From pmmm at rnl.ist.utl.pt Mon May 3 17:53:17 2004 From: pmmm at rnl.ist.utl.pt (Pedro Morais) Date: Mon, 3 May 2004 18:53:17 +0100 Subject: FC2t3: error installing Sun j2re rpm on kickstart %post Message-ID: <200405031853.17625.pmmm@rnl.ist.utl.pt> Hi! First of all this is not a "jvm does not work" problem; it works just fine, after I manage to install it. I've been checking my kickstarts against FC2 test 3; other than small changes, the only problem is that installing the j2re rpm on %post gives the following error: --------------------------------------------------------------- Retrieving http://kickstart/redhat/fc2/extra/java/j2re.rpm Preparing... ################################################## j2re ################################################## error: unpacking of archive failed on file /usr/java/j2re1.4.2_02/CHANGES;40967cf3: cpio: lsetfilecon failed - Inappropriate ioctl for device --------------------------------------------------------------- j2re.rpm is a symlink on the webserver to j2re 1.4.2_02. However, if I just give the rpm -Uvh command after the install is complete, using exactly the same http url, everything works fine (besides the file_contexts warnings). The CHANGES file mentioned on the error message is the first file on the RPM. Other RPMs are installed ok, so this is not a generic rpm-in-%post issue. On the rpm package changelog I found this: --------------------------------------------------------------- * Qua Mar 10 2004 Jeff Johnson 4.3-0.20 - add sparcv8 and enable elf32/elf64 Zon sparc64 (#117302). - fix: --querybynumber looped. - fix: ENOTSUP filter from lsetfilecon borkage. ... Sex Mar 05 2004 Jeff Johnson 4.3-0.18 - selinux: ignore ENOTSUP return from lsetfilecon. --------------------------------------------------------------- Could it be that that the 'rpm' binary in post is different (older) than the one that ends up installed? And maybe this could be related to this bug? https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=120552 All SELinux-related stuff was left at the default values. Worth putting in bugzilla? Pedro Morais From Rebecca.R.Hepper at seagate.com Mon May 3 18:15:22 2004 From: Rebecca.R.Hepper at seagate.com (Rebecca.R.Hepper at seagate.com) Date: Mon, 3 May 2004 13:15:22 -0500 Subject: RH 9: Compiling modules under BOOT kernel Message-ID: Hi All, I am using RH 9 and need to compile a module under the BOOT kernel because I need a newer module than what is available with the bootnet.img. These are my steps: 1. Install the BOOT kernel RPM and the kernel-source RPM 2. /boot/initrd-2.4.20-6BOOT.img does not exist but I'll need to specify it in the /etc/lilo.conf so I manually create one by running 'mkinitrd /boot/initrd-2.4.20-6BOOT.img 2.4.20-6.img' 3. Edit /etc/lilo.conf adding a section for the BOOT kernel 4. run /sbin/lilo 5. Reboot the system and boot into the BOOT kernel. 6. Edit /usr/src/linux-2.4.2/include/linux/version.h: set UTS_RELEASE to 2.4.20-6BOOT 7. Untar the e1000 tar file that was downloaded from the Intel site and compile the module without MODVERSIONS (i.e edit the Makefile - in the section that defines the standard flags for module builds, delete the CFLAGS definition that contains the word MODVERSIONS) While booted into the BOOT kernel, I can do an 'insmod' on the module and it will install without errors but my eth0 does not work. I guess there must be something wrong in my steps. Can anyone point me in the right direction for compiling a module under the BOOT kernel in RH 9? Thanks! From Rebecca.R.Hepper at seagate.com Mon May 3 19:14:48 2004 From: Rebecca.R.Hepper at seagate.com (Rebecca.R.Hepper at seagate.com) Date: Mon, 3 May 2004 14:14:48 -0500 Subject: RH 9: Compiling modules under BOOT kernel Message-ID: I forgot to add something that happens when I try to compile the module in step 7. When I do a 'make' everything looks good. When I do a 'make install', I see the following warning: depmod: *** Unresolved symbols in /lib/modules/2.4.20-6BOOT/kernel/drivers/net/e1000.o |---------+---------------------------------> | | Rebecca.R.Hepper at seaga| | | te.com | | | Sent by: | | | kickstart-list-bounces| | | @redhat.com | | | 952-402-3686 | | | | | | 05/03/2004 01:15 PM | | | Please respond to | | | Discussion list about | | | Kickstart | | | | |---------+---------------------------------> >------------------------------------------------------------------------------------------------------------------------------| | | | To: kickstart-list at redhat.com | | cc: | | Subject: RH 9: Compiling modules under BOOT kernel | >------------------------------------------------------------------------------------------------------------------------------| Hi All, I am using RH 9 and need to compile a module under the BOOT kernel because I need a newer module than what is available with the bootnet.img. These are my steps: 1. Install the BOOT kernel RPM and the kernel-source RPM 2. /boot/initrd-2.4.20-6BOOT.img does not exist but I'll need to specify it in the /etc/lilo.conf so I manually create one by running 'mkinitrd /boot/initrd-2.4.20-6BOOT.img 2.4.20-6.img' 3. Edit /etc/lilo.conf adding a section for the BOOT kernel 4. run /sbin/lilo 5. Reboot the system and boot into the BOOT kernel. 6. Edit /usr/src/linux-2.4.2/include/linux/version.h: set UTS_RELEASE to 2.4.20-6BOOT 7. Untar the e1000 tar file that was downloaded from the Intel site and compile the module without MODVERSIONS (i.e edit the Makefile - in the section that defines the standard flags for module builds, delete the CFLAGS definition that contains the word MODVERSIONS) While booted into the BOOT kernel, I can do an 'insmod' on the module and it will install without errors but my eth0 does not work. I guess there must be something wrong in my steps. Can anyone point me in the right direction for compiling a module under the BOOT kernel in RH 9? Thanks! _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From jamesaharrisonuk at yahoo.co.uk Mon May 3 21:48:40 2004 From: jamesaharrisonuk at yahoo.co.uk (James Harrison) Date: Mon, 3 May 2004 14:48:40 -0700 (PDT) Subject: RH 9: Compiling modules under BOOT kernel In-Reply-To: Message-ID: <20040503214840.34092.qmail@web25110.mail.ukl.yahoo.com> Is there a specific reason why you need load the module from the initrd.img file or is it ok to put an entry into /etc/modules.conf: alias eth0 e1000 James --- Rebecca.R.Hepper at seagate.com wrote: > Hi All, > > I am using RH 9 and need to compile a module under the BOOT kernel because > I need a newer module than what is available with the bootnet.img. > > These are my steps: > > 1. Install the BOOT kernel RPM and the kernel-source RPM > 2. /boot/initrd-2.4.20-6BOOT.img does not exist but I'll need to specify it > in the /etc/lilo.conf so I manually create one by running 'mkinitrd > /boot/initrd-2.4.20-6BOOT.img 2.4.20-6.img' > 3. Edit /etc/lilo.conf adding a section for the BOOT kernel > 4. run /sbin/lilo > 5. Reboot the system and boot into the BOOT kernel. > 6. Edit /usr/src/linux-2.4.2/include/linux/version.h: set UTS_RELEASE to > 2.4.20-6BOOT > 7. Untar the e1000 tar file that was downloaded from the Intel site and > compile the module without MODVERSIONS (i.e edit the Makefile - in the > section that defines the standard flags for module builds, delete the > CFLAGS definition that contains the word MODVERSIONS) > > While booted into the BOOT kernel, I can do an 'insmod' on the module and > it will install without errors but my eth0 does not work. I guess there > must be something wrong in my steps. Can anyone point me in the right > direction for compiling a module under the BOOT kernel in RH 9? > > Thanks! > > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list __________________________________ Do you Yahoo!? Win a $20,000 Career Makeover at Yahoo! HotJobs http://hotjobs.sweepstakes.yahoo.com/careermakeover From Rebecca.R.Hepper at seagate.com Tue May 4 14:05:17 2004 From: Rebecca.R.Hepper at seagate.com (Rebecca.R.Hepper at seagate.com) Date: Tue, 4 May 2004 09:05:17 -0500 Subject: RH 9: Compiling modules under BOOT kernel Message-ID: After I get an e1000 module compiled under the BOOT kernel, I will place it in the initrd.img then copy it to my USB keychain device so it can get automatically loaded during kickstart. |---------+---------------------------------> | | James Harrison | | | | | | Sent by: | | | kickstart-list-bounces| | | @redhat.com | | | No Phone Info | | | Available | | | | | | 05/03/2004 04:48 PM | | | Please respond to | | | Discussion list about | | | Kickstart | | | | |---------+---------------------------------> >------------------------------------------------------------------------------------------------------------------------------| | | | To: Discussion list about Kickstart | | cc: | | Subject: Re: RH 9: Compiling modules under BOOT kernel | >------------------------------------------------------------------------------------------------------------------------------| Is there a specific reason why you need load the module from the initrd.img file or is it ok to put an entry into /etc/modules.conf: alias eth0 e1000 James --- Rebecca.R.Hepper at seagate.com wrote: > Hi All, > > I am using RH 9 and need to compile a module under the BOOT kernel because > I need a newer module than what is available with the bootnet.img. > > These are my steps: > > 1. Install the BOOT kernel RPM and the kernel-source RPM > 2. /boot/initrd-2.4.20-6BOOT.img does not exist but I'll need to specify it > in the /etc/lilo.conf so I manually create one by running 'mkinitrd > /boot/initrd-2.4.20-6BOOT.img 2.4.20-6.img' > 3. Edit /etc/lilo.conf adding a section for the BOOT kernel > 4. run /sbin/lilo > 5. Reboot the system and boot into the BOOT kernel. > 6. Edit /usr/src/linux-2.4.2/include/linux/version.h: set UTS_RELEASE to > 2.4.20-6BOOT > 7. Untar the e1000 tar file that was downloaded from the Intel site and > compile the module without MODVERSIONS (i.e edit the Makefile - in the > section that defines the standard flags for module builds, delete the > CFLAGS definition that contains the word MODVERSIONS) > > While booted into the BOOT kernel, I can do an 'insmod' on the module and > it will install without errors but my eth0 does not work. I guess there > must be something wrong in my steps. Can anyone point me in the right > direction for compiling a module under the BOOT kernel in RH 9? > > Thanks! > > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list __________________________________ Do you Yahoo!? Win a $20,000 Career Makeover at Yahoo! HotJobs http://hotjobs.sweepstakes.yahoo.com/careermakeover _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From David.Avrahami at comverse.com Tue May 4 15:05:47 2004 From: David.Avrahami at comverse.com (Avrahami David) Date: Tue, 4 May 2004 18:05:47 +0300 Subject: Problem with reading ks.cfg from the cdrom Message-ID: Hi, I'm a newbie to kickstart and I am trying to install RHEL 3.0 by kickstart procedure from the cdrom. I did as following: 1) I prepared a bootable cdrom (http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/x8664-multi-ins tall-guide/s1-steps-install-cdrom.html#S2-STEPS-MAKE-CD )and put the ks.cfg on top-level directory (http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/sysadmin-guide/ s1-kickstart2-putkickstarthere.html ). Then I boot from cdrom and type on boot: linux ks=cdrom:/ks.cfg In this stage I got the installation menus which says that kickstart fails to read/find the ks.cfg. It seems that ks.cfg somehow can't be read. the ALT+F4 says: unable to identify cd-rom format the ALT+F3says: failed to mount /dev/hdd invalid argument trying to mount device hdd There is no such problem if I put the same ks.cfg on floppy and boot from cdrom and type boot: linux ks=floppy Any clue? David -------------- next part -------------- An HTML attachment was scrubbed... URL: From steve.mah at oracle.com Tue May 4 17:01:38 2004 From: steve.mah at oracle.com (Stephen Mah) Date: Tue, 04 May 2004 10:01:38 -0700 Subject: Problem with reading ks.cfg from the cdrom In-Reply-To: References: Message-ID: <4097CC72.5060409@oracle.com> Never used the cdrom option. I use this: append initrd=initrd.img lang= devfs=nomount ramdisk_size=8192 vga=788 ks=file:/tmp/ks.cfg then I put the kickstart file in the initrd.img file system, which is on the cdrom when I make it bootable. cd isolinux gunzip -c initrd.img >initrd.img.unzip mkdir -p loop mount -o loop initrd.img.unzip loop/ cp ks.cfg ./loop/tmp/ umount loop/ gzip -9 -c initrd.img.unzip > initrd.img rm -rf initrd.img.unzip then use mkisofs to create your cdrom. Avrahami David wrote: > Hi, > > I'm a newbie to kickstart and I am trying to install RHEL 3.0 by > kickstart procedure from the cdrom. > I did as following: > > 1) I prepared a bootable cdrom > (_http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/x8664-multi-install-guide/s1-steps-install-cdrom.html#S2-STEPS-MAKE-CD_ > )and put the ks.cfg > > on top-level directory > (_http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/sysadmin-guide/s1-kickstart2-putkickstarthere.html_ > ). > > Then I boot from cdrom and type on boot: linux ks=cdrom:/ks.cfg > In this stage I got the installation menus which says that kickstart > fails to read/find the ks.cfg. > It seems that ks.cfg somehow can't be read. > the ALT+F4 says: unable to identify cd-rom format > the ALT+F3says: failed to mount /dev/hdd invalid argument trying to > mount device hdd > > There is no such problem if I put the same ks.cfg on floppy and boot > from cdrom and type boot: linux ks=floppy > Any clue? > > David > >------------------------------------------------------------------------ > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > From pantz at lqt.ca Tue May 4 18:44:04 2004 From: pantz at lqt.ca (Paul Pianta) Date: Tue, 04 May 2004 14:44:04 -0400 Subject: problem with %pre script writing to /tmp Message-ID: <4097E474.5050409@lqt.ca> Hi people, I am experiencing an interesting problem with a %pre script in my kickstart file ... # pre-script that is run to determine partition layout %pre #!/bin/sh # get list of drives with size of each drive set $(list-harddrives) firstDiskSize=$2 if [ $firstDiskSize >= "20480" ] ; then echo " part swap --size=500 --asprimary part /boot --size=100 --fstype ext3 --asprimary part / --size=2000 --fstype ext3 --asprimary part /usr --size=2000 --fstype ext3 part /tmp --size=1000 --fstype ext3 part /var --size=4000 --fstype ext3 part /home --size=1000 --grow --fstype ext3 " > /tmp/partinfo else echo " part swap --size=500 --asprimary part /boot --size=100 --fstype ext3 --asprimary part / --size=500 --fstype ext3 --asprimary part /usr --size=1000 --fstype ext3 part /tmp --size=500 --fstype ext3 part /var --size=1000 --fstype ext3 part /home --size=100 --grow --fstype ext3 " > /tmp/partinfo fi My problem is not in the script itself - I know that it works. My problem is that anaconda bugs out on me when I run this on a brand spanking new IBM xSeries 225 server with a new raid 0 setup of 2 36Gb hard drives. It says that it cannot include the file /tmp/partinfo because it doesn't exist (btw - i use an %include in my general ks.cfg section). When I run this exact same kickstart in a vmware machine that I use for testing - it works fine. This leads me to believe it could be something to do with the IBM disks not having been initialized or something like that. Anaconda quits out and suggests a reboot too quickly for me to get hold of a shell to see if I can manually write to /tmp. Any ideas? pantz -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From j.rink at freenet.de Tue May 4 06:24:58 2004 From: j.rink at freenet.de (=?UTF-8?B?SsO2cm4=?= Rink) Date: Tue, 4 May 2004 08:24:58 +0200 Subject: Dialog while ks installation Message-ID: <20040504082458.657d2fb5.j.rink@freenet.de> Hi, anyone has experience with making dialogs for a user for a kickstart installation? I made this in the past for SuSE 7.x and the auto installation via yast. There i have to use some tricks to manage a dialog for the user. For fedora i have the text install option, where i can use dialog or the x install option, where i could use xdialog. CU Joern Rink -- Nine (not 9) Never trust a hippie From pantz at lqt.ca Tue May 4 19:07:04 2004 From: pantz at lqt.ca (Paul Pianta) Date: Tue, 04 May 2004 15:07:04 -0400 Subject: problem with %pre script writing to /tmp In-Reply-To: <4097E474.5050409@lqt.ca> References: <4097E474.5050409@lqt.ca> Message-ID: <4097E9D8.1030706@lqt.ca> oo oo I forgot to mention that just before anaconda poops itself - I can get to the ALT-F2 screen and instead of seeing the old friendly '-/bin/sh-2.05b#' prompt I have this 'friendly' message ... -bin/sh: relocation error: -/bin/sh: symbol, version GLIBC_2.0 not defined in file libc.so.6 with link time reference. This message scares me - ALOT! - can anyone suggest something to help me sleep better? pantz -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From alexander_rau at yahoo.com Tue May 4 19:20:15 2004 From: alexander_rau at yahoo.com (Alexander Rau (private)) Date: Tue, 4 May 2004 15:20:15 -0400 Subject: Dialog while ks installation In-Reply-To: <20040504082458.657d2fb5.j.rink@freenet.de> Message-ID: <200405041924.i44JOLkG000315@mx1.redhat.com> Joern: Here is what I did. Hope this helps Example %post section %post exec < /dev/tty3 > /dev/tty3 chvt 3 ############################################################################ ############################################################################ ############ # Enter password with check if input is identical # ############################################################################ ############################################################################ ############ PASSWORD="&*UY^&^YI" PASSWORD_CHECK=1 until [ "$PASSWORD" = "$PASSWORD_CHECK" ] do if [ $PASSWORD = "&*UY^&^YI" ] ; then dialog --title "Enter root (Administrator) Password" --backtitle "MyDistro (C) 2003 MyDistro" --no-cancel --passwordbox "Enter a root (Administrator) password between 6 and 8 characters" 8 60 2>/tmp/passwd.$$ SEL=$? NA5=`cat /tmp/passwd.$$` case $SEL in 0) PASSWORD=$NA5;; 1) echo "Cancel is Pressed" ;; 255) echo "[ESCAPE] key pressed" ;; esac rm -f /tmp/passwd.$$ else dialog --title "Passwords do not match" --backtitle "MyDistro (C) 2003 MyDistro" --msgbox "The passwords you entered do not match. Please re-enter passwords" 6 50 dialog --title "Enter root (Administrator) Password" --backtitle "MyDistro (C) 2003 MyDistro" --no-cancel --passwordbox "Enter a root (Administrator) password between 6 and 8 characters" 8 60 2>/tmp/passwd.$$ SEL=$? NA5=`cat /tmp/passwd.$$` case $SEL in 0) PASSWORD=$NA5;; 1) echo "Cancel is Pressed" ;; 255) echo "[ESCAPE] key pressed" ;; esac rm -f /tmp/passwd.$$ fi dialog --title "Re-enter root (Administrator) Password" --backtitle "MyDistro (C) 2003 MyDistro" --no-cancel --passwordbox "Please re-enter the root (Administrator) password" 8 60 2>/tmp/passwd.$$ SEL=$? NA4=`cat /tmp/passwd.$$` case $SEL in 0) PASSWORD_CHECK=$NA4;; 1) echo "Cancel is Pressed" ;; 255) echo "[ESCAPE] key pressed" ;; esac rm -f /tmp/passwd.$$ done echo $NA4 | passwd --stdin root >/dev/null 2>&1 ############################################################################ ############################################################################ ############ ############################################################################ ############################################################################ ############ # Enter Username # ############################################################################ ############################################################################ ############ dialog --title "Add Username" --backtitle "MyDistro (C) 2003 MyDistro" --no-cancel --inputbox "Enter an easy to remember Username please" 8 60 2>/tmp/usrname.$$ SEL=$? NA=`cat /tmp/usrname.$$` case $SEL in 0) useradd $NA >/dev/null 2>&1 ;; 1) echo "Cancel is Pressed" ;; 255) echo "[ESCAPE] key pressed" ;; esac rm -f /tmp/usrname.$$ ############################################################################ ############################################################################ ############ ############################################################################ ############################################################################ ############ # Enter Real Name # ############################################################################ ############################################################################ ############ dialog --title "Enter Name" --backtitle "MyDistro (C) 2003 MyDistro" --no-cancel --inputbox "Please enter your real name(first and last)" 8 60 2>/tmp/name.$$ SEL=$? NA2=`cat /tmp/name.$$` case $SEL in 0) chfn -f "$NA2" $NA >/dev/null 2>&1 ;; 1) echo "Cancel is Pressed" ;; 255) echo "[ESCAPE] key pressed" ;; esac rm -f /tmp/name.$$ ############################################################################ ############################################################################ ############ ############################################################################ ############################################################################ ############ # Enter password with check if input is identical # ############################################################################ ############################################################################ ############ PASSWORD="&*UY^&^YY" PASSWORD_CHECK=1 until [ "$PASSWORD" = "$PASSWORD_CHECK" ] do if [ $PASSWORD = "&*UY^&^YY" ] ; then dialog --title "Enter Password" --backtitle "MyDistro (C) 2003 MyDistro" --no-cancel --passwordbox "Enter a password between 6 and 8 characters" 8 60 2>/tmp/passwd.$$ SEL=$? NA1=`cat /tmp/passwd.$$` case $SEL in 0) PASSWORD=$NA1;; 1) echo "Cancel is Pressed" ;; 255) echo "[ESCAPE] key pressed" ;; esac rm -f /tmp/passwd.$$ else dialog --title "Password's do not match" --backtitle "MyDistro (C) 2003 MyDistro" --msgbox "The passwords you entered do not match. Please re-enter passwords" 6 50 dialog --title "Enter Password" --backtitle "MyDistro (C) 2003 MyDistro" --no-cancel --passwordbox "Enter a password between 6 and 8 characters" 8 60 2>/tmp/passwd.$$ SEL=$? NA1=`cat /tmp/passwd.$$` case $SEL in 0) PASSWORD=$NA1;; 1) echo "Cancel is Pressed" ;; 255) echo "[ESCAPE] key pressed" ;; esac rm -f /tmp/passwd.$$ fi dialog --title "Re-enter Password" --backtitle "MyDistro (C) 2003 MyDistro" --no-cancel --passwordbox "Please re-enter your password" 8 60 2>/tmp/passwd.$$ SEL=$? NA3=`cat /tmp/passwd.$$` case $SEL in 0) PASSWORD_CHECK=$NA3;; 1) echo "Cancel is Pressed" ;; 255) echo "[ESCAPE] key pressed" ;; esac rm -f /tmp/passwd.$$ done echo $NA3 | passwd --stdin $NA >/dev/null 2>&1 ############################################################################ ############################################################################ ############ ############################################################################ ############################################################################ ############ # Thank you # ############################################################################ ############################################################################ ############ dialog --title "Thank you" --backtitle "MyDistro (C) 2003 MyDistro" --msgbox "Thank you for choosing and installing MyDistro. Please press to reboot system and start MyDistro. Please remove any floppy disks and/or CD-ROMS from their respective drives." 9 50 ############################################################################ ############################################################################ ############ chvt 3 -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of J?rn Rink Sent: Tuesday, May 04, 2004 2:25 AM To: Discussion list about Kickstart Subject: Dialog while ks installation Hi, anyone has experience with making dialogs for a user for a kickstart installation? I made this in the past for SuSE 7.x and the auto installation via yast. There i have to use some tricks to manage a dialog for the user. For fedora i have the text install option, where i can use dialog or the x install option, where i could use xdialog. CU Joern Rink -- Nine (not 9) Never trust a hippie _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From pantz at lqt.ca Tue May 4 19:34:46 2004 From: pantz at lqt.ca (Paul Pianta) Date: Tue, 04 May 2004 15:34:46 -0400 Subject: problem with %pre script writing to /tmp In-Reply-To: <4097E474.5050409@lqt.ca> References: <4097E474.5050409@lqt.ca> Message-ID: <4097F056.7030202@lqt.ca> me bad - here is another error message that I found on ALT-F3 ... * WARNING - Error code 32512 encountered running a kickstart %pre%post script pantz -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From phr at doc.ic.ac.uk Tue May 4 19:49:42 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 4 May 2004 20:49:42 +0100 (BST) Subject: problem with %pre script writing to /tmp In-Reply-To: <4097E474.5050409@lqt.ca> References: <4097E474.5050409@lqt.ca> Message-ID: On Tue, 4 May 2004, Paul Pianta wrote: >This leads me to believe it could be something to do with the IBM disks >not having been initialized or something like that. Anaconda quits out >and suggests a reboot too quickly for me to get hold of a shell to see >if I can manually write to /tmp. If you get the Alt-F2 shell during %pre (and I think you do), you could insert various "sleep"s throughout your script and watch while it executes. Cheers, Phil From dtegelaar at egenera.com Tue May 4 19:47:27 2004 From: dtegelaar at egenera.com (David Tegelaar) Date: Tue, 4 May 2004 15:47:27 -0400 Subject: Problem with reading ks.cfg from the cdrom In-Reply-To: <4097CC72.5060409@oracle.com> Message-ID: Haven't done this myself but Section 7.8.2 of this doc: http://www.redhat.com/docs/manuals/linux/RHL-9-Manual/custom-guide/s1-kickstart2-putkickstarthere.html talks about copying the ks.cfg file to the isolinux/ directory before you make the iso. -dave -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Stephen Mah Sent: Tuesday, May 04, 2004 1:02 PM To: Discussion list about Kickstart Subject: Re: Problem with reading ks.cfg from the cdrom Never used the cdrom option. I use this: append initrd=initrd.img lang= devfs=nomount ramdisk_size=8192 vga=788 ks=file:/tmp/ks.cfg then I put the kickstart file in the initrd.img file system, which is on the cdrom when I make it bootable. cd isolinux gunzip -c initrd.img >initrd.img.unzip mkdir -p loop mount -o loop initrd.img.unzip loop/ cp ks.cfg ./loop/tmp/ umount loop/ gzip -9 -c initrd.img.unzip > initrd.img rm -rf initrd.img.unzip then use mkisofs to create your cdrom. Avrahami David wrote: > Hi, > > I'm a newbie to kickstart and I am trying to install RHEL 3.0 by > kickstart procedure from the cdrom. > I did as following: > > 1) I prepared a bootable cdrom > (_http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/x8664-multi-install-guide/s1-steps-install-cdrom.html#S2-STEPS-MAKE-CD_ > )and put the ks.cfg > > on top-level directory > (_http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/sysadmin-guide/s1-kickstart2-putkickstarthere.html_ > ). > > Then I boot from cdrom and type on boot: linux ks=cdrom:/ks.cfg > In this stage I got the installation menus which says that kickstart > fails to read/find the ks.cfg. > It seems that ks.cfg somehow can't be read. > the ALT+F4 says: unable to identify cd-rom format > the ALT+F3says: failed to mount /dev/hdd invalid argument trying to > mount device hdd > > There is no such problem if I put the same ks.cfg on floppy and boot > from cdrom and type boot: linux ks=floppy > Any clue? > > David > >------------------------------------------------------------------------ > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From David.Avrahami at comverse.com Wed May 5 11:28:10 2004 From: David.Avrahami at comverse.com (Avrahami David) Date: Wed, 5 May 2004 14:28:10 +0300 Subject: Problem with reading ks.cfg from the cdrom Message-ID: It seems that there is a problem with cdrom option,(case opened on RedHat,no bug has been issued yet on bugzilla). Including ks.cfg to initrd.img works. Thanks for the help, great list David > Stephen Mah wrote: > Sent: Tuesday, May 04, 2004 8:02 PM > To: Discussion list about Kickstart > Subject: Re: Problem with reading ks.cfg from the cdrom > > > Never used the cdrom option. > I use this: > append initrd=initrd.img lang= devfs=nomount > ramdisk_size=8192 vga=788 > ks=file:/tmp/ks.cfg > > then I put the kickstart file in the initrd.img file system, > which is on > the cdrom when I make it bootable. > > cd isolinux > gunzip -c initrd.img >initrd.img.unzip > mkdir -p loop > mount -o loop initrd.img.unzip loop/ > cp ks.cfg ./loop/tmp/ > umount loop/ > gzip -9 -c initrd.img.unzip > initrd.img > rm -rf initrd.img.unzip > > then use mkisofs to create your cdrom. > > Avrahami David wrote: > > > Hi, > > > > I'm a newbie to kickstart and I am trying to install RHEL 3.0 by > > kickstart procedure from the cdrom. > > I did as following: > > > > 1) I prepared a bootable cdrom > > > (_http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/ > x8664-multi-install-guide/s1-steps-install-cdrom.html#S2-STEPS > -MAKE-CD_ > > )and put the ks.cfg > > > > on top-level directory > > > (_http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/ > sysadmin-guide/s1-kickstart2-putkickstarthere.html_ > > ). > > > > Then I boot from cdrom and type on boot: linux ks=cdrom:/ks.cfg In > > this stage I got the installation menus which says that kickstart > > fails to read/find the ks.cfg. It seems that ks.cfg somehow > can't be > > read. the ALT+F4 says: unable to identify cd-rom format > > the ALT+F3says: failed to mount /dev/hdd invalid argument trying to > > mount device hdd > > > > There is no such problem if I put the same ks.cfg on floppy and boot > > from cdrom and type boot: linux ks=floppy > > Any clue? > > > > David > > > >------------------------------------------------------------- > ---------- > >- > > > >_______________________________________________ > >Kickstart-list mailing list > >Kickstart-list at redhat.com > >https://www.redhat.com/mailman/listinfo/kickstart-list > > > > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstar> t-list > > > > ______________________________________________________________________ > This email message has been scanned by PineApp Mail-Secure > and has been found clean. > -------------- next part -------------- An HTML attachment was scrubbed... URL: From j.rink at freenet.de Fri May 7 06:47:15 2004 From: j.rink at freenet.de (=?UTF-8?B?SsO2cm4=?