INSTALLATION NOTES for OpenBSD/armish 5.8 What is OpenBSD? ---------------- OpenBSD is a fully functional, multi-platform UN*X-like Operating System based on Berkeley Networking Release 2 (Net/2) and 4.4BSD-Lite. There are several operating systems in this family, but OpenBSD differentiates itself by putting security and correctness first. The OpenBSD team strives to achieve what is called a 'secure by default' status. This means that an OpenBSD user should feel safe that their newly installed machine will not be compromised. This 'secure by default' goal is achieved by taking a proactive stance on security. Since security flaws are essentially mistakes in design or implement- ation, the OpenBSD team puts as much importance on finding and fixing existing design flaws and implementation bugs as it does writing new code. This means that an OpenBSD system will not only be more secure, but it will be more stable. The source code for all critical system components has been checked for remote-access, local-access, denial- of-service, data destruction, and information-gathering problems. In addition to bug fixing, OpenBSD has integrated strong cryptography into the base system. A fully functional IPsec implementation is provided as well as support for common protocols such as SSL and SSH. Network filtering and monitoring tools such as packet filtering, NAT, and bridging are also standard, as well as several routing services, such as BGP and OSPF. For high performance demands, support for hardware cryptography has also been added to the base system. Because security is often seen as a tradeoff with usability, OpenBSD provides as many security options as possible to allow the user to enjoy secure computing without feeling burdened by it. Because OpenBSD is from Canada, the export of Cryptography pieces (such as OpenSSH and IPsec) to the world is not restricted. (NOTE: OpenBSD can not be re-exported from the US once it has entered the US. Because of this, take care NOT to get the distribution from an FTP server in the US if you are outside of Canada and the US.) A comprehensive list of the improvements brought by the 5.8 release is available on the web at http://www.OpenBSD.org/58.html. OpenBSD/armish runs on the Thecus N2100 and IODATA HDL-G ARM-based network disk appliances. Sources of OpenBSD: ------------------- Please refer to http://www.openbsd.org/ftp.html for all the ways you may acquire OpenBSD. OpenBSD 5.8 Release Contents: ----------------------------- The OpenBSD 5.8 release is organized in the following way. In the .../5.8 directory, for each of the architectures having an OpenBSD 5.8 binary distribution, there is a sub-directory. The armish-specific portion of the OpenBSD 5.8 release is found in the "armish" subdirectory of the distribution. That subdirectory is laid out as follows: .../5.8/armish/ INSTALL.armish Installation notes; this file. SHA256 Output of the sum(1) program using the option -a sha256, usable for verification of the correctness of downloaded files. SHA256.sig The above file, signed with the OpenBSD signing key for the 5.8 release, usable for verification of the integrity of the above file, and thus of the downloaded files. *.tgz armish binary distribution sets; see below. bsd A stock GENERIC armish kernel which will be installed on your system during the install. bsd.rd A compressed RAMDISK kernel; the embedded filesystem contains the installation tools. Used for simple installation from a pre-existing system. The OpenBSD/armish binary distribution sets contain the binaries which comprise the OpenBSD 5.8 release for armish systems. There are ten binary distribution sets. The binary distribution sets can be found in the "armish" subdirectory of the OpenBSD 5.8 distribution tree, and are as follows: base58 The OpenBSD/armish 5.8 base binary distribution. You MUST install this distribution set. It contains the base OpenBSD utilities that are necessary for the system to run and be minimally functional. It includes shared library support, and excludes everything described below. [ 57.5 MB gzipped, 159.6 MB uncompressed ] comp58 The OpenBSD/armish Compiler tools. All of the tools relating to C, C++ and Objective-C are supported. This set includes the system include files (/usr/include), the linker, the compiler tool chain, and the various system libraries (except the shared libraries, which are included as part of the base set). This set also includes the manual pages for all of the utilities it contains, as well as the system call and library manual pages. [ 46.7 MB gzipped, 141.6 MB uncompressed ] game58 This set includes the games and their manual pages. [ 2.5 MB gzipped, 5.7 MB uncompressed ] man58 This set includes all of the manual pages for the binaries and other software contained in the base set. Note that it does not include any of the manual pages that are included in the other sets. [ 9.6 MB gzipped, 37.3 MB uncompressed ] xbase58 This set includes the base X distribution. This includes programs, headers and libraries. [ 12.0 MB gzipped, 32.0 MB uncompressed ] xfont58 This set includes all of the X fonts. [ 38.0 MB gzipped, 49.7 MB uncompressed ] xserv58 This set includes all of the X servers. [ 5.6 MB gzipped, 14.2 MB uncompressed ] xshare58 This set includes all text files equivalent between all architectures. [ 4.2 MB gzipped, 23.2 MB uncompressed ] OpenBSD System Requirements and Supported Devices: -------------------------------------------------- OpenBSD/armish runs on the Thecus N2100 and IOData HDL-G models at the moment. Verifying the OpenBSD Installation Media: ----------------------------------------- As of OpenBSD 