August 2005 - openSUSE Autoinstall - openSUSE Mailing Lists

Display the currently loaded config-file
by Guido Kölsch 18 Aug '05

18 Aug '05

Hi, is it possible to add a label to the autoyast-module, where the currently loaded xml-file is displayed? It's a bit confusing if you have a large amount of files to edit. thanks, -- ------------------------------------------------------------------------ Guido Kölsch Max-Planck-Institut fuer Radioastronomie Phone: +49 (0)228 525 164 Auf dem Huegel 69, D-53121 Bonn, Germany FAX : +49 (0)228 525 229 email: gkoelsch(a)mpifr-bonn.mpg.de ------------------------------------------------------------------------

2 1

SLES9: Reread of partition table failed
by John Clemens 18 Aug '05

18 Aug '05

I've got fully automated SLES9 installations almost working with our software (we provide a virtualized system, with virtual network and virtual disk). The only problem I've yet to solve is that the install pauses halfway through with the above error, making the user press ok to continue. When you do press ok, the install works fine, and the right partitions are set up and installed. Well, ok, that's not the only problem, but the others (network config not getting setup even though specified in the xml, and inittab trying to run getty's on tty1-6 even though its a serial port install) I think are solvable. This one however.. Looking at the yast2-storage source, it appears it's searching the fdisk output for WARNING (which fdisk prints out if the BLKRRPART ioctl fails). however, if i run fdisk manually on the system, and set up the partitions exactly the same way, I don't see that error. I've even gone so far as to mount the root install image via loopback/chroot and run that version of fdisk, and i still don't see that error message. So I don't know why yast, which seems to be reading the output of fdisk(?) would fail and trigger this dialog. Also, looking at the <report> section of my xml file, I see that errors, messages, and warnings are all set to timeout in 10 seconds. Why does this message not time out? Any thoughts? I'm including the partition section of the XML below my sig if it helps. Checking the archives it looks like something like this has happened before, but it seems it should have been solved years ago. I'm guessing I'm doing something wrong.. john.c - -- John Clemens jclemens(a)virtualiron.com <partitioning config:type="list"> <drive> <device>/dev/sda</device> <initialize config:type="boolean">true</initialize> <partitions config:type="list"> <partition> <crypt>twofish256</crypt> <filesystem config:type="symbol">reiser</filesystem> <format config:type="boolean">true</format> <loop_fs config:type="boolean">false</loop_fs> <mount>/</mount> <partition_id config:type="integer">131</partition_id> <partition_type>primary</partition_type> <size>4.3GB</size> </partition> <partition> <crypt>twofish256</crypt> <format config:type="boolean">false</format> <loop_fs config:type="boolean">false</loop_fs> <partition_id config:type="integer">131</partition_id> <partition_type>primary</partition_type> <size>4.3GB</size> </partition> <partition> <crypt>twofish256</crypt> <format config:type="boolean">false</format> <loop_fs config:type="boolean">false</loop_fs> <partition_id config:type="integer">131</partition_id> <partition_type>primary</partition_type> <size>4.3GB</size> </partition> <partition> <crypt>twofish256</crypt> <format config:type="boolean">false</format> <loop_fs config:type="boolean">false</loop_fs> <partition_id config:type="integer">130</partition_id> <partition_type>primary</partition_type> <size>max</size> </partition> </partitions> <use>1,4</use> </drive> </partitioning>