= Rink) Date: Fri, 7 May 2004 08:47:15 +0200 Subject: Problem with reading ks.cfg from the cdrom In-Reply-To: References: Message-ID: <20040507084715.64eef509.j.rink@freenet.de> Am Tue, 4 May 2004 18:05:47 +0300 hat Avrahami David (Avrahami David) folgendes geschrieben: > Hi, > > I'm a newbie to kickstart and I am trying to install RHEL 3.0 by > kickstart procedure from the cdrom. > I did as following: Hi, i made the same wih fedora core 1. You have to put the ks.cfg file in the iso image. Mount the iso.image and copy the ks.cfg file into the dir where you mount it. Aftter that, make an new iso and burn it. When you want to make the ks a default boot option, the you have to reconfig the boot.cfg file, as i remember correctly, syslinux.cfg. No i am sitting in a train. When i am at work, i take a look at the computer in my firm, where i made the cd. CU J?rn Rink -- Nine (not 9) Never trust a hippie From pantz at lqt.ca Fri May 7 19:34:09 2004 From: pantz at lqt.ca (Paul Pianta) Date: Fri, 07 May 2004 15:34:09 -0400 Subject: problem with %pre script writing to /tmp In-Reply-To: References: <4097E474.5050409@lqt.ca> Message-ID: <409BE4B1.30100@lqt.ca> Well I found the solution to my problem. It turns out it is a well known Fedora bug and I only had to pass 'allowcddma' at the boot prompt and things were fine ... Turns out it was nothing to do with my script at all - tough one to debug there. pantz -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From jrobertson at convera.com Fri May 7 21:05:20 2004 From: jrobertson at convera.com (Joe Robertson) Date: Fri, 7 May 2004 14:05:20 -0700 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: I believe this is probably an anaconda problem but since it only shows up when I'm trying to do kickstart I will send it here. Question: Has anyone has been successful in running kickstart via NFS through an Intel 82546EB network interface? I'm fighting a problem getting kickstart to run with these NICs (using NFS). It appears to be related to the NIC in use. I have setup a PXE install environment using NFS to hold a copy of the installation tree (created by copying all of the contents of all of the CDs for a given version of linux). I can do interactive installs with RH9, Fedora Core 1, Core 2/test 1-3, as well as others. These installs work without incident on all hardware I've tried so far. The problem comes in when I try to use kickstart instead of doing the interactive installs on a machine that uses the Intel 82546EB Gigabit ethernet chipset. When the installation gets to a point where it needs to access NFS (for the kickstart file or for the installation) at which time it reports that NFS has failed and tries to load from CD (according to the ALT-F3 screen) then halts. I have tried this on several Dell 1600sc and a SuperMicro 6013P-T with the same results. I added a 3COM 3C905C-TX card to the Dell 1600sc and configured the bios so I could install from that card via PXE. The kickstart installation runs just fine with this nic - no other changes. I've tried several suggestions - I used a 3COM 3C39036 switch (instead of the netgear Gb), I've connected two computers directly with a crossover cable (no switch) but cannot get kickstart to run when using the Intel 82546EB Gb nic. Any suggestions for isolating this (or solving it) would be welcomed. Thanks, Joe -------------- next part -------------- An HTML attachment was scrubbed... URL: From jkeating at j2solutions.net Fri May 7 21:04:46 2004 From: jkeating at j2solutions.net (Jesse Keating) Date: Fri, 7 May 2004 14:04:46 -0700 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: References: Message-ID: <200405071404.46474.jkeating@j2solutions.net> On Friday 07 May 2004 14:05, Joe Robertson wrote: > Question: > Has anyone has been successful in running kickstart via NFS through > an Intel 82546EB network interface? > > I'm fighting a problem getting kickstart to run with these NICs > (using NFS). > It appears to be related to the NIC in use. > > I have setup a PXE install environment using NFS to hold a copy of > the installation tree (created > by copying all of the contents of all of the CDs for a given version > of linux). I can do interactive installs > with RH9, Fedora Core 1, Core 2/test 1-3, as well as others. These > installs work without incident on > all hardware I've tried so far. > > The problem comes in when I try to use kickstart instead of doing the > interactive installs on a machine > that uses the Intel 82546EB Gigabit ethernet chipset. When the > installation gets to a point where it > needs to access NFS (for the kickstart file or for the installation) > at which time it reports that NFS > has failed and tries to load from CD (according to the ALT-F3 screen) > then halts. I have tried this on > several Dell 1600sc and a SuperMicro 6013P-T with the same results. Is the module even able to load? We've seen the newer e1000 chipset boards require a newer driver from intel.com. Check your other ttys, to see if it even loads the e1000 driver. -- Jesse Keating RHCE (geek.j2solutions.net) Fedora Legacy Team (www.fedoralegacy.org) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) Was I helpful? Let others know: http://svcs.affero.net/rm.php?r=jkeating -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: signature URL: From joe at swelltech.com Fri May 7 21:13:42 2004 From: joe at swelltech.com (Joe Cooper) Date: Fri, 07 May 2004 16:13:42 -0500 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: References: Message-ID: <409BFC06.20201@swelltech.com> Joe Robertson wrote: > I believe this is probably an anaconda problem but since it only shows > up when I'm trying > to do kickstart I will send it here. > > Question: > Has anyone has been successful in running kickstart via NFS through an > Intel 82546EB network interface? > > I'm fighting a problem getting kickstart to run with these NICs (using > NFS). > It appears to be related to the NIC in use. I think I ran into the same issue with a new MSI 1U system. I tried to follow some documentation I found on this list and at the below link for adding a new driver to the boot kernel (which I tested successfully on an already installed machine--so the driver works), and adding the ID 0x1075 and 0x1076 for the cards in my machine to the pcitable in the kickstart image, but the install still fails when trying to load the network drivers. I found someone else who had removed the e100 driver and that solved a similar problem, so I tried that, and it also failed. I'd be pleased as punch if someone would tell me what I missed in my modifications of the kickstart initrd image, that would make this work...Even a simple list of files that must be changed when adding a new or modified driver to the kickstart image (so I could spot the one I didn't change) would be exceedingly useful. Some links that were helpful (but didn't result in a working solution): http://blog.oddbit.com/tech/new_install_kernel.html http://www.puschitz.com/Kickstart.shtml An additional data point, a Fedora Core 2 test2 PXE image worked up until the point where it complains about the boot image not matching the installation directory. So the kernel in fc2t2 works fine, but it won't let me install fc1. From joe at swelltech.com Fri May 7 21:19:05 2004 From: joe at swelltech.com (Joe Cooper) Date: Fri, 07 May 2004 16:19:05 -0500 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: <200405071404.46474.jkeating@j2solutions.net> References: <200405071404.46474.jkeating@j2solutions.net> Message-ID: <409BFD49.6090707@swelltech.com> Jesse Keating wrote: > On Friday 07 May 2004 14:05, Joe Robertson wrote: > >>Question: >>Has anyone has been successful in running kickstart via NFS through >>an Intel 82546EB network interface? > > Is the module even able to load? We've seen the newer e1000 chipset > boards require a newer driver from intel.com. Check your other ttys, > to see if it even loads the e1000 driver. If it is helpful to anyone, I have a working (moderately tested) patch of the latest version of the e1000 driver against the latest Fedora 2.4.22 kernel. It works for me, though I still have a weird, and probably completely unrelated, depmod issue in my kernels built with the patch (my kernels are pretty heavily patched, so I'm pretty confident it is unrelated). And if I ever get a new installer image with the driver to work, I'll make it available... From ccollier at viawest.net Fri May 7 21:35:51 2004 From: ccollier at viawest.net (ccollier at viawest.net) Date: Fri, 7 May 2004 15:35:51 -0600 (MDT) Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: Message-ID: I have successfully used kickstart via http with this network interface using rhel 3. It should work for rh 9 as well. (I might have used 9, but I can't recall now.) Depending on your pxe configuration there might not be much difference between this and an nfs source for the file. verification of an interface match: 03:02.0 Ethernet controller: Intel Corp. 82546EB Gigabit Ethernet Controller (Copper) (rev 01) 03:02.1 Ethernet controller: Intel Corp. 82546EB Gigabit Ethernet Controller (Copper) (rev 01) This requires that you use the images from: (yourdistribution)/images/pxeboot/ This ramdisk contains all the drivers, including the e1000, rather than the smaller subset in the normal boot images. Can you pass on what images you are using, as well as what your netboot file looks like? Are you specifying the ksdevice and other ks parameters in the append line in this file? cody On Fri, 7 May 2004, Joe Robertson wrote: > I believe this is probably an anaconda problem but since it only shows > up when I'm trying > to do kickstart I will send it here. > > Question: > Has anyone has been successful in running kickstart via NFS through an > Intel 82546EB network interface? > > I'm fighting a problem getting kickstart to run with these NICs (using > NFS). > It appears to be related to the NIC in use. > > I have setup a PXE install environment using NFS to hold a copy of the > installation tree (created > by copying all of the contents of all of the CDs for a given version of > linux). I can do interactive installs > with RH9, Fedora Core 1, Core 2/test 1-3, as well as others. These > installs work without incident on > all hardware I've tried so far. > > The problem comes in when I try to use kickstart instead of doing the > interactive installs on a machine > that uses the Intel 82546EB Gigabit ethernet chipset. When the > installation gets to a point where it > needs to access NFS (for the kickstart file or for the installation) at > which time it reports that NFS > has failed and tries to load from CD (according to the ALT-F3 screen) > then halts. I have tried this on > several Dell 1600sc and a SuperMicro 6013P-T with the same results. > > I added a 3COM 3C905C-TX card to the Dell 1600sc and configured the bios > so I could install from > that card via PXE. The kickstart installation runs just fine with this > nic - no other changes. > > I've tried several suggestions - I used a 3COM 3C39036 switch (instead > of the netgear Gb), I've > connected two computers directly with a crossover cable (no switch) but > cannot get kickstart to > run when using the Intel 82546EB Gb nic. > > Any suggestions for isolating this (or solving it) would be welcomed. > > Thanks, > Joe > > From ccollier at viawest.net Fri May 7 21:51:12 2004 From: ccollier at viawest.net (ccollier at viawest.net) Date: Fri, 7 May 2004 15:51:12 -0600 (MDT) Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: <409BFC06.20201@swelltech.com> Message-ID: On Fri, 7 May 2004, Joe Cooper wrote: > Joe Robertson wrote: --snip-- > I'd be pleased as punch if someone would tell me what I missed in my > modifications of the kickstart initrd image, that would make this > work...Even a simple list of files that must be changed when adding a > new or modified driver to the kickstart image (so I could spot the one I > didn't change) would be exceedingly useful. --snip-- I struggled with some similar issues with the e1000 and rh 7.3. Typically, for module insertion into the ramdisk you have to edit these three files found inside the ramdisk: /modules/pci-table (needs the appropriate hardware identification lines) /modules/module-info (needs an entry for e1000) /modules/modules.cgz (in here of course is another set of paths, and you need to insert the compiled e1000.o file in the appropriate path for your kernel version) I ran into some interesting situations where the installation still did not properly identify the device without user intervention. It's also important that the driver is compiled against the correct kernel/source. Note that the installation images use red hat's special boot kernel. Sometimes you can get away with just getting this step "close" but not always. hope this helps a little. cody From jrobertson at convera.com Fri May 7 21:58:20 2004 From: jrobertson at convera.com (Joe Robertson) Date: Fri, 7 May 2004 14:58:20 -0700 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: > -----Original Message----- Jesse Keating wrote: > > On Friday 07 May 2004 14:05, Joe Robertson wrote: > > Question: > > Has anyone has been successful in running kickstart via NFS through > > an Intel 82546EB network interface? ... > > The problem comes in when I try to use kickstart instead of > doing the > > interactive installs on a machine that uses the Intel > 82546EB Gigabit > > ethernet chipset. When the installation gets to a point where it > > needs to access NFS (for the kickstart file or for the installation) > > at which time it reports that NFS > > has failed and tries to load from CD (according to the > ALT-F3 screen) > > then halts. I have tried this on > > several Dell 1600sc and a SuperMicro 6013P-T with the same results. > > Is the module even able to load? We've seen the newer e1000 chipset > boards require a newer driver from intel.com. Check your other ttys, > to see if it even loads the e1000 driver. > As far as I can tell, the driver loads just fine. It is just too slow (or anaconda is too impatient?). The output on the console is as follows: Intel(R) PRO/1000 Network Driver - version 5.2.30.1-k2 Copyright (c) 1999-2004 Intel Corporation. eth0: Intel(R) PRO/1000 Network Connection If I do the installation interactively (using PXE / NFS), everything works and I have no problems with the network - only when I try to kickstart. I also use NFS during the interactive installation and at other times as well - with no problems. I have worked on isolating this by building a new initrd.img file (with help thanks to Richard Black) that can run a bash script in place of the binary /linuxrc that is in the initial file. I tried loading the e1000.ko file then adding a delay. I got all of this working - as evidenced by output to the console. However, anaconda ignores this and reloads the e1000.ko file later - just before it declares an NFS failure. Cody Collier wrote: >I have successfully used kickstart via http with this network interface >using rhel 3. It should work for rh 9 as well. (I might have used 9, but >I can't recall now.) Depending on your pxe configuration there might not >be much difference between this and an nfs source for the file. > >verification of an interface match: >03:02.0 Ethernet controller: Intel Corp. 82546EB Gigabit Ethernet Controller (Copper) (rev 01) 03:02.1 Ethernet controller: Intel Corp. >82546EB Gigabit Ethernet Controller (Copper) (rev 01)> > >This requires that you use the images from: > >(yourdistribution)/images/pxeboot/ > >This ramdisk contains all the drivers, including the e1000, rather than >the smaller subset in the normal boot images. > >Can you pass on what images you are using, as well as what your netboot >file looks like? Are you specifying the ksdevice and other ks parameters >in the append line in this file? > >cody Cody - can you also provide a sample of what parameters you use for ksdevice etc for your http setup? I haven't set up an http server for this yet but someone else told me he gave up on trying to use NFS and switched to HTTP. For my setup, I copy all of the files from all of the CDs (or from the mirror /os directory) to / (I'll use /pxe/fc2-3 for this example) I copy the vmlinuz and initrd.img files from the /pxe/fc2-3/images/pxeboot to /tftpboot/vmlinuz-fc2-3 and /tftpboot/initrd-fc2-3.img (for Fedora core 2 test 3). I created a directory and put the kickstart file there: /pxe/fc2-3/ks/ks-min.cfg The entries in the /tftpboot/pxelinux.cfg/default file are then set according to the version I want to use for installations. For example: LABEL fc2-3 KERNEL vmlinuz-fc2-3 APPEND ksdevice=eth0 console=tty0 load_ramdisk=1 initrd=initrd-fc2-3.img \ network ip=dhcp ks=nfs:10.0.124.21:/pxe/fc2-3/ks/ks-min.cfg The dhcpd.conf entry uses the MAC address of the NIC, it assigns a hostname and ip address. The e1000.ko module is present in the /modules/modules/cgz file so this is not an issue for this version (although it could still be a 'bad' module - version 5.2.30.1-k2). Joe Robertson > -- > Jesse Keating RHCE (geek.j2solutions.net) > Fedora Legacy Team (www.fedoralegacy.org) > GPG Public Key > (geek.j2solutions.net/jkeating.j2solutions.pub) > > Was I helpful? Let others know: > http://svcs.affero.net/rm.php?r=jkeating > From mlblandf at sedona.ch.intel.com Fri May 7 22:05:40 2004 From: mlblandf at sedona.ch.intel.com (Michael Blandford) Date: Fri, 07 May 2004 15:05:40 -0700 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: References: Message-ID: <409C0834.3080407@sedona.intel.com> Joe Robertson wrote: >I have worked on isolating this by building a new initrd.img file (with >help thanks to Richard Black) that can run a bash script in place of >the binary /linuxrc that is in the initial file. I tried loading the >e1000.ko file then adding a delay. I got all of this working - as >evidenced >by output to the console. However, anaconda ignores this and >reloads the e1000.ko file later - just before it declares an NFS >failure. > > Make sure portfast is enabled on your switch. I know that causes us problems on gig cards if it isn't. Michael Disclaimer: The content of this message is my personal opinion only and although I am an employee of Intel, the statements I make here in no way represent Intel's position on the issue, nor am I authorized to speak on behalf of Intel on this matter. From ccollier at viawest.net Fri May 7 22:06:27 2004 From: ccollier at viawest.net (ccollier at viawest.net) Date: Fri, 7 May 2004 16:06:27 -0600 (MDT) Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: Message-ID: On Fri, 7 May 2004, Joe Robertson wrote: --snip-- > > Cody - can you also provide a sample of what parameters you use for > ksdevice etc > for your http setup? I haven't set up an http server for this > yet but > someone else told me he gave up on trying to use NFS and > switched to HTTP. [somehost]$cat /tftpboot/linux-install/pxelinux.cfg/netboot-rhel3_es-basic default linux label linux kernel vmlinuz-rhel3_es append initrd=initrd-rhel3_es.img vnc ksdevice=eth0 ks=http://123.456.789.123/somepath/here/vwkickstart.py The images (vmlinuz-rhel3_es/initrd-rhel3_es.img) are just renamed and come from rhel3es-distribution/images/pxeboot/. I've found making the installation tree and the kickstart config file available via http to work very well. cody From jrobertson at convera.com Fri May 7 22:21:25 2004 From: jrobertson at convera.com (Joe Robertson) Date: Fri, 7 May 2004 15:21:25 -0700 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: > -----Original Message----- > From: Michael Blandford [mailto:mlblandf at sedona.ch.intel.com] > Sent: Friday, May 07, 2004 3:06 PM > To: Discussion list about Kickstart > Subject: Re: NFS kickstart fails with e1000.ko drivers > > > Joe Robertson wrote: > > >I have worked on isolating this by building a new initrd.img > file (with > >help thanks to Richard Black) that can run a bash script in place of > >the binary /linuxrc that is in the initial file. I tried > loading the > >e1000.ko file then adding a delay. I got all of this working - as > >evidenced by output to the console. However, anaconda > ignores this and > >reloads the e1000.ko file later - just before it declares an NFS > >failure. > > > > > Make sure portfast is enabled on your switch. I know that causes us > problems on gig cards if it isn't. > The switches I'm currently using are unmanaged netgear 24 port (or 5 port) cheap switches so there is no way to change this. I've tried using a crossover cable so the switch was eliminated altogether - same results. I also tried using a 3COM 3C39036 100 Mbit switch in default mode (I'll learn how - then try setting the portfast later to see if it helps.) Thanks, Joe > Michael > > > Disclaimer: The content of this message is my personal > opinion only and > although I am an employee of Intel, the statements I make > here in no way > represent Intel's position on the issue, nor am I authorized > to speak on > behalf of Intel on this matter. > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickst> art-list > From ryangolhar at verizon.net Fri May 7 23:05:30 2004 From: ryangolhar at verizon.net (Ryan Golhar) Date: Fri, 7 May 2004 19:05:30 -0400 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: Message-ID: <002601c43487$cbb25d70$f4aadb82@GOLHARMOBILE1> Yes, I'm having the same problem. It fails initially, but the kickstart prompts me for a new location. I just press Enter to accept the old one and it proceeds to mount the nfs share. I've already submitted a report on bugzilla on this: 117748 ----- Ryan Golhar Computational Biologist The Informatics Institute at The University of Medicine & Dentistry of NJ Phone: 973-972-5034 Fax: 973-972-7412 Email: golharam at umdnj.edu -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Joe Robertson Sent: Friday, May 07, 2004 5:05 PM To: kickstart-list at redhat.com Subject: NFS kickstart fails with e1000.ko drivers I believe this is probably an anaconda problem but since it only shows up when I'm trying to do kickstart I will send it here. Question: Has anyone has been successful in running kickstart via NFS through an Intel 82546EB network interface? I'm fighting a problem getting kickstart to run with these NICs (using NFS). It appears to be related to the NIC in use. I have setup a PXE install environment using NFS to hold a copy of the installation tree (created by copying all of the contents of all of the CDs for a given version of linux). I can do interactive installs with RH9, Fedora Core 1, Core 2/test 1-3, as well as others. These installs work without incident on all hardware I've tried so far. The problem comes in when I try to use kickstart instead of doing the interactive installs on a machine that uses the Intel 82546EB Gigabit ethernet chipset. When the installation gets to a point where it needs to access NFS (for the kickstart file or for the installation) at which time it reports that NFS has failed and tries to load from CD (according to the ALT-F3 screen) then halts. I have tried this on several Dell 1600sc and a SuperMicro 6013P-T with the same results. I added a 3COM 3C905C-TX card to the Dell 1600sc and configured the bios so I could install from that card via PXE. The kickstart installation runs just fine with this nic - no other changes. I've tried several suggestions - I used a 3COM 3C39036 switch (instead of the netgear Gb), I've connected two computers directly with a crossover cable (no switch) but cannot get kickstart to run when using the Intel 82546EB Gb nic. Any suggestions for isolating this (or solving it) would be welcomed. Thanks, Joe -------------- next part -------------- An HTML attachment was scrubbed... URL: From lccha-rhlist at naos.org Fri May 7 23:11:01 2004 From: lccha-rhlist at naos.org (lccha-rhlist at naos.org) Date: Fri, 7 May 2004 16:11:01 -0700 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: References: <409BFC06.20201@swelltech.com> Message-ID: <20040507231101.GA32303@turtleville.lccha.org> Once upon a time (like on May 07, 2004), ccollier at viawest.net wrote: > On Fri, 7 May 2004, Joe Cooper wrote: > > Joe Robertson wrote: > --snip-- > > I'd be pleased as punch if someone would tell me what I missed in my > > modifications of the kickstart initrd image, that would make this > > work...Even a simple list of files that must be changed when adding a > > new or modified driver to the kickstart image (so I could spot the one I > > didn't change) would be exceedingly useful. > --snip-- > > I struggled with some similar issues with the e1000 and rh 7.3. > Typically, for module insertion into the ramdisk you have to edit these > three files found inside the ramdisk: Hi - I too struggled with e1000 vs RH 7.3 in the past, and had similar symptoms to what Joe described. I found it easier to compile support for the e1000 into a new kernel image and use that for kickstart instead of trying to add the e1000 module into the ramdisk. We've had a few sets of machines with this card. When we purchased our machines, the rev of the card was changing pretty frequently, so we'd have to check for updated e1000 drivers from Intel's website pretty often. Hopefully things are more stable now. -L From jrobertson at convera.com Fri May 7 23:29:57 2004 From: jrobertson at convera.com (Joe Robertson) Date: Fri, 7 May 2004 16:29:57 -0700 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: > -----Original Message----- > From: lccha-rhlist at naos.org [mailto:lccha-rhlist at naos.org] > Sent: Friday, May 07, 2004 4:11 PM > To: Discussion list about Kickstart > Subject: Re: NFS kickstart fails with e1000.ko drivers > > > Once upon a time (like on May 07, 2004), ccollier at viawest.net wrote: > > On Fri, 7 May 2004, Joe Cooper wrote: > > > Joe Robertson wrote: > > --snip-- > > > I'd be pleased as punch if someone would tell me what I > missed in my > > > modifications of the kickstart initrd image, that would make this > > > work...Even a simple list of files that must be changed > when adding a > > > new or modified driver to the kickstart image (so I could > spot the one I > > > didn't change) would be exceedingly useful. > > --snip-- > > > > I struggled with some similar issues with the e1000 and rh 7.3. > > Typically, for module insertion into the ramdisk you have > to edit these > > three files found inside the ramdisk: > > Hi - > > I too struggled with e1000 vs RH 7.3 in the past, and had > similar symptoms to what Joe described. I found it easier to > compile support for the e1000 into a new kernel image and use > that for kickstart instead of trying to add the e1000 module > into the ramdisk. > > We've had a few sets of machines with this card. When we > purchased our machines, the rev of the card was changing > pretty frequently, so we'd have to check for updated e1000 > drivers from Intel's website pretty often. Hopefully things > are more stable now. > > -L > I'm using Fedora Core 2 test 3 and the e1000 driver is already included in the ramdisk - /modules/modules.cgz I'm not familiar with compiling the kernel image. I assume that if it is in the ramdisk, then it probably isn't compiled into the kernel in the distribution? Joe > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickst> art-list > From joe at swelltech.com Fri May 7 23:38:37 2004 From: joe at swelltech.com (Joe Cooper) Date: Fri, 07 May 2004 18:38:37 -0500 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: <20040507231101.GA32303@turtleville.lccha.org> References: <409BFC06.20201@swelltech.com> <20040507231101.GA32303@turtleville.lccha.org> Message-ID: <409C1DFD.1080201@swelltech.com> lccha-rhlist at naos.org wrote: > Once upon a time (like on May 07, 2004), ccollier at viawest.net wrote: > >>On Fri, 7 May 2004, Joe Cooper wrote: >> >>>Joe Robertson wrote: >> >>--snip-- >> >>>I'd be pleased as punch if someone would tell me what I missed in my >>>modifications of the kickstart initrd image, that would make this >>>work...Even a simple list of files that must be changed when adding a >>>new or modified driver to the kickstart image (so I could spot the one I >>>didn't change) would be exceedingly useful. >> >>--snip-- >> >>I struggled with some similar issues with the e1000 and rh 7.3. >>Typically, for module insertion into the ramdisk you have to edit these >>three files found inside the ramdisk: > > > Hi - > > I too struggled with e1000 vs RH 7.3 in the past, and had similar > symptoms to what Joe described. I found it easier to compile support > for the e1000 into a new kernel image and use that for kickstart > instead of trying to add the e1000 module into the ramdisk. > > We've had a few sets of machines with this card. When we purchased > our machines, the rev of the card was changing pretty frequently, so > we'd have to check for updated e1000 drivers from Intel's website > pretty often. Hopefully things are more stable now. Thanks for the suggestion. I've give that a spin. But I'm not extremely hopeful, as manually specifying e1000 also didn't work. From jrobertson at convera.com Fri May 7 23:47:54 2004 From: jrobertson at convera.com (Joe Robertson) Date: Fri, 7 May 2004 16:47:54 -0700 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: > -----Original Message----- > From: ccollier at viawest.net [mailto:ccollier at viawest.net] > Sent: Friday, May 07, 2004 3:06 PM > To: Discussion list about Kickstart > Subject: RE: NFS kickstart fails with e1000.ko drivers > > > On Fri, 7 May 2004, Joe Robertson wrote: > --snip-- > > > > Cody - can you also provide a sample of what parameters you use for > > ksdevice etc > > for your http setup? I haven't set up an http server > for this yet > > but > > someone else told me he gave up on trying to use NFS > and switched to > > HTTP. > > [somehost]$cat > /tftpboot/linux-install/pxelinux.cfg/netboot-rhel3_es-basic > default linux > label linux > kernel vmlinuz-rhel3_es > append initrd=initrd-rhel3_es.img vnc ksdevice=eth0 > ks=http://123.456.789.123/somepath/here/vwkickstart.py > > The images (vmlinuz-rhel3_es/initrd-rhel3_es.img) are just > renamed and > come from rhel3es-distribution/images/pxeboot/. I've found > making the > installation tree and the kickstart config file available via > http to work > very well. > > cody > I setup a configuration with the kickstart file on an http server. I created a label with the entries to match your example. When I tried to run, it failed to retrieve the file (ALT-F3 console messages follow): ... * load module set done * getting kickstart file * sending dhcp request through device eth0 * waiting for link... * 0 seconds. * doing kickstart... setting it up * ks location http://10.0.125.106/pxe/fc2-3/ks/ks-min.cfg * transferring http://10.0.125.106//./pxe/fc2-3/ks/ks-min.cfg to a fd * failed to retrieve http://10.0.125.106///pxe/fc2-3/ks/ks-min.cfg * trying to mount CD device hdc I can open a browser and enter the url specified as ks location above and it will display the contents of the kickstart configuration file. This sounds like the same issue - the NIC driver... or anaconda... I had another instance of this when I tried to use pxe to boot into a diskless environment - the dhcp lease was failing. The /linuxrc in the ramdisk was a script so I had it try again to get a lease if the first try failed. It always succeeded. This whole thing sounds like a timing issue and something is simply not giving the driver enough time before it gives up and reports a failure. Joe > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickst> art-list > From seph at directionless.org Sat May 8 02:11:49 2004 From: seph at directionless.org (seph) Date: Fri, 07 May 2004 22:11:49 -0400 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: (Joe Robertson's message of "Fri, 7 May 2004 15:21:25 -0700") References: Message-ID: > The switches I'm currently using are unmanaged netgear 24 port > (or 5 port) cheap switches so there is no way to change this. > I've tried using a crossover cable so the switch was eliminated > altogether - same results. I also tried using a 3COM 3C39036 > 100 Mbit switch in default mode (I'll learn how - then try > setting the portfast later to see if it helps.) I observed the same problem, and mentioned it here at the end of march. unfortunately, I then got buried in work, and didn't get a chance to debug it. I'd observed that kickstart failed for my dells with intel gige on the motherboard. But only when hooked up to netgear equipment. Failed on both a netgear gige switch, as well as a netgear 10/100 hub. Worked okay on a cheap linksys switch though. I also observed the same failure. Namely, I could manually walk through an install, but kickstart always failed to fetch its config file. Interestingly, it successfully used dhcp to get an ip address. I'm certainly interested in whatever results you end up with. seph From ryangolhar at verizon.net Sat May 8 03:08:02 2004 From: ryangolhar at verizon.net (Ryan Golhar) Date: Fri, 7 May 2004 23:08:02 -0400 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: Message-ID: <003f01c434a9$ad24e450$f4aadb82@GOLHARMOBILE1> I observed the same thing with a Cisco switch. The switch was set at 10 Mbps half-duplex. The fact that it was obtaining a DHCP IP address threw me off for quite awhile. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of seph Sent: Friday, May 07, 2004 10:12 PM To: Discussion list about Kickstart Subject: Re: NFS kickstart fails with e1000.ko drivers > The switches I'm currently using are unmanaged netgear 24 port > (or 5 port) cheap switches so there is no way to change this. > I've tried using a crossover cable so the switch was eliminated > altogether - same results. I also tried using a 3COM 3C39036 > 100 Mbit switch in default mode (I'll learn how - then try > setting the portfast later to see if it helps.) I observed the same problem, and mentioned it here at the end of march. unfortunately, I then got buried in work, and didn't get a chance to debug it. I'd observed that kickstart failed for my dells with intel gige on the motherboard. But only when hooked up to netgear equipment. Failed on both a netgear gige switch, as well as a netgear 10/100 hub. Worked okay on a cheap linksys switch though. I also observed the same failure. Namely, I could manually walk through an install, but kickstart always failed to fetch its config file. Interestingly, it successfully used dhcp to get an ip address. I'm certainly interested in whatever results you end up with. seph _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From jrobertson at convera.com Sun May 9 00:26:38 2004 From: jrobertson at convera.com (Joe Robertson) Date: Sat, 8 May 2004 17:26:38 -0700 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: Well, I've done quite a bit of testing to characterize this year-old problem when using Gbit ethernet interface cards. I believe the answer is somewhere in anaconda and is timing related. My next step is to load anaconda source and try to figure out how to fix this issue (I don't know anaconda nor python so that will be a slow process). This problem appears this has been in the discussion world for the past year now and is still a problem as evidenced by the number of responses I've seen to my first post. I'm pretty well convinced this is simply a timing issue and anaconda is declaring a failure too early. Perhaps someone with anaconda knowledge can point me in a direction that might help? My test effort / results to date: With considerable help from Richard Black, I've gone through modifying the initrd.img file to insert a script that loads the NIC driver then sleep for awhile before anaconda starts. The changes made to implement this idea work but they don't help - anaconda loads the driver again anyway. I then tried removing the entry for the e1000 driver for my card in the /modules/pcitable file in the initrd.img file. This didn't work since I then got a report that there was no network driver and the install terminated. Based on a suggesion by seph, I have experimented with connecting the pxe installation client to one of several different switch models. I found that some switches appear to work around the problem (Unfortunately I can't use any of the switches that do work around it - they are all 10/100). This does let me validate all of my setup and I know that the kickstart configuration is correct. After finding a switch that allowed a kickstart via network, I immediately ran into a bug reported earlier by Martin Robb (https://listman.redhat.com/archives/kickstart-list/2004-April/msg00107. html) and I cannot get Fedora Core 2 test 3 to finish a kickstart install (error installing at-3.1.8-52). Does anyone know of a fix for this yet? I was able to get Core 2 test 2 to run kickstart to completion. I tested the following switches and found some that worked and some that didn't. I tested with the kickstart file via NFS and HTTP and the installation files via NFS. Switches used when the kickstart install was successful (via network): Linksys EZXS16W 16 Port EtherFast 10/100 (Unmanaged) (HTTP ks failed once - tried again it worked) Intel InBusiness 16 Port 10/100 Switch (Unmanaged) Netgear FS108 8 Port 10/100 Switch (Unmanaged) Switches used when the install failed to get the kickstart file (NFS or HTTP): No switch (used crossover cable directly from pxe install server and client. The install server and client both use the same Gbit ethernet type) Netgear GS105 5 Port Gigabit Switch (Unmanaged) Netgear GS524T 24 Port Gigabit Switch (Unmanaged) 3COM 3C39036 36 Port 10/100 (Managed) Spanning Tree disabled Autonegotiate disabled Port set to 100half and 100full (separate tests) Note: Interactive installs work in all of these cases and I have no problems with the network once the installation is done. The only failure mode here is when anaconda loads the drivers and attempts to do an NFS or HTTP access for kickstart. I tried loading the kickstart file from floppy and the failure was simply delayed until it needed to access the NFS for the install files. A suggestion was made by lccha-rhlist at naos.org to update the e1000.ko NIC driver with the latest from Intel. I looked at their download site and their latest appears to be the same version that is in test 3. The Intel version is 5.2.39 and the Fedora Core 2 test 3 version is (reported when loaded by insmod) as 5.2.39-k2. I've been pulling my hair out on this one and really need to resolve it. I'm preparing to install a large number of machines using pxe installs and I don't want to have to do each one interactively. It would be great to get a fix for this in the Core 2 release. Thanks for any suggestions / help. Joe > -----Original Message----- > From: Ryan Golhar [mailto:ryangolhar at verizon.net] > Sent: Friday, May 07, 2004 8:08 PM > To: 'Discussion list about Kickstart' > Subject: RE: NFS kickstart fails with e1000.ko drivers > > > I observed the same thing with a Cisco switch. The switch > was set at 10 Mbps half-duplex. The fact that it was > obtaining a DHCP IP address threw me off for quite awhile. > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On > Behalf Of seph > > Sent: Friday, May 07, 2004 10:12 PM > To: Discussion list about Kickstart > Subject: Re: NFS kickstart fails with e1000.ko drivers > > > > The switches I'm currently using are unmanaged netgear 24 > port (or 5 > > port) cheap switches so there is no way to change this. > > I've tried using a crossover cable so the switch was eliminated > > altogether - same results. I also tried using a 3COM 3C39036 > > 100 Mbit switch in default mode (I'll learn how - then try > > setting the portfast later to see if it helps.) > > I observed the same problem, and mentioned it here at the end > of march. unfortunately, I then got buried in work, and > didn't get a chance to debug it. > > I'd observed that kickstart failed for my dells with intel > gige on the motherboard. But only when hooked up to netgear > equipment. Failed on both a netgear gige switch, as well as a > netgear 10/100 hub. Worked okay on a cheap linksys switch though. > > I also observed the same failure. Namely, I could manually > walk through an install, but kickstart always failed to fetch > its config file. Interestingly, it successfully used dhcp to > get an ip address. > > I'm certainly interested in whatever results you end up with. > > seph > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickst> art-list > > > > _______________________________________________ > > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickst> art-list > From lccha-rhlist at naos.org Sun May 9 01:58:50 2004 From: lccha-rhlist at naos.org (lccha-rhlist at naos.org) Date: Sat, 8 May 2004 18:58:50 -0700 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: References: Message-ID: <20040509015850.GA18013@turtleville.lccha.org> Once upon a time (like on May 07, 2004), Joe Robertson wrote: > > I too struggled with e1000 vs RH 7.3 in the past, and had > > similar symptoms to what Joe described. I found it easier to > > compile support for the e1000 into a new kernel image and use > > that for kickstart instead of trying to add the e1000 module > > into the ramdisk. ... > I'm using Fedora Core 2 test 3 and the e1000 driver is already > included in the ramdisk - /modules/modules.cgz > > I'm not familiar with compiling the kernel image. I assume > that if it is in the ramdisk, then it probably isn't > compiled into the kernel in the distribution? > > Joe Yes, that's probably true. It could still be compiled into the kernel, but it's not likely that they would bother including the module as well in that case. The question is, what version of the e1000 driver is already included in the ramdisk? Our experience has been that we needed the very latest e1000 driver to get our newer e1000 cards to work. -L From Thomas.Cameron at bankofamerica.com Sun May 9 03:19:00 2004 From: Thomas.Cameron at bankofamerica.com (Cameron, Thomas) Date: Sat, 08 May 2004 22:19:00 -0500 Subject: kickstart on Linux on zSeries? Message-ID: <1C455916B3418145B320711B21A2F04515D04C@ex2k.bankofamerica.com> Howdy all - I am working with 64-bit RHEL 3 for zSeries. We are currently thinking that we will be using a mainframe tool (ddr, similar to the dd command under *nix) to copy a single "golden image" to multiple production servers. I want to explore using kickstart to build servers instead, but I don't know if it is possible. For those of you not familiar with Linux on the mainframe, it doesn't boot with the Syslinux boot screen or even a floppy or compact disk. Instead you load the kernel and the initrd.img into a virtual punchcard reader and then load it into the mainframe CPU/memory. I have read that there is a RUNKS ("run kickstart") option listed for the mainframe parm (configuration) file, but I do not know how to actually specify the kickstart file. I know how to copy the ks.cfg into the initrd.img file, but does it go into the root of the initrd, or in /tmp, or somewhere else? Are there additional parm file options I need which are not discussed at https://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/s390-multi-install-guide/s1-s390-steps-vm.html? Anyone who has any pointers, I am all ears! Thanks! Thomas Cameron, RHCE, CNE, MCSE, MCT Assistant Vice President Linux Design and Engineering Bank of America (972) 997-9641 From seph at directionless.org Sun May 9 05:48:14 2004 From: seph at directionless.org (seph) Date: Sun, 09 May 2004 01:48:14 -0400 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: (Joe Robertson's message of "Sat, 8 May 2004 17:26:38 -0700") References: Message-ID: > Well, I've done quite a bit of testing to characterize > this year-old problem when using Gbit ethernet interface > cards. I believe the answer is somewhere in anaconda and > is timing related. To clarify my experience, I encountered these problems when using the intel e1000 card on my dell motherboard. I did NOT have any problems when using dell supplied pci broadcom gbit cards. So it's clearly not all gbit ethernet. seph From Rebecca.R.Hepper at seagate.com Mon May 10 12:16:49 2004 From: Rebecca.R.Hepper at seagate.com (Rebecca.R.Hepper at seagate.com) Date: Mon, 10 May 2004 07:16:49 -0500 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: I am seeing the same issue with RH 9 & an e1000 module I compiled under the 2.4.20-6BOOT kernel. I am kickstarting via a USB flash drive. When kickstart tries to mount the NFS server, I receive an error "That directory could not be mounted from the server". If I hit 'OK' to accept the NFS Setup information it proceeds to mount the NFS server. Should I add my information to bugzilla 117748? |---------+---------------------------------> | | "Ryan Golhar" | | | | | | Sent by: | | | kickstart-list-bounces| | | @redhat.com | | | No Phone Info | | | Available | | | | | | 05/07/2004 06:05 PM | | | Please respond to | | | golharam; Please | | | respond to Discussion | | | list about Kickstart | | | | |---------+---------------------------------> >------------------------------------------------------------------------------------------------------------------------------| | | | To: "'Discussion list about Kickstart'" | | cc: | | Subject: RE: NFS kickstart fails with e1000.ko drivers | >------------------------------------------------------------------------------------------------------------------------------| Yes, I'm having the same problem. It fails initially, but the kickstart prompts me for a new location. I just press Enter to accept the old one and it proceeds to mount the nfs share. I've already submitted a report on bugzilla on this: 117748 ----- Ryan Golhar Computational Biologist The Informatics Institute at The University of Medicine & Dentistry of NJ Phone: 973-972-5034 Fax: 973-972-7412 Email: golharam at umdnj.edu -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Joe Robertson Sent: Friday, May 07, 2004 5:05 PM To: kickstart-list at redhat.com Subject: NFS kickstart fails with e1000.ko drivers I believe this is probably an anaconda problem but since it only shows up when I'm trying to do kickstart I will send it here. Question: Has anyone has been successful in running kickstart via NFS through an Intel 82546EB network interface? I'm fighting a problem getting kickstart to run with these NICs (using NFS). It appears to be related to the NIC in use. I have setup a PXE install environment using NFS to hold a copy of the installation tree (created by copying all of the contents of all of the CDs for a given version of linux). I can do interactive installs with RH9, Fedora Core 1, Core 2/test 1-3, as well as others. These installs work without incident on all hardware I've tried so far. The problem comes in when I try to use kickstart instead of doing the interactive installs on a machine that uses the Intel 82546EB Gigabit ethernet chipset. When the installation gets to a point where it needs to access NFS (for the kickstart file or for the installation) at which time it reports that NFS has failed and tries to load from CD (according to the ALT-F3 screen) then halts. I have tried this on several Dell 1600sc and a SuperMicro 6013P-T with the same results. I added a 3COM 3C905C-TX card to the Dell 1600sc and configured the bios so I could install from that card via PXE. The kickstart installation runs just fine with this nic - no other changes. I've tried several suggestions - I used a 3COM 3C39036 switch (instead of the netgear Gb), I've connected two computers directly with a crossover cable (no switch) but cannot get kickstart to run when using the Intel 82546EB Gb nic. Any suggestions for isolating this (or solving it) would be welcomed. Thanks, Joe _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From phr at doc.ic.ac.uk Wed May 12 10:32:05 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 12 May 2004 11:32:05 +0100 (BST) Subject: RHEL3U2 anaconda changes Message-ID: https://rhn.redhat.com/errata/RHBA-2004-195.html To quote (phr-edited): " Bug fixes include: - Increase amount of time we wait for network controllers to establish a link - Include more drivers on boot images " Cheers, Phil From jrobertson at convera.com Wed May 12 15:26:01 2004 From: jrobertson at convera.com (Joe Robertson) Date: Wed, 12 May 2004 08:26:01 -0700 Subject: RHEL3U2 anaconda changes Message-ID: Philip, This sounds like good news. Hopefully these changes will address the problems I've been seeing with the e1000 NIC drivers. However, the fixes are only listed for the Enterprise versions of Redhat... I'm using Fedora Core 2. Hopefully these changes can find their way into the final release of Core 2!! If not - I'm trying to learn how to go through the process of building new anaconda rpms from src (I've done that) then putting these new rpms into an NFS installation tree - including new pxeboot images (I have not done this yet). Does anyone have a pointer as to how I can do this? Some instructions I've seen refer to installing anaconda - but I'm trying to put it into the installation tree and tftpboot image directory so that new installations make use of it during install. Thanks, Joe > -----Original Message----- > From: Philip Rowlands [mailto:phr at doc.ic.ac.uk] > Sent: Wednesday, May 12, 2004 3:32 AM > To: kickstart-list at redhat.com > Subject: RHEL3U2 anaconda changes > > > https://rhn.redhat.com/errata/RHBA-2004-195.html > > To quote (phr-edited): > > " > Bug fixes include: > - Increase amount of time we wait for network controllers to > establish a link > - Include more drivers on boot images > " > > > Cheers, > Phil > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickst> art-list > From Avi.Koski at comverse.com Thu May 13 11:27:38 2004 From: Avi.Koski at comverse.com (Koski Avi) Date: Thu, 13 May 2004 14:27:38 +0300 Subject: Creating custom CD for AS3.0 & WS3.0 Message-ID: <32B823C1CD4FD5119C0D0002A560F78E0CC48F3C@ismail3.comverse.com> Hi list, Can I get some howto tips, for creation of a custom CD ? Especially, how can I update the RedHat/base/comps.rpm file ? There are some excellent docs for customizing RH7.2 CD. Is there any docs for AS3.0 & WS3.0 ? Thanks, Avi Koski -------------- next part -------------- An HTML attachment was scrubbed... URL: From blin at na.uni-tuebingen.de Fri May 14 10:05:25 2004 From: blin at na.uni-tuebingen.de (Kai Blin) Date: Fri, 14 May 2004 12:05:25 +0200 Subject: sshd during anaconda install Message-ID: <200405141205.25430.blin@na.uni-tuebingen.de> Hi folks, I noticed a strange behavior of sshd during a FC1 kickstart install. We use rsync/ssh to copy over config files and the like to our boxes in the %post script. For some reason, with FC, we're unable to log into a sshd running in the /mnt/sysimage changeroot using hostbased authentication. ssh -v shows that it gets stuck when trying to set up the hostbased authentication: debug1: Host 'na8' is known and matches the RSA host key. debug1: Found key in /etc/ssh/ssh_known_hosts2:146 debug1: ssh_rsa_verify: signature correct debug1: SSH2_MSG_NEWKEYS sent debug1: expecting SSH2_MSG_NEWKEYS debug1: SSH2_MSG_NEWKEYS received debug1: SSH2_MSG_SERVICE_REQUEST sent debug1: SSH2_MSG_SERVICE_ACCEPT received debug1: Authentications that can continue: publickey,password,keyboard-interactive,hostbased debug1: Next authentication method: hostbased After that, it doesn't print out anything else. I'm also attaching the sshd_config file from that box. Any clue what could be wrong? Thanks, Kai -- Kai Blin, Sysop Dept. of Numerical Algebra, University of T?bingen, Germany -------------- next part -------------- # This is ssh server systemwide configuration file. Port 22 #Protocol 1,2 Protocol 2 ListenAddress 0.0.0.0 HostKey /etc/ssh/ssh_host_key HostKey /etc/ssh/ssh_host_rsa_key HostKey /etc/ssh/ssh_host_dsa_key ServerKeyBits 768 LoginGraceTime 600 KeyRegenerationInterval 3600 X11Forwarding yes X11DisplayOffset 10 KeepAlive yes # Logging SyslogFacility LOCAL0 LogLevel INFO IgnoreRhosts no IgnoreUserKnownHosts no PermitEmptyPasswords no #PermitRootLogin without-password StrictModes yes # configurations for both protocol versions PasswordAuthentication yes # Protocol version 1 configurations RhostsAuthentication no RhostsRSAAuthentication yes RSAAuthentication yes # Protocol version 2 configurations HostbasedAuthentication yes PubkeyAuthentication yes # Uncomment to disable s/key passwords # SkeyAuthentication no # To change Kerberos options # KerberosAuthentication no # KerberosOrLocalPasswd yes # AFSTokenPassing no # KerberosTicketCleanup no # Kerberos TGT Passing does only work with the AFS kaserver # KerberosTgtPassing yes #CheckMail no UseLogin no PrintMotd no #Subsystem sftp /usr/local/sbin/sftpd From pantz at lqt.ca Fri May 14 14:27:27 2004 From: pantz at lqt.ca (Paul Pianta) Date: Fri, 14 May 2004 10:27:27 -0400 Subject: %post trouble Message-ID: <40A4D74F.7070001@lqt.ca> Hi people I have a question about kickstart in Fedora Core 1. In %post I am trying to copy a file from the cd onto the newly installed system (it is a modified mysqld init script - can't be bothered rebuilding mysql rpm just for that) This is what I have tried so far for tests but none of it has worked: %post --nochroot echo "in nochroot post" echo "doing first mkdir" mkdir -p /mnt/source echo "doing first mount" mount /tmp/cdrom /mnt/source echo "trying first copy" cp -f /mnt/source/Fedora/RPMS/extras/mysqld.init /mnt/sysimage/etc/rc.d/init.d/mysqld echo "trying first unmount" umount /mnt/source %post echo "in regular post" echo "doing second mkdir" mkdir -p /mnt/source echo "doing second mount" mount /tmp/cdrom /mnt/source echo "trying second copy" cp -f /mnt/source/Fedora/RPMS/extras/mysqld.init /mnt/sysimage/etc/rc.d/init.d/mysqld echo "trying second unmount" umount /mnt/source I had done this in redhat 7.3 with just a simple: %post --nochroot cp -f /mnt/source/Fedora/RPMS/extras/mysqld.init /mnt/sysimage/etc/rc.d/init.d/mysqld This doesn't work for me with Fedora Core 1 - any ideas how I can run my copy command while the cd is mounted - either mounted by me or still mounted from install? thanks -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From ebrown at lanl.gov Fri May 14 14:48:21 2004 From: ebrown at lanl.gov (Ed Brown) Date: Fri, 14 May 2004 08:48:21 -0600 Subject: %post trouble In-Reply-To: <40A4D74F.7070001@lqt.ca> References: <40A4D74F.7070001@lqt.ca> Message-ID: <1084546101.15522.12.camel@edbrown.lanl.gov> My favorite, often overlooked, debugging and development tool: open a shell in %post: %post chvt 3 exec < /dev/tty3 > /dev/tty3 sh Now you are in a chrooted shell, and you can alt+F2 to the no-chrooted shell console. When you exit the shell, on tty3, the %post script continues. Also, the "chvt 3" would enable you to see the output of your echo statements. hth, Ed On Fri, 2004-05-14 at 08:27, Paul Pianta wrote: > Hi people > > I have a question about kickstart in Fedora Core 1. In %post I am trying > to copy a file from the cd onto the newly installed system (it is a > modified mysqld init script - can't be bothered rebuilding mysql rpm > just for that) > > This is what I have tried so far for tests but none of it has worked: > %post --nochroot > echo "in nochroot post" > echo "doing first mkdir" > mkdir -p /mnt/source > echo "doing first mount" > mount /tmp/cdrom /mnt/source > echo "trying first copy" > cp -f /mnt/source/Fedora/RPMS/extras/mysqld.init > /mnt/sysimage/etc/rc.d/init.d/mysqld > echo "trying first unmount" > umount /mnt/source > > %post > echo "in regular post" > echo "doing second mkdir" > mkdir -p /mnt/source > echo "doing second mount" > mount /tmp/cdrom /mnt/source > echo "trying second copy" > cp -f /mnt/source/Fedora/RPMS/extras/mysqld.init > /mnt/sysimage/etc/rc.d/init.d/mysqld > echo "trying second unmount" > umount /mnt/source > > > I had done this in redhat 7.3 with just a simple: > %post --nochroot > cp -f /mnt/source/Fedora/RPMS/extras/mysqld.init > /mnt/sysimage/etc/rc.d/init.d/mysqld > > This doesn't work for me with Fedora Core 1 - any ideas how I can run my > copy command while the cd is mounted - either mounted by me or still > mounted from install? > > thanks From James_Martin at ao.uscourts.gov Fri May 14 14:49:52 2004 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Fri, 14 May 2004 10:49:52 -0400 Subject: %post trouble In-Reply-To: <40A4D74F.7070001@lqt.ca> Message-ID: Try doing it this way... should save you much work.. %post #mount the cdrom updfstab mount /mnt/cdrom cp /mnt/cdrom/Fedora/RPMS/extras/mysqld.init /etc/rc.d/init.d/mysqld That should be all you need.. the updfstab updates the /etc/fstab with /mnt/cdrom and points it to the correct device. I'm sure you can figure out the rest :). Good luck! James James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 05/14/2004 10:27:27 AM: > Hi people > > I have a question about kickstart in Fedora Core 1. In %post I am trying > to copy a file from the cd onto the newly installed system (it is a > modified mysqld init script - can't be bothered rebuilding mysql rpm > just for that) > > This is what I have tried so far for tests but none of it has worked: > %post --nochroot > echo "in nochroot post" > echo "doing first mkdir" > mkdir -p /mnt/source > echo "doing first mount" > mount /tmp/cdrom /mnt/source > echo "trying first copy" > cp -f /mnt/source/Fedora/RPMS/extras/mysqld.init > /mnt/sysimage/etc/rc.d/init.d/mysqld > echo "trying first unmount" > umount /mnt/source > > %post > echo "in regular post" > echo "doing second mkdir" > mkdir -p /mnt/source > echo "doing second mount" > mount /tmp/cdrom /mnt/source > echo "trying second copy" > cp -f /mnt/source/Fedora/RPMS/extras/mysqld.init > /mnt/sysimage/etc/rc.d/init.d/mysqld > echo "trying second unmount" > umount /mnt/source > > > I had done this in redhat 7.3 with just a simple: > %post --nochroot > cp -f /mnt/source/Fedora/RPMS/extras/mysqld.init > /mnt/sysimage/etc/rc.d/init.d/mysqld > > This doesn't work for me with Fedora Core 1 - any ideas how I can run my > copy command while the cd is mounted - either mounted by me or still > mounted from install? > > thanks > > -- > Before you criticize someone, walk a mile in their shoes ... > That way when you do criticize them, you're a mile away and you have > their shoes! > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From throwaway5549 at earthlink.net Fri May 14 19:17:08 2004 From: throwaway5549 at earthlink.net (Gary Ford) Date: Fri, 14 May 2004 13:17:08 -0600 Subject: Newbie Needs Help with Kickstart Fedora Installations Message-ID: <4AFB5F82-A5DB-11D8-A7CC-000393C353D0@earthlink.net> I am a Fedora newbie who needs help preparing for a Kickstart installation. I hope to be able to create several different Kickstart configuration files to test several different Fedora configurations. My test environment is an Ethernet network with one Windows 2000 Pro box, one Mac G4 OS X box, and the Fedora installation target, a Dell Dimension 4100. Only the Mac has a CD writer. The Dell box already has appropriate partitions from a previous Red Hat 9 installation. I would like to use the following scenario for my tests: 1. The Fedora installation CD-ROMs will have their contents transferred to an appropriate directory structure on the Win2000 box. The files will be served via HTTP (the Apache web server is running on that host). 2. If needed, Win2000 box can run a DHCP server to assign a static IP address to the Dell box during Fedora installation. 3. The Dell box will boot from Fedora installation CD 1. 4. I would like to have all the various Kickstart configuration files on the Win2000 box and have the installer access them via HTTP/Apache. 5. I would eventually like the option of adding additional RPMs to the Fedora/RPMS directory on the Win2000 server, with the goal of being able to test software not included in the Fedora core distribution. I need help with the following items: 1. Is my scenario feasible? 2. So far, I've had no luck getting a Kickstart installation to run. The Apache server log shows that the ks.cfg file is being requested and sent, but the installer seems to ignore it and proceed with a completely interactive installation. I'm using the boot option "linux ks=http://192.168.0.1/Linux/ks.cfg". Do I need additional boot options? 