5.5, installations are able to verify files for the next release. The OpenBSD 5.8 release was signed with the /etc/signify/openbsd-58-base.pub release key. If you have an existing OpenBSD 5.5 or higher installation, you can run signify(1) to verify the signature and checksum. For example, run the following to verify that the cd58.iso file was distributed by the OpenBSD team: signify -C -p /etc/signify/openbsd-58-base.pub -x SHA256.sig cd58.iso If you are unable to run or compile signify(1), use sha256(1) with the SHA256 file to see if a file was corrupt during the transfer. Getting the OpenBSD System onto Useful Media: --------------------------------------------- Installation is supported from several media types, including: CD-ROM FFS partitions DOS (FAT) partitions EXT2 partitions HTTP The steps necessary to prepare the distribution sets for installation depend on which method of installation you choose. Some methods require a bit of setup first that is explained below. The installation allows installing OpenBSD directly from FTP mirror sites over the internet, however you must consider the speed and reliability of your internet connection for this option. It may save much time and frustration to use ftp get/reget to transfer the distribution sets to a local server or disk and perform the installation from there, rather than directly from the internet. If you are upgrading OpenBSD, you also have the option of installing OpenBSD by putting the new distribution sets somewhere in your existing file system, and using them from there. To do that, do the following: Place the distribution sets you wish to upgrade somewhere in your current file system tree. At a bare minimum, you must upgrade the "base" binary distribution, and so must put the "base58" set somewhere in your file system. It is recommended that you upgrade the other sets, as well. Preparing your System for OpenBSD Installation: ----------------------------------------------- To be able to boot the OpenBSD/armish installation program, you will need to learn a bit about RedBoot, the low-level process that controls the microprocessor after hardware initialization. The enabled features between IOData HDL-G and Thecus N2100 vary, so different methods will be needed on both machines. All these machines use RedBoot as their firmware and boot loader interface, with varying limitations. Thecus systems cannot load a file from disk, but can load an image from flash or network. IOData systems however are more restricted, and can only load an image from ext2fs partitions on disk, while the commands to load from flash are not available, and network booting does not appear to work. Thecus N2100 setup: Talking to the Thecus The Thecus N2100 needs a serial cable. On the back of the disk drive circuit board, there is either a 10-pin header (with one pin removed), or 9 holes for you to solder in your own header. The ribbon cable that connects from there to your serial cable or the serial port on your computer is of the same type as used on older i386 PCs, but be aware that there are two different types of 10-pin IDC to DB-9M ribbon cable in use, which look identical (the differences are hidden inside the DB-9 connector). The cable you need is wired using the "AT-Everex" or "Asus" pinout, as described here: http://www.pccables.com/07120.htm and *not* the more straightforward to solder "crossed" type: http://www.pccables.com/07121.htm. If you wish to check a cable, or make your own, the pinouts are: DB9 IDC10 === ===== 1 2 3 4 5 1 3 5 7 9 6 7 8 9 2 4 6 8 10 Wire pin 1 to 1, 2 to 2 etc, with 10 not connected. On the Thecus N2100, it is also necessary to move jumper J3 to J4 to properly route the serial port interrupt to allow serial to work under OpenBSD. This jumper is located under the disk drives, so you may wish to ensure its correct setting while you are installing the disk(s). For instructions on assembling/disassembling your N2100, see the N2100 Manual at http://www.thecus.com/download/manual/N2100%20UM%20EN.pdf Booting Thecus N2100 When you have connected your computer, a command such as "cu -l cuaU0 -s 115200" (assuming cuaU0 is your serial port device) should connect you to the Thecus console. Now apply power to the Thecus and start it. After some device probe information, you should see this prompt: == Executing boot script in 3.000 seconds - enter ^C to abort To enter interactive RedBoot, it is necessary to press ^C within the given number of seconds of the prompt. It will typically accept the ^C if typed during the diagnostics that precede this prompt. The recommended way to boot OpenBSD on Thecus is to configure networking and load bsd.rd from a TFTP server or HTTP server. Update the IP address, netmask and server address with the commands listed below to allow booting from the network. TFTP and HTTP are supported. (BOOTP configuration _may_ not work) Using ip_address only allows you to set the local IP address and the IP address of the server where bsd.rd is located. This command takes effect immediately but will not persist across a reboot. RedBoot> ip_address -l 10.0.0.21 -h 10.0.0.7 IP: 10.0.0.21/255.255.255.0, Gateway: 0.0.0.0 Default server: 10.0.0.7, DNS server IP: 0.0.0.0 Once you have done this, and have set up your TFTP server on the gateway machine (10.0.0.7 in the above example), you can boot manually. RedBoot> load /bsd.rd Using default protocol (TFTP) Entry point: 0x00200000, address range: 0x00200000-0x0066a49c RedBoot> go Once OpenBSD is installed on the hard disk, you will only need to load the OpenBSD/armish boot loader (/boot) the same way, and just hit "return" to boot OpenBSD. Automatic Booting However, for easier booting, you may wish to store the OpenBSD boot loader in the flash memory, so that the Thecus will boot automatically into OpenBSD when it is powered up. Doing so disables the devices' built-in Linux kernel, however, and makes it difficult or impossible to apply future RedBoot firmware upgrades. There are two general approaches: 1) Just change the boot configuration (using 'fconfig') to automate the above booting (your TFTP or HTTP boot server will still be needed). 