From trsmith at mail.nih.gov Wed Mar 1 13:59:09 2006 From: trsmith at mail.nih.gov (Smith, Troy (NIH/NCI) [C]) Date: Wed, 1 Mar 2006 08:59:09 -0500 Subject: How do I specify i386 & x86_64 pkgs in rhel 4 64 bit kickstart Message-ID: <86C31097FBD0B549855F1209D0C2183E0BFD23@NIHCESMLBX11.nih.gov> Hello. Please forgive me if this issue has been discussed in a previous post. We are installing several instances of 64 bit rhel es 4 using kickstart, and have found that we need to go back after the installation and manually install the i386 version (specific to 64 bit rhel es 4 from dvd or rhn) of certain pkgs that already have the x86_64 bit version, in order to satisfy library dependencies for 3rd party apps. Is it possible to specify the i386 and x86_64 version of the same pkg in the %post section? How? Any and all assistance is greatly appreciated. Thanks Troy -------------- next part -------------- An HTML attachment was scrubbed... URL: From rspaulding at mail.arc.nasa.gov Wed Mar 1 16:27:57 2006 From: rspaulding at mail.arc.nasa.gov (Ryan C. Spaulding) Date: Wed, 1 Mar 2006 08:27:57 -0800 Subject: How do I specify i386 & x86_64 pkgs in rhel 4 64 bit kickstart In-Reply-To: <86C31097FBD0B549855F1209D0C2183E0BFD23@NIHCESMLBX11.nih.gov> References: <86C31097FBD0B549855F1209D0C2183E0BFD23@NIHCESMLBX11.nih.gov> Message-ID: <20FBC54F-D030-4AFF-A60A-D921580D2568@mail.arc.nasa.gov> Hi Troy, Add this to your packages section: %packages --resolvedeps @Compatibility Arch Support @Compatibility Arch Development Support Thank you, Ryan On Mar 1, 2006, at 5:59 AM, Smith, Troy (NIH/NCI) [C] wrote: > Hello. Please forgive me if this issue has been discussed in a > previous post. > > > > We are installing several instances of 64 bit rhel es 4 using > kickstart, and have found that we need to go back after the > installation and manually install the i386 version (specific to > 64 bit rhel es 4 from dvd or rhn) of certain pkgs that already have > the x86_64 bit version, in order to satisfy library dependencies > for 3rd party apps. > > > > Is it possible to specify the i386 and x86_64 version of the same > pkg in the %post section? How? > > > > Any and all assistance is greatly appreciated. > > > > Thanks > > Troy > > > > > > _______________________________________________ > 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 trsmith at mail.nih.gov Wed Mar 1 18:32:39 2006 From: trsmith at mail.nih.gov (Smith, Troy (NIH/NCI) [C]) Date: Wed, 1 Mar 2006 13:32:39 -0500 Subject: How do I specify i386 & x86_64 pkgs in rhel 4 64 bit kickstart Message-ID: <86C31097FBD0B549855F1209D0C2183E0BFD30@NIHCESMLBX11.nih.gov> Thanks, we'll give it a shot. We appreciate the help. Thanks Troy Hi Troy, Add this to your packages section: %packages --resolvedeps @Compatibility Arch Support @Compatibility Arch Development Support Thank you, Ryan On Mar 1, 2006, at 5:59 AM, Smith, Troy (NIH/NCI) [C] wrote: Hello. Please forgive me if this issue has been discussed in a previous post. We are installing several instances of 64 bit rhel es 4 using kickstart, and have found that we need to go back after the installation and manually install the i386 version (specific to 64 bit rhel es 4 from dvd or rhn) of certain pkgs that already have the x86_64 bit version, in order to satisfy library dependencies for 3rd party apps. Is it possible to specify the i386 and x86_64 version of the same pkg in the %post section? How? Any and all assistance is greatly appreciated. Thanks Troy _______________________________________________ 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 ks at puzzled.xs4all.nl Thu Mar 2 15:28:07 2006 From: ks at puzzled.xs4all.nl (Patrick) Date: Thu, 02 Mar 2006 16:28:07 +0100 Subject: New to list and question Message-ID: <1141313287.5514.67.camel@guru.puzzled.xs4all.nl> Hi all, Last time I was on this list was iirc in 2002 so I'm sort of new to the list. Recently I setup a PXE/kickstart install system for a bunch of servers. Works very well (pxe/kickstart rocks!) but it is incomplete so I would like to iron out more details. Once the install is finished I would like to: * add extra firewall rules to /etc/sysconfig/iptables or replace it * change or replace some configs like /etc/ntp.conf, /etc/hosts * turn off a ton of unneeded services (chkconfig --level 2345 foo off) * turn off IPv6 in (iirc) /etc/modules.conf * turn off zerconf in /etc/sysconfig/network * install a bunch of updates * install some extra rpms I assume I should specify this in the %post section of the kickstart file but I don't know what the required syntax is. Any suggestions or docs how I go about this? Thanks and regards, Patrick From dan at half-asleep.com Thu Mar 2 16:02:56 2006 From: dan at half-asleep.com (Daniel Segall) Date: 2 Mar 2006 11:02:56 -0500 