April 2008 - openSUSE Autoinstall - openSUSE Mailing Lists

linuxrc - netdevice=<hwaddress>
by John Willis 18 Dec '12

18 Dec '12

Does anyone know if any of the SLES releases 9 or 10 support passing the linuxrc option netdevice=<hardware mac address> ? I noticed Red Hat has had this for sometime and it seems totally missing in SuSE which seems really odd. I'm still scratching my head if we really are still passing only by kernel interface name. I read in the change log for linuxrc-1.8.6-3 a change that "sounds" like this feature. * Mon Mar 14 2005 - snwint(a)suse.de - s390: fix segfault in net_list_s390_devs() - extra root image was sometimes not used - allow netdevices to be specified by hwaddress But it looks like linuxrc-1.8.6-3 is only available in SuSE Professional 9.3. "any" commentary would be sincerely appreciated the feature also seems to be fairly minor since any delimiter seperating the octets in the hwaddr string could be used to identify the difference between an interface name and interface hw address if i could figure out a generic method of overriding the linuxrc with a custom script and daisy chain it back into linuxrc after setting up the interface name, I would.. and have a slight idea on how to do that.. but i'd rather find out i'm just missing something and the feature is already there Thanks!

3 2

[opensuse-autoinstall] /root/autoinst.xml
by LDB 08 May '08

08 May '08

When you install manually you can choose to leave a /root/autoinst.xml. When I am using autoinstallation, what do I do? Thanks, LDB -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

2 3

[opensuse-autoinstall] Custom autoInstall boot CD for multiple distributions (SLES 9 & SLES10)
by Adam Hawks 05 May '08

05 May '08

Is it possible with AutoYast and CD install source to have different options in the isolinux.cfg that will autoinstall DIFFERENT releases. I have a product that installs on top of SLES and supports both SLES 9 and SLES 10. My customer base is mostly hardware maintenance folks not computer users so automation in necessary I'm trying to make my software install using autoinstall with a custom boot CD. I plan to have the "isolinux.cfg" present a list of machine hardware models and SLES version combination for them to select from and the rest be automated. I'm having trouble getting my custom boot CD to boot and start yast and then ask for the CD1 of SLES 9 or SLES 10 depending on which "isolinux.cfg" option was selected. It correctly boots and starts yast with the correct config file but then fails when it reaches "Configure Software Selections" and pops up the "Initializing catalogs" window. The error window contains: Error occurred while creating the catalog. cd:///?devices=/dev/hdc Source rejected by the user Retry Abort I was expecting it to ask for the SLES 9 CD #1 at this point but it does not. I have tried various combinations of options in the "content" files and in the "yast/order" and "yast/instorder" files to no avail. At this point I'm in need of help on how or IF this is even possible. My CD layout is as follows with some contents specific to the SLES 9 selection /media.1/media products /XPONSLES10/content 686850U.xml 68685AU.xml 8639PTW.xml /XPONSLES9/content 686850U.xml 68685AU.xml 8639PTW.xml /SLES9/yast/order instorder /boot/i386/root.fonts root loader/linux text4c.pcx 16001200.spl isolinux.bin isolinux.cfg initrd 16x16.fnt text2b.pcx text4a.pcx back.jpg text1.pcx bootlogo kroete.dat text3.pcx message text2c.pcx text4b.pcx memtest welcome.jpg en.tr 10240768.spl directory.yast timer_a.jpg text2a.pcx en.hlp 12801024.spl 14001050.spl 08000600.spl The files in the boot/i386/ folders are from SLES 10 SP1 CD #1 except for isolinux.cfg which contains default Install_3890XP_on_6868_50U_with_SLES9 # install label Install_3890XP_on_6868_50U_with_SLES9 kernel linux append initrd=initrd splash=verbose install=cd autoyast=file:///XPONSLES9/686850U.xml showopts # install label Install_3890XP_on_6868_50U_with_SLES10 kernel linux append initrd=initrd splash=verbose install=cd autoyast=file:///XPONSLES10/686850U.xml showopts ... ... ... The media.1/media file contains IBM 3890XP Control Program 20080317141526 1 The media.1/product file contains /XPONSLES9 IBM-3890XP-SLES9 2.0-21 /XPONSLES10 IBM-3890XP-SLES10 2.0-21 The SLES9/yast/order file contains / /SLES9/ /XPONSLES9 /XPONSLES9/ The SLES9/yast/instorder file contains /SLES9 /XPONSLES9 Thanks Adam W. Hawks T/L: 794-4033 Office: 704 594-4033 -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