2) Remove a file from the flash and replace it with OpenBSD's /boot; this configuration can then be booted standalone. Thecus RedBoot provides the 'fconfig' (not ifconfig) command to change the boot configuration of the device. Also 'fis' is provided to perform flash operations. Due to a change made by Thecus to the firmware, fconfig does not work with Thecus firmware version "1.93 (Feb 2007)" or later, and *it is thus not possible to make the Thecus auto-boot into OpenBSD* (worse, it will let you enter the "script" commands below, but ignore them on boot). To correct this, you must "upgrade" your Thecus down to firmware version "1.93 Nov 29 2005" using the file n2100-downgrade-reboot.rom *before* you remove any files from flash. You can get this file from Thecus; they have given permission for it to be mirrored at http://www.darwinsys.com/openbsd/thecus.html However, if you have removed files from flash before running this downgrade, it is not believed to be possible to get your machine to auto-boot. Once the downgrade is done, if your TFTP machine will always be around, you can autoboot just by changing the boot script with fconfig, as follows: (in this example, 192.168.1.254 is both the default IP gateway and also the TFTP server): RedBoot> fconfig Run script at boot: true Boot script: .. load /boot.armish .. go Enter script, terminate with empty line >> load /boot.armish >> go >> Boot script timeout (1000ms resolution): 3 Use BOOTP for network configuration: false Gateway IP address: 192.168.1.254 Local IP address: 192.168.1.253 Local IP address mask: 255.255.255.0 Default server IP address: 192.168.1.254 Console baud rate: 115200 DNS server IP address: GDB connection port: 9000 HW Status: 0 Force console for special debug messages: false MAC address: 0x00:0x14:0xFD:0x30:0x25:0x14 MAC address 2: 0x00:0x14:0xFD:0x30:0x25:0x15 Network debug at boot time: false Reset default: Normal Serial number: N2100 V2.1.06 ... Unlock from 0xf0fc0000-0xf0fc1000: . ... Erase from 0xf0fc0000-0xf0fc1000: . ... Program from 0x07fd2000-0x07fd3000 at 0xf0fc0000: . ... Lock from 0xf0fc0000-0xf0fc1000: . RedBoot> This configuration can easily be set back to the default, as the Linux system does not get erased from the flash memory. To set it back, just change the boot script with fconfig back to what it was, for example: thecus_setip fis load ramdisk fis load kernel exec -c "console=ttyS0,115200 root=/dev/ram0 initrd=0xa0800000,42M mem=128M@0xa0000000" You could also remove the boot script altogether by not re-entering it, so you would be dropped to the RedBoot shell by default. Recheck the new values and issue the 'reset' command. For the standalone booting option, you will have to load /boot into the flash memory. To overwrite the Thecus ROM behaviour, it is necessary to delete a 'rom file' and load the OpenBSD bootloader image there. With networking configured (as above), remove a region and load the 'boot' program. Initially the machine will contain much like the following: RedBoot> fis list Name FLASH addr Mem addr Length Entry point RedBoot 0xF0000000 0xF0000000 0x00040000 0x00000000 RedBoot config 0xF0FC0000 0xF0FC0000 0x00001000 0x00000000 FIS directory 0xF0FE0000 0xF0FE0000 0x00020000 0x00000000 ramdisk 0xF0040000 0x00800000 0x00D00000 0x00800000 kernel 0xF0D40000 0x00200000 0x00160000 0x00200000 user 0xF0EA0000 0xF0EA0000 0x00120000 0x00200000 There is no free space in the rom for the 'boot' program, so something must be removed. WARNING: this will make the machine unbootable into the original Thecus server mode and it is quite difficult to restore to the original software. EXTREME WARNING: Do not touch the RedBoot, RedBoot config, or FIS directory regions, as doing so will probably brick the device. Remove (with confirmation) a region to make space, eg 'user' (the non-root files needed to run Linux from Flash Memory). RedBoot> fis delete user Delete image 'user' - continue (y/n)? y ... Erase from 0xf0ea0000-0xf0fc0000: ......... ... Unlock from 0xf0fe0000-0xf1000000: . ... Erase from 0xf0fe0000-0xf1000000: . ... Program from 0x07fdf000-0x07fff000 at 0xf0fe0000: . ... Lock from 0xf0fe0000-0xf1000000: . Load the OpenBSD/armish bootloader 'boot' via TFTP or HTTP: RedBoot> load -m http /boot Entry point: 0x00100000, address range: 0x00100000-0x001067e0 Save that loaded file to flash: RedBoot> fis create boot ... Erase from 0xf0ea0000-0xf0ec0000: . ... Program from 0x00100000-0x001067e0 at 0xf0ea0000: . ... Unlock from 0xf0fe0000-0xf1000000: . ... Erase from 0xf0fe0000-0xf1000000: . ... Program from 0x07fdf000-0x07fff000 at 0xf0fe0000: . ... Lock from 0xf0fe0000-0xf1000000: . Confirm its addition: RedBoot> fis list Name FLASH addr Mem addr Length Entry point RedBoot 0xF0000000 0xF0000000 0x00040000 0x00000000 RedBoot config 0xF0FC0000 0xF0FC0000 0x00001000 0x00000000 FIS directory 0xF0FE0000 0xF0FE0000 0x00020000 0x00000000 ramdisk 0xF0040000 0x00800000 0x00D00000 0x00800000 kernel 0xF0D40000 0x00200000 0x00160000 0x00200000 boot 0xF0EA0000 0x00100000 0x00020000 0x00100000 Now that the bootloader is in place, update fconfig to autoboot OpenBSD: RedBoot> fconfig Run script