Red Hat Linux S/390 7.2 (Enigma) ======================================= The contents of this CD-ROM are Copyright (C) 1995-2001 Red Hat, Inc. and others. Please see the individual copyright notices in each source package for distribution terms. The distribution terms of the tools copyrighted by Red Hat, Inc. are as noted in the file COPYING. Red Hat and RPM are trademarks of Red Hat, Inc. ============================================================================ DIRECTORY ORGANIZATION Red Hat Linux is delivered on two CDROMs (disc 1 and disc 2). Disc 1 contains the following directory structure: /mnt/redhat |----> RedHat | |----> RPMS -- binary packages | `----> base -- information on this release of Red Hat | Linux used by the installation process |----> images -- boot and ramdisk images |----> COPYING -- copyright information |----> README -- this file `----> RPM-GPG-KEY -- GPG signature for packages from Red Hat The directory layout of disc 2 is as follows: /mnt/redhat |----> RedHat | `----> RPMS -- additional binary packages | `----> installer -- sources and specfiles for the installer |----> SRPMS -- source packages (more are on the dedicated | SRPMS CD) |----> COPYING -- copyright information |----> README -- this file `----> RPM-GPG-KEY -- GPG signature for packages from Red Hat Installation options -------------------- There are two installation programs available for Red Hat Linux for use on the IBM S/390 architecture. Depending on your requirements you can now either use the 'rhsetup' installation program which was developed especially for the S/390 and zSeries or you can use the standard Red Hat Linux installation program 'anaconda'. Both methods allow installation via FTP, HTTP, NFS or from a local hard disk. The advantages of the 'rhsetup' installation program are that it has a minimum RAM requirement (32MB) and lets you quickly install a machine using either a simple text interface or a non-interactive installation method. The disadvantage is that rhsetup itself does not support the configuration of various services during installation, so it has to be done manually after the installation is complete. It also cannot partition a hard disk (DASD) into several partitions, and instead creates one partition for each available hard disk. Anaconda may be preferred if you have experience performing Red Hat Linux installations on other architectures or if you would like to configure services during installation or select individual packages to install, as Anaconda allows you to configure such items during installation. Anaconda also offers a graphical user interface for NFS installations. The memory requirement during installation under Anaconda is 128MB, but this is required only during installation and can be reduced after the installation is complete. For a more detailed description of how to use the two installation programs please refer to the section "How to Install" located in this document. Resources for S/390 from Red Hat and others: ------------------------------------------------------ - You can subscribe to a public mailing list regarding S/390 and zSeries development of Red Hat Linux at the following URL: https://listman.redhat.com/mailman/listinfo/redhat-s390-list Please report any problems or any other feedback to this public mailing list at . - You can report problems about specific RPMs to our bug database: http://bugzilla.redhat.com/bugzilla - You can find general Linux and S/390 information from IBM at the following URL: http://www.s390.ibm.com/linux - Download current development versions of Red Hat Linux for the S/390 from the following sites: ftp://ftp.redhat.com/pub/redhat/linux/rawhide/s390/ ftp://ftp.redhat.de/pub/s390/ rsync://ftp.redhat.de/s390/ - Other resources relating to Linux and the S/390: http://linuxvm.org/ http://linux.zseries.org/ - The directory /usr/src/linux/Documentation/s390/ are part of the current kernel sources available from: ftp://ftp.kernel.org/pub/linux/kernel/ We are very interested in your feedback and thank you for your participation and time to help us make this product as good as our releases for other platforms. How To Install: =============== Requirements for installation: ------------------------------ You can install the Red Hat Linux distribution via four different installation methods: via FTP, NFS, HTTP or from a local DASD hard disk. Installation with the 'rhsetup' installation program requires 32 MB of RAM. The 'Anaconda' installation program with an advanced set of configuration options requires at least 128 MB of RAM, but you can reduce the amount of RAM used for the installed Linux system once the installation is complete. For the installation you must start a basic Linux system that is structured