4 3

[opensuse-autoinstall] pre-install scripts before or after partitioning?
by Jim Michaud 30 Apr '08

30 Apr '08

In the custom user scripts section of the Autoyast documentation, it says that the pre-install scripts run "(Before partitioning and package installation but after the hardware detection)" However I believe for suse 10.3 this is incorrect. Can anyone confirm that? I was having a problem with an 10.3 install that I traced back to be with an incorrect MBR from a previous install, so I put the following line in the pre-install script in order to wipe out the MBR. test -b /dev/sda && dd if=/dev/zero of=/dev/sda bs=512 count=1 I realize this also wipes out the partition table, but I was assuming this was done before partitioning, so the install process would write the new partition table. But I get an error part way through the install complaining the disk has no partitioning (which I did confirm). So is this a difference in 10.3 or do I need some other syntax in the autoyast xml file ---snippet from autoyast xml file --- <pre-scripts config:type="list"> <script> <debug config:type="boolean">true</debug> <filename>pre-install.sh</filename> <interpreter>shell</interpreter> <location>http://10.0.0.20/sys/pre-install.sh</location> </script> </pre-scripts> ------ -- Jim Michaud SiCortex IT Director -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

3 3

[opensuse-autoinstall] Multiple custom variables together?
by Mike Marion 25 Apr '08

25 Apr '08

I wrote up a quick custom rule to solve my last issues (about matching different multiple disk layouts) but I seem to have bumped into another one. Is it not possible to use multiple custom variables on the same line? i.e. <profile>@custom1@_@custom2@.xml</profile. is matching between the farthest @ signs so the variable it's trying to expand is "custom1@_@custom2" Patching the regexp in the autoyast module would be my next step if the above is true and mixing variables like that isn't possible. This is on sles10-sp1. Google searching didn't turn up much beyond the autoyast docs and some threads on this list I was already in, about 1 variable. -- Mike Marion-Unix SysAdmin/Staff IT Engineer-http://www.qualcomm.com [Homer left a 'Dear Jane' tape for Marge] Homer: "Marge, if you're watching this then it means I figured out how to work the camera." ==> Simpsons -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

2 3

[opensuse-autoinstall] rules in rules.xml not and-ing
by Mike Marion 25 Apr '08

25 Apr '08

I just ran into my first case where I had a situation that seems to be triggering a bug. I have several rules to get partition layout files based on the disk(s) seen, and have a case where a vendor replaced a drive and put it on the 2nd ata channel so it's hdc vs hda. This shouldn't technically matter and it should come through as a single disk, however the following rule: <rule> <disksize> <match>/dev/hda 10000</match> <match_type>greater</match_type> </disksize> <disksize> <match>/dev/hdc 10000</match> <match_type>greater</match_type> </disksize> <result> <profile>@custom1@_dideac.xml</profile> <continue config:type="boolean">false</continue> </result> <operator>and</operator> </rule> Is matching purely based on the size/existence of hdc. I just added the operator tag to see if it helped even though it's supposed to default to that. y2log shows: 2007-05-21 15:05:40 <1> 10.222.149.124(3424) [YCP] AutoInstallRules.ycp:390 Bash return: if ( [ "$disksize_size0" -gt "10000" -a "$disksize_device0" = "/dev/hdc" ] ) ; then exit 0; else exit 1; fi ($["exit":0, "stderr":"", "stdout":""]) ($["custom1":"compute", "disksize_device0":"/dev/hdc", "disksize_size0":38166, "memsize":4032]) (note: yeah the clock is way off at this point) I never even see it looking at hda... in either the above test or an identical one that looks for both hda and hdb. If I do have a host with both hda and hdc, it works as I would expect. Or at least it always did, but I haven't done one of those in ages so I'm not entirely sure the rules looked at both or just compared hdc like above and it just worked since the resulting partition file was valid for the disks in the host. I realize now that it's completely throwing away the first <disksize> in each rule... apparently you can't do the same type of attribute in the same rule? Does someone have an example of checking more then 1 disksize in the same rule? I don't see this in the autoyast docs. I freely admit I might just be missing it though. I guess I've been lucky that the rules have been working like this thus far. -- Mike Marion-Unix SysAdmin/Staff IT Engineer-http://www.qualcomm.com The Matrix is going down for reboot now! Stopping reality: ....OK The system is halted. -- yet another sig stolen from /. -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