at boot: true Boot script: .. fis load ramdisk .. fis load kernel .. exec ... Enter script, terminate with empty line >> fis load boot >> go >> Boot script timeout (1000ms resolution): 1 ... Update RedBoot non-volatile configuration - continue (y/n)? y ... Unlock from 0xf0fc0000-0xf0fc1000: . ... Erase from 0xf0fc0000-0xf0fc1000: . ... Program from 0x0ffd2000-0x0ffd3000 at 0xf0fc0000: . ... Lock from 0xf0fc0000-0xf0fc1000: . The important part is that 'Run script' is true and that the commands are 'fis load boot' followed with 'go' ('boot' being the file created by 'fis create boot' earlier). Confirm the update and type 'reset' to auto boot. The bootloader is command driven with timeout like many other OpenBSD architectures. IOData HDL-G setup: The hda1 partition is the boot partition for Linux; it contains zImage and initrd. Since RedBoot on this machine can only boot from the ext2fs partition, it is very important that this partition be preserved and not be reformatted. If this partition is destroyed, it will be necessary to remove the disk drive and recreate the partition on another machine. (RedBoot should be able to boot from the network, however that has been unsuccessful on IODATA HDL-G so far) To install OpenBSD, boot linux, and copy 'boot' and 'bsd.rd' into the partition mounted on /dev/hda1. After that has been done, it is possible to reboot into OpenBSD and install the device. Installation can then proceed as on other machines, however care must be take to not remove wd0i/hda1 partition (fdisk partition 0). DO NOT chose the 'full disk installation' option. Unfortunately, since IODATA did not provide the 'fconfig' command, it is not possible to change the boot commands run when RedBoot starts, thus it not possible to autoboot OpenBSD; the OpenBSD/armish bootloader is currently unable to work when loaded as `zImage'. Futhermore, to disable autobooting into Linux, at least one of the zImage and initrd files need to be renamed or removed. This can be done from the shell in bsd.rd. To load the ramdisk kernel to install use the commands RedBoot> load -m disk hda1:/bsd.rd RedBoot> go to boot normally, load the bootloader which will give options to boot other kernels. RedBoot> load -m disk hda1:/boot RedBoot> go Since this bootloader is able to boot any kernel from the OpenBSD root partition, it is not necessary to keep bsd.rd on the hda partition after OpenBSD is installed. Installing the OpenBSD System: ------------------------------ Installing OpenBSD is a relatively simple process. If you take your time and are careful to read the information presented by the installer, you shouldn't have any trouble. You should now be ready to install OpenBSD. The following is a walk-through of the steps you will take while getting OpenBSD installed on your hard disk. The installation procedure is designed to gather as much information about your system setup as possible at the beginning, so that no human interaction is required as soon as the questions are over. The order of these questions might be quite disconcerting if you are used to other installation procedures, including older OpenBSD versions. If any question has a default answer, it will be displayed in brackets ("[]") after the question. If you wish to stop the installation, you may hit Control-C at any time, but if you do, you'll have to begin the installation process again from scratch. Using Control-Z to suspend the process may be a better option, or at any prompt enter `!' to get a shell, from which 'exit' will return you back to that prompt (no refresh of the prompt will occur, though). Once the kernel has loaded, you will be presented with the OpenBSD kernel boot messages which contain information about the hardware that was detected and supported by OpenBSD. After the kernel is done initialising, you will be asked whether you wish to do an "(A)utoinstall", "(I)nstall" or an "(U)pgrade". Enter 'I' for a fresh install or 'U' to upgrade an existing installation. Enter 'A' to start an unattended installation where all of your answers are supplied in a response file (more on that in "Preparing an unattended installation of OpenBSD"). You will next be asked for your terminal type. You should choose the terminal type from amongst those listed. (If your terminal type is xterm, just use vt220). The first question you will be asked is the system hostname. Reply with the name of the system, without any domain part. You will now be given an opportunity to configure the network. The network configuration you enter (if any) can then be used to do the install from another system using HTTP or FTP, and will also be the configuration used by the system after the installation is complete. The install program will give you a list of network interfaces you can configure. For each network interface you select to configure, you will be asked for: - the symbolic host name to use (except for the first interface setup, which will reuse the host name entered at the beginning of the installation). - the IPv4 settings: address and netmask. If the IP address should be obtained from a DHCP server, simply enter ``dhcp'' when asked for the address. - the IPv6 settings (address, prefix length, and default router). You may enter ``rtsol'' when asked for the address for the interface to configure automatically via router solicitation messages. After all interfaces have been configured, if there have been any IPv4 interfaces setup, you will be asked for the IPv4 default route. This step is skipped if you only have one IPv4 interface setup, and it is configured with DHCP. The install program will also ask you for your DNS domain