in several files: - the Linux kernel (kernel.img or tapekrnl.img) - an "initrd" consisting of a basic Linux system and our installation program (initrd.img or tapeinrd.img) - an optional second "initrd" consisting of IBM proprietary object code only modules required for use with certain hardware (ocord.img) - a parameter file with information about your hardware configuration This basic Linux setup creates a TCP/IP network that you can log in to via ssh or telnet to begin the installation of Red Hat Linux. Common steps needed for installing Red Hat Linux to a VM or LPAR: ----------------------------------------------------------------- Installing on systems that are not currently running Linux ----------------------------------------------------------- - Use this procedure ONLY if you DO NOT have a working Linux system available on the target system. - Set up an FTP, HTTP or NFS server on a separate machine. We will name the server RPMSERVER for the purposes of this document. FTP, HTTP and NFS installation methods require that you copy the complete installation CD-ROM (CD 1) into a new directory on the server. * Insert disc 1 * mount /mnt/cdrom * cp -a /mnt/cdrom/RedHat /target/directory * umount /mnt/cdrom Next, you must copy the contents of the /RedHat/RPMS/ directory of the second CD-ROM into the subdirectory RedHat/RPMS on the server: * Replace disc 1 with disc 2 * mount /mnt/cdrom * cp -a /mnt/cdrom/RedHat /target/directory * umount /mnt/cdrom - Setup up RPMSERVER to export the target directory you created. - If you intend to use an NFS server, add the target directory to /etc/exports and restart your NFS daemon after making this change. The correct entry in /etc/exports on a Red Hat Linux system would be: /target/directory ip.of.your.s390(ro) Then, restart your nfs server: /sbin/service nfs restart - Verify that this is working from another machine by either mounting the directory (NFS). Setup of an FTP or HTTP server for an installation with anaconda: ----------------------------------------------------------------- This is an alternative to the above setup of a FTP/HTTP server (only usable with anaconda). You can either mount both CDs into the directories of the server (requires two CDROM drives) or mount the ISO files over loopback. - FTP installations access the CD-ROMs/mounted ISO images using URLs in this format: * CD1 accessible from ftp://.../disc1 * CD2 accessible from ftp://.../disc2 - HTTP installations access the CD-ROMs/mounted ISO images using URLs in this format: * CD1 accessible from http://.../disc1 * CD2 accessible from http://.../disc2 When prompted for the URL, enter the URL for the first CD-ROM without a trailing slash. Installing on systems with a working Linux installed ---------------------------------------------------- If you already have a working Linux system on your S/390, you can install from the CD images using the target system's own DASD. - Copy images of the two installation CD-ROMs to the running DASD, make it available to the new VM/LPAR, and then install from the running DASD. Depending on how you acquired a copy of Red Hat Linux for S/390, you either have CD images or a set of CD-ROMs. If you have CD-ROMs, you must dump the images on your Linux workstation. Insert the first CD-ROM into your CD-ROM drive and run: dd if=/dev/cdrom of=/tmp/CD1.iso (This command will raise an error message, which should be ignored, when it reaches the end of data on the CD-ROM.) Installing in VM: ----------------- - Log in to VM as the user maintaining Linux installations. - You can use x3270 or c3270 (from the x3270-text package in Red Hat Linux) to log in to VM from Linux systems. Alternatively, use the OS/2 3270 terminal emulator on the S/390 management console. If you use x3270, the recommended character set is 'U.S.English' If you do not use this character set, you will notice some garbled characters. Add these lines to /usr/X11R6/lib/X11/app-defaults/X3270 or $HOME/.Xdefaults : x3270.charset: us.intl x3270.keymap: circumfix x3270.keymap.circumfix: asciicircum: Key("^")\n - If you are not already in CMS mode, enter it now: i cms - Ensure that you have enough free space available on your 191 disk to hold ~8MB of additional data. - Connect via FTP to a machine containing the Red Hat Linux boot images and log in. - Execute the following commands: cd /where/the/boot/images/are ascii mget *.prm *.exec (repl locsite fix 80 bin mget kernel.img initrd.img (repl if using optional OCO second ram disk: cd /where/the/second/ramdisk/is mget ocord.img (repl quit - If you want to, edit the parameter file "redhat.prm". For details, refer to the section "Customizing the parameter file" in this document. If you do not customize the parameter file, you will be