2 1

[opensuse-autoinstall] Corrupt RPM database after autoyast installation
by Ragnar Kjørstad 25 Apr '08

25 Apr '08

Hi We're experiencing a problem with RPM database corruption after installing a system with autoyast. So far we've only been able to trigger the problem with SLES9 SP4, but the problem only occurs once in a while so it's possible it's present with the other SUSE distributions only we haven't been able to trigger it yet. The problem only occurs if we install additional RPM packages in the chroot hook in the autoyast file. E.g.: if we do: rpm -ivh http://172.11.0.111/scali/repository/os_sles9_sp4_i386/sp4-i386/CD1/suse/i5… in the chroot section, after installation "rpm -qa" report ypserv beeing installed but "rpm -q ypsev" says it's not installed: # rpm -q $(rpm -qa) | grep "not installed" package ypserv-2.12.1-44.5 is not installed # Is this a known bug? Are you supposed to be able to install rpm packages like this in the chroot section, or could there be problems with autoyast and rpm both making changes to the rpm database at the same time -- Ragnar Kjørstad Software Engineer Platform Computing -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

1 0

[opensuse-autoinstall] How to get local patch repository
by Beat Rubischon 23 Apr '08

23 Apr '08

Hello! I was asked by a customer if it's possible to create a site local patch repository of SLES10 packages. Is there an official way to "mirror" the distribution? Is a special contract like a site license with Novell needed? Beat -- \|/ Beat Rubischon <beat(a)0x1b.ch> ( 0^0 ) http://www.0x1b.ch/~beat/ oOO--(_)--OOo--------------------------------------------------- Meine Erlebnisse, Gedanken und Traeume: http://www.0x1b.ch/blog/ -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

2 1

[opensuse-autoinstall] remove-packages section
by Hans-Robert Vermeulen 10 Apr '08

10 Apr '08

Hi all, Does anyone know how the remove-packages section is determined? We are installing OES2 and SLES servers and the default generated autoyast profile contained the section at the end of this email. All we did was remark the NCS package, as stated in the documentation. This has been running fine on all my customers hardware and on VMware in both 32 and 64 bits machines. Now however, we have a new DL360 G5 machine and the installer fails. It tunrs out I need to remark the following 3 packages as well to fix it: <package>numactl</package> <package>mcelog</package> <package>linux32</package> Needless to say, this took some time to figure out.... My question is how this list is determined and if it is save to leave the entire <remove-packages config:type="list"> section out, so the server can determine it for itself. HRV <remove-packages config:type="list"> <package>mDNSResponder</package> <package>gnome2-SLES</package> <package>evolution</package> <package>gnome2-SuSE</package>  <package>pfmon</package> <package>sles-preparation-power_en</package> <package>sles-preparation-zseries_en</package> <package>powerpc32</package> <package>s390-32</package> <package>s390-tools</package> <package>salinfo</package> <package>pdisk</package> <package>numactl</package> <package>mcelog</package> <package>libpfm</package> <package>linux32</package> </remove-packages> Hans-Robert Vermeulen -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

3 2

[opensuse-autoinstall] Need an Opensuse 10.2 Autoyast Installer 'expert' for a project
by Tim Stoner 08 Apr '08

