:--:}OSrev.tgz misc{:--:}OSrev.tgz xserv{:--:}OSrev.tgz bsd game{:--:}OSrev.tgz xbase{:--:}OSrev.tgz xshare{:--:}OSrev.tgz comp{:--:}OSrev.tgz man{:--:}OSrev.tgz xfont{:--:}OSrev.tgz solaris{:-#-:} gunzip < base{:--:}OSrev.tgz | (cd /mnt; tar xvpf -) [...] for each set And finally copy the OpenBSD kernel onto your disk. solaris# cp bsd /mnt/bsd The GNU gunzip program is not distributed as part of Solaris, but may be present in your /usr/local/bin. If not, you will need to obtain it from a GNU archive and install before proceeding. After the files have been extracted, setup /mnt/etc/fstab to match your actual disk layout. (Minus the "/mnt" component of each path, of course :-) Now proceed to reboot the machine and the customize your installation. Net Boot or Diskless Setup Information: The set up is similar to the diskless setup, but not identical, because the Sun setup assumes that the bootblocks load a kernel image, which then uses NFS to access the exported root partition, while the OpenBSD bootblocks use internal NFS routines to load the kernel image directly from the exported root partition. Please understand that no one gets this right the first try, since there is a lot of setup and all the host daemons must be running and configured correctly. If you have problems, extract the diskless(8) manpage, find someone who's been through it before and use the host syslog and tcpdump(8) to get visibility of what's happening (or not). Your UltraSPARC expects to be able to download a second stage bootstrap program via TFTP after having acquired its IP address through RevARP when instructed to boot "over the net". It will look for a filename composed of the machine's IP address. For example, a machine which has been assigned IP address 130.115.144.11, will make an TFTP request for `8273900B'. Normally, this file is a symbolic link to an appropriate second-stage boot program, which should be located in a place where the TFTP daemon can find it (remember, many TFTP daemons run in a chroot'ed environment). You can find the boot program in `/usr/mdec/boot' in the OpenBSD/MACHINE distribution. After the boot program has been loaded into memory and given control by the PROM, it starts locating the machine's remote root directory through the BOOTPARAM protocol. First a BOOTPARAM WHOAMI request is broadcast on the local net. The answer to this request (if it comes in) contains the client's name. This name is used in next step, a BOOTPARAM GETFILE request -- sent to the server that responded to the WHOAMI request -- requesting the name and address of the machine that will serve the client's root directory, as well as the path of the client's root on that server. Finally, this information (if it comes in) is used to issue a REMOTE MOUNT request to the client's root filesystem server, asking for an NFS file handle corresponding to the root filesystem. If successful, the boot program starts reading from the remote root filesystem in search of the kernel which is then read into memory. You will want export the miniroot{:--:}OSrev.fs filesystem to the client. You can dd this filesystem image to some spare partition, mount and export that partition or use tar to copy the contents to a more convenient spot. Alternatively you can build a bootable partition from the distribution sets as follows: Unpack `base{:--:}OSrev.tgz' and `etc{:--:}OSrev.tgz' on the server in the root directory for your target machine. If you elect to use a separately NFS-mounted filesystem for `/usr' with your diskless setup, make sure the "./usr" base files in base{:--:}OSrev.tgz end up in the correct location. One way to do this is to temporarily use a loopback mount on the server, re-routing /usr to your server's exported OpenBSD "/usr" directory. Also put the kernel and the install/upgrade scripts into the root directory. A few configuration files need to be edited: /etc/hosts Add the IP addresses of both server and client. /etc/myname This files contains the client's hostname; use the same name as in /etc/hosts. /etc/fstab Enter the entries for the remotely mounted filesystems. For example: server:/export/root/client / nfs rw 0 0 server:/export/exec/MACHINE.OpenBSD /usr nfs rw 0 0 Now you must populate the the `/dev' directory for your client. If you server runs SunOS 4.x, you can simply change your working directory to `/dev' and run the MAKEDEV script: `sh MAKEDEV all' (this might require the edition of MAKEDEV to change the PATH for it to work properly). On Solaris (SunOS 5.x) systems, MAKEDEV can also be used, but there'll be error messages about unknown user and groups. These errors are inconsequential for the purpose of installing OpenBSD. However, you may want to correct them if you plan to the diskless setup regularly. In that case, you may re-run MAKEDEV on your OpenBSD machine once it has booted.