February 2004 - openSUSE Autoinstall - openSUSE Mailing Lists

How to create a single CD for SLES (+ SP3) installation?
by Robert Pollak 09 Feb '04

09 Feb '04

Hi, to install a SLES "Minimum system" with SP3, normally three CDs are needed: - UnitedLinux 1.0 SP3 CD1 (to boot with) - SLES8 Installation CD - UnitedLinux 1.0 CD1. I want to create a single CD that combines the necessary data s.t. installing only from this CD gives the same "Minimum system" of SLES-SP3. (Adding autoinst.xml would then yield an install process without any necessary interaction, without the need of a network connection (as with nfs install).) Can this be done with AutoYaST2's "Create ISO Image", or do i have to collect the data and call mkisofs manually (as sketched in http://lists.suse.com/archive/suse-autoinstall/2003-Apr/0047.html "Self made CD-ROMS")? Anas Nashif wrote in "Re: No Kernel SLES8 SP3 autoinstallation", 2004-01-16: > The autoyast2 interface does not support creating CD for SLES8 > installation or in general terms, a CD from multiple sources. So therefore i tried to combine the three CDs' contents into one source (similar to "Create a Packagedatabase", http://lists.suse.com/archive/suse-autoinstall/2003-Oct/0067.html ): Working on a SLES-SP3 minimum system as described at the beginning, I created suse/{i586,i686,noarch} and filled it by copying the contents of - UL CD1: UnitedLinux/{i586,i686,noarch} - SLES CD: suse/{i586, noarch}, - UL SP3 CD1: UnitedLinux/{i586, i686, noarch} . Then i called cd suse; /usr/lib/YaST2/bin/create_package_descr -d . This gives me "processed 1069 packages", and the setup/descr/packages file. Next step would be to add the suse directory as Software Source to yast2, but this only gives me: ERROR(InstSrc:E_no_instsrc_on_media). How do i have to prepare the directory for use as software source? -Robert P.S.: Or should i forget about SP3 and try to manually extend the SLES8 installation CD contents by data from UL-CD1 to get a self-contained CD? (This would give me a single installation interaction: accepting the license.)

2 1

Network Card autoinstall problem
by Bernhard Nöggerath 07 Feb '04

07 Feb '04

Hi, I use SuSE Linux9 AutoYast for installation systems from my own created CD. The destination systems have different network cards, graphic cards, etc. Installation and configuration of the new system works fine incl. Graphic card, X11 etc. During the configuration also the network card is found, configured and useable. But after rebooting the system, the network is not useable since the module was not loaded. It seems to me that just the /etc/modules.conf was not setup right during the autoinstallation. Is there something wrong, or is there a way that I can do "repair" this with one of my post(chroot)-install scripts ? Regards, Bernhard -- | SAM Electronics GmbH | Dipl.-Ing. Bernhard Nöggerath | Software Navigation Systems, Dept. MST 21 | Kurfürstenallee 130 | D-28211 Bremen, Germany | | Phone: +49 421 1784 3513 | Fax: +49 421 1784 2026 | mailto:Bernhard.Noeggerath@sam-electronics.de

2 1

Driver update media problems with SLES 8 SP3
by Andrew Wray 07 Feb '04

07 Feb '04

All, I've been trying to setup a network install (over NFS) for SLES 8 with service pack 3 integrated. I've updated the boot directory with the SP3 files in the root of the install source, and also put the linux directory from SP3 CD1 into an 02/linux directory in the root of the install source. The driverupdates file is there as well. The driver that I need is megaide.o, for an LSI IDE RAID adapter. I've taken the package that LSI provides for SP2a, recompiled the drivers against SP3, repackaged the update.tar.gz file, and then placed it under the 01/linux directory (the BootGRUB.ycp and inst_finish.ycp files were included in the megaide driver update media package, so I thought I should install the SP3 linux directory updates after, so its BootGRUB.ycp and inst_finish.ycp files would overwrite the older ones from the megaide package). The install goes OK, the megaide.o driver is loaded, and the RAID device is seen and installed to, and the SP3 files are installed as well, however, when the system reboots, the megaide module is not on the initrd, so it fails to mount root. I checked the output from mk_initrd during the install, and it reports "Cannot determine dependencies of module megaide\Is modules.dep up to date?". I checked into this further, and it looks like the mk_initrd from SP3 looks at /lib/modules/<kernel-version>/modules.dep, which is not setup by the update.pre script (which is basically just a copy from the example given in the driver update howto), or the update.post script. It is setup in the /etc/modules.dep file though. I modified the update.post script to append the right dependency line in the /lib/modules/<kernel-version>/modules.dep file, and then the install completed OK. However, my question is, does the driver update howto need to be updated to explain how to do this for SP3, or is this a mk_initrd bug? It seems that the way I've set this up, just outputting to the specific kernel that I know will wind up on the system, post-install, is not very portable either--it would fail if a different kernel was installed (on a different system perhaps). Thanks, Andy Wray