3. If I add RPMs to the Fedora/RPMS directory, how does that affect files in Fedora/base? I understand "comps.xml" and have a tool on the Mac OS X box that can generate it and a Kickstart configuration file after "interviewing" the user to determine the packages that the user is likely to need. I do not understand the "hdlist", "hdlist2", "hdstg2.img", "netstg2.img", or "stage2.img" files, their roles in installation, and how to create them. 4. The RH9 documentation for an HTTP installation of Linux makes reference to the "installation tree". The only definition of that term I could find says it means the entire contents of all the installation CD-ROMs. That seems extreme (does the installer really need the DOS utilities or the boot diskette image file?). What does "installation tree" really mean? 5. Is it okay to boot the Dell host from Fedora CD 1, or is there a better choice? Thanks for any help you can provide. I'd especially appreciate specific references to existing, beginner-oriented documentation. Gary Ford Plaid Flannel Software From jbrash at us.ibm.com Fri May 14 19:45:44 2004 From: jbrash at us.ibm.com (Jason Brashear) Date: Fri, 14 May 2004 14:45:44 -0500 Subject: Create a Kickstart cdrom for AS 3.0 In-Reply-To: <4AFB5F82-A5DB-11D8-A7CC-000393C353D0@earthlink.net> Message-ID: Hello I am new to Kickstart. I wonder if anyone knows how to create a kickstart CDROM. I am even having trouble creating a Kickstart Disk. Thank you in advance for all your help! Best Regards, Jason Brashear jbrash at us.ibm.com Websphere Lab Administrator IBM Corporation 11501 Burnet Rd., Austin, Texas 78758 LWWS/7G , Bldg. 903, 4H-004 512-838-4776 T/L 678-4776 FAX: 512-838-0330 (5C-007) -------------- next part -------------- An HTML attachment was scrubbed... URL: From rick_dennis at earthlink.net Fri May 14 21:41:59 2004 From: rick_dennis at earthlink.net (Rick Dennis) Date: Fri, 14 May 2004 17:41:59 -0400 (EDT) Subject: Providing ks.cfg with dhcp In-Reply-To: References: <1083280401.2206.692.camel@apathy.liquidweb.com> Message-ID: <3049.192.168.1.4.1084570919.squirrel@192.168.1.2> I don't think this will work. You can't put your ks.cfg file in the dhcpd.conf. That "filename" that's passed back from dhcp is typically used by pxe, but to my knowledge is not used by "loader" to reference the ks.cfg file. If that's what you want, you're better off setting up a kickstart+pxe environment. --Rick > Assuming that you didn't do a typo with the rh9.cfg/ks.cfg thing... > > You might want to try this in your dhcpd.conf > host your_hostname { hardware ethernet your_ethernet_address; > filename "/home/kickstart/ks.cfg"; } > > dhcpd should log to your messages log with syslog so if it's not working > it might give you a clue. > > The alternate file it's looking for is a default location that ks looks > for so you could always create a link there and leave all of file stuff > out of dhcp. > > Also you don't need to type 'linux ks' apparently it works but 'ks' is > enough. > > Hope this helps. > -Kelly > > On Thu, 29 Apr 2004, Joseph Glass wrote: > >> Hi everyone, >> >> I'm trying to provide my ks.cfg over dhcp so I only need to type 'linux >> ks' at the boot prompt. The dhcpd/kickstart server are on the same >> machine (192.168.0.1.) dhcp works, as well as kickstart if I run the >> following command manually: >> >> linux ks=nfs:192.168.0.1:/home/kickstart/rh9.cfg >> >> In my dhcpd.conf file, I added following line in the subnet {} section: >> >> filename "/home/kickstart/ks.cfg"; >> >> My /etc/exports file contains the following: >> >> /home/kickstart *(ro) >> >> The error I get is: >> >> bootp: no bootfile received >> >> Then it proceeds to look for the kickstart file at >> nfs:192.168.0.1:/kickstart/kickstart-192.168.0.18 >> >> Any suggestions? >> >> Thanks for helping me with the question about using perl in %pre. Since >> you can't use perl in %pre from just the install CD, I ended up porting >> my script to python which works well. And I learned a new programming >> language in the process :). >> >> Thank you. >> >> -- >> Joseph Glass >> Systems Administrator >> Liquid Web Inc. >> 800.580.4985 x227 >> >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list >> > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From kbbrown at cs.ucsd.edu Fri May 14 22:05:39 2004 From: kbbrown at cs.ucsd.edu (Kelly Brown) Date: Fri, 14 May 2004 15:05:39 -0700 (PDT) Subject: Providing ks.cfg with dhcp In-Reply-To: <3049.192.168.1.4.1084570919.squirrel@192.168.1.2> References: <1083280401.2206.692.camel@apathy.liquidweb.com> <3049.192.168.1.4.1084570919.squirrel@192.168.1.2> Message-ID: It works for me. I can't always use pxe because alot of my hardware is really old. That's the way I have it in my dhcpd.conf and all I do is type ks after booting off of a cd. On Fri, 14 May 2004, Rick Dennis wrote: > I don't think this will work. You can't put your ks.cfg file in the > dhcpd.conf. That "filename" that's passed back from dhcp is typically > used by pxe, but to my knowledge is not used by "loader" to reference the > ks.cfg file. > > If that's what you want, you're better off setting up a kickstart+pxe > environment. > > --Rick > > > Assuming that you didn't do a typo with the rh9.cfg/ks.cfg thing... > > > > You might want to try this in your dhcpd.conf > > host your_hostname { hardware ethernet your_ethernet_address; > > filename "/home/kickstart/ks.cfg"; } > > > > dhcpd should log to your messages log with syslog so if it's not working > > it might give you a clue. > > > > The alternate file it's looking for is a default location that ks looks > > for so you could always create a link there and leave all of file stuff > > out of dhcp. > > > > Also you don't need to type 'linux ks' apparently it works but 'ks' is > > enough. > > > > Hope this helps. > > -Kelly > > > > On Thu, 29 Apr 2004, Joseph Glass wrote: > > > >> Hi everyone, > >> > >> I'm trying to provide my ks.cfg over dhcp so I only need to type 'linux > >> ks' at the boot prompt. The dhcpd/kickstart server are on the same > >> machine (192.168.0.1.) dhcp works, as well as kickstart if I run the > >> following command manually: > >> > >> linux ks=nfs:192.168.0.1:/home/kickstart/rh9.cfg > >> > >> In my dhcpd.conf file, I added following line in the subnet {} section: > >> > >> filename "/home/kickstart/ks.cfg"; > >> > >> My /etc/exports file contains the following: > >> > >> /home/kickstart *(ro) > >> > >> The error I get is: > >> > >> bootp: no bootfile received > >> > >> Then it proceeds to look for the kickstart file at > >> nfs:192.168.0.1:/kickstart/kickstart-192.168.0.18 > >> > >> Any suggestions? > >> > >> Thanks for helping me with the question about using perl in %pre. Since > >> you can't use perl in %pre from just the install CD, I ended up porting > >> my script to python which works well. And I learned a new programming > >> language in the process :). > >> > >> Thank you. > >> > >> -- > >> Joseph Glass > >> Systems Administrator > >> Liquid Web Inc. > >> 800.580.4985 x227 > >> > >> > >> _______________________________________________ > >> Kickstart-list mailing list > >> Kickstart-list at redhat.com > >> https://www.redhat.com/mailman/listinfo/kickstart-list > >> > > > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From pantz at lqt.ca Mon May 17 13:21:21 2004 From: pantz at lqt.ca (Paul Pianta) Date: Mon, 17 May 2004 09:21:21 -0400 Subject: %post trouble In-Reply-To: References: Message-ID: <40A8BC51.9090102@lqt.ca> James_Martin at ao.uscourts.gov wrote: >Try doing it this way... should save you much work.. > >%post > >#mount the cdrom >updfstab >mount /mnt/cdrom >cp /mnt/cdrom/Fedora/RPMS/extras/mysqld.init /etc/rc.d/init.d/mysqld > > >That should be all you need.. the updfstab updates the /etc/fstab with >/mnt/cdrom and points it to the correct device. I'm sure you can figure >out the rest :). > >Good luck! > > Thanks heaps James! it worked as you said it would :) pantz -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From mlblandf at sedona.ch.intel.com Mon May 17 21:04:52 2004 From: mlblandf at sedona.ch.intel.com (Michael Blandford) Date: Mon, 17 May 2004 14:04:52 -0700 Subject: Issue with %include on AW2.1 Message-ID: <40A928F4.3080009@sedona.intel.com> I have a python script that I run in the %pre section to generate partitioning information. It dumps this information into /tmp/partition. in the ks.cfg, I am adding %include /tmp/partition. If the partitioning is different that what is already on the machine, the build will stop and complain the /usr is too small. If I choose 'no' to continue, the machine reboots and starts the build again. This time it will suceed. It seems like the machine doesn't notice the partitioning has changed during the build and is trying to continue using the old partitions. This is on ia64. I am also using the same process on ia32 with RedHat 7.3 without seeing this problem. Any thoughts on what might be wrong? Michael Disclaimer: The content of this message is my personal opinion only and although I am an employee of Intel, the statements I make here in no way represent Intel's position on the issue, nor am I authorized to speak on behalf of Intel on this matter. From phr at doc.ic.ac.uk Tue May 18 10:26:02 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 18 May 2004 11:26:02 +0100 (BST) Subject: Newbie Needs Help with Kickstart Fedora Installations In-Reply-To: <4AFB5F82-A5DB-11D8-A7CC-000393C353D0@earthlink.net> References: <4AFB5F82-A5DB-11D8-A7CC-000393C353D0@earthlink.net> Message-ID: On Fri, 14 May 2004, Gary Ford wrote: >I need help with the following items: > >1. Is my scenario feasible? Yes. >2. So far, I've had no luck getting a Kickstart installation to run. >The Apache server log shows that the ks.cfg file is being requested and >sent, but the installer seems to ignore it and proceed with a >completely interactive installation. I'm using the boot option "linux >ks=http://192.168.0.1/Linux/ks.cfg". Do I need additional boot options? What do the other VCs (virtual consoles) say? (Press Alt+F2, Alt+F3 etc.) >3. If I add RPMs to the Fedora/RPMS directory, how does that affect >files in Fedora/base? I understand "comps.xml" and have a tool on the >Mac OS X box that can generate it and a Kickstart configuration file >after "interviewing" the user to determine the packages that the user >is likely to need. I do not understand the "hdlist", "hdlist2", >"hdstg2.img", "netstg2.img", or "stage2.img" files, their roles in >installation, and how to create them. The .img files *are* the Redhat installer, aka Anaconda. The hdlist* files are mapping files from a package name in comps.xml or the kickstart file to real .rpm files in Fedora/RPMS. They're build by genhdlist, and need to be rebuilt every time the contents of Fedora/RPMS change. >4. The RH9 documentation for an HTTP installation of Linux makes >reference to the "installation tree". The only definition of that term >I could find says it means the entire contents of all the installation >CD-ROMs. That seems extreme (does the installer really need the DOS >utilities or the boot diskette image file?). What does "installation >tree" really mean? It means the Fedora/base and Fedora/RPMS directories. Kickstarting doesn't require dosutils, images, isolinux etc., but they're small and sometimes useful to keep around. >5. Is it okay to boot the Dell host from Fedora CD 1, or is there a >better choice? For kickstarts, you'll definitely want to boot from the same kernel and initrd that the second-stage installer expects. In other words, if installing FC1 (Fedora Core 1), boot from FC1. Hope this helps, Phil From pantz at lqt.ca Tue May 18 15:35:17 2004 From: pantz at lqt.ca (Paul Pianta) Date: Tue, 18 May 2004 11:35:17 -0400 Subject: Issue with %include on AW2.1 In-Reply-To: <40A928F4.3080009@sedona.intel.com> References: <40A928F4.3080009@sedona.intel.com> Message-ID: <40AA2D35.3060606@lqt.ca> Michael Blandford wrote: > I have a python script that I run in the %pre section to generate > partitioning information. It dumps this information into /tmp/partition. > > in the ks.cfg, I am adding %include /tmp/partition. > > If the partitioning is different that what is already on the machine, > the build will stop and complain the /usr is too small. If I choose > 'no' to continue, the machine reboots and starts the build again. > This time it will suceed. > > It seems like the machine doesn't notice the partitioning has changed > during the build and is trying to continue using the old partitions. > > This is on ia64. I am also using the same process on ia32 with RedHat > 7.3 without seeing this problem. > > Any thoughts on what might be wrong? Have you tried this kickstart option for clearing all existing partitions? I also add the option for initialising a disk if it is new ... # clear all partitions and initialise disk if need be clearpart --all --initlabel hth pantz -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From mlblandf at sedona.ch.intel.com Tue May 18 15:45:50 2004 From: mlblandf at sedona.ch.intel.com (Michael Blandford) Date: Tue, 18 May 2004 08:45:50 -0700 Subject: Issue with %include on AW2.1 In-Reply-To: <40AA2D35.3060606@lqt.ca> References: <40A928F4.3080009@sedona.intel.com> <40AA2D35.3060606@lqt.ca> Message-ID: <40AA2FAE.9050409@sedona.intel.com> Paul Pianta wrote: > > Have you tried this kickstart option for clearing all existing > partitions? I also add the option for initialising a disk if it is new > ... > > # clear all partitions and initialise disk if need be > clearpart --all --initlabel Yes, here is what I am doing. zerombr yes clearpart --all --initlabel %include /tmp/partition Michael Disclaimer: The content of this message is my personal opinion only and although I am an employee of Intel, the statements I make here in no way represent Intel's position on the issue, nor am I authorized to speak on behalf of Intel on this matter. From barjunk at attglobal.net Tue May 18 17:31:47 2004 From: barjunk at attglobal.net (mbox mbarsalou) Date: Tue, 18 May 2004 09:31:47 -0800 Subject: Fedora Kickstart In-Reply-To: References: Message-ID: <1084901507.5918.285.camel@localhost> There used be a wiki that had information on how to setup kickstart with Fedora....anyone know where it went? I think the website was rau.homedns.org. Mike From mikem at redhat.com Tue May 18 20:27:37 2004 From: mikem at redhat.com (Mike McLean) Date: Tue, 18 May 2004 16:27:37 -0400 Subject: kickstart on Linux on zSeries? In-Reply-To: <1C455916B3418145B320711B21A2F04515D04C@ex2k.bankofamerica.com> References: <1C455916B3418145B320711B21A2F04515D04C@ex2k.bankofamerica.com> Message-ID: <40AA71B9.4070309@redhat.com> Cameron, Thomas wrote: > Howdy all - > > I am working with 64-bit RHEL 3 for zSeries. We are currently thinking that we will be using a mainframe tool (ddr, similar to the dd command under *nix) to copy a single "golden image" to multiple production servers. I want to explore using kickstart to build servers instead, but I don't know if it is possible. > > For those of you not familiar with Linux on the mainframe, it doesn't boot with the Syslinux boot screen or even a floppy or compact disk. Instead you load the kernel and the initrd.img into a virtual punchcard reader and then load it into the mainframe CPU/memory. > > I have read that there is a RUNKS ("run kickstart") option listed for the mainframe parm (configuration) file, but I do not know how to actually specify the kickstart file. I know how to copy the ks.cfg into the initrd.img file, but does it go into the root of the initrd, or in /tmp, or somewhere else? Are there additional parm file options I need which are not discussed at https://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/s390-multi-install-guide/s1-s390-steps-vm.html? > > Anyone who has any pointers, I am all ears! you need to specify a ks= option on the kernel command line (i.e. the parm file). The installer can pull ks.cfg from many different places. Look at the kickstart docs in the sysadmin guide. https://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/sysadmin-guide/s1-kickstart2-startinginstall.html From forrestx.taylor at intel.com Tue May 18 17:55:47 2004 From: forrestx.taylor at intel.com (Taylor, ForrestX) Date: Tue, 18 May 2004 10:55:47 -0700 Subject: Fedora Kickstart In-Reply-To: <1084901507.5918.285.camel@localhost> References: <1084901507.5918.285.camel@localhost> Message-ID: <1084902947.7765.59.camel@bad.jf.intel.com> On Tue, 2004-05-18 at 10:31, mbox mbarsalou wrote: > There used be a wiki that had information on how to setup kickstart with > Fedora....anyone know where it went? I think the website was > rau.homedns.org. Yes, it is: http://rau.homedns.org/twiki/bin/view/Anaconda/AnacondaDocumentationProject Although it is mostly about anaconda and rebuilding a custom distro. Forrest From barjunk at attglobal.net Tue May 18 22:39:37 2004 From: barjunk at attglobal.net (mbox mbarsalou) Date: Tue, 18 May 2004 14:39:37 -0800 Subject: Fedora Kickstart In-Reply-To: <1084902947.7765.59.camel@bad.jf.intel.com> References: <1084901507.5918.285.camel@localhost> <1084902947.7765.59.camel@bad.jf.intel.com> Message-ID: <1084919977.5918.300.camel@localhost> Thanks...it seemed to be done for a while...anyway. I was trying to use the Fedora info to setup a single Custom CD. I keep running into little problems. I will start again from the top and post to the list what I find. Since I wasn't trying to customize Anaconda, I just skipped to Section V. Customizing Fedora Core. Thanks. Mike On Tue, 2004-05-18 at 09:55, Taylor, ForrestX wrote: > On Tue, 2004-05-18 at 10:31, mbox mbarsalou wrote: > > There used be a wiki that had information on how to setup kickstart with > > Fedora....anyone know where it went? I think the website was > > rau.homedns.org. > > Yes, it is: > http://rau.homedns.org/twiki/bin/view/Anaconda/AnacondaDocumentationProject > > Although it is mostly about anaconda and rebuilding a custom distro. > > Forrest > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From barjunk at attglobal.net Tue May 18 23:21:36 2004 From: barjunk at attglobal.net (mbox mbarsalou) Date: Tue, 18 May 2004 15:21:36 -0800 Subject: getfullcomps errors In-Reply-To: References: Message-ID: <1084922496.5918.305.camel@localhost> I ran the getfullcomps as described in the Wiki and I get back a few CRITICAL ERROR messages. I realize that this means that those particular packages are missing, but this is the comps.xml file that comes with the CD set. Why would it do that? Since the machine is working just fine the way it is, how could I create a list of packages that would go into the comps.xml file? Basically, I want a snapshot of the packages from the current machine and produce that to the comps.xml so I can make my single CD. Mike From barjunk at attglobal.net Wed May 19 00:24:15 2004 From: barjunk at attglobal.net (mbox mbarsalou) Date: Tue, 18 May 2004 16:24:15 -0800 Subject: getfullcomps errors In-Reply-To: <1084922496.5918.305.camel@localhost> References: <1084922496.5918.305.camel@localhost> Message-ID: <1084926254.5918.309.camel@localhost> Although it probably wasn't the right step to do, I removed the packages that gave critical errors from the comps.xml file. This may break something in the future, but the packages did not exist on any of the Three CD's I was using...FC 1. Mike On Tue, 2004-05-18 at 15:21, mbox mbarsalou wrote: > I ran the getfullcomps as described in the Wiki and I get back a few > CRITICAL ERROR messages. I realize that this means that those > particular packages are missing, but this is the comps.xml file that > comes with the CD set. > > Why would it do that? > > Since the machine is working just fine the way it is, how could I create > a list of packages that would go into the comps.xml file? > > Basically, I want a snapshot of the packages from the current machine > and produce that to the comps.xml so I can make my single CD. > > Mike > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From joden at lee.k12.nc.us Wed May 19 03:09:06 2004 From: joden at lee.k12.nc.us (James Olin Oden) Date: Tue, 18 May 2004 23:09:06 -0400 (EDT) Subject: getfullcomps errors In-Reply-To: <1084922496.5918.305.camel@localhost> References: <1084922496.5918.305.camel@localhost> Message-ID: On Tue, 18 May 2004, mbox mbarsalou wrote: > I ran the getfullcomps as described in the Wiki and I get back a few > CRITICAL ERROR messages. I realize that this means that those > particular packages are missing, but this is the comps.xml file that > comes with the CD set. > > Why would it do that? > It will also give critical errors for packages that require the non-standard kernel (probably not the best name, but things like kernel-smp and kernel-bigmem being required by others). The patch is trivial: 152,153c152 < if (h['name'] == "glibc-debug" or < h['name'].startswith("kernel-")): --- > if (h['name'] == "glibc-debug"): At least that is the patch agaisnt the RH 9 version. I asked about this a while back, but got no response. This may not be your problem, but I figured I would mention it just in case. Cheers...james P.S. I would really like to know why this limitation is there. Course I probably should have bugzilla'd it, and let the developers sort it out whether its a NOFIX or patch gleefully accepted. From jrobertson at convera.com Wed May 19 03:20:10 2004 From: jrobertson at convera.com (Joe Robertson) Date: Tue, 18 May 2004 20:20:10 -0700 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: Well, my hopes were dashed today. I just installed the Final version of Fedora Core 2 and it does not address the issue related to the e1000.ko drivers. I still get the nfs (or http) failed message the first time anaconda attempts to access the network to get the ks file. I have now added yet another switch to the list of switches that do not work in this environment (Extreme Networks Summit 400-48t). Perhaps someone can suggest a change to the switch - someone previously mentioned enabling 'fast port' but I have not seen anything like that in the switch documentation. The switch settings are all default and I have verified that Spanning Tree (SMP) is disabled as is the Ethernet Automatic Protection Switching (EAPS). I connected to a netgear FS108 (cheap 100 Mb unmanaged switch) and kickstart via PXE / NFS works just fine. Replace that switch with the Extreme Networks Switch and it fails when it tries to access the ks file. The version of anaconda in this is 10.0-5 which is later than the version in the recent updates to Redhat Enterprise versions. Apparently those fixes either didn't get into this or simply don't fix the problem. Joe From seph at directionless.org Wed May 19 03:47:27 2004 From: seph at directionless.org (seph) Date: Tue, 18 May 2004 23:47:27 -0400 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: (Joe Robertson's message of "Tue, 18 May 2004 20:20:10 -0700") References: Message-ID: > I have now added yet another switch to the list of switches that do not > work in this environment (Extreme Networks Summit 400-48t). Perhaps > someone can suggest a change to the switch - someone previously > mentioned enabling 'fast port' but I have not seen anything like that in > the switch documentation. All the switches I'm currently playing with are unmanaged dumb things, but I suspect the person was suggesting fiddling the STP settings. Full STP can take some time to settle, as it checks for network loops. So most switches provide some sort of fast STP mode where the switch assumes there's just a single host. I don't think this is likely my problem, as dhcp is successfully completing. seph From barjunk at attglobal.net Wed May 19 04:49:31 2004 From: barjunk at attglobal.net (mbox mbarsalou) Date: Tue, 18 May 2004 20:49:31 -0800 Subject: pkgorder and scripting In-Reply-To: <200405041924.i44JOLkG000315@mx1.redhat.com> References: <200405041924.i44JOLkG000315@mx1.redhat.com> Message-ID: <1084942171.5918.316.camel@localhost> I was trying to script the use of the python script pkgorder with bash but recieve no output. For example if I use: pkgorder I get some output. However, if I put that same command in a bash script, I get no output... Is there some special command that I need to do to get it to produce output when run from a Bash script? Mike From anaconda at maisondubonheur.com Wed May 19 08:08:15 2004 From: anaconda at maisondubonheur.com (Alain TAUCH) Date: Wed, 19 May 2004 10:08:15 +0200 Subject: pkgorder and scripting In-Reply-To: <1084942171.5918.316.camel@localhost> References: <200405041924.i44JOLkG000315@mx1.redhat.com> <1084942171.5918.316.camel@localhost> Message-ID: <20040519100815.3ecc088a.anaconda@maisondubonheur.com> On Tue, 18 May 2004 20:49:31 -0800 mbox mbarsalou [mbox] wrote: mbox> I was trying to script the use of the python script pkgorder with bash mbox> but recieve no output. For example if I use: mbox> mbox> pkgorder mbox> mbox> I get some output. mbox> mbox> However, if I put that same command in a bash script, I get no output... mbox> mbox> Is there some special command that I need to do to get it to produce mbox> output when run from a Bash script? AFAIK, no. It's working fine with my RH9 scripts. -- Alain From pantz at lqt.ca Wed May 19 13:37:00 2004 From: pantz at lqt.ca (Paul Pianta) Date: Wed, 19 May 2004 09:37:00 -0400 Subject: Issue with %include on AW2.1 In-Reply-To: <40AA2FAE.9050409@sedona.intel.com> References: <40A928F4.3080009@sedona.intel.com> <40AA2D35.3060606@lqt.ca> <40AA2FAE.9050409@sedona.intel.com> Message-ID: <40AB62FC.2020802@lqt.ca> Michael Blandford wrote: > Paul Pianta wrote: > >> >> Have you tried this kickstart option for clearing all existing >> partitions? I also add the option for initialising a disk if it is >> new ... >> >> # clear all partitions and initialise disk if need be >> clearpart --all --initlabel > > > Yes, here is what I am doing. > > zerombr yes > clearpart --all --initlabel > %include /tmp/partition This seems to be ok - it is the same as what I do (except the 'zerombr yes' bit). And if the 'clearpart --all' option is working - you shouldn't ever see a complaint about the /usr partition being too small - UNLESS - your %pre python script is not creating the /usr partition with the right size. Maybe paste your python %pre script here so we can have a look at it - my guess would be that there is something funny going on there ... pantz -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From pantz at lqt.ca Wed May 19 14:26:29 2004 From: pantz at lqt.ca (Paul Pianta) Date: Wed, 19 May 2004 10:26:29 -0400 Subject: pkgorder and scripting In-Reply-To: <1084942171.5918.316.camel@localhost> References: <200405041924.i44JOLkG000315@mx1.redhat.com> <1084942171.5918.316.camel@localhost> Message-ID: <40AB6E95.4090208@lqt.ca> mbox mbarsalou wrote: >I was trying to script the use of the python script pkgorder with bash >but recieve no output. For example if I use: > >pkgorder > >I get some output. > >However, if I put that same command in a bash script, I get no output... > >Is there some special command that I need to do to get it to produce >output when run from a Bash script? > > are you a golfer? if not you can try using a 'tee' anyway - it always helps to give you a cleaner swing at the ball!! :) pantz snipppet from my bash script echo '** Running pkgorder ...' /usr/lib/anaconda-runtime/pkgorder /data/Fedora-Custom/i386 i386 Fedora | tee /data/Fedora-Custom/pkgorder.txt -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From mlblandf at sedona.ch.intel.com Wed May 19 18:26:17 2004 From: mlblandf at sedona.ch.intel.com (Michael Blandford) Date: Wed, 19 May 2004 11:26:17 -0700 Subject: Issue with %include on AW2.1 In-Reply-To: <40AB62FC.2020802@lqt.ca> References: <40A928F4.3080009@sedona.intel.com> <40AA2D35.3060606@lqt.ca> <40AA2FAE.9050409@sedona.intel.com> <40AB62FC.2020802@lqt.ca> Message-ID: <40ABA6C9.4010505@sedona.intel.com> Paul Pianta wrote: > > This seems to be ok - it is the same as what I do (except the 'zerombr > yes' bit). > And if the 'clearpart --all' option is working - you shouldn't ever > see a complaint about the /usr partition being too small - UNLESS - > your %pre python script is not creating the /usr partition with the > right size. The script is pretty big so I am not going to post it, however, I can post what the output looks like. Here is an example of /tmp/partition on a machine that fails. # Disk: sda 70001 megs # Ram: 4034 megs # Arch: ia64 part / --size 300 --maxsize 375 --grow part /boot/efi --size 100 --fstype vfat --asprimary part /scratch --size 400 --grow part /tmp --size 512 --maxsize 1024 --grow part /usr --size 2400 --maxsize 2550 --grow part /var --size 512 # Swap requested: 8068 part swap --size 8068 --ondisk sda /usr is more than big enough for the disto I am putting on. > > Maybe paste your python %pre script here so we can have a look at it - > my guess would be that there is something funny going on there ... If I run it again ( by rebooting ) it all works fine the second time. That is what is bizaar. Michael Disclaimer: The content of this message is my personal opinion only and although I am an employee of Intel, the statements I make here in no way represent Intel's position on the issue, nor am I authorized to speak on behalf of Intel on this matter. From forrestx.taylor at intel.com Wed May 19 18:49:52 2004 From: forrestx.taylor at intel.com (Taylor, ForrestX) Date: Wed, 19 May 2004 11:49:52 -0700 Subject: Issue with %include on AW2.1 In-Reply-To: <40ABA6C9.4010505@sedona.intel.com> References: <40A928F4.3080009@sedona.intel.com> <40AA2D35.3060606@lqt.ca> <40AA2FAE.9050409@sedona.intel.com> <40AB62FC.2020802@lqt.ca> <40ABA6C9.4010505@sedona.intel.com> Message-ID: <1084992592.14818.7.camel@bad.jf.intel.com> On Wed, 2004-05-19 at 11:26, Michael Blandford wrote: > Paul Pianta wrote: > > > > > This seems to be ok - it is the same as what I do (except the 'zerombr > > yes' bit). > > And if the 'clearpart --all' option is working - you shouldn't ever > > see a complaint about the /usr partition being too small - UNLESS - > > your %pre python script is not creating the /usr partition with the > > right size. > > The script is pretty big so I am not going to post it, however, I can > post what the output looks like. > Here is an example of /tmp/partition on a machine that fails. > > # Disk: sda 70001 megs > # Ram: 4034 megs > # Arch: ia64 > part / --size 300 --maxsize 375 --grow > part /boot/efi --size 100 --fstype vfat --asprimary > part /scratch --size 400 --grow > part /tmp --size 512 --maxsize 1024 --grow > part /usr --size 2400 --maxsize 2550 --grow > part /var --size 512 > > # Swap requested: 8068 > part swap --size 8068 --ondisk sda > > /usr is more than big enough for the disto I am putting on. Are you absolutely certain that 2.4 GB is enough room in /usr? Since you have a 70GB hard drive, why don't you try 6 GB for /usr? You could also reduce your swap a bit. Forrest From mlblandf at sedona.ch.intel.com Wed May 19 21:02:16 2004 From: mlblandf at sedona.ch.intel.com (Michael Blandford) Date: Wed, 19 May 2004 14:02:16 -0700 Subject: Issue with %include on AW2.1 In-Reply-To: <1084992592.14818.7.camel@bad.jf.intel.com> References: <40A928F4.3080009@sedona.intel.com> <40AA2D35.3060606@lqt.ca> <40AA2FAE.9050409@sedona.intel.com> <40AB62FC.2020802@lqt.ca> <40ABA6C9.4010505@sedona.intel.com> <1084992592.14818.7.camel@bad.jf.intel.com> Message-ID: <40ABCB58.7060603@sedona.intel.com> > > >Are you absolutely certain that 2.4 GB is enough room in /usr? Since >you have a 70GB hard drive, why don't you try 6 GB for /usr? You could >also reduce your swap a bit. > > > I am absolutely sure. I have increased /usr to >30 GB and it says the same thing. Simply rebooting and trying again and it all works. Here is what it looks like after it is built. /dev/sda3 2569552 2113192 325832 87% /usr Michael Disclaimer: The content of this message is my personal opinion only and although I am an employee of Intel, the statements I make here in no way represent Intel's position on the issue, nor am I authorized to speak on behalf of Intel on this matter. From barjunk at attglobal.net Thu May 20 02:12:47 2004 From: barjunk at attglobal.net (mbox mbarsalou) Date: Wed, 19 May 2004 18:12:47 -0800 Subject: pkgorder and scripting In-Reply-To: <40AB6E95.4090208@lqt.ca> References: <200405041924.i44JOLkG000315@mx1.redhat.com> <1084942171.5918.316.camel@localhost> <40AB6E95.4090208@lqt.ca> Message-ID: <1085019167.5918.318.camel@localhost> Thanks Paul. I wonder if that is what Alain is doing too? Mike On Wed, 2004-05-19 at 06:26, Paul Pianta wrote: > mbox mbarsalou wrote: > > >I was trying to script the use of the python script pkgorder with bash > >but recieve no output. For example if I use: > > > >pkgorder > > > >I get some output. > > > >However, if I put that same command in a bash script, I get no output... > > > >Is there some special command that I need to do to get it to produce > >output when run from a Bash script? > > > > > are you a golfer? if not you can try using a 'tee' anyway - it always > helps to give you a cleaner swing at the ball!! > > :) > > pantz > > snipppet from my bash script > > echo '** Running pkgorder ...' > /usr/lib/anaconda-runtime/pkgorder /data/Fedora-Custom/i386 i386 Fedora > | tee /data/Fedora-Custom/pkgorder.txt From barjunk at attglobal.net Thu May 20 02:16:03 2004 From: barjunk at attglobal.net (mbox mbarsalou) Date: Wed, 19 May 2004 18:16:03 -0800 Subject: Issue with %include on AW2.1 In-Reply-To: <40ABCB58.7060603@sedona.intel.com> References: <40A928F4.3080009@sedona.intel.com> <40AA2D35.3060606@lqt.ca> <40AA2FAE.9050409@sedona.intel.com> <40AB62FC.2020802@lqt.ca> <40ABA6C9.4010505@sedona.intel.com> <1084992592.14818.7.camel@bad.jf.intel.com> <40ABCB58.7060603@sedona.intel.com> Message-ID: <1085019363.5918.321.camel@localhost> Although this probably doesn't apply to you, I was getting not enough disk space when installing from the FC 1 disks using a SCSI drive. The solution was to run the CD disk check routine at the start of the install. I realize how strange this sounds, but I attempted the install multiple times and it would only work if I did the disk check first. Go figure. Mike On Wed, 2004-05-19 at 13:02, Michael Blandford wrote: > > > > > >Are you absolutely certain that 2.4 GB is enough room in /usr? Since > >you have a 70GB hard drive, why don't you try 6 GB for /usr? You could > >also reduce your swap a bit. > > > > > > > I am absolutely sure. I have increased /usr to >30 GB and it says the > same thing. Simply rebooting and trying again and it all works. > > Here is what it looks like after it is built. > > /dev/sda3 2569552 2113192 325832 87% /usr > > Michael > > > Disclaimer: The content of this message is my personal opinion only and > although I am an employee of Intel, the statements I make here in no way > represent Intel's position on the issue, nor am I authorized to speak on > behalf of Intel on this matter. > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From stuart.grigg at baesystems.com Thu May 20 04:34:05 2004 From: stuart.grigg at baesystems.com (GRIGG, Stuart) Date: Thu, 20 May 2004 14:04:05 +0930 Subject: Do not eject CDROM after kickstart install Message-ID: I do not want to eject the CDROM (DVD) after a kickstart install. Is there a way to do it. I tried adding the following in the POST section. mv /usr/sbin/eject /usr/sbin/dont_eject but the eject file must be in the boot kernel not on the fresh install The reason I do not want it to eject is that the CDROM is a slim line (laptop like) and it does not get sucked back in on computer boot. Any help would be appreciated. From shorne at softhome.net Thu May 20 05:39:13 2004 From: shorne at softhome.net (Stafford Horne) Date: Thu, 20 May 2004 13:39:13 +0800 Subject: Do not eject CDROM after kickstart install