prompted for the values during setup. - Log out with: #cp log - Reconnect and log in as the machine you wish to install. - If you are not in CMS mode, enter it now: i cms - If you are choosing to use the second initial ramdisk (containing the proprietary object code only drivers), you may need to adjust your exec script to include the additional ramdisk: for example, a redhat exec script that wouldn't use the second ramdisk may look similar to: /* */ 'cl rdr' 'purge rdr all' 'spool punch * rdr' 'PUNCH KERNEL IMG A (NOH' 'PUNCH REDHAT PRM A (NOH' 'PUNCH INITRD IMG A (NOH' 'ch rdr all keep nohold' 'i 00c' but the same exec script including the second ramdisk would have an additional line: /* */ 'cl rdr' 'purge rdr all' 'spool punch * rdr' 'PUNCH KERNEL IMG A (NOH' 'PUNCH REDHAT PRM A (NOH' 'PUNCH INITRD IMG A (NOH' 'PUNCH OCORD IMG A (NOH' 'ch rdr all keep nohold' 'i 00c' - Run the redhat script: redhat - Answer any questions the script asks (if you have set up all parameters in the redhat.prm file, there should be no questions). - When the boot process is complete, use ssh or telnet to log in to your S/390 machine as root (any password will work) and run 'rhsetup' for the simple installation program or 'loader' for the Anaconda installation program. You may run 'rhsetup' from the 3270 console as well, however the installation will proceed in text-mode only. If you encounter problems using the graphical Anaconda interface, you must unset the DISPLAY variable (unset DISPLAY). You will then use the text mode interface of Anaconda. - The remaining installation process should proceed normally. - When the installation finishes, set up your linux virtual machine to boot from the DASD (virtual hard disk) rather than entering CMS or CP - This is done by issuing the following command: #cp i nnn nnn is the address of the DASD containing your Linux boot image. - Once the installation is complete and the installed system is running (after booting Linux), disconnect from 3270 using: #cp disc DO NOT log out, as logging out will stop the machine. Installing in an LPAR using the Red Hat Linux CD ------------------------------------------------ - Use Single Object Operations on the HMC to access the Support Element Workplace (SEW) as a user with sufficient privileges to install a new OS to an LPAR. - Select "Images", then select the LPAR you wish to install to. - Use the arrows in the frame on the right side to navigate to the "CPC Recovery" menu. - Double-click on "Load from CD-ROM or Server". - In the dialog box that follows, select "Local CD-ROM" and click on "Continue". - In the dialog that follows, keep the default selection of redhat.ins and click continue. - Continue on to the section "Installing in an LPAR (common steps)" below. Installing in an LPAR without the Red Hat Linux CD -------------------------------------------------- - Use Single Object Operations on the HMC to access the Support Element Workplace (SEW) as a user with sufficient privileges to install a new OS to an LPAR. - Select "Images", then select the LPAR you wish to install to. - Use the arrows in the frame on the right side to navigate to the "CPC Recovery" menu. - Double-click on "Load from CD-ROM or Server". - In the dialog box that follows, select "FTP Source", and enter the following information: * Host Computer: hostname or IP address of the ftp server you wish to install from (e.g. ftp.redhat.com) * User ID: Your user name on the ftp server (or "anonymous") * Password: Your password (use your email address if you're logging in as anonymous). * Account (can be blank): Leave this field empty * File location (can be blank): Directory on the ftp server holding Red Hat Linux for S/390 (e.g. /pub/redhat/linux/rawhide/s390) - Click "Continue". - In the dialog that follows, keep the default selection of redhat.ins and click continue. - Continue on to the section "Installing in an LPAR (common steps)" below. Installing in on an MP3000 without the Red Hat Linux CD: --------------------------------------------------------------------------- Virtual tape is part of the emulated I/O support found only on an MP3000. - Transfer the redhat.tdf, tapekrnl.img, redhat.prm and tapeinrd.img files to the OS/2 machine, in the directory used for virtual tapes. This is usually F:\AWSOMA. - If the directory is not F:\AWSOMA, you will have to edit the redhat.tdf file. - A usable editor on OS/2 is "epm". - You may customize the parameter file "redhat.prm", if you wish. For details, see the "Customizing the parameter file" section. If you omit any of the parameters, you will be prompted for them at installation time. - Create a virtual tape containing the Red Hat Linux images: awsmount 080 F:\AWSOMA\REDHAT.TDF awsmount 080 /D /R /REW - Log