2 4

network driver built into kernel?
by John Clemens 17 Aug '05

17 Aug '05

Hi, another quick question. Our virtual network driver is built into the kernel that we install. I'm like to set this up to use DHCP but default and come up on boot. I have this section in my XML file: <networking> <dhcp_options> <dhclient_additional_options>-B</dhclient_additional_options> <dhclient_hostname_option>AUTO</dhclient_hostname_option> </dhcp_options> <dns> <dhcp_hostname config:type="boolean">true</dhcp_hostname> <dhcp_resolv config:type="boolean">true</dhcp_resolv> <domain>virtualiron.com</domain> <hostname>linux</hostname> </dns> <interfaces config:type="list"> <interface> <bootproto>dhcp</bootproto> <device>eth0</device> <startmode>onboot</startmode> </interface> </interfaces> <routing> <ip_forward config:type="boolean">false</ip_forward> </routing> </networking> I have also tried the above with this included: <modules config:type="list"> <module_entry> <device>eth0</device> <module>kveth</module> <options></options> </module_entry> </modules> In both cases, The autoinstall completes, but on first boot, it fails to configure eth0 (startup scripts after initial yast run, show 'eth0 ... unused'). If I manually go in after bootup and use yast to configure eth0 for dhcp, then it works fine. Looking at /var/log/YaST2/y2log (snippet below) it appears that all of these options are passed to yast2 correctly, and that dhcpcd finds the correct IP information, but because I don't have a module name specified it refuses to write out the configuration...? How can i specify that the network driver is really in the kernel and not to bother with trying to load a module? or is that not the problem? thanks, john.c 2005-09-17 15:55:22 <1> linux(509) [Interpreter] network/install.ycp:389 Calling YaST client lan_auto (arguments: ["Import", $["dhcp_options":$["dhclient_additi onal_options":"-B", "dhclient_hostname_option":"AUTO"], "dns":$["dhcp_hostname": true, "dhcp_resolv":true, "domain":"virtualiron.com", "hostname":"linux"], "inte rfaces":[$["bootproto":"dhcp", "device":"eth0", "startmode":"onboot"]], "routing ":$["ip_forward":false]]]) 2005-09-17 15:55:22 <1> linux(509) [YCP] clients/lan_auto.ycp:14 --------------- ------------------------- 2005-09-17 15:55:22 <1> linux(509) [YCP] clients/lan_auto.ycp:15 Lan autoinst cl ient started 2005-09-17 15:55:22 <1> linux(509) [YCP] clients/lan_auto.ycp:106 InstallInf($["bootproto":"dhcp", "gateway":"10.1.5.1", "hwaddr":"00:0f:4b:00:00:22", "installation":"yes", "ipaddr":"10.1.5.12", "nameserver":"10.1.1.3", "netmask":"255.255.255.0", "type":"eth"]) 2005-09-17 15:55:22 <1> linux(509) [YCP] Lan.ycp:1089 hw=[] 2005-09-17 15:55:22 <3> linux(509) [YCP] Lan.ycp:483 No HW info found: $[] 2005-09-17 