In-Reply-To: References: Message-ID: <1085031553.1082.20.camel@localhost.localdomain> I don't think you can do this without modifying the Anaconda source code. You can modify the code to skip the doMethodComplete step. If you want to do this, please let us know. Stafford On Thu, 2004-05-20 at 12:34, GRIGG, Stuart wrote: > I do not want to eject the CDROM (DVD) after a kickstart install. Is there a > way to do it. > > I tried adding the following in the POST section. > mv /usr/sbin/eject /usr/sbin/dont_eject > but the eject file must be in the boot kernel not on the fresh install > > The reason I do not want it to eject is that the CDROM is a slim line > (laptop like) and it does not get sucked back in on computer boot. > > Any help would be appreciated. > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From shorne at softhome.net Thu May 20 06:42:11 2004 From: shorne at softhome.net (Stafford Horne) Date: Thu, 20 May 2004 14:42:11 +0800 Subject: Do not eject CDROM after kickstart install In-Reply-To: References: Message-ID: <1085035330.1082.49.camel@localhost.localdomain> That is correct, There is a script that comes in the anaconda-runtime RPM called buildinstall. This will build the new installation images (PXE and Cdrom) from a set of RPMs you place in a certain directory. This process is described here: https://linux.co.uk/Pages/howtos/RedHat-CD-HOWTO.html There are also other great sources you can find on google. after you find out how to rebuild the system using the buildinstall script you can get the anaconda source RPM. Unpack it with rpm -Uvh (anaconda src rpm) unpack the tar.bz2 repack the tar.bz2 rebuild the rpm with: rpmbuild -bb anaconda.spec now rebuild the Anaconda CDs using the new anaconda RPMs This is a very rough discription of what you need to do. but I hope you can get the general Idea Stafford On Thu, 2004-05-20 at 13:22, Joe Robertson wrote: > Stafford, > > Does this mean you have a script or something to repackage the > system after modifying Anaconda? If so, I'm very interested. > I need to modify Anaconda then do a complete repackage so I > can use the Anaconda build results in the pxeboot images. > > Thanks, > Joe > > > > > > > I don't think you can do this without modifying the Anaconda > > source code. You can modify the code to skip the > > doMethodComplete step. If you want to do this, please let us know. > > > > Stafford > > From anaconda at maisondubonheur.com Thu May 20 09:38:18 2004 From: anaconda at maisondubonheur.com (Alain TAUCH) Date: Thu, 20 May 2004 11:38:18 +0200 Subject: pkgorder and scripting In-Reply-To: <1085019167.5918.318.camel@localhost> References: <200405041924.i44JOLkG000315@mx1.redhat.com> <1084942171.5918.316.camel@localhost> <40AB6E95.4090208@lqt.ca> <1085019167.5918.318.camel@localhost> Message-ID: <20040520113818.1ef2bddd.anaconda@maisondubonheur.com> On Wed, 19 May 2004 18:12:47 -0800 mbox mbarsalou [Mike] wrote: Mike> Thanks Paul. Mike> Mike> I wonder if that is what Alain is doing too? well, it's quite the same: pkgorder options > pkgorder.txt 2> logfile Alain Mike> Mike> Mike Mike> On Wed, 2004-05-19 at 06:26, Paul Pianta wrote: [...] Mike> > snipppet from my bash script Mike> > Mike> > echo '** Running pkgorder ...' Mike> > /usr/lib/anaconda-runtime/pkgorder /data/Fedora-Custom/i386 i386 Fedora Mike> > | tee /data/Fedora-Custom/pkgorder.txt From ajay at cedar.Buffalo.EDU Thu May 20 14:05:26 2004 From: ajay at cedar.Buffalo.EDU (Ajay Shekhawat) Date: Thu, 20 May 2004 10:05:26 -0400 Subject: Do not eject CDROM after kickstart install In-Reply-To: References: Message-ID: <20040520140526.GZ24536@zaurak.cedar.Buffalo.EDU> Or you could do eject -t /dev/cdrom during boot, that'll close the CDROM. (Assuming you don't want to poke around inside Anaconda...) On Thu, May 20, 2004 at 02:04:05PM +0930, GRIGG, Stuart wrote: > I do not want to eject the CDROM (DVD) after a kickstart install. Is there a > way to do it. > > I tried adding the following in the POST section. > mv /usr/sbin/eject /usr/sbin/dont_eject > but the eject file must be in the boot kernel not on the fresh install > > The reason I do not want it to eject is that the CDROM is a slim line > (laptop like) and it does not get sucked back in on computer boot. > > Any help would be appreciated. From jbrash at us.ibm.com Thu May 20 15:14:13 2004 From: jbrash at us.ibm.com (Jason Brashear) Date: Thu, 20 May 2004 10:14:13 -0500 Subject: Create a Kickstart cdrom for AS 3.0 In-Reply-To: Message-ID: Does anyone know how to create a Kickstart CDROM? Best Regards, Jason Brashear jbrash at us.ibm.com Websphere Lab Administrator IBM Corporation 11501 Burnet Rd., Austin, Texas 78758 LWWS/7G , Bldg. 903, 4H-004 512-838-4776 T/L 678-4776 FAX: 512-838-0330 (5C-007) Jason Brashear/Austin/Contr/IBM at IBMUS Sent by: kickstart-list-bounces at redhat.com 05/14/2004 02:45 PM Please respond to Discussion list about Kickstart To Discussion list about Kickstart cc Subject Create a Kickstart cdrom for AS 3.0 Hello I am new to Kickstart. I wonder if anyone knows how to create a kickstart CDROM. I am even having trouble creating a Kickstart Disk. Thank you in advance for all your help! Best Regards, Jason Brashear jbrash at us.ibm.com Websphere Lab Administrator IBM Corporation 11501 Burnet Rd., Austin, Texas 78758 LWWS/7G , Bldg. 903, 4H-004 512-838-4776 T/L 678-4776 FAX: 512-838-0330 (5C-007)_______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list -------------- next part -------------- An HTML attachment was scrubbed... URL: From tibbs at math.uh.edu Thu May 20 15:37:29 2004 From: tibbs at math.uh.edu (Jason L Tibbitts III) Date: 20 May 2004 10:37:29 -0500 Subject: Create a Kickstart cdrom for AS 3.0 In-Reply-To: References: Message-ID: >>>>> "JB" == Jason Brashear writes: JB> Does anyone know how to create a Kickstart CDROM? Sure, what do you want on it? If you just want to use it to boot into the installer so you can do a network install, just grab isolinux and make yourself a CD. I have one CD that boots all different versions o the installer, for both i386 and x86_64, and also has memtest86 and various hard drive diagnostics programs on it. To boot one of the Red Hat/Fedora installers, all you need is the vmlinuz and initrd.img files and an entry in isolinux.cfg. Here's my FC2 for x86_64 entry: label ks64-2 kernel vml642 append initrd=ird642.img devfs=nomount ramdisk_size=9216 ksdevice=link ks=http://url.for/kickstart.file - J< From James_Martin at ao.uscourts.gov Thu May 20 16:32:05 2004 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Thu, 20 May 2004 12:32:05 -0400 Subject: Create a Kickstart cdrom for AS 3.0 In-Reply-To: Message-ID: That is a very broad, general question. Have you tried using google? You can even read the manual that's on Red Hat's web page that documents how to do this. Nobody's going to sit and tell you step for step on how to do this, it's already documented elsewhere. If you have a specific problem/question, it is more likely to get answered on this list. Good luck, James James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 05/20/2004 11:14:13 AM: > > Does anyone know how to create a Kickstart CDROM? > > > Best Regards, > Jason Brashear > jbrash at us.ibm.com > Websphere Lab Administrator > IBM Corporation > 11501 Burnet Rd., Austin, Texas 78758 > LWWS/7G , Bldg. 903, 4H-004 > 512-838-4776 T/L 678-4776 > FAX: 512-838-0330 (5C-007) > > > Jason Brashear/Austin/Contr/IBM at IBMUS > Sent by: kickstart-list-bounces at redhat.com > 05/14/2004 02:45 PM > > Please respond to > Discussion list about Kickstart > > To > > Discussion list about Kickstart > > cc > > Subject > > Create a Kickstart cdrom for AS 3.0 > > > > > > Hello I am new to Kickstart. > I wonder if anyone knows how to create a kickstart CDROM. > I am even having trouble creating a Kickstart Disk. > Thank you in advance for all your help! > > > Best Regards, > Jason Brashear > jbrash at us.ibm.com > Websphere Lab Administrator > IBM Corporation > 11501 Burnet Rd., Austin, Texas 78758 > LWWS/7G , Bldg. 903, 4H-004 > 512-838-4776 T/L 678-4776 > FAX: 512-838-0330 (5C-007)_______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From jbrash at us.ibm.com Thu May 20 18:01:11 2004 From: jbrash at us.ibm.com (Jason Brashear) Date: Thu, 20 May 2004 13:01:11 -0500 Subject: Create a Kickstart cdrom for AS 3.0 In-Reply-To: Message-ID: Thank you. Yep I want to be able to boot off of the cdrom and do a network install via http or nfs. I have x86 and x86_64"AMD". I dont have a CDR in a linux system. can I make these cd's from my windows system that has the cdr? If so will nero work to creat the cd? I tried once and to create a bootable cd in nero it puts its own version of dos on it to boot with.. That is not what I was needing. Thanks for you reply! Best Regards, Jason Brashear jbrash at us.ibm.com Websphere Lab Administrator IBM Corporation 11501 Burnet Rd., Austin, Texas 78758 LWWS/7G , Bldg. 903, 4H-004 512-838-4776 T/L 678-4776 FAX: 512-838-0330 (5C-007) Jason L Tibbitts III Sent by: kickstart-list-bounces at redhat.com 05/20/2004 10:37 AM Please respond to Discussion list about Kickstart To Discussion list about Kickstart cc Subject Re: Create a Kickstart cdrom for AS 3.0 >>>>> "JB" == Jason Brashear writes: JB> Does anyone know how to create a Kickstart CDROM? Sure, what do you want on it? If you just want to use it to boot into the installer so you can do a network install, just grab isolinux and make yourself a CD. I have one CD that boots all different versions o the installer, for both i386 and x86_64, and also has memtest86 and various hard drive diagnostics programs on it. To boot one of the Red Hat/Fedora installers, all you need is the vmlinuz and initrd.img files and an entry in isolinux.cfg. Here's my FC2 for x86_64 entry: label ks64-2 kernel vml642 append initrd=ird642.img devfs=nomount ramdisk_size=9216 ksdevice=link ks=http://url.for/kickstart.file - J< _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list -------------- next part -------------- An HTML attachment was scrubbed... URL: From tibbs at math.uh.edu Thu May 20 18:24:17 2004 From: tibbs at math.uh.edu (Jason L Tibbitts III) Date: 20 May 2004 13:24:17 -0500 Subject: Create a Kickstart cdrom for AS 3.0 In-Reply-To: References: Message-ID: >>>>> "JB" == Jason Brashear writes: JB> I dont have a CDR in a linux system. can I make these cd's from my JB> windows system that has the cdr? Sorry, I don't use windows so I can't really say. You could build an ISO under Linux and transfer that to your windows machine, assuming you have software that will burn a plain ISO file. Try this: make a directory (doesn't matter what you call it) and CD to it. Create this little one-liner script; I called it "makeit": ------- #!/bin/sh mkisofs -v -v -o boot.iso -b isolinux/isolinux.bin -c isolinux/boot.cat -no-emul-boot -boot-load-size 4 -boot-info-table -R -J -V -T image ------- Now mkdir -p image/isolinux; cd image/isolinux Copy your vmlinuz and initrd.img files here; I called them "vml642" and "ird642.img". isolinux severely resticts the naming you can use, so don't get fancy or it won't work. Create isolinux.cfg. You can copy in or create a boot image and various messages; I don't know what the minimum you have to specify is. But you will at least need a stanza for these images: ------- label ks64-2 kernel vml642 append initrd=ird642.img devfs=nomount ramdisk_size=9216 ksdevice=link ks=http://url.to/kickstart.file ------- cd ../.. run the "makeit" script, which will give you boot.iso. Burn that however you can. - J< From jbrash at us.ibm.com Thu May 20 18:35:50 2004 From: jbrash at us.ibm.com (Jason Brashear) Date: Thu, 20 May 2004 13:35:50 -0500 Subject: Create a Kickstart cdrom for AS 3.0 In-Reply-To: Message-ID: Thank you.. I will try this you are a great help! -Jason Best Regards, Jason Brashear jbrash at us.ibm.com Websphere Lab Administrator IBM Corporation 11501 Burnet Rd., Austin, Texas 78758 LWWS/7G , Bldg. 903, 4H-004 512-838-4776 T/L 678-4776 FAX: 512-838-0330 (5C-007) Jason L Tibbitts III Sent by: kickstart-list-bounces at redhat.com 05/20/2004 01:24 PM Please respond to Discussion list about Kickstart To Discussion list about Kickstart cc Subject Re: Create a Kickstart cdrom for AS 3.0 >>>>> "JB" == Jason Brashear writes: JB> I dont have a CDR in a linux system. can I make these cd's from my JB> windows system that has the cdr? Sorry, I don't use windows so I can't really say. You could build an ISO under Linux and transfer that to your windows machine, assuming you have software that will burn a plain ISO file. Try this: make a directory (doesn't matter what you call it) and CD to it. Create this little one-liner script; I called it "makeit": ------- #!/bin/sh mkisofs -v -v -o boot.iso -b isolinux/isolinux.bin -c isolinux/boot.cat -no-emul-boot -boot-load-size 4 -boot-info-table -R -J -V -T image ------- Now mkdir -p image/isolinux; cd image/isolinux Copy your vmlinuz and initrd.img files here; I called them "vml642" and "ird642.img". isolinux severely resticts the naming you can use, so don't get fancy or it won't work. Create isolinux.cfg. You can copy in or create a boot image and various messages; I don't know what the minimum you have to specify is. But you will at least need a stanza for these images: ------- label ks64-2 kernel vml642 append initrd=ird642.img devfs=nomount ramdisk_size=9216 ksdevice=link ks=http://url.to/kickstart.file ------- cd ../.. run the "makeit" script, which will give you boot.iso. Burn that however you can. - J< _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list -------------- next part -------------- An HTML attachment was scrubbed... URL: From throwaway5549 at earthlink.net Fri May 21 01:58:29 2004 From: throwaway5549 at earthlink.net (Gary Ford) Date: Thu, 20 May 2004 19:58:29 -0600 Subject: Fatal Error (was: Newbie Needs Help with Kickstart Fedora Installations) In-Reply-To: Message-ID: <5B3EC116-AACA-11D8-BED6-000393C353D0@earthlink.net> Thanks to Philip Rowlands for substantial help with my previous questions. I'm now running a kickstart installation via HTTP from an Apache web server running under Windows 2000 Pro. The contents of the server were created from the Fedora Core 2 test 3 CD-ROMs. The CDs pass the media checks, and an interactive installation works as expected. However, a kickstart installation always generates a fatal error when trying to install at-3.1.8-52.i386.rpm. I'm installing a smallish Fedora subset onto a 12 GB partition. I've replaced the at-3.1.8 file on the server a couple of times, and I've replaced all the files in the Fedora/base directory with fresh copies. The kickstart configuration file package section is taken from the file produced by anaconda during an interactive installation, so the dependencies are probably correct. Any help would be appreciated. Gary Ford Plaid Flannel Software From phr at doc.ic.ac.uk Fri May 21 07:11:29 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Fri, 21 May 2004 08:11:29 +0100 (BST) Subject: Fatal Error (was: Newbie Needs Help with Kickstart Fedora Installations) In-Reply-To: <5B3EC116-AACA-11D8-BED6-000393C353D0@earthlink.net> References: <5B3EC116-AACA-11D8-BED6-000393C353D0@earthlink.net> Message-ID: On Thu, 20 May 2004, Gary Ford wrote: >I'm now running a kickstart installation via HTTP from an Apache web >server running under Windows 2000 Pro. The contents of the server were >created from the Fedora Core 2 test 3 CD-ROMs. It'll take a few hours to download, but now that Fedora Core 2 (final) has been released, the beta version isn't the best choice. >However, a kickstart installation always generates a fatal error when >trying to install at-3.1.8-52.i386.rpm. Please elaborate on "fatal error"? What happens, or what is displayed? There is a logged bug for at: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=118040 but it doesn't look severe enough to crash the install. Cheers, Phil From brilong at cisco.com Fri May 21 13:06:11 2004 From: brilong at cisco.com (Brian Long) Date: Fri, 21 May 2004 09:06:11 -0400 Subject: Kickstart and SuSE? Message-ID: <1085144771.13440.110.camel@grimlock.cisco.com> Has anyone on this list used anaconda/kickstart to install SuSE? I've heard it can be done and I wonder if it makes sense to use kickstart instead of learning AutoYast. Thanks. /Brian/ -- Brian Long | | | Americas IT Hosting Sys Admin | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From throwaway5549 at earthlink.net Fri May 21 16:24:40 2004 From: throwaway5549 at earthlink.net (Gary Ford) Date: Fri, 21 May 2004 10:24:40 -0600 Subject: Fatal Error (was: Newbie Needs Help with Kickstart Fedora Installations) In-Reply-To: Message-ID: <5BFE4BEE-AB43-11D8-BA6B-000393C353D0@earthlink.net> Here are some additional details about the fatal error during a kickstart installation of Fedora Core 2 test 3: When the installation fails, anaconda presents a dialog box containing: "There was an error installing at-3.1.18-52. This can indicate media failure, lack of disk space, and/or hardware problems. This is a fatal error and your install will be aborted. Please verify your media and try your install again." The last few lines on the F3 console are: * moving (1) to step installpackages * setting file_context_path to nil * Initial install time estimate = 21.xxxxxxx The F4 console shows a huge number of messages that begin with: <3> audit(...): avc: denied {XXX} ... where "XXX" is "gettattr" or "use" or "write" or "searc"h or several others. If I perform a kickstart installation with a configuration file that is identical to the one causing the problems, but with "-at" in the packages section, the installation is successful. However, if I then (as root) try to install the at-3.1.18-52 package, I get well over 1000 lines of messages; here are some examples: /etc/security/selinux/file_contexts: invalid context system_u:object_r:smbd_var_run_t on line number 1535 /etc/security/selinux/file_contexts: invalid context system_u:object_r:nmbd_var_run_t on line number 1536 /etc/security/selinux/file_contexts: invalid context system_u:object_r:samba_spool_t on line number 1537 /etc/security/selinux/file_contexts: invalid context system_u:object_r:scannerdaemon_exec_t on line number 1539 /etc/security/selinux/file_contexts: invalid context system_u:object_r:scannerdaemon_etc_t on line number 1540 /etc/security/selinux/file_contexts: invalid context system_u:object_r:scannerdaemon_log_t on line number 1541 /etc/security/selinux/file_contexts: invalid context system_u:object_r:screen_exec_t on line number 1543 /etc/security/selinux/file_contexts: invalid context system_u:object_r:user_home_screen_t on line number 1544 /etc/security/selinux/file_contexts: invalid context system_u:object_r:user_home_screen_t on line number 1545 /etc/security/selinux/file_contexts: invalid context system_u:object_r:screen_dir_t on line number 1546 /etc/security/selinux/file_contexts: invalid context system_u:object_r:screensaver_exec_t on line number 1549 /etc/security/selinux/file_contexts: invalid context system_u:object_r:bin_t on line number 1550 /etc/security/selinux/file_contexts: invalid context system_u:object_r:user_screensaver_rw_t on line number 1551 /etc/security/selinux/file_contexts: invalid context system_u:object_r:user_screensaver_rw_t on line number 1552 The at-3.1.18 package seems to install despite these messages, based on the report from "rpm -qi at". I'm not doing anything overt to enable or disable SELinux (I'm using whatever is the default). My selection of packages is about 2.2 GB, and I'm installing on a 12 GB partition. I've defragmented the disk on the HTTP server host and reloaded the files from the Fedora Core CDs. The CDs pass the media check, and an interactive installation does not object to the at-3.1.18 package. I hope to solve this problem without getting the new Fedora Core iso images, because I live in an area with only two choices for Internet access: dialup at roughly 9600 baud, or two-way satellite connection with severe limitations on bandwidth once a download exceeds 150 MB. Each iso image requires >40 hours to download. Any help would be appreciated. Thanks to Philip Rowlands for his very helpful previous comments. Gary Ford Plaid Flannel Software From rreed at ops.sgp.arm.gov Fri May 21 16:48:10 2004 From: rreed at ops.sgp.arm.gov (Ronald Reed) Date: Fri, 21 May 2004 11:48:10 -0500 Subject: ByPass the supported chipset on the installer Message-ID: <1085158090.4267.56.camel@ronnote> Is there a way to make the redhat installer skip the chipset check? I have some MediaGX computers that I would like to install RHE 3 on, but the installer stops dead. Any sugguestions or helpful links to read? -- =========================== Ron Reed RedHat Certified Engineer SGP Computer Department Manager Unix Systems Administrator ARM SGP CART Site (580)388-4053 ron.reed at arm.gov From phr at doc.ic.ac.uk Fri May 21 17:02:30 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Fri, 21 May 2004 18:02:30 +0100 (BST) Subject: Fatal Error (was: Newbie Needs Help with Kickstart Fedora Installations) In-Reply-To: <5BFE4BEE-AB43-11D8-BA6B-000393C353D0@earthlink.net> References: <5BFE4BEE-AB43-11D8-BA6B-000393C353D0@earthlink.net> Message-ID: On Fri, 21 May 2004, Gary Ford wrote: >When the installation fails, anaconda presents a dialog box containing: > >"There was an error installing at-3.1.18-52. This can indicate media >failure, lack of disk space, and/or hardware problems. This is a fatal >error and your install will be aborted. Please verify your media and >try your install again." > >The last few lines on the F3 console are: > >* moving (1) to step installpackages >* setting file_context_path to nil >* Initial install time estimate = 21.xxxxxxx > >The F4 console shows a huge number of messages that begin with: > ><3> audit(...): avc: denied {XXX} ... SElinux is spoiling your day :( >However, if I then (as root) try to install the at-3.1.18-52 package, I >get well over 1000 lines of messages; here are some examples: > >/etc/security/selinux/file_contexts: invalid context >system_u:object_r:smbd_var_run_t on line number 1535 I'd not hesitate to recommend FC2 in this case. SElinux is still present, but disabled by default (probably due to too many problems such as this). >I hope to solve this problem without getting the new Fedora Core iso >images, because I live in an area with only two choices for Internet >access: dialup at roughly 9600 baud, or two-way satellite connection >with severe limitations on bandwidth once a download exceeds 150 MB. >Each iso image requires >40 hours to download. Hmm; I don't know where you are in the world, but this UK-based company sells cheap CD images of popular distros (Fedora Core included, but not FC2 final as of right now it appears). Slow downloads aren't something this list can help with much, I'm afraid. Cheers, Phil From phr at doc.ic.ac.uk Fri May 21 17:20:38 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Fri, 21 May 2004 18:20:38 +0100 (BST) Subject: Fatal Error (was: Newbie Needs Help with Kickstart Fedora Installations) In-Reply-To: References: <5BFE4BEE-AB43-11D8-BA6B-000393C353D0@earthlink.net> Message-ID: On Fri, 21 May 2004, Philip Rowlands wrote: >this UK-based company sells cheap CD images of popular distros (Fedora >Core included, but not FC2 final as of right now it appears). URL'd probably be useful :) http://www.linuxemporium.co.uk/products/general/ Cheers, Phil From kbbrown at cs.ucsd.edu Fri May 21 20:02:44 2004 From: kbbrown at cs.ucsd.edu (Kelly Brown) Date: Fri, 21 May 2004 13:02:44 -0700 (PDT) Subject: Fatal Error (was: Newbie Needs Help with Kickstart Fedora Installations) In-Reply-To: References: <5BFE4BEE-AB43-11D8-BA6B-000393C353D0@earthlink.net> Message-ID: A suggestion for the download speeds, download bittorrent and use that... http://bitconjurer.org/BitTorrent/ On Fri, 21 May 2004, Philip Rowlands wrote: > On Fri, 21 May 2004, Gary Ford wrote: > > >When the installation fails, anaconda presents a dialog box containing: > > > >"There was an error installing at-3.1.18-52. This can indicate media > >failure, lack of disk space, and/or hardware problems. This is a fatal > >error and your install will be aborted. Please verify your media and > >try your install again." > > > >The last few lines on the F3 console are: > > > >* moving (1) to step installpackages > >* setting file_context_path to nil > >* Initial install time estimate = 21.xxxxxxx > > > >The F4 console shows a huge number of messages that begin with: > > > ><3> audit(...): avc: denied {XXX} ... > > SElinux is spoiling your day :( > > >However, if I then (as root) try to install the at-3.1.18-52 package, I > >get well over 1000 lines of messages; here are some examples: > > > >/etc/security/selinux/file_contexts: invalid context > >system_u:object_r:smbd_var_run_t on line number 1535 > > I'd not hesitate to recommend FC2 in this case. SElinux is still > present, but disabled by default (probably due to too many problems such > as this). > > >I hope to solve this problem without getting the new Fedora Core iso > >images, because I live in an area with only two choices for Internet > >access: dialup at roughly 9600 baud, or two-way satellite connection > >with severe limitations on bandwidth once a download exceeds 150 MB. > >Each iso image requires >40 hours to download. > > Hmm; I don't know where you are in the world, but this UK-based company > sells cheap CD images of popular distros (Fedora Core included, but not > FC2 final as of right now it appears). Slow downloads aren't something > this list can help with much, I'm afraid. > > > Cheers, > Phil > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From throwaway5549 at earthlink.net Sat May 22 01:24:35 2004 From: throwaway5549 at earthlink.net (Gary Ford) Date: Fri, 21 May 2004 19:24:35 -0600 Subject: Fatal Error (was: Newbie Needs Help with Kickstart Fedora Installations) In-Reply-To: Message-ID: This is slightly off-topic; apologies in advance. Kelly Brown noted my Internet connectivity limitations and said: "A suggestion for the download speeds, download bittorrent and use that..." My understanding is that BitTorrent addresses the slow server issue, not the limitations imposed by my satellite connection (provided by Earthlink as a VAR of DirecWay). Although I can get burst download speeds of 200 to 300 kilobytes/second, my upload speed is always about 1 to 2 kilobytes/second. That reduces my suitability as a BitTorrent host. Philip Rowlands also noted my connectivity limitations and that he didn't know where I was located. I'm in rural Colorado (United States). Sometimes it is good for us in the open-source software community to remember that grass-roots efforts to adopt Linux need to reach all kinds of users in all kinds of locales. I live in an area where there is no DSL, no cable TV and hence no cable modems, and 40 miles of copper wire between my telephone and the ISP point of presence. It is about seven miles from my house to the nearest paved road. Most of the county I live in did not get telephone service until 1994 (yes, one nine nine four). Many rural areas of the United States have similar limitations; think what it must be like in the developing nations. Now a technical question... I'm trying to run Fedora Core 2 test 3 after a kickstart installation. I had assumed that SELinux was disabled by default. When I install, I get messages like: <3> audit(...): avc: denied {XXX} ... where "XXX" is "gettattr" or "use" or "write" or "search " or several others. When I run rpm to install additional packages, I get thousands of error messages like: /etc/security/selinux/file_contexts: invalid context system_u:object_r:samba_spool_t on line number 1537 In either case, is this evidence that SELinux is enabled and is doing strange things? How do I guarantee that SELinux is not enabled when doing a kickstart installation. Thanks. Gary Ford Plaid Flannel Software From phr at doc.ic.ac.uk Sun May 23 01:02:02 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Sun, 23 May 2004 02:02:02 +0100 (BST) Subject: Fatal Error (was: Newbie Needs Help with Kickstart Fedora Installations) In-Reply-To: References: Message-ID: On Fri, 21 May 2004, Gary Ford wrote: >I'm trying to run Fedora Core 2 test 3 after a kickstart installation. >I had assumed that SELinux was disabled by default. The release notes agree with you. http://download.fedora.redhat.com/pub/fedora/linux/core/test/1.92/i386/os/RELEASE-NOTES-en.html > When I install, I get messages like: > ><3> audit(...): avc: denied {XXX} ... >where "XXX" is "gettattr" or "use" or "write" or "search " or several >others. > >When I run rpm to install additional packages, I get thousands of error >messages like: > >/etc/security/selinux/file_contexts: invalid context >system_u:object_r:samba_spool_t on line number 1537 > >In either case, is this evidence that SELinux is enabled and is doing >strange things? Prima facie, I'd say. Odd... >How do I guarantee that SELinux is not enabled when doing a kickstart >installation. selinux is active from boot-time onwards, so to disable it you'd need to use boot-time arguments. Hopefully section 1.2 of the FAQ contains sufficient detail: http://people.redhat.com/kwade/fedora-docs/selinux-faq-en/ Cheers, Phil From jkeating at j2solutions.net Sun May 23 04:09:09 2004 From: jkeating at j2solutions.net (Jesse Keating) Date: Sat, 22 May 2004 21:09:09 -0700 Subject: Fatal Error (was: Newbie Needs Help with Kickstart Fedora Installations) In-Reply-To: References: Message-ID: <200405222109.09132.jkeating@j2solutions.net> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday 21 May 2004 18:24, Gary Ford wrote: > I'm trying to run Fedora Core 2 test 3 after a kickstart > installation. I had assumed that SELinux was disabled by default. > When I install, I get messages like: FC2 Test3 "disabled" SELinux in a way that wasn't quite fixed yet. SELinux didn't completely unload itself. This was fixed in the final FC2 release, which has been out since Monday. - -- Jesse Keating RHCE (http://geek.j2solutions.net) Fedora Legacy Team (http://www.fedoralegacy.org) GPG Public Key (http://geek.j2solutions.net/jkeating.j2solutions.pub) Was I helpful? Let others know: http://svcs.affero.net/rm.php?r=jkeating -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFAsCPl4v2HLvE71NURAlDMAJ99kvS9qhtbBxKD8q3pheQWzRgfNACfYxnx h9OpZk60KduyvKspJ42rMY4= =Z4S2 -----END PGP SIGNATURE----- From bakins at web.turner.com Mon May 24 13:34:47 2004 From: bakins at web.turner.com (Brian Akins) Date: Mon, 24 May 2004 09:34:47 -0400 Subject: NFS kickstart fails with e1000.ko drivers In-Reply-To: References: Message-ID: <40B1F9F7.3090602@web.turner.com> seph wrote: > >To clarify my experience, I encountered these problems when using the >intel e1000 card on my dell motherboard. I did NOT have any problems >when using dell supplied pci broadcom gbit cards. So it's clearly not >all gbit ethernet. > > > I have the same issue with Fedora Core 2 and broadcom Gig cards. -- Brian Akins Senior Systems Engineer CNN Internet Technologies From egan at sense.net Mon May 24 15:28:08 2004 From: egan at sense.net (Egan Ford) Date: Mon, 24 May 2004 09:28:08 -0600 Subject: Kickstart and SuSE? In-Reply-To: <1085144771.13440.110.camel@grimlock.cisco.com> Message-ID: <029301c441a3$b84ac8c0$0183a8c0@titan> IMHO, it does not make sense to use kickstart. Autoyast is easy and the native installer. It works and works well. I have yet to see a project that successfully replaces the native installer of the OS (short of cloning) that actually works well and works for every version. The appeal of kickstart is that when a new version of RH is out little to no changes need to be made in your environment to leverage it. The same is true for Autoyast. Netbooting and installing SuSE with AY is very similar to kickstart. Just read the autoyast PDF documentation, use "yast2 autoyast" command from a SuSE box to help create an XML-based config script (this is where I prefer kickstart), and that's about it. If you are interested in a complex frontend to kickstart and autoyast, look at xCAT (xcat.org). > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Brian Long > Sent: Friday, May 21, 2004 7:06 AM > To: kickstart-list at redhat.com > Subject: Kickstart and SuSE? > > > Has anyone on this list used anaconda/kickstart to install SuSE? I've > heard it can be done and I wonder if it makes sense to use kickstart > instead of learning AutoYast. > > Thanks. > > /Brian/ > -- > Brian Long | | | > Americas IT Hosting Sys Admin | .