in on the Support Element Workplace using a user ID with sufficient privileges to install a new OS to an LPAR. The "SYSPROG" user is recommended. - Select "Images", then select the LPAR you wish to install to. - Use the arrows in the frame on the right side to navigate to the "CPC Recovery" menu. - Double-click on "Load". - In the dialog box that follows, select "Clear" and enter the address of the virtual tape you just generated. Usually, this will be "fc80". - Click OK. Red Hat Linux will be booted. - If you receive any "Load error" messages or something similar on a Multiprise 3000 or another machine with CD-ROM support, try installing something from any CD, interrupting, stopping the LPAR, and retrying the "Load" step. This appears to be a bug in some versions of the Support Element Workplace. - Continue to the section "Installing in an LPAR (common steps)" below. Installing in an LPAR (common steps): ------------------------------------- - Once the Red Hat Linux installation program has started (the red field behind the LPAR icon will disappear), select the LPAR and double-click on "Operating System Messages". - Answer any questions the script asks you (if you have set up all parameters in the redhat.prm file, you should not be asked anything). When the boot process is complete, telnet or ssh to your S/390 machine as "root" and run "rhsetup" or "loader" (anaconda). - The remaining installation process should proceed normally. Customizing the parameter file: =============================== - You may also customize the parameter file "redhat.prm" with the following parameters: * required parameters: root=/dev/ram0 ro ip=off These values are given directly to the Linux kernel and should be left as provided DASD=x-y Range of addresses of your DASD devices. 200-20f should be sufficient in most cases. Be careful when the addresses of your VM partitions are in this range: They show up in the list of available DASDs and you can select them to be formatted (DO NOT DO THAT). CHANDEV=device,input,output device: Type of device (e.g. ctc0,escon0) If you are using IBM's OCO modules, this can also be lcs0 or qeth0. See the section on OCO modules in RELEASE-NOTES for details. input: the input port for this device output: the output port for this device eg: CHANDEV=ctc0,0x600,0x601 or CHANDEV=escon0,0x605,0x608 To force a specific CTC protocol, additionals parameters can be added: CHANDEV=ctc0,0x600,0x601,0,XXX XXX should be either 0 for compatibility mode, used with NON-Linux peers other than OS/390 and z/OS 1 for extended mode 2 for 'CTC-based tty', only supported on Linux-Linux connections 3 for compatibility mode with OS/390 and z/OS * optional parameters: HOST=<...> The values for the HOST statement depend on whether you want to install this machine via an Ethernet driver or via ctc/escon/iucv devices: HOST=fqdn:device:ipaddr[:mtu] (for Ethernet devices) fqdn: your full qualified domain name of the virtual machine device: eth0, tr0, hsi0 ipaddr: the fixed IP-address for this virtual machine mtu: maximum transfer units (optional, should be 1492 or 1500) HOST=fqdn:device:ipaddr:gateway[:mtu] (for ctc/escon devices) fqdn: your full qualified domain name of the virtual machine device: ctc0, escon0 ipaddr: the fixed IP-address for this virtual machine gateway: IP-address of the gateway mtu: maximum transfer units (optional. should be 1492 or 1500) HOST=fqdn:device:ipaddr:host:gateway[:mtu] (for iucv devices) fqdn: your full qualified domain name of the virtual machine device: iucv0 ipaddr: the fixed IP-address for this virtual machine host: VM host name of IUCV partner (usually TCPIP for VM or $TCPIP for VIF) gateway: IP-address of the IUCV partner mtu: maximum transfer units (optional) NETWORK=IP:netmask:broadcast[:gw] IP: The adress of your network netmask: the netmask broadcast: the broadcast address gw: the gateway-IP for your eth device (for eth-device only) DNS=list:of:dns:servers the list of DNS servers, separated by colons e.g. DNS=10.0.0.1:10.0.0.2 will use the DNS servers 10.0.0.1 and 10.0.0.2 SEARCHDNS=list:of:search:domains the list of the search domains, separated by colons eg: SEARCHDNS=redhat.de:redhat.com RPMSERVER=ftp://your.ftp.server/your.s390.rpm.dir RPMSERVER=ftp://user:password@your.ftp.server//your.s390.rpm.dir RPMSERVER=http://your.http.server/your.s390.rpm.path RPMSERVER=IP:/dir RPMSERVER=dasd[a-z][1]:/,dasd[a-z]1:/ Your Red Hat Linux for S/390 binary RPM packages may be located on any of your FTP or HTTP servers in the given directories, or you may give an IP-address and a directory of a NFS server (e.g. a mount of your Red Hat Linux for S/390 CD-ROM). You can also install from a local DASD. Copy the ISO files of the installation CD-ROMs onto the ext2 