15:55:22 <1> linux(509) [YCP] Lan.ycp:488 No detailed HW info: 2005-09-17 15:55:22 <4> linux(509) [YCP] Lan.ycp:503 hw=$[] 2005-09-17 15:55:22 <1> linux(509) [YCP] clients/lan_auto.ycp:218 hw from type = [] 2005-09-17 15:55:22 <3> linux(509) [YCP] network/hardware.ycp:308 Item not found in Hardware: 1 (0) 2005-09-17 15:55:22 <3> linux(509) [YCP] Lan.ycp:483 No HW info found: $[] 2005-09-17 15:55:22 <1> linux(509) [YCP] Lan.ycp:488 No detailed HW info: 2005-09-17 15:55:22 <4> linux(509) [YCP] Lan.ycp:503 hw=$[] 2005-09-17 15:55:22 <1> linux(509) [YCP] Lan.ycp:504 device= 2005-09-17 15:55:22 <1> linux(509) [YCP] network/routines.ycp:266 input=$["confi g":$["dhcp":$["DHCLIENT_ADDITIONAL_OPTIONS":"-B", "DHCLIENT_HOSTNAME_OPTION":"AU TO", "DHCLIENT_MODIFY_RESOLV_CONF":true, "DHCLIENT_SET_HOSTNAME":true]], "device s":$["":$["":$["BOOTPROTO":"dhcp", "STARTMODE":"onboot", "UNIQUE":"", "_type":"" ]]], "dhcp_options":$["dhclient_additional_options":"-B", "dhclient_hostname_opt ion":"AUTO"], "dns":$["dhcp_hostname":true, "dhcp_resolv":true, "domain":"virtualiron.com", "hostname":"linux"], "hwcfg":$["":$["MODULE":"", "MODULE_OPTIONS":"", "STARTMODE":"auto"]], "interfaces":[$["bootproto":"dhcp", "device":"eth0", "startmode":"onboot"]], "routing":$["ip_forward":false]] 2005-09-17 15:55:22 <1> linux(509) [YCP] clients/lan_auto.ycp:388 Lan auto finished . . . YaST client lan_auto (arguments: ["Write", $["dhcp_options":$["dhclient_additio nal_options":"-B", "dhclient_hostname_option":"AUTO"], "dns":$["dhcp_hostname":t rue, "dhcp_resolv":true, "domain":"virtualiron.com", "hostname":"linux"], "inter faces":[$["bootproto":"dhcp", "device":"eth0", "startmode":"onboot"]], "routing" :$["ip_forward":false]]]) 2005-09-17 15:55:24 <1> linux(509) [YCP] clients/lan_auto.ycp:14 --------------- ------------------------- 2005-09-17 15:55:24 <1> linux(509) [YCP] clients/lan_auto.ycp:15 Lan autoinst cl ient started 2005-09-17 15:55:24 <1> linux(509) [YCP] Lan.ycp:291 Writing configuration 2005-09-17 15:55:24 <1> linux(509) [YCP] NetworkModules.ycp:183 Writing configur ation 2005-09-17 15:55:24 <1> linux(509) [YCP] NetworkModules.ycp:189 OriginalMods=nil 2005-09-17 15:55:24 <1> linux(509) [YCP] NetworkModules.ycp:190 Mods=$["":$["MOD ULE":"", "MODULE_OPTIONS":"", "STARTMODE":"auto"]] 2005-09-17 15:55:24 <1> linux(509) [YCP] NetworkModules.ycp:199 Deleted=[] 2005-09-17 15:55:24 <1> linux(509) [YCP] NetworkDevices.ycp:262 Writing configur ation 2005-09-17 15:55:24 <1> linux(509) [YCP] NetworkDevices.ycp:268 OriginalDevs=nil 2005-09-17 15:55:24 <1> linux(509) [YCP] NetworkDevices.ycp:269 Devs=$[] 2005-09-17 15:55:24 <1> linux(509) [YCP] NetworkDevices.ycp:278 Deleted=[] 2005-09-17 15:55:24 <1> linux(509) [YCP] NetworkConfig.ycp:86 Writing configuration