|||. .|||. > Cisco Linux Developer | ..:|||||||:...:|||||||:.. > Phone: (919) 392-7363 | C i s c o S y s t e m s > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From Matt.Fahrner at coat.com Mon May 24 16:54:10 2004 From: Matt.Fahrner at coat.com (Matt Fahrner) Date: Mon, 24 May 2004 12:54:10 -0400 Subject: Kickstart and SuSE? In-Reply-To: <1085144771.13440.110.camel@grimlock.cisco.com> References: <1085144771.13440.110.camel@grimlock.cisco.com> Message-ID: <40B228B2.7020004@coat.com> Sorry I didn't answer this earlier (probably someone already has as well - haven't been paying attention). The tool is "Alice", which can be found here: http://www.suse.de/~fabian/alice/ God knows how to use it though (we need to figure out as well)... - Matt Brian Long wrote: > Has anyone on this list used anaconda/kickstart to install SuSE? I've > heard it can be done and I wonder if it makes sense to use kickstart > instead of learning AutoYast. > > Thanks. > > /Brian/ -- --------------------------------------------------------------------- Matt Fahrner 2 South Park St. Manager of Networking Willis House Burlington Coat Factory Warehouse Lebanon, N.H. 03766 TEL: (603) 448-4100 xt 5150 USA FAX: (603) 443-6190 Matt.Fahrner at COAT.COM --------------------------------------------------------------------- From jglass at liquidweb.com Mon May 24 17:23:11 2004 From: jglass at liquidweb.com (Joseph Glass) Date: Mon, 24 May 2004 13:23:11 -0400 Subject: Preventing /root/anaconda-ks.cfg Message-ID: <1085419391.3048.29.camel@apathy.liquidweb.com> Hi everyone, Is there a way to prevent the creation of /root/anaconda-ks.cfg during a kickstart install? Thanks in advance! -- Joseph Glass Systems Administrator Liquid Web Inc. 800.580.4985 x227 From phr at doc.ic.ac.uk Mon May 24 17:29:37 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Mon, 24 May 2004 18:29:37 +0100 (BST) Subject: Preventing /root/anaconda-ks.cfg In-Reply-To: <1085419391.3048.29.camel@apathy.liquidweb.com> References: <1085419391.3048.29.camel@apathy.liquidweb.com> Message-ID: On Mon, 24 May 2004, Joseph Glass wrote: >Is there a way to prevent the creation of /root/anaconda-ks.cfg during a >kickstart install? There's no setting (AFAIK) to disable it, so either hack anaconda to suppress it or use a %post script to remove it. Cheers, Phil From egan at sense.net Mon May 24 17:36:37 2004 From: egan at sense.net (Egan Ford) Date: Mon, 24 May 2004 11:36:37 -0600 Subject: Kickstart and SuSE? In-Reply-To: <40B228B2.7020004@coat.com> Message-ID: <02e001c441b5$aa8e1220$0183a8c0@titan> Use autoyast, alice is legacy and is more than just an installer, it is more of a management environment. autoyast is the equivalent of kickstart. > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Matt Fahrner > Sent: Monday, May 24, 2004 10:54 AM > To: Discussion list about Kickstart > Subject: Re: Kickstart and SuSE? > > > Sorry I didn't answer this earlier (probably someone already has as > well - haven't been paying attention). > > The tool is "Alice", which can be found here: > http://www.suse.de/~fabian/alice/ God knows how to use it though (we need to figure out as well)... - Matt Brian Long wrote: > Has anyone on this list used anaconda/kickstart to install SuSE? I've > heard it can be done and I wonder if it makes sense to use kickstart > instead of learning AutoYast. > > Thanks. > > /Brian/ -- --------------------------------------------------------------------- Matt Fahrner 2 South Park St. Manager of Networking Willis House Burlington Coat Factory Warehouse Lebanon, N.H. 03766 TEL: (603) 448-4100 xt 5150 USA FAX: (603) 443-6190 Matt.Fahrner at COAT.COM --------------------------------------------------------------------- _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From jeroen at science.uva.nl Tue May 25 06:15:38 2004 From: jeroen at science.uva.nl (Jeroen Roodhart) Date: Tue, 25 May 2004 08:15:38 +0200 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: <1085465737.2667.10.camel@heksenbrij.demon.nl> Hi, Just a "me too" report. I can confirm this to be a problem on most of our Gigabit Cisco switches (I have to owe you the exact type id-s, I'm now working from home) and our Foundry core router. Tried all configurations with Spanning tree disabled etc. No avail. We get our kickstart stages through TFTP/etherboot/DHCP. Same problem: First stage (load PXE image) OK Second stage (load etherboot image) OK Third stage (get filename paramater from DHCP to NFS repo) FAIL On the server side we see (DHCP-REQUEST --> DHCP-ACK)* The e1000 just doesn't want to see the ACK... Put a 10/100 HUB in between and everything works fine. I'm not so sure if this isn't a driver problem instead of kickstart related. Cheers, Jeroen -- Jeroen Roodhart University of Amsterdam jeroen at science.uva.nl Faculty of Science / ICT-Group Systeem- en netwerkbeheer Tel. 020 525 7203 / 06 51338165 -- See http://www.science.uva.nl/~jeroen for openPGP public key -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part URL: From jeroen at science.uva.nl Tue May 25 07:19:25 2004 From: jeroen at science.uva.nl (Jeroen Roodhart) Date: Tue, 25 May 2004 09:19:25 +0200 Subject: NFS kickstart fails with e1000.ko drivers Message-ID: <1085469565.2667.20.camel@heksenbrij.demon.nl> Hi, As an extra data point: This doesn't seem to be Dell specific, I can confirm the same issue on Supermicro systems which also use the e1000.ko driver. With kind regards, Jeroen -- Jeroen Roodhart University of Amsterdam jeroen at science.uva.nl Faculty of Science / ICT-Group Systeem- en netwerkbeheer Tel. 020 525 7203 / 06 51338165 -- See http://www.science.uva.nl/~jeroen for openPGP public key -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part URL: From pantz at lqt.ca Tue May 25 13:28:58 2004 From: pantz at lqt.ca (Paul Pianta) Date: Tue, 25 May 2004 09:28:58 -0400 Subject: Preventing /root/anaconda-ks.cfg In-Reply-To: <1085419391.3048.29.camel@apathy.liquidweb.com> References: <1085419391.3048.29.camel@apathy.liquidweb.com> Message-ID: <40B34A1A.8050000@lqt.ca> Joseph Glass wrote: >Hi everyone, > >Is there a way to prevent the creation of /root/anaconda-ks.cfg during a >kickstart install? > >Thanks in advance! > > > I remove it in my %post script ... %post # clean-up root directory rm -f /root/anaconda-ks.cfg pantz -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From hzhu at wesleyan.edu Tue May 25 13:49:56 2004 From: hzhu at wesleyan.edu (Hong Zhu) Date: Tue, 25 May 2004 09:49:56 -0400 Subject: can't start with kickstart In-Reply-To: <40B34A1A.8050000@lqt.ca> Message-ID: Hi all, We would like to use kickstart to install RH WS 3 on a blade, which only has one USB port for either CDROM drive or floppy drive but not at same time. I have the ks.cfg on the floppy diskette, and ok to boot the blade from RH CDROM to the boot command. How am I able to perform kickstart without cdrom and floppy being available at same time? BTW, I also tried to store the ks.cfg on the NFS server (NAS), then booted from RH CDROM to boot prompt, and issued: ks=nfs:nfs-server-ip-address:/path/to/ks.cfg but it just hung. I think it requires the use of DHCP server however we don't have this available for kickstart use, and so I'm wondering how we're able to perform kickstart without using DHCP. Thanks a lot, Hong From MartinRobb at ieee.org Mon May 24 22:04:21 2004 From: MartinRobb at ieee.org (Martin Robb) Date: Mon, 24 May 2004 18:04:21 -0400 Subject: genhdlist problems under RHEL3 Message-ID: <40B27165.2040800@ieee.org> I have a script which takes a ks.cfg file and a set of install CDs, reads the package and rpm list from the ks.cfg file, and extracts the infrastructure and the corresponding RPMs from the install CDs into an ISO file which can be used to roast an install CD. It all works very nicely under Fedora Core 2 using the FC2 install CDs. The same script can be used with Red Hat Enterprise Linux 3 install CDs to build an ISO file. The only tweaks were 1) replacing references to the Fedora subdirectory with the RedHat subdirectory, 2) omitting the --productpath argument to genhdlist since RHEL3 anaconda doesn't understand it. Although everything appears to work properly, during the install I get an error window: "Error: An error occurred unmounting the CD. Please make sure you're not accessing /mnt/source from the shell on tty2 and then click OK to retry." When I switch to tty2 I find that /mnt/source does not exist as a directory and nothing is mounted there. I can create the directory and mount /dev/hda as /mnt/source, but the install continues to give errors. The difficulty appears to be introduced by genhdlist. If I comment that out and leave the RHEL3 hdlist file, the resulting CD is useable as a first install CD, though the other RHEL disks are necessary since hdlist thinks some of the RPMs are on them. The genhdlist call isn't giving me any errors and the same logic is working under FC2. Any ideas what the problem may be? Thanks, Martin Robb PS: once I get it firing on all cylinders, I can post the script to the list if there is interest. It uses the comps.xml file to convert package names to a list of rpms (though not recursively), it understands the rather large set of rpms that seem to be "implied" under RHEL3 and FC2 without being explicitly listed in comps.xml, and it can handle local additions or overrides to the rpm list. From phr at doc.ic.ac.uk Tue May 25 14:04:42 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 25 May 2004 15:04:42 +0100 (BST) Subject: can't start with kickstart In-Reply-To: References: Message-ID: On Tue, 25 May 2004, Hong Zhu wrote: >We would like to use kickstart to install RH WS 3 on a blade, which >only has one USB port for either CDROM drive or floppy drive but not at >same time. Not even with a USB hub? >I have the ks.cfg on the floppy diskette, and ok to boot the >blade from RH CDROM to the boot command. How am I able to perform >kickstart without cdrom and floppy being available at same time? - Pack all the network and storage drivers you need onto a single floppy (see list archives) - Put the kickstart file on the CD (see list archives) - Put the kickstart file on an NFS/HTTP server, and read it over the network. >BTW, I also tried to store the ks.cfg on the NFS server (NAS), >then booted from RH CDROM to boot prompt, and issued: >ks=nfs:nfs-server-ip-address:/path/to/ks.cfg > >but it just hung. I think it requires the use of DHCP server Try passing the IP address on the kernel command line, e.g. linux ks=nfs:n.n.n.n:/path ip=x.x.x.x netmask=y.y.y.y gateway=z.z.z.z dns=d.d.d.d See /usr/share/doc/anaconda-9.1.1 (assuming RHEL3) for full documentation. Cheers, Phil From phr at doc.ic.ac.uk Tue May 25 14:17:32 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 25 May 2004 15:17:32 +0100 (BST) Subject: genhdlist problems under RHEL3 In-Reply-To: <40B27165.2040800@ieee.org> References: <40B27165.2040800@ieee.org> Message-ID: On Mon, 24 May 2004, Martin Robb wrote: >"Error: An error occurred unmounting the CD. Please make sure you're >not accessing /mnt/source from the shell on tty2 and then click OK to >retry." >When I switch to tty2 I find that /mnt/source does not exist as a >directory and nothing is mounted there. I can create the directory and >mount /dev/hda as /mnt/source, but the install continues to give errors. What is mounted, and can you unmount it by hand? >The difficulty appears to be introduced by genhdlist. If I comment that >out and leave the RHEL3 hdlist file, the resulting CD is useable as a >first install CD, though the other RHEL disks are necessary since hdlist >thinks some of the RPMs are on them. The genhdlist call isn't giving me >any errors and the same logic is working under FC2. How are you calling genhdlist? IIRC there are special options to indicate multiple CDs (--withnumbers??). Cheers, Phil From michael_lynn at ml.com Tue May 25 14:02:48 2004 From: michael_lynn at ml.com (Lynn, Michael (IDS DM&DS)) Date: Tue, 25 May 2004 10:02:48 -0400 Subject: genhdlist problems under RHEL3 Message-ID: Martin, Just registering my extreme interest in your script... thanks for the offer and associated effort. Regards, Mike -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Martin Robb Sent: Monday, May 24, 2004 6:04 PM To: kickstart-list at redhat.com Subject: genhdlist problems under RHEL3 I have a script which takes a ks.cfg file and a set of install CDs, reads the package and rpm list from the ks.cfg file, and extracts the infrastructure and the corresponding RPMs from the install CDs into an ISO file which can be used to roast an install CD. It all works very nicely under Fedora Core 2 using the FC2 install CDs. The same script can be used with Red Hat Enterprise Linux 3 install CDs to build an ISO file. The only tweaks were 1) replacing references to the Fedora subdirectory with the RedHat subdirectory, 2) omitting the --productpath argument to genhdlist since RHEL3 anaconda doesn't understand it. Although everything appears to work properly, during the install I get an error window: "Error: An error occurred unmounting the CD. Please make sure you're not accessing /mnt/source from the shell on tty2 and then click OK to retry." When I switch to tty2 I find that /mnt/source does not exist as a directory and nothing is mounted there. I can create the directory and mount /dev/hda as /mnt/source, but the install continues to give errors. The difficulty appears to be introduced by genhdlist. If I comment that out and leave the RHEL3 hdlist file, the resulting CD is useable as a first install CD, though the other RHEL disks are necessary since hdlist thinks some of the RPMs are on them. The genhdlist call isn't giving me any errors and the same logic is working under FC2. Any ideas what the problem may be? Thanks, Martin Robb PS: once I get it firing on all cylinders, I can post the script to the list if there is interest. It uses the comps.xml file to convert package names to a list of rpms (though not recursively), it understands the rather large set of rpms that seem to be "implied" under RHEL3 and FC2 without being explicitly listed in comps.xml, and it can handle local additions or overrides to the rpm list. _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list ============================================================================== If you are not an intended recipient of this e-mail, please notify the sender, delete it and do not read, act upon, print, disclose, copy, retain or redistribute it. Click here for important additional terms relating to this e-mail. ============================================================================== From James_Martin at ao.uscourts.gov Tue May 25 14:36:50 2004 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Tue, 25 May 2004 10:36:50 -0400 Subject: genhdlist problems under RHEL3 In-Reply-To: <40B27165.2040800@ieee.org> Message-ID: Sometimes it helps to read the archives :). Check out the thread starting 5/13/2004 titled "An error occurred unmounting the CD..." Basically it comes down doing a "genhdlist --withnumbers". I spent a couple days with this.. Red Hat changes things significantly in RHEL3U2.. Here's basically the steps I had to use. genhdlist --withnumbers --hdlist base/hdlist temp_distro/ #i generate my hdlist in a working directory and then later copy it over to the directory i turn into an ISO. export PYTHONDIR=/usr/lib/anaconda /usr/lib/anaconda-runtime/pkgorder temp_distro/ "" rhe3qu2 > pkgorder.txt genhdlist --withnumbers --fileorder pkgorder.txt --hdlist base/hdlist temp_distro/ After this I copy over the hdlist into the temp_distro/ dir. I could not get around doing it twice because of the way my directory structure is set up, but it only takes a second and works fine. If you don't do the pkgorder, anaconda will not pay any attention to depedencies and install everything in alphabetical order (it really screws things up). Good luck. James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 05/24/2004 06:04:21 PM: > I have a script which takes a ks.cfg file and a set of install CDs, > reads the package and rpm list from the ks.cfg file, and extracts the > infrastructure and the corresponding RPMs from the install CDs into an > ISO file which can be used to roast an install CD. It all works very > nicely under Fedora Core 2 using the FC2 install CDs. > > The same script can be used with Red Hat Enterprise Linux 3 install CDs > to build an ISO file. The only tweaks were 1) replacing references to > the Fedora subdirectory with the RedHat subdirectory, 2) omitting the > --productpath argument to genhdlist since RHEL3 anaconda doesn't > understand it. > > Although everything appears to work properly, during the install I get > an error window: > "Error: An error occurred unmounting the CD. Please make sure you're > not accessing /mnt/source from the shell on tty2 and then click OK to > retry." > When I switch to tty2 I find that /mnt/source does not exist as a > directory and nothing is mounted there. I can create the directory and > mount /dev/hda as /mnt/source, but the install continues to give errors. > > The difficulty appears to be introduced by genhdlist. If I comment that > out and leave the RHEL3 hdlist file, the resulting CD is useable as a > first install CD, though the other RHEL disks are necessary since hdlist > thinks some of the RPMs are on them. The genhdlist call isn't giving me > any errors and the same logic is working under FC2. > > Any ideas what the problem may be? > > Thanks, > Martin Robb > > PS: once I get it firing on all cylinders, I can post the script to the > list if there is interest. It uses the comps.xml file to convert > package names to a list of rpms (though not recursively), it understands > the rather large set of rpms that seem to be "implied" under RHEL3 and > FC2 without being explicitly listed in comps.xml, and it can handle > local additions or overrides to the rpm list. > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From rreed at ops.sgp.arm.gov Tue May 25 15:17:42 2004 From: rreed at ops.sgp.arm.gov (Ronald Reed) Date: Tue, 25 May 2004 10:17:42 -0500 Subject: genhdlist problems under RHEL3 In-Reply-To: References: Message-ID: <1085498262.4239.0.camel@ronnote> Add me to the list of interested people. I would love to see that script when you are done with it. Ron Reed On Tue, 2004-05-25 at 09:02, Lynn, Michael (IDS DM&DS) wrote: > Martin, > > Just registering my extreme interest in your script... thanks for the offer and associated effort. > > Regards, > Mike > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Martin Robb > Sent: Monday, May 24, 2004 6:04 PM > To: kickstart-list at redhat.com > Subject: genhdlist problems under RHEL3 > > > I have a script which takes a ks.cfg file and a set of install CDs, > reads the package and rpm list from the ks.cfg file, and extracts the > infrastructure and the corresponding RPMs from the install CDs into an > ISO file which can be used to roast an install CD. It all works very > nicely under Fedora Core 2 using the FC2 install CDs. > > The same script can be used with Red Hat Enterprise Linux 3 install CDs > to build an ISO file. The only tweaks were 1) replacing references to > the Fedora subdirectory with the RedHat subdirectory, 2) omitting the > --productpath argument to genhdlist since RHEL3 anaconda doesn't > understand it. > > Although everything appears to work properly, during the install I get > an error window: > "Error: An error occurred unmounting the CD. Please make sure you're > not accessing /mnt/source from the shell on tty2 and then click OK to > retry." > When I switch to tty2 I find that /mnt/source does not exist as a > directory and nothing is mounted there. I can create the directory and > mount /dev/hda as /mnt/source, but the install continues to give errors. > > The difficulty appears to be introduced by genhdlist. If I comment that > out and leave the RHEL3 hdlist file, the resulting CD is useable as a > first install CD, though the other RHEL disks are necessary since hdlist > thinks some of the RPMs are on them. The genhdlist call isn't giving me > any errors and the same logic is working under FC2. > > Any ideas what the problem may be? > > Thanks, > Martin Robb > > PS: once I get it firing on all cylinders, I can post the script to the > list if there is interest. It uses the comps.xml file to convert > package names to a list of rpms (though not recursively), it understands > the rather large set of rpms that seem to be "implied" under RHEL3 and > FC2 without being explicitly listed in comps.xml, and it can handle > local additions or overrides to the rpm list. > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > ============================================================================== > > If you are not an intended recipient of this e-mail, please notify > the sender, delete it and do not read, act upon, print, disclose, > copy, retain or redistribute it. > > Click here for important additional terms relating to this e-mail. > > > ============================================================================== > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -- =========================== Ron Reed RedHat Certified Engineer SGP Computer Department Manager Unix Systems Administrator ARM SGP CART Site (580)388-4053 ron.reed at arm.gov From forrestx.taylor at intel.com Tue May 25 16:11:50 2004 From: forrestx.taylor at intel.com (Taylor, ForrestX) Date: Tue, 25 May 2004 09:11:50 -0700 Subject: can't start with kickstart In-Reply-To: References: Message-ID: <1085501509.11092.8.camel@bad.jf.intel.com> On Tue, 2004-05-25 at 06:49, Hong Zhu wrote: > Hi all, > > We would like to use kickstart to install RH WS 3 on a blade, > which only has one USB port for either CDROM drive or floppy > drive but not at same time. > > I have the ks.cfg on the floppy diskette, and ok to boot the > blade from RH CDROM to the boot command. How am I able to perform > kickstart without cdrom and floppy being available at same time? > > BTW, I also tried to store the ks.cfg on the NFS server (NAS), > then booted from RH CDROM to boot prompt, and issued: > ks=nfs:nfs-server-ip-address:/path/to/ks.cfg > > but it just hung. I think it requires the use of DHCP server > however we don't have this available for kickstart use, and so > I'm wondering how we're able to perform kickstart without using DHCP. Create a new disc1 by copying the contents to a directory and adding your ks.cfg. Use mkisofs to create an iso and burn it. Your new ks line would be something like this: ks=cdrom:/ks.cfg Forrest From gordon.mcdowall at fasthosts.co.uk Tue May 25 10:26:35 2004 From: gordon.mcdowall at fasthosts.co.uk (Gordon McDowall) Date: Tue, 25 May 2004 11:26:35 +0100 Subject: Kickstart partition woes Message-ID: <040BC5194880BA4DBAF15C0C9D5A11BA0522E5BF@fhexchange01.gl.fasthosts.co.uk> -----Original Message----- Fr Hi I recently sent the following email to the general Redhat list but got no reply, I was hoping that someone on this list might be able to help out...please! Apologies to anyone who may have already read this post in the general mailing list. I am currently trying to upgrade a server from Redhat 7.2 to Redhat Enterprise 3.0, the server is running raid and I would like to keep the /home partition intact and basically do a full install on the remaining / /boot and /var partitions. When I try the install it always errors out with the error "unable to locate raid device None for /home" but as far as I can see the partition section of the kickstart file is fine, does anyone have any suggestions as to what may be the issue. Here is a copy of the partition section of the ks.cfg file #Disk partitioning information part raid.01 --onpart sda1 part raid.03 --onpart sda2 part swap --onpart sda3 part raid.05 --onpart sda6 --noformat part raid.07 --onpart sda5 part raid.02 --onpart sdb1 part raid.04 --onpart sdb2 part swap --onpart sdb3 part raid.06 --onpart sdb6 --noformat part raid.08 --onpart sdb5 raid /boot --level=1 --device=md0 --fstype ext2 raid.01 raid.02 raid / --level=1 --device=md1 --fstype ext3 raid.03 raid.04 raid /var --level=1 --device=md2 --fstype ext3 raid.07 raid.08 raid /home --level=1 --device=md3 --fstype ext3 --noformat raid.05 raid.06 Thanks, any suggestions appreciated Gordon McDowall -- redhat-list mailing list unsubscribe mailto:redhat-list-request at redhat.com?subject=unsubscribe https://www.redhat.com/mailman/listinfo/redhat-list From egan at sense.net Tue May 25 16:39:02 2004 From: egan at sense.net (Egan Ford) Date: Tue, 25 May 2004 10:39:02 -0600 Subject: can't start with kickstart In-Reply-To: Message-ID: <05e201c44276$cd895480$0183a8c0@titan> You can put the ks.cfg on the CD. Google for tips on how to do this and make custom RH cds. > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Hong Zhu > Sent: Tuesday, May 25, 2004 7:50 AM > To: Discussion list about Kickstart > Subject: can't start with kickstart > > > Hi all, > > We would like to use kickstart to install RH WS 3 on a blade, > which only has one USB port for either CDROM drive or floppy > drive but not at same time. > > I have the ks.cfg on the floppy diskette, and ok to boot the > blade from RH CDROM to the boot command. How am I able to perform > kickstart without cdrom and floppy being available at same time? > > BTW, I also tried to store the ks.cfg on the NFS server (NAS), > then booted from RH CDROM to boot prompt, and issued: > ks=nfs:nfs-server-ip-address:/path/to/ks.cfg > > but it just hung. I think it requires the use of DHCP server > however we don't have this available for kickstart use, and so > I'm wondering how we're able to perform kickstart without using DHCP. > > Thanks a lot, > Hong > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From James_Martin at ao.uscourts.gov Tue May 25 20:01:10 2004 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Tue, 25 May 2004 16:01:10 -0400 Subject: can't start with kickstart In-Reply-To: <05e201c44276$cd895480$0183a8c0@titan> Message-ID: You're approach is all wrong here.. If you are using blades, you are handicapping yourself if you are using a floppy or CD. You need to learn how to use pxelinux, which is part of the syslinux family. Basically you set up a DHCP and tftp server. When the blades PXE is activated, it grabs an IP address and is told to pull a kernel & initrd from the tftp server. It is also told where to find a kickstart file. More information can by found on syslinux.zytor.com. James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 05/25/2004 12:39:02 PM: > You can put the ks.cfg on the CD. Google for tips on how to do this and make > custom RH cds. > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com > > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Hong Zhu > > Sent: Tuesday, May 25, 2004 7:50 AM > > To: Discussion list about Kickstart > > Subject: can't start with kickstart > > > > > > Hi all, > > > > We would like to use kickstart to install RH WS 3 on a blade, > > which only has one USB port for either CDROM drive or floppy > > drive but not at same time. > > > > I have the ks.cfg on the floppy diskette, and ok to boot the > > blade from RH CDROM to the boot command. How am I able to perform > > kickstart without cdrom and floppy being available at same time? > > > > BTW, I also tried to store the ks.cfg on the NFS server (NAS), > > then booted from RH CDROM to boot prompt, and issued: > > ks=nfs:nfs-server-ip-address:/path/to/ks.cfg > > > > but it just hung. I think it requires the use of DHCP server > > however we don't have this available for kickstart use, and so > > I'm wondering how we're able to perform kickstart without using DHCP. > > > > Thanks a lot, > > Hong > > > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From MartinRobb at ieee.org Tue May 25 21:47:40 2004 From: MartinRobb at ieee.org (Martin Robb) Date: Tue, 25 May 2004 17:47:40 -0400 Subject: genhdlist problems solved under RHEL3 -- buildIso script enclosed (sorry, long) In-Reply-To: <20040525160101.4829174EA1@hormel.redhat.com> References: <20040525160101.4829174EA1@hormel.redhat.com> Message-ID: <40B3BEFC.8030900@ieee.org> Many thanks for the assistance from the list, especially Philip Rowlands and James Martin. My "buildIso" script is now working under both FC2 and RHEL3u2. It may even work to cross-target RHEL from an FC2 environment -- I don't have time left today to check, but since my problems were tracking 1 for 1 between targeting RHEL from RHEL and targeting RHEL from FC2, I think it's likely. For the record, using --withnumbers on the genhdlist command got me past the "/mnt/source" problem but ultimately grub failed at stage 1. Using the pkgorder command got me past the grub problem. I am enclosing the buildIso script for those interested. It is still a work in progress but is reasonably functional. It requires some supporting files which I will explain as I go. This would work a WHOLE lot better as an html message with attachments, but I don't know how well that would encapsulate in the digest or to all list members with plain mail clients. So let me apologize in advance for the lengthy plain text message. It seemed the most portable if not the most elegant. The heart of the script looks at the packages section of the ks.cfg file and expands it into a list of rpms, using the comps.xml file to expand package names into the corresponding rpms. It does NOT recursively find packages within packages. Along the way I found that both FC2 and RHEL3 have a sizeable list of rpms that are included "automagically" even in a minimal install -- even though they don't belong to any packages and aren't even listed in comps.xml. The script looks for the list of implicit rpms in the files "implicit.Fedora" and "implicit.RedHat" respectively. BTW, I obtained the list by doing a minimal install and running "rpm --query --all" and comparing the results to what I was extracting for the Base, Core and Dialup Networking Support packages. I don't know if there are other implicit RPMs for other packages. For my purposes I am always building off a minimal install. If anyone knows a more direct way to obtain this list I'd be interested in learning it. The rpms should be listed one per line, but I will join them together a bit here to save some space: The implicit.Fedora file contains: beecrypt bind-libs bzip2-libs chkconfig comps cracklib cracklib-dicts cyrus-sasl cyrus-sasl-md5 db4 dev device-mapper elfutils elfutils-libelf findutils gawk gdbm glib glib2 glibc-common gmp gnupg grep groff gzip hesiod hwdata info iptables isdn4k-utils kernel-smp krb5-libs krbafs less libacl libattr libselinux libstdc++ libuser libwvstreams libxml2 libxml2-python lockdev lrzsz lvm2 mailx make MAKEDEV mingetty mkinitrd mktemp