formatted DASD and supply the device names, the directories and the file names of the ISO images (The images can be on different DASD). * optional parameter for installations using a qeth device: QETHPARM= Use this option to give additional parameters for your QETH device (for chandev) such as: add_parms,0x10,{lo_devno,hi_devno},portname:port_name * optional parameters for installations using rhsetup only: DTZ=timezone default time zone as returned by the tzselect-utility, eg: DTZ=Europe/Berlin or DTZ=America/Chicago VERBOSEWGET= Use this option to supply additional parameters to the wget command. If this parameter is non-empty, the downloading of files from the FTP server will be more verbose. This may help with tracking down errors, but the parameter should be omitted if you do not have any problems with the installation. Setting VERBOSEWGET to a value which wget (see 'man wget') does not understand will result in an error during the installation. This parameter only applies to the FTP/HTTP installation methods. INSTALL=package selection defines the pre-selection of packages. It must be one of minimal / desktop/ default / server / all. MOUNTS=dev:mountpoint,dev2:mountpoint2 defines a comma-delimited list of DASD device-partitions and where they should be mounted by default. eg: MOUNTS=/dev/dasda1:/,/dev/dasdb1:/usr,/dev/dasdc1:/tmp SWAP=list:of:swap:devices the list of initial swap devices, delimited by colons ROOTPW=my_secret_password the default unencrypted root password. eg: ROOTPW=sEcrEt CROOTPW=my_crypted_password the encrypted password from /etc/shadow (taken from an other machine) DEBUG= if set to any value you will get a shell during installation to debug if you encounter problems. FORCEDASDFORMAT= if set to 'yes', the rhsetup will not ask for confirmation if the DASDs from the MOUNT variable should be formatted INTERACTIVE= if set to 0 (zero) and all needed parameters are given in the .prm file, the rhsetup script will be called automatically, therefore the telnet/login step may be omitted and a complete installation without user interaction is possible. - The parameter file should contain NO MORE THAN 80 characters per line. - The vm reader has a limit of 11 lines for the parameter file (880 characters). - The DASD and the tape IPL have a limit of 896 characters (include/asm/setup.h and arch/s390/kernel/head.S). - If you omit any of these parameters, you will be prompted for them at installation time. Mounting a DASD in the installation program / Rescue mode: ========================================================== Sometimes you may have to mount a DASD in the installation program manually. This is especially true when the Linux system does not boot because because of a boot loader error. You do not have to reinstall the entire system. Simply start the installation program and log on via telnet or ssh, then load the dasd_mod and the dasd_eckd_mod modules: insmod dasd_mod insmod dasd_eckd_mod Now you can mount your DASDs on /mnt/sysimage (or a newly created directory) and run chroot to that directory. Complete your tasks and exit from the chroot environment. Umount the DASDs and IPL the machine. Example For Parameter Files: ============================ Install in a LPAR: ------------------ root=/dev/ram0 ro ip=off HOST=lpar.redhat.de:eth0:172.16.2.176 NETWORK=172.16.2.0:255.255.255.0:172.16.2.255:172.16.2.1 DNS=172.16.2.2:172.16.2.15 SEARCHDNS=redhat.de:redhat.com RPMSERVER=172.16.2.140:/ftp/pub/redhat/linux/s390 MOUNTS=/dev/dasda1:/,/dev/dasdb1:/usr/share INSTALL=default DTZ=CET ROOTPW=sEcrEt FORCEDASDFORMAT=yes INTERACTIVE=0 Install in VM: -------------- root=/dev/ram0 ro ip=off DASD=200-20f CHANDEV=ctc0,0x600,0x601 HOST=s390.redhat.de:ctc0:172.16.3.225:192.168.20.1:1492 DNS=172.16.2.2:172.16.2.15 SEARCHDNS=redhat.de:redhat.com RPMSERVER=ftp://172.16.2.140/pub/redhat/linux/s390 MOUNTS=/dev/dasda1:/,/dev/dasdb1:/usr/share INSTALL=default DTZ=CET Common System Administration Tasks After Installation: ====================================================== - "chkconfig" can enable services including xinetd - "setup" can be used to configure your machine. Check especially for "locale_config" and "timetool" after the installation. - Configure XDMCP for remote X: * Set the desktop to 'AnotherLevel'. Edit /etc/sysconfig/desktop and change the line DESKTOP=xxxx to DESKTOP=AnotherLevel or * Edit /etc/X11/prefdm and search for the line if [ -n "$preferred" ] && which $preferred >/dev/null 2>&1; then Add the following line: preferred=xdm ABOVE it. * Edit /etc/X11/xdm/xdm-config and comment out the line DisplayManager.requestPort: 0 by placing a ! in front of it. * Edit /etc/X11/xdm/Xaccess and remove the first # from the following line # * #any host can get a login window