name, and the domain name servers, unless this information has already been obtained from a DHCP server during interface setup. You will also be presented with an opportunity to do more manual configuration. If you accept, you will be dropped to a shell; when you are done, enter `exit' to return to the installation program. You will then be asked to enter the initial root password of the system, twice. Although the install program will only check that the two passwords match, you should make sure to use a strong password. As a minimum, the password should be at least eight characters long and a mixture of both lower and uppercase letters, numbers and punctuation characters. You will then be asked whether you want to start sshd(8) by default, as well as ntpd(8). If you choose to start ntpd(8), you will be asked for your ntp server; if you don't have any preferred ntp server, press enter to confirm the default setting of using the pool.ntp.org servers. You will now be given the possibility to setup a user account on the forthcoming system. This user will be added to the `wheel' group. Enter the desired login name, or `n' if you do not want to add a user account at this point. Valid login names are sequences of digits and lowercase letters, and must start with a lowercase letter. If the login name matches this criteria, and doesn't conflict with any of the administrative user accounts (such as `root', `daemon' or `ftp'), you will be prompted with the users descriptive name, as well as its password, twice. As for the root password earlier, the install program will only check that the two passwords match, but you should make sure to use a strong password here as well. If you have chosen to setup a user account, and you had chosen to start sshd(8) on boot, you will be given the possibility to disable sshd(8) logins as root. You may now be given the opportunity to configure the time zone your system will be using (this depends on the installation media you are using). If the installation program skips this question, do not be alarmed, the time zone will be configured at the end of the installation. The installation program will now tell you which disks it can install on, and ask you which it should use. Reply with the name of your root disk. You will the be asked if you want to use DUID notation in /etc/fstab, instead of traditional device names. You are strongly advised to use DUIDs, as they allow you to move your disks to different controllers, or change their bus identifiers, without having to modify /etc/fstab every time your configuration changes. Disks on OpenBSD/armish are partitioned using the ``MBR'' partitioning scheme. You will need to create one MBR partition, in which all the real OpenBSD partitions will be created. The installation program will ask you if you want to use the whole disk for OpenBSD. If you don't need to or don't intend to share the disk with other operating systems, answer `w' here. The installation program will then create a single MBR partition spanning the whole disk, dedicated to OpenBSD. Otherwise, fdisk(8) will be invoked to let you to edit your MBR partitioning. The current MBR partitions defined will be displayed and you will be allowed to modify them, add new partitions, and change which partition to boot from by default. Note that you should make the OpenBSD partition the active partition at least until the install has been completed. After your OpenBSD MBR partition has been setup, the real partition setup can follow. Next the disk label which defines the layout of the OpenBSD partitions must be set up. Each file system you want will require a separate partition. You will be proposed a default partition layout, trying to set up separate partitions, disk size permitting. You will be given the possibility to either accept the proposed layout, or edit it, or create your own custom layout. These last two choices will invoke the disklabel(8) interactive editor, allowing you to create your desired layout. Within the editor, you will probably start out with only the 'c' partition of fstype 'unused' that represents the whole disk. This partition can not be modified. If you have Linux partitions defined on the disk, these will usually show up as partition 'i', 'j' and so on. You must create partition 'a' as a native OpenBSD partition, i.e. one with "4.2BSD" as the fstype, to hold the root file system. In addition to partition 'a' you should create partition 'b' with fstype "swap", and native OpenBSD partitions to hold separate file systems such as /usr, /tmp, /var, and /home. You will need to provide a mount point for all partitions you define. Partitions without mount points, or not of 4.2BSD fstype, will neither be formatted nor mounted during the installation. For quick help while in the interactive editor, enter '?'