modutils ncurses net-tools newt nscd openldap openssh openssl pam pcre perl perl-Filter popt portmap ppp procmail psmisc pyOpenSSL python python-optik pyxf86config rhnlib rhpl rmt rpm-python sed slang syslinux tar tzdata usbutils usermode which words yp-tools zlib The implicit.RedHat file contains: ash basesystem bash beecrypt binutils bzip2-libs chkconfig comps coreutils cpio cracklib cracklib-dicts cups cups-libs cyrus-sasl cyrus-sasl-md5 cyrus-sasl-plain db4 dev e2fsprogs ed elfutils elfutils-libelf ethtool expat file filesystem findutils fontconfig freetype gawk gdbm gettext glib glib2 glibc glibc-common gmp gnupg grep groff gzip hdparm hesiod hotplug htmlview hwdata info initscripts iproute iputils isdn4k-utils jfsutils jwhois kbd krb5-libs krbafs krbafs-utils less libacl libattr libgcc libgcj libjpeg libpng libstdc++ libtermcap libtiff libtool-libs libuser libwvstreams lockdev losetup lslk lsof lvm m4 mailx make MAKEDEV mingetty minicom mkinitrd mktemp modutils mount mt-st ncurses net-tools newt nscd openldap openssh openssh-server openssl pam passwd patch pcre perl perl-Filter popt portmap ppp procmail procps psmisc pspell pyOpenSSL python python-optik pyxf86config raidtools rdist readline redhat-config-mouse redhat-menus redhat-release rhnlib rhpl rmt rootfiles rpm rpm-python rp-pppoe sed sendmail setserial setup shadow-utils slang sudo sysklogd syslinux SysVinit tar tcpdump termcap traceroute tzdata up2date usbutils usermode utempter util-linux vim-minimal wget which wireless-tools words wvdial XFree86-libs XFree86-libs-data XFree86-Mesa-libGL xinetd yp-tools zlib Both FC2 and RHEL3 appear to have some "missing" files that are listed in the comps.xml file but don't exist on the disks. The installer doesn't complain but my buildIso script needs to know about them to know when to stop looking for RPMs. It looks for these in the files "missing.Fedora" and "missing.RedHat." The missing.Fedora file contains: efibootmgr elilo losetup openCryptoki ppc64-utils prctl run s390utils yaboot The missing.RedHat file contains: acpid efibootmgr elilo iprutils iscsi lrzsz openCryptoki ppc64-utils prctl reiserfs-utils s390utils statserial yaboot Finally, there may be some RPMs you want added in or left out depending on which target you've selected. These are in extra.Fedora and extra.RedHat. The extra.Fedora file contains: -at checkpolicy policy-sources The at rpm was not installing properly under FC2t3 and the -at leaves it out. This may be fixed under FC2 -- I haven't had time to check yet. The checkpolicy and policy-sources files are only of interest if you are using selinux. The extra.RedHat file contains: bridge-utils brltty I don't know that I really need either of these, but they are installed in a minimal install and they don't seem to be picked up implicitly like those listed in implicit.RedHat. So for the moment I'm including them explicitly. Before showing the buildIso script itself, I will also include an example ks.cfg file that seems to be working under both FC2 and RHEL3 and should work with this script. I've left off the %post commands and hashed out the rootpw. What's left of the ks.cfg file contains: install cdrom lang en_US.UTF-8 langsupport --default en_US.UTF-8 en_US.UTF-8 keyboard us mouse generic skipx rootpw --iscrypted ########################## firewall --disabled #selinux --disabled authconfig --enableshadow --enablemd5 timezone America/New_York bootloader --location=mbr --append hda=ide-scsi clearpart --all --initlabel part / --fstype ext3 --size 2048 part /sandbox --fstype ext3 --size 1 --grow part swap --size 1024 reboot %packages @ Core @ Dialup Networking Support @ Base # GENERIC ADDITIONS m4 libsafe tripwire # NEEDED FOR NTP libcap ntp tcl expect amhsfilter %post The buildIso script looks first for custom rpms in the CUSTOM_RPMS directory (default ../RPMs), then for replacement rpms in the PACKAGES directory (default ../Packages), and then for standard rpms on the install CDs. It lays out the ISO file directory structure under ROOTS (default ../Roots) is ${ROOTS}/IsoRoot. The script contains a lot of commented mutterings about where things resided under various RH versions. There are many apparent redundancies and out of desperation I tended to replace them all where possible. Much of this could/should be cleaned up -- but it works. The buildIso script contains: #! /bin/sh -- # # Interesting files and where they (should?) reside: # boot.iso: # images/boot.iso # (in 7.1 this was images/boot.img) # initrd.img: (first two are md5 identical in FC2t2) # isolinux/initrd.img # images/boot.iso[/isolinux/initrd.img] # images/pxeboot/initrd.img # (in 7.1 this was images/boot.img[initrd.img]) # isolinux.cfg: (identical) # isolinux/isolinux.cfg # images/boot.iso[isolinux/isolinux.cfg] # (in 7.1 this was images/boot.img[/syslinux.cfg] # ks.cfg: # /ks.cfg # isolinux/ks.cfg # isolinux/initrd.img[(gzipped)/isolinux/ks.cfg] # isolinux/initrd.img[(gzipped)/ks.cfg] # images/pxeboot/initrd.img[(gzipped)/isolinux/ks.cfg] # images/pxeboot/initrd.img[(gzipped)/ks.cfg] # (in 7.1 this was images/boot.img[(gzipped)/ks.cfg] # case `uname --kernel-release` in 2.4*) DEFAULT_TARGET=RedHat ;; 2.6*) DEFAULT_TARGET=Fedora ;; *) echo "Unrecognized kernel release `uname --kernel-release`" exit 1 ;; esac TARGET=${TARGET:-$DEFAULT_TARGET} ROOTS=${ROOTS:-`(cd ../Roots ; pwd)`} PACKAGES=${PACKAGES:-`(cd ../Packages ; pwd)`} CUSTOM_RPMS=${CUSTOM_RPMS:-`(cd ../RPMs ; pwd)`} ISO_LABEL=PinnacleCSIBastionInstall ISO_ROOT=${ROOTS}/IsoRoot ISO_NAME=${1:-BastionInstall.iso} RPMLIST_NAME=/tmp/rpmlist TMPLIST_NAME=/tmp/tmplist RPM_PATH=${TARGET}/RPMS COMPS_NAME=comps.xml COMPS_PATH=${TARGET}/base/${COMPS_NAME} HDLIST_NAME=hdlist HDLIST_PATH=${TARGET}/base/${HDLIST_NAME} # N.B. - the IsoRoot ks.cfg has extra rpm names not in the local ks.cfg KICKSTART_NAME=ks.cfg KICKSTART_PATH=${ISO_ROOT}/${KICKSTART_NAME} BOOTCAT_NAME=boot.cat BOOTCAT_PATH=isolinux/${BOOTCAT_NAME} CD_ROOT=/mnt/cdrom BOOTIMG_NAME=boot.iso BOOTIMG_PATH=images/${BOOTIMG_NAME} BOOTIMG_MOUNT=/mnt/bootimg BOOTIMG_ROOT=${ROOTS}/BootImage CONFIG_NAME=isolinux.cfg CONFIG_PATH=isolinux/${CONFIG_NAME} INITRD_NAME=initrd.img INITRD_PATH=isolinux/${INITRD_NAME} INITRD_MOUNT=/mnt/initrd Cleanup() { for FS_ROOT in ${ISO_ROOT} ${BOOTIMG_ROOT} do if [ -e ${FS_ROOT} ] then rm -rf ${FS_ROOT} fi mkdir ${FS_ROOT} done } ReplicateNonRpms() { if [ -e ${CD_ROOT} ] then sudo umount ${CD_ROOT} 2>/dev/null else sudo mkdir ${CD_ROOT} fi eject echo -n "Please insert ${TARGET} CD 1 and press ENTER: " read line sudo mount ${CD_ROOT} ( cd ${CD_ROOT} ; find . -print | grep -v RPMS | cpio -pdmvu ${ISO_ROOT} ) sudo umount ${CD_ROOT} eject chmod 644 ${ISO_ROOT}/${HDLIST_PATH} } UnpackBootIso() { if [ -e ${BOOTIMG_MOUNT} ] then sudo umount ${BOOTIMG_MOUNT} 2>/dev/null else sudo mkdir ${BOOTIMG_MOUNT} fi sudo mount -o loop -t iso9660 ${ISO_ROOT}/${BOOTIMG_PATH} ${BOOTIMG_MOUNT} mkdir ${BOOTIMG_ROOT} ( cd ${BOOTIMG_MOUNT} ; find . -print | cpio -pdmvu ${BOOTIMG_ROOT} ) sudo umount ${BOOTIMG_MOUNT} } RepackBootIso() { chmod 644 ${ISO_ROOT}/${BOOTIMG_PATH} sudo mkisofs -o ${ISO_ROOT}/${BOOTIMG_PATH} -V "${ISO_LABEL}" -b isolinux/isolinux.bin -c isolinux/boot.cat -no-emul-boot -boot-load-size 4 -boot-info-table -R -J -V -T ${BOOTIMG_ROOT} } # initrd.img: (first two are md5 identical in FC2t2) # isolinux/initrd.img # images/boot.iso[/isolinux/initrd.img] # images/pxeboot/initrd.img UnpackInitRd() { if [ -e ${INITRD_MOUNT} ] then sudo umount ${INITRD_MOUNT} 2>/dev/null else sudo mkdir ${INITRD_MOUNT} fi zcat ${ISO_ROOT}/${INITRD_PATH} >/tmp/${INITRD_NAME} sudo mount -o loop /tmp/${INITRD_NAME} ${INITRD_MOUNT} } RepackInitRd() { sudo umount ${INITRD_MOUNT} rm -f /tmp/${INITRD_NAME}.gz 2>/dev/null gzip /tmp/${INITRD_NAME} sudo cp /tmp/${INITRD_NAME}.gz ${ISO_ROOT}/${INITRD_PATH} sudo cp /tmp/${INITRD_NAME}.gz ${BOOTIMG_ROOT}/${INITRD_PATH} sudo cp /tmp/${INITRD_NAME}.gz ${ISO_ROOT}/images/pxeboot/${INITRD_NAME} } # isolinux.cfg: (they are identical) # isolinux/isolinux.cfg # images/boot.iso[isolinux/isolinux.cfg] # # allowcddma sometimes helps with CD problems - hasn't made a difference here # mem=exactmap mem=640k at 0m mem=1023M at 1M sometimes helps Proliant memory # recognition - hasn't made a difference here # UpdateLinuxConfig() { chmod 644 ${ISO_ROOT}/${CONFIG_PATH} ed - ${ISO_ROOT}/${CONFIG_PATH} << '!' H /default/s/linux/ks/p /append ks/s|ks|ks=cdrom:/ks.cfg|p w q ! chmod 644 ${BOOTIMG_ROOT}/${CONFIG_PATH} cp ${ISO_ROOT}/${CONFIG_PATH} ${BOOTIMG_ROOT}/${CONFIG_PATH} } # ks.cfg: # /ks.cfg # isolinux/ks.cfg # isolinux/initrd.img[(gzipped)/ks.cfg] # images/boot.iso[isolinux/initrd.img[(gzipped)/ks.cfg]] AddKickstartConfig() { cp -f ${KICKSTART_NAME} ${KICKSTART_PATH} ed - ${KICKSTART_PATH} < ${ISO_ROOT}/BOOTMSG.TXT <<'!' Test boot message ! } RpmsInPackage() { sed -n -e "/${1}<\/name>/,/<\/group>/p" <${ISO_ROOT}/${COMPS_PATH} | grep ']*[>]//g' | sed -e 's/ *//' } DetermineRpmList() { # These RPMS seem to be implied without explicit comps.xml entries cat implicit.${TARGET} > ${RPMLIST_NAME} sed -n -e '/%packages/,/%post/p' <${KICKSTART_PATH} | grep -v '^ *#' | grep -v '^%' | grep -v '^ *$' | while read LINE do case ${LINE} in [@]*) RpmsInPackage "`echo $LINE | sed -e 's/@ *//'`" >>${RPMLIST_NAME} ;; [-]*) ;; *) echo ${LINE} >>${RPMLIST_NAME} ;; esac done sort -u -o ${RPMLIST_NAME} ${RPMLIST_NAME} >${TMPLIST_NAME} while read PACKAGE do if ! grep -q '^ *'"${PACKAGE}"' *$' missing.${TARGET} then echo ${PACKAGE} >> ${TMPLIST_NAME} fi done < ${RPMLIST_NAME} mv ${TMPLIST_NAME} ${RPMLIST_NAME} } CopyRPMsFromDirectory() { DIRECTORY=${1:?"usage: $0 "} >${TMPLIST_NAME} while read PACKAGE do FILES=`ls ${DIRECTORY}/${PACKAGE}-*.rpm 2>/dev/null | grep ${PACKAGE}-[a-zA-Z0-9._]*-[a-zA-Z0-9._]*.rpm` if [ -n "${FILES}" ] then for FILE in ${FILES} do echo ${FILE} cp ${FILE} ${ISO_ROOT}/${RPM_PATH} done else echo ${PACKAGE} >> ${TMPLIST_NAME} fi done <${RPMLIST_NAME} mv ${TMPLIST_NAME} ${RPMLIST_NAME} } CopyRPMs() { if [ -e ${ISO_ROOT}/${RPM_PATH} ] then rm -rf ${ISO_ROOT}/${RPM_PATH} fi mkdir ${ISO_ROOT}/${RPM_PATH} CopyRPMsFromDirectory ${CUSTOM_RPMS} CopyRPMsFromDirectory ${PACKAGES} REMAINING_RPMS=`cat ${RPMLIST_NAME} | wc -l` DISK=1 while [ ${REMAINING_RPMS} -gt 0 ] do echo "${REMAINING_RPMS} RPMs remaining." eject echo -n "Please insert ${TARGET} Disk ${DISK} and press ENTER: " read line sudo mount ${CD_ROOT} CopyRPMsFromDirectory ${CD_ROOT}/${TARGET}/RPMS sudo umount ${CD_ROOT} eject REMAINING_RPMS=`cat ${RPMLIST_NAME} | wc -l` DISK=`expr ${DISK} + 1` done } MakeIsoFile() { ( cd ${ROOTS} ; find src -print | cpio -pdmvu ${ISO_ROOT} ) if [ "$TARGET" = RedHat ] then # RHEL3 genhdlist doesn't have --productpath /usr/lib/anaconda-runtime/genhdlist --withnumbers ${ISO_ROOT} PYTHONPATH=/usr/lib/anaconda /usr/lib/anaconda-runtime/pkgorder ${ISO_ROOT}/ '' rhe3qu2 >/tmp/pkgorder.txt /usr/lib/anaconda-runtime/genhdlist --withnumbers --fileorder /tmp/pkgorder.txt ${ISO_ROOT} else /usr/lib/anaconda-runtime/genhdlist --productpath ${TARGET} ${ISO_ROOT} fi sudo mkisofs -o ${ISO_NAME} -V "${ISO_LABEL}" -b isolinux/isolinux.bin -c isolinux/boot.cat -no-emul-boot -boot-info-table -J -r -nobak -T ${ISO_ROOT} } Cleanup ReplicateNonRpms UnpackBootIso #UnpackInitRd UpdateLinuxConfig AddKickstartConfig #AddBootMessage #RepackInitRd RepackBootIso DetermineRpmList CopyRPMs MakeIsoFile From barjunk at attglobal.net Wed May 26 05:14:28 2004 From: barjunk at attglobal.net (mbox mbarsalou) Date: Tue, 25 May 2004 21:14:28 -0800 Subject: genhdlist problems under RHEL3 In-Reply-To: <40B27165.2040800@ieee.org> References: <40B27165.2040800@ieee.org> Message-ID: <1085548468.4834.54.camel@localhost> I'm Interested. On Mon, 2004-05-24 at 14:04, Martin Robb wrote: > I have a script which takes a ks.cfg file and a set of install CDs, > reads the package and rpm list from the ks.cfg file, and extracts the > infrastructure and the corresponding RPMs from the install CDs into an > ISO file which can be used to roast an install CD. It all works very > nicely under Fedora Core 2 using the FC2 install CDs. > > The same script can be used with Red Hat Enterprise Linux 3 install CDs > to build an ISO file. The only tweaks were 1) replacing references to > the Fedora subdirectory with the RedHat subdirectory, 2) omitting the > --productpath argument to genhdlist since RHEL3 anaconda doesn't > understand it. > > Although everything appears to work properly, during the install I get > an error window: > "Error: An error occurred unmounting the CD. Please make sure you're > not accessing /mnt/source from the shell on tty2 and then click OK to > retry." > When I switch to tty2 I find that /mnt/source does not exist as a > directory and nothing is mounted there. I can create the directory and > mount /dev/hda as /mnt/source, but the install continues to give errors. > > The difficulty appears to be introduced by genhdlist. If I comment that > out and leave the RHEL3 hdlist file, the resulting CD is useable as a > first install CD, though the other RHEL disks are necessary since hdlist > thinks some of the RPMs are on them. The genhdlist call isn't giving me > any errors and the same logic is working under FC2. > > Any ideas what the problem may be? > > Thanks, > Martin Robb > > PS: once I get it firing on all cylinders, I can post the script to the > list if there is interest. It uses the comps.xml file to convert > package names to a list of rpms (though not recursively), it understands > the rather large set of rpms that seem to be "implied" under RHEL3 and > FC2 without being explicitly listed in comps.xml, and it can handle > local additions or overrides to the rpm list. > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From sinamdar at cisco.com Wed May 26 06:30:16 2004 From: sinamdar at cisco.com (Sarika Inamdar) Date: Wed, 26 May 2004 12:00:16 +0530 Subject: genhdlist problems under RHEL3 In-Reply-To: <1085548468.4834.54.camel@localhost> Message-ID: <023c01c442ea$e93aecb0$cc064d0a@apac.cisco.com> I'm interested in the script too !!! > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of mbox mbarsalou > Sent: Wednesday, May 26, 2004 10:44 AM > To: Discussion list about Kickstart > Subject: Re: genhdlist problems under RHEL3 > > > I'm Interested. > > On Mon, 2004-05-24 at 14:04, Martin Robb wrote: > > I have a script which takes a ks.cfg file and a set of install CDs, > > reads the package and rpm list from the ks.cfg file, and > extracts the > > infrastructure and the corresponding RPMs from the install > CDs into an > > ISO file which can be used to roast an install CD. It all > works very > > nicely under Fedora Core 2 using the FC2 install CDs. > > > > The same script can be used with Red Hat Enterprise Linux 3 install > > CDs > > to build an ISO file. The only tweaks were 1) replacing > references to > > the Fedora subdirectory with the RedHat subdirectory, 2) > omitting the > > --productpath argument to genhdlist since RHEL3 anaconda doesn't > > understand it. > > > > Although everything appears to work properly, during the > install I get > > an error window: > > "Error: An error occurred unmounting the CD. Please make > sure you're > > not accessing /mnt/source from the shell on tty2 and then > click OK to > > retry." > > When I switch to tty2 I find that /mnt/source does not exist as a > > directory and nothing is mounted there. I can create the > directory and > > mount /dev/hda as /mnt/source, but the install continues to > give errors. > > > > The difficulty appears to be introduced by genhdlist. If I comment > > that > > out and leave the RHEL3 hdlist file, the resulting CD is > useable as a > > first install CD, though the other RHEL disks are necessary > since hdlist > > thinks some of the RPMs are on them. The genhdlist call > isn't giving me > > any errors and the same logic is working under FC2. > > > > Any ideas what the problem may be? > > > > Thanks, > > Martin Robb > > > > PS: once I get it firing on all cylinders, I can post the script to > > the > > list if there is interest. It uses the comps.xml file to convert > > package names to a list of rpms (though not recursively), > it understands > > the rather large set of rpms that seem to be "implied" > under RHEL3 and > > FC2 without being explicitly listed in comps.xml, and it can handle > > local additions or overrides to the rpm list. > > > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickst> art-list > From jrobertson at convera.com Wed May 26 06:45:24 2004 From: jrobertson at convera.com (Joe Robertson) Date: Tue, 25 May 2004 23:45:24 -0700 Subject: can't start with kickstart Message-ID: James_Martin at ao.uscourts.gov said: > You're approach is all wrong here.. If you are using blades, you are > handicapping yourself if you are using a floppy or CD. You > need to learn > how to use pxelinux, which is part of the syslinux family. > Basically you > set up a DHCP and tftp server. When the blades PXE is > activated, it grabs > an IP address and is told to pull a kernel & initrd from the > tftp server. > It is also told where to find a kickstart file. More > information can by > found on syslinux.zytor.com. I agree with James about using PXE. I just hope you don't have the problem of having one of the Gigabit NIC's that inhibit your use of kickstart via network. The e1000 drivers have this problem - depending on the switch you connect it to. Joe Robertson From wtriest at chemistry.ohio-state.edu Wed May 26 21:17:43 2004 From: wtriest at chemistry.ohio-state.edu (Bill (William) Triest) Date: Wed, 26 May 2004 17:17:43 -0400 Subject: Probing Monitor from Postconfig Message-ID: <1085606263.12717.5.camel@meitnerium.chemistry.ohio-state.edu> Everytime I kickstart our test box for RHEL 3, I have to go into system settings, tell it to probe the monitor, and then I can set the resolution (if I do it before then it ignores the change in resolution). This might be a bug in the installer, since during install it posts the correct monitor information, OR I inherited the postconfigure scripts from the guy before me, and he may just be overiding it on accident (basically RHEL postconfigure is modification of the RH9 postconfigure which is....). I have looked into it and don't see where the postconfigure might be overiding it, but it is possible. I think the easiest solution would be to re-probe at the end and have the setting changed. I think this is realistic as I don't have to tell RHEL what monitor I have, I just have to tell it to probe. Links are definately welcome, as I googled the subject to attempt to RTFM and came up empty handed. Thanks, Wiliam E. T. Student Worker - Linux Chemistry Department The Ohio State University From James_Martin at ao.uscourts.gov Wed May 26 22:02:30 2004 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Wed, 26 May 2004 18:02:30 -0400 Subject: Probing Monitor from Postconfig In-Reply-To: <1085606263.12717.5.camel@meitnerium.chemistry.ohio-state.edu> Message-ID: You shouldn't need to do anything in post to configure the monitor.. Why not post your %post section here? What monitor/video card do you have? All I do in my kickstart file to make it automatically probe for video info is have this line: xconfig James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 05/26/2004 05:17:43 PM: > Everytime I kickstart our test box for RHEL 3, I have to go into system > settings, tell it to probe the monitor, and then I can set the > resolution (if I do it before then it ignores the change in resolution). > > This might be a bug in the installer, since during install it posts the > correct monitor information, OR I inherited the postconfigure scripts > from the guy before me, and he may just be overiding it on accident > (basically RHEL postconfigure is modification of the RH9 postconfigure > which is....). I have looked into it and don't see where the > postconfigure might be overiding it, but it is possible. I think the > easiest solution would be to re-probe at the end and have the setting > changed. I think this is realistic as I don't have to tell RHEL what > monitor I have, I just have to tell it to probe. > > Links are definately welcome, as I googled the subject to attempt to > RTFM and came up empty handed. > > Thanks, > Wiliam E. T. > Student Worker - Linux > Chemistry Department > The Ohio State University > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From paul at rh.sbrk.co.uk Thu May 27 09:04:46 2004 From: paul at rh.sbrk.co.uk (paul at rh.sbrk.co.uk) Date: Thu, 27 May 2004 10:04:46 +0100 Subject: FC2 network kickstart not working Message-ID: <20040527090446.GB24793@strut.sbrk.co.uk> Hi, I've configured DHCP and PXE for a network kickstart install. DHCP works, pxe works, the kernel (from images/pxeboot) and initrd.img load, I've passed the ks option as a boot parameter, the machine sends a DHCP request, gets a response, but then makes no attempt to contact the NFS server for a ks.cfg file. Any idea what's missing? I'm not passing any argument to ks, therefore I believe it should be passing default parameters to the NFS server (192.168.8.2 below). Here's the response from my DHCP server for the anaconda DHCP request: 08:30:26.893575 0:0:39:89:be:15 0:6:1b:da:fe:b9 ip 342: 192.168.8.1.bootps > 192.168.8.10.bootpc: [udp sum ok] xid:0xee401bda secs:768 Y:192.168.8.10 S:192.168.8.2 sname "hut2" file "/kickstart/" vend-rfc1048 DHCP:ACK SID:81.2.102.1 LT:0 SM:255.255.255.0 NS:192.168.8.1 DN:"sbrk.co.uk" BR:192.168.8.255 HN:"pxetest" [tos 0x10] (ttl 16, id 0, len 328) After this no network traffic is seen and anaconda goes into an interactive install. Paul From phr at doc.ic.ac.uk Thu May 27 09:12:32 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 27 May 2004 10:12:32 +0100 (BST) Subject: FC2 network kickstart not working In-Reply-To: <20040527090446.GB24793@strut.sbrk.co.uk> References: <20040527090446.GB24793@strut.sbrk.co.uk> Message-ID: On Thu, 27 May 2004 paul at rh.sbrk.co.uk wrote: >I've configured DHCP and PXE for a network kickstart install. DHCP >works, pxe works, the kernel (from images/pxeboot) and initrd.img >load, I've passed the ks option as a boot parameter, the machine >sends a DHCP request, gets a response, but then makes no attempt >to contact the NFS server for a ks.cfg file. Any idea what's >missing? Look to the other virtual consoles for error messages (Alt+F2, Alt+F3 etc.), and at the log output of the NFS server itself (/var/log/messages) to see whether/why the mount is failing. Cheers, Phil From paul at rh.sbrk.co.uk Thu May 27 09:18:46 2004 From: paul at rh.sbrk.co.uk (paul at rh.sbrk.co.uk) Date: Thu, 27 May 2004 10:18:46 +0100 Subject: FC2 network kickstart not working In-Reply-To: References: <20040527090446.GB24793@strut.sbrk.co.uk> Message-ID: <20040527091845.GC24793@strut.sbrk.co.uk> On Thu, May 27, 2004 at 10:12:32AM +0100, Philip Rowlands wrote: > On Thu, 27 May 2004 paul at rh.sbrk.co.uk wrote: > >I've configured DHCP and PXE for a network kickstart install. DHCP > >works, pxe works, the kernel (from images/pxeboot) and initrd.img > >load, I've passed the ks option as a boot parameter, the machine > >sends a DHCP request, gets a response, but then makes no attempt > >to contact the NFS server for a ks.cfg file. Any idea what's > >missing? > Look to the other virtual consoles for error messages (Alt+F2, Alt+F3 > etc.), and at the log output of the NFS server itself > (/var/log/messages) to see whether/why the mount is failing. The NFS server isn't contacted as I can see from tcpdump. After the DHCP request there is no traffic at all, no DNS, no arp, no NFS, no nothing. I'll check the other consoles, maybe I missed something. Cheers, Paul From okatzke at stadthagen.faurecia.com Thu May 27 09:22:47 2004 From: okatzke at stadthagen.faurecia.com (KATZKE Oliver) Date: Thu, 27 May 2004 11:22:47 +0200 Subject: FC2 network kickstart not working Message-ID: > > >I've configured DHCP and PXE for a network kickstart install. DHCP > > >works, pxe works, the kernel (from images/pxeboot) and initrd.img > > >load, I've passed the ks option as a boot parameter, the machine > > >sends a DHCP request, gets a response, but then makes no attempt > > >to contact the NFS server for a ks.cfg file. Any idea what's > > >missing? > > > Look to the other virtual consoles for error messages > (Alt+F2, Alt+F3 > > etc.), and at the log output of the NFS server itself > > (/var/log/messages) to see whether/why the mount is failing. > > The NFS server isn't contacted as I can see from tcpdump. After > the DHCP request there is no traffic at all, no DNS, no arp, no > NFS, no nothing. I'll check the other consoles, maybe I missed > something. Do you have two or more network interfaces in your system? Maybe kickstart tries to use another device than the one where the cable is plugged in? Oliver From paul at rh.sbrk.co.uk Thu May 27 09:25:14 2004 From: paul at rh.sbrk.co.uk (paul at rh.sbrk.co.uk) Date: Thu, 27 May 2004 10:25:14 +0100 Subject: FC2 network kickstart not working In-Reply-To: References: Message-ID: <20040527092514.GD24793@strut.sbrk.co.uk> On Thu, May 27, 2004 at 11:22:47AM +0200, KATZKE Oliver wrote: > Do you have two or more network interfaces in your system? > Maybe kickstart tries to use another device than the one > where the cable is plugged in? No, only one. Paul From wtriest at chemistry.ohio-state.edu Thu May 27 13:23:08 2004 From: wtriest at chemistry.ohio-state.edu (Bill (William) Triest) Date: Thu, 27 May 2004 09:23:08 -0400 Subject: Probing Monitor from Postconfig In-Reply-To: References: Message-ID: <1085664188.13767.26.camel@meitnerium.chemistry.ohio-state.edu> I hesitate to post postconfig since its VERY LONG and it really needs to be cleaned up (I just started this job in April). I don't even see where they are setting up x, so I think that might be the problem (I keep assuming of course they would setup x). If I run xconfig from a terminal (after su - and typing the password) I get a command not found. Is that really the program I run for Red Hat Enterprise? If not can anyone tell me what it is for Enterprise? Thanks for the help, at least now I'm getting some where. William E. T. Student Worker Chemistry Department The Ohio State University On Wed, 2004-05-26 at 18:02, James_Martin at ao.uscourts.gov wrote: > You shouldn't need to do anything in post to configure the monitor.. Why > not post your %post section here? What monitor/video card do you have? > All I do in my kickstart file to make it automatically probe for video > info is have this line: > > xconfig > > > > James S. Martin, RHCE > Contractor > Administrative Office of the United States Courts > Washington, DC > (202) 502-2394 > > kickstart-list-bounces at redhat.com wrote on 05/26/2004 05:17:43 PM: > > > Everytime I kickstart our test box for RHEL 3, I have to go into system > > settings, tell it to probe the monitor, and then I can set the > > resolution (if I do it before then it ignores the change in resolution). > > > > This might be a bug in the installer, since during install it posts the > > correct monitor information, OR I inherited the postconfigure scripts > > from the guy before me, and he may just be overiding it on accident > > (basically RHEL postconfigure is modification of the RH9 postconfigure > > which is....). I have looked into it and don't see where the > > postconfigure might be overiding it, but it is possible. I think the > > easiest solution would be to re-probe at the end and have the setting > > changed. I think this is realistic as I don't have to tell RHEL what > > monitor I have, I just have to tell it to probe. > > > > Links are definately welcome, as I googled the subject to attempt to > > RTFM and came up empty handed. > > > > Thanks, > > Wiliam E. T. > > Student Worker - Linux > > Chemistry Department > > The Ohio State University > > > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From James_Martin at ao.uscourts.gov Thu May 27 13:22:35 2004 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Thu, 27 May 2004 09:22:35 -0400 Subject: FC2 network kickstart not working In-Reply-To: <20040527090446.GB24793@strut.sbrk.co.uk> Message-ID: Are you sending the kickstart file via DHCP or retrieving via NFS? I've never had much luck retrieving it via DHCP. My system writes the pxelinux.cfg files on the fly and uses syntax like: label ks kernel images/rhel/3.0/as/qu1/vmlinuz append ksdevice=link initrd=images/rhel/3.0/as/qu1/initrd.img ks=nfs:156.132.110.110:/data/linux_build/kickstart/seahawk.imd.lab/ks.cfg lang= lowres devfs=nomount ramdisk_size=8192 James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 05/27/2004 05:04:46 AM: > Hi, > > I've configured DHCP and PXE for a network kickstart install. DHCP > works, pxe works, the kernel (from images/pxeboot) and initrd.img > load, I've passed the ks option as a boot parameter, the machine > sends a DHCP request, gets a response, but then makes no attempt > to contact the NFS server for a ks.cfg file. Any idea what's > missing? I'm not passing any argument to ks, therefore I believe > it should be passing default parameters to the NFS server (192.168.8.2 > below). > > Here's the response from my DHCP server for the anaconda DHCP > request: > > 08:30:26.893575 0:0:39:89:be:15 0:6:1b:da:fe:b9 ip 342: 192.168.8.1. > bootps > 192.168.8.10.bootpc: [udp sum ok] xid:0xee401bda secs:768 > Y:192.168.8.10 S:192.168.8.2 sname "hut2" file "/kickstart/" vend- > rfc1048 DHCP:ACK SID:81.2.102.1 LT:0 SM:255.255.255.0 NS:192.168.8.1 > DN:"sbrk.co.uk" BR:192.168.8.255 HN:"pxetest" [tos 0x10] (ttl 16, > id 0, len 328) > > After this no network traffic is seen and anaconda goes into an > interactive install. > > Paul > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From cdmaest at sandia.gov Thu May 27 13:28:06 2004 From: cdmaest at sandia.gov (Maestas, Christopher Daniel) Date: Thu, 27 May 2004 07:28:06 -0600 Subject: FC2 network kickstart not working Message-ID: <8C457A052EDF6F49AA11EA7E3750A581022614CF@es09snlnt.sandia.gov> I have a similar problem using kickstart using qu2 on rhel3 now. Have you tried with the pxe kernels from this update? -----Original Message----- From: James_Martin at ao.uscourts.gov [mailto:James_Martin at ao.uscourts.gov] Sent: Thursday, May 27, 2004 7:23 AM To: Discussion list about Kickstart Subject: Re: FC2 network kickstart not working Are you sending the kickstart file via DHCP or retrieving via NFS? I've never had much luck retrieving it via DHCP. My system writes the pxelinux.cfg files on the fly and uses syntax like: label ks kernel images/rhel/3.0/as/qu1/vmlinuz append ksdevice=link initrd=images/rhel/3.0/as/qu1/initrd.img ks=nfs:156.132.110.110:/data/linux_build/kickstart/seahawk.imd.lab/ks.cfg lang= lowres devfs=nomount ramdisk_size=8192 James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 05/27/2004 05:04:46 AM: > Hi, > > I've configured DHCP and PXE for a network kickstart install. DHCP > works, pxe works, the kernel (from images/pxeboot) and initrd.img > load, I've passed the ks option as a boot parameter, the machine sends > a DHCP request, gets a response, but then makes no attempt to contact > the NFS server for a ks.cfg file. Any idea what's missing? I'm not > passing any argument to ks, therefore I believe it should be passing > default parameters to the NFS server (192.168.8.2 below). > > Here's the response from my DHCP server for the anaconda DHCP > request: > > 08:30:26.893575 0:0:39:89:be:15 0:6:1b:da:fe:b9 ip 342: 192.168.8.1. > bootps > 192.168.8.10.bootpc: [udp sum ok] xid:0xee401bda secs:768 > Y:192.168.8.10 S:192.168.8.2 sname "hut2" file "/kickstart/" vend- > rfc1048 DHCP:ACK SID:81.2.102.1 LT:0 SM:255.255.255.0 NS:192.168.8.1 > DN:"sbrk.co.uk" BR:192.168.8.255 