1 1

Re: [suse-autoinstall] SLES9: Reread of partition table failed
by Hans-Joachim Ehlers 17 Aug '05

17 Aug '05

If you are starting from scratch without a partition table or with an invalid one you should use: <use>all</use> This should wipe the partition table. Also all partions should be formated. If you going to install additional Suse linux version to the other partitions then you there is no NEED anymore for partitioning. Of cource then you have to use the "rules.xml" Also if you define "/" you should also define swap ! ... -> <format config:type="boolean">true</format> .... .... <partition> <crypt>twofish256</crypt> -> <filesystem config:type="symbol">swap</filesystem> -> <format config:type="boolean">true</format> <loop_fs config:type="boolean">false</loop_fs> -> <mount>swap</mount> <partition_id config:type="integer">130</partition_id> <partition_type>primary</partition_type> <size>max</size> </partition> After these steps you should have 3 formated "/" partitions and a swap. where Partition 1 will be used as "/" and Partition 4 as swap. Partition 2 & 3 are free for use. hth Hajo >>> John Clemens <jclemens(a)virtualiron.com> 08/17/05 5:13 pm >>> On Wed, 17 Aug 2005, Uwe Gansert wrote: > On Wednesday 17 August 2005 16:46, John Clemens wrote: > >> The only problem I've yet to solve is that the install >> pauses halfway through with the above error, making the user press ok to > > I've never seen that error or heard of it until now. Thanks for responding.. >> <partitioning config:type="list"> >> <drive> >> <device>/dev/sda</device> >> <initialize config:type="boolean">true</initialize> > > you can try to set <initialize ...>false</initialize> > or do you really need it? Err, I assumed initialize meant wipe the partition table and start from scratch.. I'll go back to the documentation and see what it really does.. http://www.suse.de/~nashif/autoinstallation/9.1/html/CreateProfile.Partitio… doesn't seem to mention it. what exactly does it do? It was a checkbox in the yast autoinstall program i used to create the file. After trying it, it didn't give the reread failure, which is a nice step forward; but it did give me a timed out error message saying "error, partitions are not contiguous". and then another one claiming there was an error setting up the partitions, and the dumped me back to the linuxrc menu. >> <use>1,4</use> > > why is that? local convention.. in our test environment we have OS's on partition 1-3 and swap on partition 4. so I want to tell SLES to use partition 1 for the OS, and 4 for swap, and ignore the part 2 and 3. maybe that's causing the above problem. Let me try a few things now that I'm getting further. Thanks for the help. john.c -- To unsubscribe, e-mail: suse-autoinstall-unsubscribe(a)suse.com For additional commands, e-mail: suse-autoinstall-help(a)suse.com

1 0

Incorporating SLES-SP2 into an AutoYast installation source?
by Michael T. Halligan 16 Aug '05

16 Aug '05

Can anybody recommend a guide as to how I'd go about incorporating SLES SP2 into a SLES9 repository? I'm about to image a number of servers, and ideally would have them all come up fully patched, and ready to go. -- Michael T. Halligan Chief Technology Officer ------------------- BitPusher, LLC http://www.bitpusher.com/ 1.888.9PUSHER 415.724.7998 (Mobile) 415.520.0876 (Fax)

2 3

Adding New Package to Install
by Sean Siemsen 16 Aug '05

16 Aug '05

What Im Trying to do: Adding a package "scli-1.06.16-20.i386.rpm" to the install. What Ive done so far: I am using AutoYast to install SuSE 9.2. I have copied the rpm to the path2CD/suse/i586 folder. Deleted MD5SUMS, ran 'md5sum * > MD5SUMS'. Then I backed up to the path2CD/suse folder and ran 'create_package_descr -x path2CD/suse/setup/descr/EXTRA_PROV' command to rebuild the package database. I also read somewhere about updating the INDEX.gz and ls-lR.gz files in the main folder. (Do I need to do this every time I add a package to my custom CD with these commands? ie. update MD5SUMS, INDEX.gz, and ls-lR.gz files?) --------- cd /suse_9.2 rm INDEX.gz ls-lR.gz ls – lR > ls-lR gzip ls-lR find . > INDEX gzip INDEX --------- Problem: When I run the install it says "Package scli-1.06.16-20.i386 not found on medium. Details- Err 11283:ERROR(Medium: file not found)" I have read through the previous postings and found a simular posting. I followed the same steps in the posting and received the same error. He was trying to put his packages in the noarch folder. It worked for him when he put them in the i586 folder. He wanted to know why it didnt work in the noarch folder and the poster said maybe that was because the rpms were not noarch. I noticed my rpm is a i386. Does this have anything to do with my problem? I also noticed it didnt say "Package scli-1.06.16-20.i386.rpm" (with the .rpm) Thanks Sean Siemsen