. The `z' command (which deletes all partitions and starts with a clean label), the `A' command (which performs the automatic partition layout) and the `n' command (to change mount points) are of particular interest. Although the partitions position and size are written in exact sector values, you do not need a calculator to create your partitions! Human-friendly units can be specified by adding `k', `m' or `g' after any numbers to have them converted to kilobytes, megabytes or gigabytes. Or you may specify a percentage of the disk size using `%' as the suffix. Enter 'M' to view the entire manual page (see the info on the ``-E'' flag). To exit the editor enter 'q'. After the layout has been saved, new filesystems will be created on all partitions with mount points. This will DESTROY ALL EXISTING DATA on those partitions. After configuring your root disk, the installer will return to the list of available disks to configure. You can choose the other disks to use with OpenBSD in any order, and will get to setup their layout similarly to the root disk above. However, for non-root disks, you will not be proposed a default partition layout. When all your disks are configured, simply hit return at the disk prompt. After these preparatory steps have been completed, you will be able to extract the distribution sets onto your system. There are several install methods supported: FTP, HTTP, CD-ROM, or a local disk partition. To install via FTP or HTTP: To begin an FTP or HTTP install you will need the following pieces of information: 1) Proxy server URL if you are using a URL-based FTP or HTTP proxy (squid, CERN FTP, Apache 1.2 or higher). You need to define a proxy if you are behind a firewall that blocks outgoing FTP or HTTP connections (assuming you have a proxy available to use). 2) The IP address (or hostname if you configured DNS servers earlier in the install) of an FTP or HTTP server carrying the OpenBSD 5.8 distribution. The installation program will try to fetch a list of such servers; depending on your network settings, this might fail. If the list could be fetched, it will be displayed, and you can choose an entry from the list (the first entries are expected to be the closest mirrors to your location). 3) The directory holding the distribution sets. The default value of pub/OpenBSD/5.8/armish is almost always correct on FTP servers; for HTTP servers there is no standard location for this. 4) For FTP installs only, the login and password for the FTP account. You will only be asked for a password for non-anonymous FTP. Then refer to the section named "installation set selection" below. To install from CD-ROM: When installing from a CD-ROM, you will be asked which device holds the distribution sets. This will typically be "cd0". If there is more than one partition on the CD-ROM, you will be asked which partition the distribution is to be loaded from. This is normally partition "a". You will also have to provide the relative path to the directory on the CD-ROM which holds the distribution, for the armish this is "5.8/armish". Then refer to the section named "installation set selection" below. To install from a local disk partition: When installing from a local disk partition, you will first have to identify which disk holds the distribution sets. This is normally "wdN" or "sdN", where N is a number. Next you will have to identify the partition within that disk that holds the distribution; this is a single letter between 'a' and 'p'. You will also have to identify the type of file system residing in the partition identified. Currently, you can install from partitions that have been formatted as the Berkeley fast file system (ffs), Linux (ext2) or MS-DOS. You will also have to provide the relative path to the directory on the file system where the distribution sets are located. Note that this path should not be prefixed with a '/'. Then refer to the next section. Installation set selection: A list of available distribution sets found on the given location will be listed. You may individually select distribution sets to install, by entering their name, or wildcards (e.g. `*.tgz' or `base*|comp*', or `all' to select all the sets (which is what most users will want to do). You may also enter `abort' to deselect everything and restart the selection from scratch, or unselect sets by entering their name prefixed with `-' (e.g. `-x*'). It is also possible to enter an arbitrary filename and have it treated as a file set. When you are done selecting distribution sets, enter `done'. The files will begin to extract. After the files have been extracted, you will be given the choice to select a new location from which to install distribution sets. If there have been errors extracting the sets from the previous location, or if some sets have been missing, this allows you to select a better source. Also, if the installation program complains that the distribution sets you have been using do not match their recorded checksums, you might want to check your installation source (although this can happen between releases, if a snapshot is being updated on an FTP or HTTP server with newer files while you are installing). The last thing you might need to configure, if you did not get the chance to earlier, is the time zone your system will be using. For this work properly, it is expected that you have installed at least the "base58" and "bsd" distribution sets. The installation program will then proceed to save the system configuration, create all the device nodes needed by the installed system, and will install bootblocks on the root disk. Finally, you will be asked whether you would like to install non-free firmware files (which can't be tightly integrated to the OpenBSD system) on first boot, by invoking fw_update(8) on the next boot. Congratulations, you have successfully installed OpenBSD 5.8. When you reboot into OpenBSD, you should log in as "root" at the login prompt. You should create yourself an account and protect it and the "root" account with good passwords. The install program leaves root an initial mail message. We recommend you read it, as it contains answers to basic questions you might have about OpenBSD, such as configuring your system, installing packages, getting more information about OpenBSD, sending in your dmesg output and more. To do this, run mail and then just enter "more 1" to get the first message. You quit mail by entering "q". Some of the files in the OpenBSD 