HN:"pxetest" [tos 0x10] (ttl 16, id > 0, len 328) > > After this no network traffic is seen and anaconda goes into an > interactive install. > > Paul > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From teh at phy.anl.gov Thu May 27 13:44:34 2004 From: teh at phy.anl.gov (Ken Teh) Date: Thu, 27 May 2004 08:44:34 -0500 (CDT) Subject: FC2 network kickstart not working In-Reply-To: <20040527090446.GB24793@strut.sbrk.co.uk> Message-ID: I had the same problem until I switched to using DNS names for the server both at the syslinux boot prompt, that is, for example, linux ks=nfs:fcserver.mynet:/disks/fc2/ks/bugs where bugs is the ks.cfg for the bugs installee. And, for the 'nfs' statement in the kickstart file itself, that is, nfs --server=fcserver.mynet --dir=/disks/fc2/cdrom The symptoms were exactly as you described. The DCHP part goes off without problems, but the installee never contacts the NFS server. Hope this helps. I'm guessing they changed anaconda. Live and learn. Ken On Thu, 27 May 2004 paul at rh.sbrk.co.uk wrote: > Hi, > > I've configured DHCP and PXE for a network kickstart install. DHCP > works, pxe works, the kernel (from images/pxeboot) and initrd.img > load, I've passed the ks option as a boot parameter, the machine > sends a DHCP request, gets a response, but then makes no attempt > to contact the NFS server for a ks.cfg file. Any idea what's > missing? I'm not passing any argument to ks, therefore I believe > it should be passing default parameters to the NFS server (192.168.8.2 > below). > > Here's the response from my DHCP server for the anaconda DHCP > request: > > 08:30:26.893575 0:0:39:89:be:15 0:6:1b:da:fe:b9 ip 342: 192.168.8.1.bootps > 192.168.8.10.bootpc: [udp sum ok] xid:0xee401bda secs:768 Y:192.168.8.10 S:192.168.8.2 sname "hut2" file "/kickstart/" vend-rfc1048 DHCP:ACK SID:81.2.102.1 LT:0 SM:255.255.255.0 NS:192.168.8.1 DN:"sbrk.co.uk" BR:192.168.8.255 HN:"pxetest" [tos 0x10] (ttl 16, id 0, len 328) > > After this no network traffic is seen and anaconda goes into an > interactive install. > > Paul > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From phr at doc.ic.ac.uk Thu May 27 13:45:33 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 27 May 2004 14:45:33 +0100 (BST) Subject: Probing Monitor from Postconfig In-Reply-To: <1085664188.13767.26.camel@meitnerium.chemistry.ohio-state.edu> References: <1085664188.13767.26.camel@meitnerium.chemistry.ohio-state.edu> Message-ID: On Thu, 27 May 2004, Bill (William) Triest wrote: >If I run xconfig from a terminal (after su - and typing the password) I >get a command not found. Is that really the program I run for Red Hat >Enterprise? If not can anyone tell me what it is for Enterprise? redhat-config-xfree86 From paul at rh.sbrk.co.uk Thu May 27 13:47:12 2004 From: paul at rh.sbrk.co.uk (paul at rh.sbrk.co.uk) Date: Thu, 27 May 2004 14:47:12 +0100 Subject: FC2 network kickstart not working In-Reply-To: References: <20040527090446.GB24793@strut.sbrk.co.uk> Message-ID: <20040527134712.GE24793@strut.sbrk.co.uk> On Thu, May 27, 2004 at 09:22:35AM -0400, James_Martin at ao.uscourts.gov wrote: > Are you sending the kickstart file via DHCP or retrieving via NFS? I've > never had much luck retrieving it via DHCP. My system writes the > pxelinux.cfg files on the fly and uses syntax like: I've never heard of sending of sending the kickstart file via DHCP. Or do you mean naming the kickstart file with DHCP? > label ks > kernel images/rhel/3.0/as/qu1/vmlinuz > append ksdevice=link initrd=images/rhel/3.0/as/qu1/initrd.img > ks=nfs:156.132.110.110:/data/linux_build/kickstart/seahawk.imd.lab/ks.cfg > lang= lowres devfs=nomount ramdisk_size=8192 I'm just doing: append acpi=off initrd=initrd.img ks DHCP has: if substring (option vendor-class-identifier, 0, 8) = "anaconda" { filename "/kickstart/"; next-server 192.168.8.2; } Paul From wtriest at chemistry.ohio-state.edu Thu May 27 13:50:28 2004 From: wtriest at chemistry.ohio-state.edu (Bill (William) Triest) Date: Thu, 27 May 2004 09:50:28 -0400 Subject: Probing Monitor from Postconfig In-Reply-To: References: <1085664188.13767.26.camel@meitnerium.chemistry.ohio-state.edu> Message-ID: <1085665827.13767.29.camel@meitnerium.chemistry.ohio-state.edu> On Thu, 2004-05-27 at 09:45, Philip Rowlands wrote: > On Thu, 27 May 2004, Bill (William) Triest wrote: > > >If I run xconfig from a terminal (after su - and typing the password) I > >get a command not found. Is that really the program I run for Red Hat > >Enterprise? If not can anyone tell me what it is for Enterprise? > > redhat-config-xfree86 That's what I thought. If I run that from Postconfig is it going to bring up the screen for prompts, or will it just autoprobe stuff like it does for the instalation. Thanks, William E. T. Student Worker - Linux Chemistry Department The Ohio State University From wtriest at chemistry.ohio-state.edu Thu May 27 13:56:09 2004 From: wtriest at chemistry.ohio-state.edu (Bill (William) Triest) Date: Thu, 27 May 2004 09:56:09 -0400 Subject: Probing Monitor from Postconfig In-Reply-To: <1085665827.13767.29.camel@meitnerium.chemistry.ohio-state.edu> References: <1085664188.13767.26.camel@meitnerium.chemistry.ohio-state.edu> <1085665827.13767.29.camel@meitnerium.chemistry.ohio-state.edu> Message-ID: <1085666169.13767.34.camel@meitnerium.chemistry.ohio-state.edu> On Thu, 2004-05-27 at 09:50, Bill (William) Triest wrote: > On Thu, 2004-05-27 at 09:45, Philip Rowlands wrote: > > On Thu, 27 May 2004, Bill (William) Triest wrote: > > > > >If I run xconfig from a terminal (after su - and typing the password) I > > >get a command not found. Is that really the program I run for Red Hat > > >Enterprise? If not can anyone tell me what it is for Enterprise? > > > > redhat-config-xfree86 > > That's what I thought. If I run that from Postconfig is it going to > bring up the screen for prompts, or will it just autoprobe stuff like it > does for the instalation. Sorry about all the e-mails I wanted to expand. Basically I test run all commands, and this brings up the GUI. I had looked into this before and there is an option to set keys ( --set-=value according to redhat-config-xfree86). Since this will run on many comptuers I really want the autoprobing as on out test boxes it works reliably, I just can't figure out how to automate it. Thanks for all the help, I'm really not a newbie I just left redhat a long time ago (I currently run gentoo on my main box, but I have reverted to fedora on my other one so I can get use to redhat's configuration tools instead of manually editing files.) (just pretend I put my thanks there again) > Thanks, > William E. T. > Student Worker - Linux > Chemistry Department > The Ohio State University From Thomas.Cameron at bankofamerica.com Thu May 27 21:28:48 2004 From: Thomas.Cameron at bankofamerica.com (Cameron, Thomas) Date: Thu, 27 May 2004 16:28:48 -0500 Subject: Fedora Core 2 Test 2 kickstart problems Message-ID: <1C455916B3418145B320711B21A2F045A69476@ex2k.bankofamerica.com> You might try one other variation: linux mem=1008M at 16M I seem to recall having to do that with a Proliant 2500R with a gig of memory. What Proliant are you using? -- Thomas Cameron, RHCE, CNE, MCSE, MCT Assistant Vice President Linux Design and Engineering Bank of America (972) 997-9641 The opinions expressed in this message are mine alone and do not necessarily reflect the opinions of my employer, Bank of America. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Martin Robb Sent: Thursday, April 22, 2004 10:53 AM To: Discussion list about Kickstart Subject: Re: Fedora Core 2 Test 2 kickstart problems Thomas, I appear to be dealing with a different issue on my Proliants -- if I boot from the install CD with: linux mem=exactmap mem=640k at 0 mem=1023M at 1M the boot actually dies even earlier -- on "loading linux" rather than on "loading the CCISS drivers". However, the fact that others are having success with the Proliant for RHL9 and FC1 gives me hope. Maybe I just need to play with my firmware settings -- what worked for me on RHL7.1 may not be working for later versions. The Proliant certainly gives you ENOUGH firmware settings to play with :-). Thanks for the tip. Marty Cameron, Thomas wrote: Actually the Proliant platform is my desired target, but I've had a good deal of trouble getting it to work with anything beyond RHL 7.x. Yeah, I had a bit of trouble after 7.x too. I think that there was some memory line that I used for 8.0 and 9, but IIRC, FC1 worked fine for me. See http://groups.google.com/groups?hl=en &lr=&ie=UTF-8&oe=UTF-8&edition=us&frame=right&th=9f79eae9cb81c7b&seekm=3BEF3A9D.1ECFA7A1%40rownetco.com#link2 for the correct syntax for the mem entry in grub.conf. -------------- next part -------------- An HTML attachment was scrubbed... URL: From paul at rh.sbrk.co.uk Thu May 27 21:32:26 2004 From: paul at rh.sbrk.co.uk (paul at rh.sbrk.co.uk) Date: Thu, 27 May 2004 22:32:26 +0100 Subject: FC2 network kickstart not working In-Reply-To: References: <20040527090446.GB24793@strut.sbrk.co.uk> Message-ID: <20040527213225.GF24793@strut.sbrk.co.uk> On Thu, May 27, 2004 at 08:44:34AM -0500, Ken Teh wrote: > I had the same problem until I switched to using DNS names for the server > both at the syslinux boot prompt, that is, for example, > linux ks=nfs:fcserver.mynet:/disks/fc2/ks/bugs > The symptoms were exactly as you described. The DCHP part goes off > without problems, but the installee never contacts the NFS server. Hope > this helps. I'm guessing they changed anaconda. Live and learn. Hmm. I've just looked at the alt-f3 and see this: * url is 192.168.8.2:/kickstart/ * file location: nfs://192.168.8.2:/kickstart/192.168.8.10-kickstart * failed to mount nfs source I tried modifying dhcpd.conf to specify a name but it just turns it back into an IP address there and then. So I specified ks=nfs://pxenfs.sbrk.co.uk:/kickstart/ on the boot command line instead of just ks. On alt-f3 I see: * url is pxenfs.sbrk.co.uk:/kickstart/ * file location: nfs://pxenfs.sbrk.co.uk:/kickstart/192.168.8.10-kickstart Bingo! It works! Thank you Ken. Paul From paul at rh.sbrk.co.uk Fri May 28 07:19:03 2004 From: paul at rh.sbrk.co.uk (paul at rh.sbrk.co.uk) Date: Fri, 28 May 2004 08:19:03 +0100 Subject: FC2 network kickstart not working In-Reply-To: <20040527213225.GF24793@strut.sbrk.co.uk> References: <20040527090446.GB24793@strut.sbrk.co.uk> <20040527213225.GF24793@strut.sbrk.co.uk> Message-ID: <20040528071902.GG24793@strut.sbrk.co.uk> On Thu, May 27, 2004 at 10:32:26PM +0100, paul at rh.sbrk.co.uk wrote: > On Thu, May 27, 2004 at 08:44:34AM -0500, Ken Teh wrote: > > I had the same problem until I switched to using DNS names for the server > > both at the syslinux boot prompt, that is, for example, > > linux ks=nfs:fcserver.mynet:/disks/fc2/ks/bugs > > The symptoms were exactly as you described. The DCHP part goes off > > without problems, but the installee never contacts the NFS server. Hope > > this helps. I'm guessing they changed anaconda. Live and learn. > Hmm. I've just looked at the alt-f3 and see this: > * url is 192.168.8.2:/kickstart/ > * file location: nfs://192.168.8.2:/kickstart/192.168.8.10-kickstart > * failed to mount nfs source > > I tried modifying dhcpd.conf to specify a name but it just > turns it back into an IP address there and then. So I specified > ks=nfs://pxenfs.sbrk.co.uk:/kickstart/ on the boot command line I actually didn't include "//" here, just nfs:pxenfs... > instead of just ks. On alt-f3 I see: > * url is pxenfs.sbrk.co.uk:/kickstart/ > * file location: nfs://pxenfs.sbrk.co.uk:/kickstart/192.168.8.10-kickstart To clarify, I've now tried just about every combination: ks=nfs:pxenfs.sbrk.co.uk:/kickstart/ works ks=nfs:pxenfs:/kickstart/ fails due to domain not used and DNS lookup of "pxenfs." ks=nfs:192.168.8.2:/kickstart/ fails - no DNS lookup, no ARP, nothing Perhaps I'll download the source and have a look into this. Paul From jeroen at science.uva.nl Fri May 28 07:27:09 2004 From: jeroen at science.uva.nl (Jeroen Roodhart) Date: Fri, 28 May 2004 09:27:09 +0200 Subject: FC2 network kickstart not working Message-ID: <40B6E9CD.80400@science.uva.nl> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi! |I've configured DHCP and PXE for a network kickstart install. DHCP |works, pxe works, the kernel (from images/pxeboot) and initrd.img |load, I've passed the ks option as a boot parameter, the machine |sends a DHCP request, gets a response, but then makes no attempt |to contact the NFS server for a ks.cfg file. Are you using by any chance a NIC that needs the e1000.ko driver? If so (or even if not so ;)) try it with a hub in between the client and the switch and see if your problem goes away... There are some _issues_ with this driver and anaconda :) Good luck, Jeroen - -- Jeroen Roodhart University of Amsterdam jeroen at science.uva.nl Faculty of Science / ICT-Group Systeem- en netwerkbeheer Tel. 020 525 7203 / 06 51338165 - -- See http://www.science.uva.nl/~jeroen for openPGP public key -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFAtunN37AP1zFtDU0RAjBjAKDvA0dio66s5P9ncHOWTMgH4nQ1nQCcDt+Q VQc07F7lpFT6cpoJ343znRU= =Fpvj -----END PGP SIGNATURE----- From douglee at arsene.com.tw Sat May 29 09:00:52 2004 From: douglee at arsene.com.tw (Doug Lee) Date: Sat, 29 May 2004 09:00:52 +0000 Subject: Fedora Kickstart Message-ID: <20040529090052.GA29598@ns.kevlo.org> On Tue, May 18, 2004 at 10:55:47AM -0700, Taylor, ForrestX wrote: > On Tue, 2004-05-18 at 10:31, mbox mbarsalou wrote: > > There used be a wiki that had information on how to setup kickstart with > > Fedora....anyone know where it went? I think the website was > > rau.homedns.org. > > Yes, it is: > http://rau.homedns.org/twiki/bin/view/Anaconda/AnacondaDocumentationProject > > Although it is mostly about anaconda and rebuilding a custom distro. The above url helps me a lot. But I have a question, I'm using Fedora Core 2, it seems that comps.rpm should be rebuild on Fedora Core 2, otherwise, I will get "cannot find CD" while booting the cdrom. Did I miss something? thanks. > Forrest Doug. From alexander_rau at yahoo.com Sat May 29 01:35:27 2004 From: alexander_rau at yahoo.com (Alexander Rau (private)) Date: Fri, 28 May 2004 21:35:27 -0400 Subject: Fedora Kickstart In-Reply-To: <20040529090052.GA29598@ns.kevlo.org> Message-ID: Have you checked that the .discinfo file is present? AR -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Doug Lee Sent: Saturday, May 29, 2004 5:01 AM To: kickstart-list at redhat.com Subject: Re: Fedora Kickstart On Tue, May 18, 2004 at 10:55:47AM -0700, Taylor, ForrestX wrote: > On Tue, 2004-05-18 at 10:31, mbox mbarsalou wrote: > > There used be a wiki that had information on how to setup kickstart with > > Fedora....anyone know where it went? I think the website was > > rau.homedns.org. > > Yes, it is: > http://rau.homedns.org/twiki/bin/view/Anaconda/AnacondaDocumentationProject > > Although it is mostly about anaconda and rebuilding a custom distro. The above url helps me a lot. But I have a question, I'm using Fedora Core 2, it seems that comps.rpm should be rebuild on Fedora Core 2, otherwise, I will get "cannot find CD" while booting the cdrom. Did I miss something? thanks. > Forrest Doug. _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From paul at rh.sbrk.co.uk Sat May 29 07:14:54 2004 From: paul at rh.sbrk.co.uk (paul at rh.sbrk.co.uk) Date: Sat, 29 May 2004 08:14:54 +0100 Subject: FC2 network kickstart not working In-Reply-To: <20040528071902.GG24793@strut.sbrk.co.uk> References: <20040527090446.GB24793@strut.sbrk.co.uk> <20040527213225.GF24793@strut.sbrk.co.uk> <20040528071902.GG24793@strut.sbrk.co.uk> Message-ID: <20040529071453.GH24793@strut.sbrk.co.uk> On Fri, May 28, 2004 at 08:19:03AM +0100, paul at rh.sbrk.co.uk wrote: > Perhaps I'll download the source and have a look into this. Well, that was pretty obvious: anaconda-10.0/isys/nfsmount.c #if 1 /* old libc's do not have inet_aton() -- change 1 to 0 */ if (!inet_aton(hostname, &server_addr.sin_addr)) #endif { if (mygethostbyname(hostname, &server_addr.sin_addr)) { myerror = ERROR_HOSTNAME; goto fail; } else { server_addr.sin_family = AF_INET; } } Paul From jrobertson at convera.com Sat May 29 07:24:35 2004 From: jrobertson at convera.com (Joe Robertson) Date: Sat, 29 May 2004 00:24:35 -0700 Subject: FC2 network kickstart not working Message-ID: Paul, Can you elaborate a little on this - for someone who doesn't know anaconda and doesn't see the 'obvious'? I'm very interested in learning about this. Thanks, Joe > > Well, that was pretty obvious: > > anaconda-10.0/isys/nfsmount.c > > #if 1 /* old libc's do not have inet_aton() -- > change 1 to 0 */ > if (!inet_aton(hostname, &server_addr.sin_addr)) > #endif > { > if (mygethostbyname(hostname, > &server_addr.sin_addr)) { > myerror = ERROR_HOSTNAME; > goto fail; > } else { > server_addr.sin_family = AF_INET; > } > } > > Paul > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickst> art-list > From paul at rh.sbrk.co.uk Sat May 29 07:38:10 2004 From: paul at rh.sbrk.co.uk (paul at rh.sbrk.co.uk) Date: Sat, 29 May 2004 08:38:10 +0100 Subject: FC2 network kickstart not working In-Reply-To: References: Message-ID: <20040529073810.GI24793@strut.sbrk.co.uk> On Sat, May 29, 2004 at 12:24:35AM -0700, Joe Robertson wrote: > paul wrote: > > anaconda-10.0/isys/nfsmount.c > > #if 1 /* old libc's do not have inet_aton() -- > > change 1 to 0 */ > > if (!inet_aton(hostname, &server_addr.sin_addr)) > > #endif > > { > > if (mygethostbyname(hostname, > > &server_addr.sin_addr)) { > > myerror = ERROR_HOSTNAME; > > goto fail; > > } else { > > server_addr.sin_family = AF_INET; > > } > > } > Can you elaborate a little on this - for someone who doesn't > know anaconda and doesn't see the 'obvious'? I'm very interested > in learning about this. Anaconda doesn't allow an IP address in an NFS mount, it has to be a resolvable host name. inet_aton() converts a dotted address such as 1.2.3.4 into a network format binary structure, but it has been commented out. Paul From martinr at hotkey.net.au Sat May 29 21:38:24 2004 From: martinr at hotkey.net.au (Martin Rheumer) Date: Sat, 29 May 2004 21:38:24 Subject: Updated RPMS. Message-ID: <3.0.5.32.20040529213824.01918e80@pop.hotkey.net.au> Peeps, Flame all you want, but I think 2 hours of searching the list might be enough especially if somone can flame me the FAQ section or any info would be appreciated. My dilema is simple. I am installing via kickstart on redhat ES 3.0 I want to add things like the new kernel and perl packages from a redhat update ( Redhat network subscribed ) to the original list of files in my http /Redhat/RPMS directory and get my fully functional kickstart routines to install them. 1. Is this possible ? 2. Has anyone got a how to handy. ? 3. Is this the right list to ask on ? Flame away if you think I didnt look hard enough, but kindly spam me the correct way to ask / look for this info.. Google is not my friend Thanks Martin From David_French at intuit.com Sat May 29 12:30:53 2004 From: David_French at intuit.com (French, David) Date: Sat, 29 May 2004 05:30:53 -0700 Subject: Updated RPMS. Message-ID: <9839AD0F1396F845A28E594EF28B2BA80B5EDAFD@sdgex01.sd.intuit.com> I've attached a ksh script that looks in the Updates directory of disk one of a RHEL release. It then copies a new version of an rpm in place of an older one in my custom build directory. It also saves the old file to a directory and updates any boot related files under isolinux, while ignoring my custom files like isolinux.cfg. Anyway, it was a quick and dirty script but works for my needs. You are welcome to start with it. Just remember you will need to rerun genhdlist after replacing the RPMS. To run: cd base_of_your_cd_build_area # Directory that isolinux and RedHat is in UpdateRPMS /cdrom/cdrom1 # where is the new CD root so it can find RedHat/Updates Good luck. --Dave -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Martin Rheumer Sent: Saturday, May 29, 2004 2:38 PM To: kickstart-list at redhat.com Subject: Updated RPMS. Peeps, Flame all you want, but I think 2 hours of searching the list might be enough especially if somone can flame me the FAQ section or any info would be appreciated. My dilema is simple. I am installing via kickstart on redhat ES 3.0 I want to add things like the new kernel and perl packages from a redhat update ( Redhat network subscribed ) to the original list of files in my http /Redhat/RPMS directory and get my fully functional kickstart routines to install them. 1. Is this possible ? 2. Has anyone got a how to handy. ? 3. Is this the right list to ask on ? Flame away if you think I didnt look hard enough, but kindly spam me the correct way to ask / look for this info.. Google is not my friend Thanks Martin _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list -------------- next part -------------- A non-text attachment was scrubbed... Name: UpdateRPMS Type: application/octet-stream Size: 2519 bytes Desc: not available URL: From jim at rossberry.com Sat May 29 12:31:48 2004 From: jim at rossberry.com (Jim Wildman) Date: Sat, 29 May 2004 08:31:48 -0400 (EDT) Subject: Updated RPMS. In-Reply-To: <3.0.5.32.20040529213824.01918e80@pop.hotkey.net.au> References: <3.0.5.32.20040529213824.01918e80@pop.hotkey.net.au> Message-ID: On Sat, 29 May 2004, Martin Rheumer wrote: > 1. Is this possible ? > 2. Has anyone got a how to handy. ? > 3. Is this the right list to ask on ? Simplest way is to add "yum -y update" in the %post section (after installing yum of course). If you want to rebuild the install media then the buildinstall script from the anaconda-runtime package will do the job. I just did this for RHES3 and will complete my howto next week. ------------------------------------------------------------------------ Jim Wildman, CISSP, RHCE jim at rossberry.com http://www.rossberry.com From jrobertson at convera.com Sat May 29 15:19:30 2004 From: jrobertson at convera.com (Joe Robertson) Date: Sat, 29 May 2004 08:19:30 -0700 Subject: FC2 network kickstart not working Message-ID: Paul, ummm... I'll have to investigate this more. I've been using IP addresses all along with kickstart and the only time I've had any problem is when I have ti use the e1000.ko driver. In this case, I get exactly the same symptoms you describe in your case. I wonder if it is inconsistent as to when this is called. I'll be investigating this over the weekend. Thanks, Joe > > On Saturday, May 29, 2004 12:38 AM, Paul wrote: > On Sat, May 29, 2004 at 12:24:35AM -0700, Joe Robertson wrote: > > paul wrote: > > > > anaconda-10.0/isys/nfsmount.c > > > > #if 1 /* old libc's do not have inet_aton() -- > > > change 1 to 0 */ > > > if (!inet_aton(hostname, &server_addr.sin_addr)) #endif > > > { > > > if (mygethostbyname(hostname, > > > &server_addr.sin_addr)) { > > > myerror = ERROR_HOSTNAME; > > > goto fail; > > > } else { > > > server_addr.sin_family = AF_INET; > > > } > > > } > > > Can you elaborate a little on this - for someone who doesn't > > know anaconda and doesn't see the 'obvious'? I'm very interested > > in learning about this. > > Anaconda doesn't allow an IP address in an NFS mount, it has > to be a resolvable host name. inet_aton() converts a dotted > address such as 1.2.3.4 into a network format binary > structure, but it has been commented out. > > Paul > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickst> art-list > From phr at doc.ic.ac.uk Sat May 29 15:26:33 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Sat, 29 May 2004 16:26:33 +0100 (BST) Subject: Updated RPMS. In-Reply-To: <3.0.5.32.20040529213824.01918e80@pop.hotkey.net.au> References: <3.0.5.32.20040529213824.01918e80@pop.hotkey.net.au> Message-ID: On Sat, 29 May 2004, Martin Rheumer wrote: >My dilema is simple. A dilemma is a choice between [two] alternatives. This isn't one. >I want to add things like the new kernel and perl packages from a >redhat update ( Redhat network subscribed ) to the original list of >files in my http /Redhat/RPMS directory and get my fully functional >kickstart routines to install them. > >1. Is this possible ? Sure. >2. Has anyone got a how to handy. ? Depends which media you install from. At a high level: 1. Replace old packages with new in RedHat/RPMS directory 2. Run genhdlist (from anaconda-runtime package) to rebuild file in RedHat/base (BTW, genhdlist is not statically linked any more - anyone know why?) HTTP/NFS is easier than rebuilding CDs. >3. Is this the right list to ask on ? The above is not the only way - anaconda now supports an "Updates" directory, which was used for RHEL3U1 but not RHEL3U2. Another reply suggests doing the updates in a %post, but that has time and storage implications for the install. I prefer to pre-patch. If you've got a RHEL3 subscription, be sure to download Update 2, which fixes a few issues in the installer. Cheers, Phil From vince.skahan at boeing.com Sat May 29 17:41:07 2004 From: vince.skahan at boeing.com (Skahan, Vince) Date: Sat, 29 May 2004 10:41:07 -0700 Subject: FC2 network kickstart not working Message-ID: <1C16569DFF0D3D4699EFA6C64D37DAE002D6F488@xch-nw-07.nw.nos.boeing.com> Why shouldn't it accept anything valid for NFS ? Seems like a reasonable thing to ask for the anaconda changes for FC3.... -- vince.skahan at boeing.com -- Cabin Network -- connexion by Boeing sm www.connexionbyboeing.com -----Original Message----- From: paul at rh.sbrk.co.uk [mailto:paul at rh.sbrk.co.uk] Sent: Saturday, May 29, 2004 12:38 AM To: Discussion list about Kickstart Subject: Re: FC2 network kickstart not working Anaconda doesn't allow an IP address in an NFS mount, it has to be a resolvable host name. inet_aton() converts a dotted address such as 1.2.3.4 into a network format binary structure, but it has been commented out. Paul _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From ryangolhar at verizon.net Sat May 29 21:02:16 2004 From: ryangolhar at verizon.net (Ryan Golhar) Date: Sat, 29 May 2004 17:02:16 -0400 Subject: Updated RPMS. In-Reply-To: <3.0.5.32.20040529213824.01918e80@pop.hotkey.net.au> Message-ID: <015001c445c0$3958e120$2e00a8c0@GOLHARMOBILE1> The the list of new packages using up2date -l. Download the new packages using up2date -u (I think). Copy the new package to whereever you put the CDs and delete the older packages. I started doing this but turned out to be too much trouble. Its easier to just let the machines download the packages themselves. Sometimes it might be slow...but less work on your part. Ryan -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Martin Rheumer Sent: Saturday, May 29, 2004 5:38 PM To: kickstart-list at redhat.com Subject: Updated RPMS. Peeps, Flame all you want, but I think 2 hours of searching the list might be enough especially if somone can flame me the FAQ section or any info would be appreciated. My dilema is simple. I am installing via kickstart on redhat ES 3.0 I want to add things like the new kernel and perl packages from a redhat update ( Redhat network subscribed ) to the original list of files in my http /Redhat/RPMS directory and get my fully functional kickstart routines to install them. 1. Is this possible ? 2. Has anyone got a how to handy. ? 3. Is this the right list to ask on ? Flame away if you think I didnt look hard enough, but kindly spam me the correct way to ask / look for this info.. Google is not my friend Thanks Martin _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From jeroen at science.uva.nl Sat May 29 23:05:18 2004 From: jeroen at science.uva.nl (Jeroen Roodhart) Date: Sun, 30 May 2004 01:05:18 +0200 Subject: FC2 network kickstart not working Message-ID: <40B9172E.30701@science.uva.nl> >paul wrote: > >>> > > anaconda-10.0/isys/nfsmount.c >>> > > #if 1 /* old libc's do not have inet_aton() -- >>> > > change 1 to 0 */ >>> > > if (!inet_aton(hostname, &server_addr.sin_addr)) #endif >>> > > { ... > > Anaconda doesn't allow an IP address in an NFS mount, it has > to be a resolvable host name. inet_aton() converts a dotted > address such as 1.2.3.4 into a network format binary > structure, but it has been commented out. Aah, I'm sorry Paul but that's not commented out. It's inside an #if 1 - #endif pair. So if you have the old libc, this will bomb on compile and you have to change 1 to 0 so the precompiler leaves it out... Regards, Jeroen -- -- Toevallig wel! -- Jeroen Roodhart (jeroen at tuxwielder.demon.nl) | Registered Linux User Brouwersgracht 214 III | No. 73309 1013 HD, Amsterdam, The Netherlands | -- See http://www.science.uva.nl/~jeroen for openPGP public key From David.Avrahami at comverse.com Mon May 31 10:53:47 2004 From: David.Avrahami at comverse.com (Avrahami David) Date: Mon, 31 May 2004 13:53:47 +0300 Subject: Auto detection option question Message-ID: Hi , I am thinking of using kick start for installing the Linux(RHEL 3) on either ide and scsii boxes. I would like to use the same installation disk(ks.cfg) for both installation. I was wondering if kick start has an ability of ide/scsii auto detection. Thanks in advance, David Avrahami Linux Platform Tel: +972-3-6452374 Mobile: +972-54-382374 Email: David.Avrahami at Comverse.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From pantz at lqt.ca Mon May 31 13:30:14 2004 From: pantz at lqt.ca (Paul Pianta) Date: Mon, 31 May 2004 09:30:14 -0400 Subject: Updated RPMS. In-Reply-To: References: <3.0.5.32.20040529213824.01918e80@pop.hotkey.net.au> Message-ID: <40BB3366.6030502@lqt.ca> Philip Rowlands wrote: >On Sat, 29 May 2004, Martin Rheumer wrote: > > > >>My dilema is simple. >> >> > > >A dilemma is a choice between [two] alternatives. This isn't one. > > > I guess you could call his situation a choice between two alternatives: 1. Successfully add new rpm packages and get kickstart to install them ... or 2. Unsuccessfully add new rpm packages and not get kickstart to install them ... :) -- Before you criticize someone, walk a mile in their shoes ... That way when you do criticize them, you're a mile away and you have their shoes! From paul at rh.sbrk.co.uk Mon May 31 13:31:39 2004 From: paul at rh.sbrk.co.uk (paul at rh.sbrk.co.uk) Date: Mon, 31 May 2004 14:31:39 +0100 Subject: FC2 network kickstart not working In-Reply-To: <40B9172E.30701@science.uva.nl> References: <40B9172E.30701@science.uva.nl> Message-ID: <20040531133139.GJ24793@strut.sbrk.co.uk> On Sun, May 30, 2004 at 01:05:18AM +0200, Jeroen Roodhart wrote: > >paul wrote: > >>> > > anaconda-10.0/isys/nfsmount.c > >>> > > #if 1 /* old libc's do not have inet_aton() -- > >>> > > change 1 to 0 */ > >>> > > if (!inet_aton(hostname, &server_addr.sin_addr)) #endif > >>> > > { > Aah, I'm sorry Paul but that's not commented out. It's inside an #if 1 - > #endif pair. So if you have the old libc, this will bomb on compile and > you have to change 1 to 0 so the precompiler leaves it out... Aagh. I'm so used to reading #if 0 that I didn't even notice that. Oh well, shall investigate further... Paul From phr at doc.ic.ac.uk Mon May 31 16:31:17 2004 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Mon, 31 May 2004 17:31:17 +0100 (BST) Subject: Auto detection option question In-Reply-To: References: Message-ID: Please don't send HTML to mailing lists. On Mon, 31 May 2004, David Avrahami wrote: >I am thinking of using kick start for installing the Linux(RHEL 3) on >either ide and scsii boxes. I would like to use the same installation >disk(ks.cfg) for both installation. I was wondering if kick start has >an ability of ide/scsii auto detection. Yes, with the caveat that any "--ondisk" settings will probably not work, as the "hda"/"sda" designation is tied to the underlying storage. Cheers, Phil