1 0

Re: [suse-autoinstall] Cannot mount NFS repository automatically ? SOLVED (maybe)
by francois.blaes＠eu.effem.com 05 Feb '04

05 Feb '04

The problem has just been solved, apparently The server is fitted with a bcm5700 card, which takes a lot of time to come up (more than 40 seconds) probably because of un-optimized switch settings. That timeout is too long for linuxrc to get the network and map the NFS server. However, after a while when the network finally comes up, the NFS volume *can* be mounted and the manual install proceed, as long as one answers "no" when prompted to reconfigure the NIC by linuxrc ("Your network is already configured. Do you want to change this ?") Otherwise, should one answer "yes I want to change my settings" , it looks like linuxrc would switch down and reconfig the NIC, thereby inducing another delay and screw up the NFS mount again. Francois BLAES ? Global Unix Team ? Mars ISI DDI *170 1639 ? Tel +33 (0) 3 88 05 16 39 ? Fax +33 (0) 3 88 05 16 79

2 1

slec8 & xd2
by Grant Williamson 05 Feb '04

05 Feb '04

Hi, I am wondering if anyone has a autoyast config to install a base slec8 & xd2 system. -- Grant Williamson <traxtopel(a)fastmail.fm>

1 0

does yast/instorder matter? Syntax errors in examples.
by Hal Wine 04 Feb '04

04 Feb '04

In some recent verification tests of my http installation server for SLES8 with Service Pack (both 2a & 3), it appears that the contents of the yast/instorder file do not affect the resulting system image. NOTE: this file is documented as having only one field. Many of the examples show 2 tab delimited fields. When you use a tab delimited file, the yast/instorder file appears to be ignored, and the displayed order of sources is as described in the yast/order file. While the listing of source order, as displayed during installation, does differ between a valid yast/instorder and an invalid one, there are no differences in the resulting systems (as measured by comparing 'rpm -qa' output from both variations). As I had read the documentation's description of the purpose of the yast/instorder file, I did expect some differences w.r.t. the yast2 packages. --Hal

3 2

rules.xml and merging
by Voss Stefan 04 Feb '04

04 Feb '04

Hi, at the moment our auto installation does read a non-machine specific generic xml file (via autoyast=...) and a machine specific xml file determined by the following rules.xml: <?xml version="1.0"?> <!DOCTYPE autoinstall SYSTEM "/usr/share/autoinstall/dtd/rules.dtd"> <autoinstall xmlns="http://www.suse.com/1.0/yast2ns" xmlns:config="http://www.suse.com/1.0/configns"> <rules config:type="list"> <rule> <custom1> <script> <![CDATA[ #!/bin/sh # Try to determine value of boot parameter "sysname" sysname=$(cat /proc/cmdline | perl -ne 'print /sysname=(\w+)/') echo -n $sysname ]]> </script> <match>machine1</match> <match_type>exact</match_type> </custom1> <result> <profile>machine1.xml</profile> <continue config:type="boolean">false</continue> </result> </rule> </rules> </autoinstall> Without rules.xml (i.e. with the generic xml file alone) the installation works ok. With rules.xml (and thus with the machine specific machine1.xml) autoyast2 suddenly complains about wrong partition plans though there is no partition plan in the machine specific xml file. The partition plans are only in the generic xml file. Here's the machine specific xml file: <?xml version="1.0"?> <!DOCTYPE profile SYSTEM "/usr/share/autoinstall/profile.dtd"> <profile xmlns="http://www.suse.com/1.0/yast2ns" xmlns:config="http://www.suse.com/1.0/configns"> <configure> <networking> <dns> <hostname>machine1</hostname> </dns> </networking> </configure> </profile> So the merge of the generic xml file and the machine specific xml file is somehow wrong. How can the result of the merge be checked? Where is the result of the merge stored? Regards, Stefan Voss