5.8 distribution might need to be tailored for your site. We recommend you run: man afterboot which will tell you about a bunch of the files needing to be reviewed. If you are unfamiliar with UN*X-like system administration, it's recommended that you buy a book that discusses it. Preparing an unattended installation of OpenBSD: ------------------------------------------------ If '(A)utoinstall' is choosen at the install prompt or if the installation system detects that it booted from the network, and isn't interrupted within 5 seconds, it attempts a fully-automatic installation. The installer runs dhclient(8) on the network interface the system booted from, or in case of multiple interfaces it will ask which one to use. Upon success it retrieves a response file via HTTP. If that fails, the installer asks for the response file location which can be either an url or a local path and retrieves the response file from there. The "next-server" DHCP option specifies the hostname part of the URL, as in "http:///install.conf". The "filename" DHCP parameter specifies the installer mode, e.g. "auto_install". On architectures where this parameter is used for netbooting, create a symbolic link named "auto_install" pointing to the boot program. The response file contains lines with key/value pairs separated by an equals sign '=', where the key is a non-ambiguous part (up to the question mark) of the installer question, consisting of whitespace separated words. The value is what would have been entered at the interactive prompt. Empty lines and lines beginning with a '#' character are ignored. The installer uses default answers in case of missing answers. Here is a response file example that uses a hashed password (see encrypt(1) for more details) for root and a public ssh key for the user that is created during the installation. System hostname = openbsd Password for root = $2a$14$Z4xRMg8vDpgYH...GVot3ySoj8yby Setup a user = puffy Password for user = ************* Public ssh key for user = ssh-ed25519 AAAAC3NzaC1...g3Aqre puffy@ai What timezone are you in = Europe/Stockholm Location of sets = http HTTP Server = ftp.eu.openbsd.org The "System hostname" key above matches the following full question asked during an interactive installation: System hostname? (short form, e.g. 'foo') While the installation is in progress the installer writes all output to the file /ai.log, which is available as mail on the freshly installed system after the initial reboot. If the installation is successful the system will reboot automatically; otherwise, you will be dropped back into the shell where you can look at the /ai.log file or try again. Upgrading a previously-installed OpenBSD System: ------------------------------------------------ Warning! Upgrades to OpenBSD 5.8 are currently only supported from the immediately previous release. The upgrade process will also work with older releases, but might not execute some migration tasks that would be necessary for a proper upgrade. The best solution, whenever possible, is to backup your data and reinstall from scratch. As a minimum, if the toolchain (the ``comp'' set) was installed, you should remove all files within /usr/include before attempting to upgrade. To upgrade OpenBSD 5.8 from a previous version, start with the general instructions in the section "Installing OpenBSD". Boot from your usual boot device. When prompted, select the (U)pgrade option rather than the (I)nstall option at the prompt in the install process. You will be presented with a welcome message and asked if you really wish to upgrade. The upgrade script will ask you for the existing root partition, and will use the existing filesystems defined in /etc/fstab to install the new system in. It will also use your existing network parameters. From then, the upgrade procedure is very close to the installation procedure described earlier in this document. However, it is strongly advised that you unpack the etc.tgz and the xetc.tgz files found in /usr/share/sysmerge in a temporary directory and merge changes by hand, or with the help of the sysmerge(8) helper script, since all components of your system may not function correctly until your files in `/etc' are updated. Getting source code for your OpenBSD System: -------------------------------------------- Now that your OpenBSD system is up and running, you probably want to get access to source code so that you can recompile pieces of the system. A few methods are provided. If you have an OpenBSD CD-ROM, the source code is provided. Otherwise, you can get the pieces over the Internet using anonymous CVS, CVSync or FTP. For more information, see http://www.OpenBSD.org/anoncvs.html http://www.OpenBSD.org/cvsync.html http://www.OpenBSD.org/ftp.html Using online OpenBSD documentation: ----------------------------------- Documentation is available if you first install the manual pages distribution set. Traditionally, the UN*X "man pages" (documentation) are denoted by 'name(section)'. Some examples of this are intro(1), man(1), apropos(1), passwd(1), passwd(5) and afterboot(8). The section numbers group the topics into several categories, but three are of primary interest: user commands are in section 1, file formats are in section 5, and administrative information is in section 8. The 'man' command is used to view the documentation on a topic, and is started by entering 'man [section] topic'. The brackets [] around the section should not be entered, but rather indicate that the section is optional. If you don't ask for a particular section, the topic with the least-numbered section name will be displayed. For instance, after logging in, enter man passwd to read the documentation for passwd(1). To view the documentation for passwd(5), enter man 5 passwd instead. If you are unsure of what man page you are looking for, enter apropos subject-word where "subject-word" is your topic of interest; a list of possibly related man pages will be displayed. Adding third party software; ``packages'' and ``ports'': -------------------------------------------------------- As complete as your OpenBSD system is, you may want to add any of several excellent third party software applications. There are several ways to do this. You can: 1) Use the OpenBSD ``package'' collection to grab a pre-compiled and tested version of the application for your hardware. 