1 0

Multiple issues with autoyast.
by GrueMaster 15 Aug '05

15 Aug '05

I need to be able to make an autoinstall for installing a minimal OS on multiple systems for testing. We are testing multiple Linux distributions, along with other OS's, averaging 10 installs/day. I don't have much control over the server, so we have to install via FTP. Getting the autoyast config is not the problem (verified Our hardware varies greatly, and is usually very bleeding edge (just got a new Pentium D EE with dual SLI NVidia 6600's). We also have workstations and servers, all with different drive controllers, etc (IDE, SATA, SCSI). I need to have an auto script that will just nuke & pave the existing drive, install, and run our generic postinstall script. I have Redhat, Fedora, and Mandriva working fine (although the varied hardware creates issues). With Suse 9.3, I am really having a rough time. I have finally gotten past all the errors from xmllint, but the install still hangs on "Error while configuriing partitions. Try again". Nothing seems to indicate an issue in the y2log, with and without the debug boot option. To start, I just want to complete an install. Here's my basic.xml file (sans post-install script). I have also tried several auto scripts that I have found on this mailing list. Any help would be greatly appreciated. Tobin --------------------------------- Yahoo! Mail for Mobile Take Yahoo! Mail with you! Check email on your mobile phone.

2 3

using autoyast to install SLES8 and SLES9 on one machine
by André Weidemann 12 Aug '05

12 Aug '05

Hello list, I'm trying to install SLES8 and SLES9 on one machine. I'm starting with a SLES8 installation where all partitions are being created. (for both SLES8 and SLES9) The machine will have the following partitions: 2 x /boot 1 x swap 1 x extended 2 x / (root) 2 x /usr 2 x /var LVM1 LVM2 LMV3 The autoinstall for SLES8 creates all the above partitions and is running through the autoinstall process without any hassle. After this I'm trying to install SLES9 with autoyast on the same machine. Yast reads it's config and is starting with the partition process. Everything is set up so that yast should use all existing partitions and thus not create any new ones. I only set up 3 or 4 partitions to be formated. When the partition process starts, yast complains that the boot partition is not within the first 1024 cylinders although it ranges from 26-50. After this point yast is displaying the progress bar for setting up the partititions but the bar does not move a single percent. Has anyone else ever tried to do something similar? Any hints on what I can do to make the installation go past the partitioning? Thanx a lot in advance. André.

3 2

Re: [suse-autoinstall] Tracking an error during installation
by Mitko Haralanov 11 Aug '05

11 Aug '05

On Wed, 10 Aug 2005 09:14:17 +0200 "Hans-Joachim Ehlers" <HansJoachim.Ehlers(a)eumetsat.int> wrote: > 1) Set installation mode to text or press F2 on the initial boot > screen > Reason: During network installation of Suse 9.3 ( but booting from DVD > ) i have got every time a error message saying something about > ./var/...X11/.. missing > Solution is to use Textmode ( F2) or F5 but NOT F3(640*480 ) - F5 i > never testet. > > 2) > use a simple "default" XML so your autoyast entry should look like: > .../default Well, I found out something interesting today! I am pretty sure it is not the XML file that is the problem. I tried using previous revisions of the XML file (ones that I know work) and I get the same error. Could the problem be somewhere in the installer or kernel/initrd images? -- Mitko Haralanov mitko(a)pathscale.com http://www.pathscale.com ========================================== I bet the human brain is a kludge. -- Marvin Minsky

1 1

Latest news from a Suse 9.3 Network installation ( no interface error message )
by Hans-Joachim Ehlers 11 Aug '05

11 Aug '05

After configuring the Section "Network Devices" - Netowork Card and starting the installation again: Error Message: ~ DHCP-Server can not be configured - no interface ~ LDAP Client can not be configured .. ~ Kerberos can not be configured .. The point is: I newer configured these Services. Taking a look into the xml file: These service has been added to the xml file by autoyast. Its looks like that they are just dummy entries BUT they are added before the Section <networking> subsection <interfaces> Its real fun to install Suse 9.3 Hajo ============================ Hans-Joachim Ehlers UNIX Systems Engineer EUMETSAT Am Kavalleriesand, 31 64295, Darmstadt, Germany Tel : +49 6151 807-7307 eMail : hansjoachim.ehlers(a)eumetsat.int iNet : www.eumetsat.de

3 2