1 0

SLES8 (with SP3), /boot on /dev/md0 with grub
by Andreas Kurz 04 Feb '04

04 Feb '04

Hello all! I´m having some problems installing the grub-bootmanager on my boot-device (is a software-raid1). /boot is also missing in the /etc/fstab once the installation is complete but the raid1 is correctly configured and formated. Does someone know how I can get the 'chroot-scripts' working? I also found no way to prevent the reboot at the end of the installation .... very strange Here are some parts of my xml file: ...... <chroot-scripts config:type="list"> <script> <filename>/tmp/grub.sh</filename> <interpreter>shell</interpreter> <source> <![CDATA[ #!/bin/sh mkinitrd -m "raid1 lvm-mod xfs" -d /dev/rootvg/rootlv -b /boot / cat << MENU > /boot/grub/menu.lst color white/blue black/light-gray default 0 timeout 10 title linux kernel (hd0,0)/vmlinuz root=/dev/rootvg/root_lv acpi=oldboot initrd (hd0,0)/initrd title failsafe kernel (hd0,0)/vmlinuz.shipped root=/dev/rootvg/root_lv ide=nodma apm=off acpi=off vga=normal nosmp noapic maxcpus=0 3 initrd (hd0,0)/initrd.shipped MENU cat << GRUB > /etc/grub.conf root (hd0,0) setup (hd0) root (hd1,0) setup (hd1) quit GRUB grub --batch < /etc/grub.conf exit 0 ]]> </source> </script> </chroot-scripts> ...... <install> <bootloader> <write_bootloader config:type="boolean">false</write_bootloader> </bootloader> .... <mode> <confirm config:type="boolean">false</confirm> <reboot config:type="boolean">false</reboot> <forceboot config:type="boolean">false</forceboot> <interactive_boot config:type="boolean">false</interactive_boot> </mode> Regards, Andreas Kurz -- Andreas Kurz | BearingPoint INFONOVA GmbH | Systems Engineer, Systemintegration | Seering 6, 8141 Unterpremstaetten/Graz, Austria Phone +43 316 8003 2933 | Fax +43 316 8003 1080 mailto:andreas.kurz@bearingpoint.com | www.bearingpoint.com | www.infonova.com --- BearingPoint INFONOVA GmbH is a subsidiary of BearingPoint GmbH, Frankfurt (formerly KPMG Consulting AG). The information in this email is confidential and may be legally privileged. Access to this email by anyone other than the intended addressee is unauthorized. If you are not the intended recipient of this message, any review, disclosure, copying, distribution, retention, or any action taken or omitted to be taken in reliance on it is prohibited and may be unlawful. If you are not the intended recipient, please reply to or forward a copy of this message to the sender and delete the message, any attachments, and any copies thereof from your system.

1 0

Install program exit code is 134
by Wallace, Andrew 04 Feb '04

04 Feb '04

Hi, I'm trying to get autoyast working on suse-8.2, using DHCP/PXELinux/NFS, and I'm having problems. The PXE boot works OK, the NFS mount works, I assume the autoinst.xml is valid as I wrote it using Yast2... Following the progress on tty3, I see it loading up a load of modules, then it gets this far (see below) and stops. OK, going to mount 192.168.0.1:/builds/suse... NFS mount ok Setting up localhost...done going for automatic install mount: /var/adm/mount/boot/root: we need a loop device mount: using /dev/loop0 rmmod usb-storage starting /mounts/instsys/sbin/inst_setup yast [pauses for a minute or 2] install program exit code is 134 sync...ok That's it - installer dies, no idea why. Anyone here know? Cheers, Andy Wallace

1 0

Cannot mount NFS repository automatically ?
by francois.blaes＠eu.effem.com 04 Feb '04

04 Feb '04

Hi all I am facing a fairly weird issue with autoyast and its NFS repository ... any help appreciated I use a custom boot CD, mostly based on the SLES8 one, just a few pre-configured isolinux options to specify the NFS repository location : autoyast=nfs://<ip address of the repository server>/<full path to directory holding the XML configuration files> install=nfs://<ip address of the repository server>/<full path to packages repository> At the prompt, I add the following ad-hoc parameters : hostip=<ip address of the server to be installed> netmask=<netmask of the server to be installed> gateway=<gateway of the server to be installed> The boot process goes well, and linuxrc starts, but fails after a few seconds with "an error occured while trying to mount the nfs directory : reason given by server unknown error -1" I retried multiple times using the manual install mode, but unsuccesssfully However, when I switch to the console with Alt F9, I can ping the NFS server I can even **manually** mount anything from the NFS server using mount -t nfs /<ip address of the repository server>/<full path> /mnt which assumingly proves that all network and NFS server settings are correct ? On other hosts, the same install runs smoothly without any problem So it can't be a question of permission, can it ? How can linuxrc not succeed to mount the NFS volume whilst I can ? What would linuxrc use as mount point to mount the NFS volumes passed as parameters ? Francois BLAES ? Global Unix Team ? Mars ISI DDI *170 1639 ? Tel +33 (0) 3 88 05 16 39 ? Fax +33 (0) 3 88 05 16 79

1 0