2) Use the OpenBSD ``ports'' collection to automatically get any needed source file, apply any required patches, create the application, and install it for you. 3) Obtain the source code and build the application based upon whatever installation procedures are provided with the application. If you purchased the OpenBSD CD-ROM set you already have a few popular ``packages'', and the ``ports'' collection. Instructions for installing applications from the various sources using the different installation methods follow. You should also refer to the packages(7) manual page. Installing applications from the CD-ROM package collection: The OpenBSD CD-ROM ships with the most commonly installed third-party applications pre-built for various hardware architectures. Limited disk space on the CD-ROM unfortunately limits the number of such packages. Check the directory 5.8/packages/arm to see which packages are available for your hardware architecture. That directory will be on the same CD-ROM containing the OS installation files for your architecture. To install one or more of these packages you must: 1) become the superuser (root). 2) mount the appropriate CD-ROM. 3) use the ``pkg_add'' command to install the software. Example (in which we use su(1) to get superuser privileges, thus you have to be in group "wheel", see the manual page for su(1)). $ su Password: # mkdir -p /cdrom # mount /dev/cd0a /cdrom # pkg_add /cdrom/5.8/packages/arm/ # # umount /cdrom Package names are usually the application name and version with .tgz appended, e.g. bzip2-1.0.6p0.tgz Installing applications from the ftp.OpenBSD.org package collection: All available packages for your architecture have been placed on ftp.OpenBSD.org in the directory pub/OpenBSD/5.8/packages/arm/ You may want to peruse this to see what packages are available. The packages are also on the OpenBSD FTP mirror sites. See http://www.OpenBSD.org/ftp.html for a list of current FTP mirror sites. Installation of a package is very easy. 1) become the superuser (root) 2) use the ``pkg_add'' command to install the software ``pkg_add'' is smart enough to know how to download the software from the OpenBSD FTP server. Example: $ su Password: # pkg_add \ ftp://ftp.OpenBSD.org/pub/OpenBSD/5.8/packages/arm/emacs-21.4p23.tgz Installing applications from the CD-ROM ports collection: The CD-ROM ``ports'' collection is a set of Makefiles, patches, and other files used to control the building and installation of an application from source files. Creating an application from sources can require a lot of disk space, sometimes 50 megabytes or more. The first step is to determine which of your disks has enough room. Once you've made this determination, read the file PORTS located on the CD-ROM which contains the ports tree. To build an application you must: 1) become the superuser (root) 2) have network access, or obtain the actual source files by some other means. 3) cd to the ports directory containing the port you wish to build. To build samba, for example, where you'd previously copied the ports files into the /usr/ports directory: cd /usr/ports/net/samba 4) make 5) make install 6) make clean Installing applications from the OpenBSD ports collection: See http://www.OpenBSD.org/faq/ports/ports.html for current instructions on obtaining and installing OpenBSD ports. You should also refer to the ports(7) manual page. Installing other applications: If an OpenBSD package or port does not exist for an application you're pretty much on your own. The first thing to do is ask if anyone is working on a port -- there may be one in progress. If no such port exists, you might want to look at the FreeBSD ports or NetBSD pkgsrc for inspiration. If you can't find an existing port, try to make your own and feed it back to OpenBSD. That's how our ports collection grows. Some details can be found in the OpenBSD Porter's Handbook at http://www.openbsd.org/faq/ports/ with more help coming from the mailing list, . Administrivia: -------------- There are various mailing lists available via the mailing list server at . To get help on using the mailing list server, send mail to that address with an empty body, and it will reply with instructions. More information about the various OpenBSD mailing list and proper netiquette is available at http://www.OpenBSD.org/mail.html To report bugs, use the 'sendbug' command shipped with OpenBSD, and fill in as much information about the problem as you can. Good bug reports include lots of details. Additionally, bug reports can be sent by mail to: bugs@OpenBSD.org As a favor, please avoid mailing huge documents or files to the mailing lists. Instead, put the material you would have sent on a web server, then mail the appropriate list about it, or if you'd rather not do that, mail the list saying you'll send the data to those who want it. For more information about reporting bugs, see http://www.OpenBSD.org/report.html