08 Apr '08

Hi OpenSUSE List readers, I need to have an installer app made that automates all the steps below. Nobody at my work knows much about autoyast and I would prefer to have a 'expert' just work it out. Do you have experience doing this kind of thing and are possibly willing to do an installer? If so, write me back. I hope that I am not off-etiquette posting a request like on this list. Its the obvious place to go to look for someone who knows how to do this. Thanks! Tim tim123 -a-t- asentria.com Note: the 'DAN' and the 'EN' refer to two network interfaces on the Suse system, one connects to the internet and the other to the LAN. Here is the install I want to automate (this is the install instructions written up by a developer): Initial boot 1. Boot with openSUSE 10.2 installation media. Keep in mind your machine BIOS needs to be told to boot from the optical media drive if the hard drive is blank. Most machines do this by default but if you find that the OS installation media is not booting, you will likely have to resolve the problem by modifying the BIOS settings. 2. Upon boot, at the next screen, select "Installation" 3. At the next screen, select the language and click Next. 4. At the next screen, accept the license agreement and click Next. 5. At the next screen, select "New Installation" and click Next. 6. At the next screen, select region and time zone, ensure the clock is set to the correct time, and click next. 7. At the next screen, select the Desktop (KDE) and click next. 8. The next screen is titled "Installation Settings". In the "Overview" tab, there are two links for "Partitioning" and "Software". You configure these two areas by clicking on those words. First click on Partitioning. 9. Select "Base Partition Setup on This Proposal" and click Next. 10. Click on the partition with the "Mount" entry set to "/home", and then click on the Edit button. 11. In the dialog box that appears, look for the "Mount Point" drop-down box ("/home" should be selected) and change this selection to "/opt". 12. Click the OK button to proceed. 13. Click the Accept button. This completes the partitioning setup. 14. You should be back at the "Installation Settings" screen where you can now click on "Software". Click on "Software" now to configure the software selection. 15. Click on the Details button (near bottom left of screen) and then make the following changes: Discard unneeded packages 16. In the right pane, click on the checkmark next to iptables such that it is "never installed"; i.e., instead of a check next to iptables, there is a red circle with a white bar through it. 17. In the left pane, never install Enterprise Software Management (this is in the "Base Technologies" section). 18. In the left pane, scroll the the "Proprietary Software" section and never install both Java Environment and Misc. Proprietray Packages. 19. In the left pane, highlight Misc. Proprietary Packages by clicking on it (not the red circle with the line through ti) 20. In the right pane, never install AdobeICCProfiles and agfa-fonts. Select required packages 21. In the left pane, scroll to the "Development" section and click on the checkboxes next to Linux Kernel Development and C/C++ Developement, so that a check appears next to each. 22. In the left pane, scroll to the "Server Functions" section and select Network Administration -- don't click on the checkbox next to Network Administration, just click on the phrase "Network Administration" so it is highlighted. This is what is meant by "select Network Administration". 23. In the right pane, click on the checkbox next to wireshark. 24. In the left pane, scroll to the "Base Technologies" section and select Console Tools. 25. In the right pane, click on the checkbox next to nano. Check dependencies 26. Click on the Check button at the bottom of the screen. 27. A dialog box appears with dependcency conflicts. For all of them, select "ignore this requirement just here" and then click the "Ok, Try Again" button. 28. In the "Filter" drop-down box, select "Package Groups" 29. A tree appears in the left pane. In this tree, navigate to the Development-Languages-Java branch. In the right pane, never install java-1_4_2-gcj-compat. 30. In the left pane, navigate to the Development->Tools branch. In the right pane, click the checkbox next to "expect" to install it. 31. In the left pane, navigate to the Productivity->Security branch. In the right pane, click the checkbox next to "pwgen" to install it. 32. Click on the Accept button at the bottom right of the screen. 33. A dialog box appears with a dependcency conflict. Select "ignore this requirement just here" and then click the "Ok, Try Again" button. Continue with installation 34. In the dialog box "Changed Packages" that appears, click the Continue button. 35. At the next screen (the Installation Settings screen), click the Accept button. 36. In the dialog box "Confirm Installation" that appears, click the Install button. 37. Wait while the machine performs installation; the machine automatically reboots at the end of this step. 38. At the next screen, choose a root password and click Next. 39. At the next screen, choose a hostname and domain and click Next 40. At the next screen, "Network Configuration", do the following: 41. Disable the firewall by clicking on "enabled" in the Firewall section (The applications controls firewall functions; this is why the OS firewall should be disabled). 42. Disable IPv6 by clicking on "Disable IPv6" in the IPv6 section. 43. Click on Network Interfaces to configure the two network interfaces of the machine: one of the interfaces sits on the EN (Enterprise Network) and the other interface sits on the DAN (network reachable by deployed units). This may be the Internet or some other network; what is required in this case is that the IP 44. address of the DAN interface is the same one used by deployed units to reach the Server. 45. When back at the Network Configuration screen, click Next. 46. At the next screen, select "No, Skip This Test" and click Next. Select No on the Additional Installation Sources prompt. 47. At the next screen, select your authentication method and click Next (Local is default, leave it at that if you are unsure). 48. At the next screen, fill in the info for the non-root user and disable automatic login. Keep in mind that this machine runs administrative serivces only, and those can only be maintained by root. Unless you want to use sudo all the time, having a non-root user has no practical purpose other than to be needed to proceed past this installation step. After the installation you may disable this user. 49. At the next screen (the Release Notes), click Next. 50. At the next screen (Hardware Configuration), review any specific configuration requirements (there should not be anything you must change here) and click Next. 51. At the next screen (Installation Completed), click Finish. The next screen is a login screen; log in as root and proceed to the next section. Configure Fundamentals 52. Now that the OS is installed it is time to set up certain fundamentals before installing the server application. Setup time service and default route 53. Both of these are configured with the YaST program. Start YaST (after logging in as root) by clicking on the right-most icon among the icons in the left side of the taskbar. Time service 54. When YaST is started, in the left pane, select Network Services, then in the right pane, select Network Service (xinetd). 55. Select the Enable radio button near the top of the screen. 56. In the list, scroll down to the "time" service of "stream" type and highlight it. 57. Click the "Toggle Status (On or Off)" button. 58. Click the Finish button. Default route 59. If deployed units go through a router to communicate with the server application then you likely need to enter the default route of the DAN interface of the application. Do this with the following steps. 60. When YaST is started, in the left pane, click Network Devices, then in the right pane, click Network Card. 61. A choice apears to select the method of configuration; leave it on the default of "Tranditional Method with ifup" and click Next. 62. Your network interfaces appear here. Click the Edit buttton; it does not matter which interface is selected. 63. Click the Routing button 64. Fill in the default route where prompted. -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

1 0

[opensuse-autoinstall] dns resolution during install
by Michael Welle 02 Apr '08

02 Apr '08

Hi, is there a way to use server names instead of the ip addresses for the installation source? Snipset from linuxrc.conf: #Nameserver: 192.168.198.100 #Gateway: 192.168.198.100 #Netmask: 255.255.255.0 #Broadcast: 192.168.198.255 Install: smb://DEMOSRV.demo.dom/share/foo AutoYaST: file:///autoyast.xml This results in the message 'dns: what is demosrv.demo.dom'. Even pinging the server doesn't work. If I exchange the server name with its ip address everything works as expected. The client gets its network parameters via dhcp. /etc/resolv.conf etc. seems to be properly filled. Next I added some network parameters to linuxrc.conf. Doesn't change the behaviour. Even worse in the summary printed on the console the broadcast address is now noted as 0.0.0.255. Any hints? Michael -- biff4emacsen - A biff-like tool for (X)Emacs http://www.c0t0d0s0.de/biff4emacsen/biff4emacsen.html Drop spots? http://dropspots.org/spot/1812 -- To unsubscribe, e-mail: opensuse-autoinstall+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-autoinstall+help(a)opensuse.org

2 3