Subject: New to list and question In-Reply-To: <1141313287.5514.67.camel@guru.puzzled.xs4all.nl> Message-ID: On 3/2/2006, "Patrick" wrote: >Hi all, > >Last time I was on this list was iirc in 2002 so I'm sort of new to the >list. Recently I setup a PXE/kickstart install system for a bunch of >servers. Works very well (pxe/kickstart rocks!) but it is incomplete so >I would like to iron out more details. Once the install is finished I >would like to: > >* add extra firewall rules to /etc/sysconfig/iptables or replace it >* change or replace some configs like /etc/ntp.conf, /etc/hosts >* turn off a ton of unneeded services (chkconfig --level 2345 foo off) >* turn off IPv6 in (iirc) /etc/modules.conf >* turn off zerconf in /etc/sysconfig/network >* install a bunch of updates >* install some extra rpms > >I assume I should specify this in the %post section of the kickstart >file but I don't know what the required syntax is. Any suggestions or >docs how I go about this? > >Thanks and regards, >Patrick > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > Hi Patrick, I do most of this type of stuff throuigh a shell script that gets called in %post (you could do it all in %post as well, but I find it easier to maintain a single script that can be called from any of my kickstart configs, or even run standalone). I've found that the easiest way to do things is to just create new files to replace the old, rather than manually inserting things into them. For example, if you wanted to have a certain set of iptables rules on all machines, you could just create the rules file, and keep it on an NFS server, and have your %post grab that file and replace the one in /etc/sysconfig. You can also run interactive commands during post, so it would be possible to write something that would prompt you for input. I use "dialog" for this, but you could use anything. Anyhow, this is how my %post looks: %post #%post --interpreter /bin/sh exec < /dev/tty3 > /dev/tty3 chvt 3 cat <>/etc/motd ####################################### ## KICKSTART STILL IN PROGRESS!!!!!! ## ####################################### `cat /etc/redhat-release` - Installed `date` EOF echo " * Mounting our NFS share..." mkdir /mnt/ks mount -r -o nolock 10.1.1.1:/kickstart /mnt/ks echo " * Copying config files from NFS..." cp -a /mnt/ks/linux-postinstall.sh /root/ cp -a /mnt/ks/redhatconfig.tar.gz /root/ # Grab our gdm-conf cp -a /mnt/ks/fedora-gdm.conf /root/ # Import the RHN GPG key rpm --import /usr/share/rhn/RPM-GPG-KEY-fedora # Grab the NVIDIA driver installer in case we need it cp -a /mnt/ks/NVIDIA-Linux-x86-1.0-7676-pkg1.run /root/ # Grab our yum.conf cp -f /mnt/ks/yum.conf /etc/yum.conf # Unmount the share & delete the mount point umount /mnt/ks rm -rf /mnt/ks # Change our motd to a "normal" value rm -f /etc/motd cat <>/etc/motd `cat /etc/redhat-release` - Installed `date` EOF # Execute our post install if [ -f /root/linux-postinstall.sh ]; then echo " * Executing Post-Install..." chmod +x /root/linux-postinstall.sh else echo "Can't find linux-postinstall.sh." fi # We done. echo " * Post-Install complete. System rebooting." >>/root/ks.log echo " * Post-Install complete. System rebooting." chvt 3 sleep 10 exit 0 #EOF Hope that helps a bit. -Dan From dan at half-asleep.com Thu Mar 2 16:08:35 2006 From: dan at half-asleep.com (Daniel Segall) Date: 2 Mar 2006 11:08:35 -0500 Subject: New to list and question In-Reply-To: Message-ID: On 3/2/2006, "Daniel Segall" wrote: ># Execute our post install >if [ -f /root/linux-postinstall.sh ]; then > echo " * Executing Post-Install..." > chmod +x /root/linux-postinstall.sh >else > echo "Can't find linux-postinstall.sh." >fi > ># We done. >echo " * Post-Install complete. System rebooting." >>/root/ks.log >echo " * Post-Install complete. System rebooting." >chvt 3 >sleep 10 >exit 0 >#EOF > > >Hope that helps a bit. > > >-Dan > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > I seemed to have missed a line here. It's obviously missing the exec line. It should look like this: # Execute our post install if [ -f /root/linux-postinstall.sh ]; then echo " * Executing Post-Install..." chmod +x /root/linux-postinstall.sh exec /root/linux-postinstall.sh else echo "Can't find linux-postinstall.sh." fi # We done. echo " * Post-Install complete. System rebooting." chvt 3 sleep 10 exit 0 #EOF -Dan From anaconda at maisondubonheur.com Thu Mar 2 16:12:30 2006 From: anaconda at maisondubonheur.com (Alain TAUCH) Date: Thu, 2 Mar 2006 17:12:30 +0100 Subject: New to list and question In-Reply-To: <1141313287.5514.67.camel@guru.puzzled.xs4all.nl> References: <1141313287.5514.67.camel@guru.puzzled.xs4all.nl> Message-ID: <20060302171230.7669dc41.anaconda@maisondubonheur.com> On Thu, 02 Mar 2006 16:28:07 +0100 Patrick [Patrick] wrote: [...] Patrick> I would like to iron out more details. Once the install is finished I Patrick> would like to: Patrick> Patrick> * add extra firewall rules to /etc/sysconfig/iptables or replace it Patrick> * change or replace some configs like /etc/ntp.conf, /etc/hosts Patrick> * turn off a ton of unneeded services (chkconfig --level 2345 foo off) Patrick> * turn off IPv6 in (iirc) /etc/modules.conf Patrick> * turn off zerconf in /etc/sysconfig/network Patrick> * install a bunch of updates Patrick> * install some extra rpms Patrick> Patrick> I assume I should specify this in the %post section of the kickstart Patrick> file but I don't know what the required syntax is. Any suggestions or Patrick> docs how I go about this? The kickstart %post section syntax is documented in your favorite redhat linux documentation. You can use it like a regular shell script, eg: %post /sbin/chkconfig --level 2345 sendmail off -- Alain TAUCH From ks at puzzled.xs4all.nl Thu Mar 2 16:32:55 2006 From: ks at puzzled.xs4all.nl (Patrick) Date: Thu, 02 Mar 2006 17:32:55 +0100 Subject: New to list and question In-Reply-To: <20060302171230.7669dc41.anaconda@maisondubonheur.com> References: <1141313287.5514.67.camel@guru.puzzled.xs4all.nl> <20060302171230.7669dc41.anaconda@maisondubonheur.com> Message-ID: <1141317175.5514.83.camel@guru.puzzled.xs4all.nl> On Thu, 2006-03-02 at 17:12 +0100, Alain TAUCH wrote: > The kickstart %post section syntax is documented in your favorite > redhat > linux documentation. Well, man system-config-kickstart doesn't really provide much info. The stuff in /usr/share/doc/system-config-kickstart-*/ was a little better but still not much to go on. > You can use it like a regular shell script, eg: > > %post > > /sbin/chkconfig --level 2345 sendmail off Now that is helpful. Thanks for the info. Regards, Patrick From ks at puzzled.xs4all.nl Thu Mar 2 16:34:04 2006 From: ks at puzzled.xs4all.nl (Patrick) Date: Thu, 02 Mar 2006 17:34:04 +0100 Subject: New to list and question In-Reply-To: References: Message-ID: <1141317244.5514.86.camel@guru.puzzled.xs4all.nl> On Thu, 2006-03-02 at 11:02 -0500, Daniel Segall wrote: [snip] > Anyhow, this is how my %post looks: [snip the good stuff] That was excellent information. Thanks! Regards, Patrick From alovell at as.arizona.edu Thu Mar 2 16:31:45 2006 From: alovell at as.arizona.edu (Alex Lovell-Troy) Date: Thu, 2 Mar 2006 09:31:45 -0700 Subject: New to list and question In-Reply-To: <1141313287.5514.67.camel@guru.puzzled.xs4all.nl> References: <1141313287.5514.67.camel@guru.puzzled.xs4all.nl> Message-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Mar 2, 2006, at 8:28 AM, Patrick wrote: > Hi all, > > Last time I was on this list was iirc in 2002 so I'm sort of new to > the > list. Recently I setup a PXE/kickstart install system for a bunch of > servers. Works very well (pxe/kickstart rocks!) but it is > incomplete so > I would like to iron out more details. Once the install is finished I > would like to: > > * add extra firewall rules to /etc/sysconfig/iptables or replace it > * change or replace some configs like /etc/ntp.conf, /etc/hosts > * turn off a ton of unneeded services (chkconfig --level 2345 foo off) > * turn off IPv6 in (iirc) /etc/modules.conf > * turn off zerconf in /etc/sysconfig/network > * install a bunch of updates > * install some extra rpms > > I assume I should specify this in the %post section of the kickstart > file but I don't know what the required syntax is. Any suggestions or > docs how I go about this? > > Thanks and regards, > Patrick > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list Patrick, I have found that a fairly minimal postinstall script with cfengine handling all the other changes later. The advantage of this is that cfengine can help maintain the state of the machine long after the installation. Another benefit of this method is that any changes you want to make to a group of machines will happen immediately and any newly kickstarted machines will get the updated config. For things like iptables blacklists and whitelists, it is really useful. So, my post install script downloads and installs yum with my custom configurations and then uses yum to apply security patches and install some third party applications that I rely on and then it calls cfengine to handle the rest of it. Check it out at http://www.cfengine.org - -alex Alex Lovell-Troy Computer Systems Engineer Large Binocular Telescope alt at arizona.edu -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) iD8DBQFEBx3xdmWtRNAkjzERAnNGAJ4hF+feg0rFbTKynJeibm54vszejQCfRuh1 JdD1l42Wm1rlE4oBenaNGdk= =CTd5 -----END PGP SIGNATURE----- From anaconda at maisondubonheur.com Thu Mar 2 17:28:24 2006 From: anaconda at maisondubonheur.com (Alain TAUCH) Date: Thu, 2 Mar 2006 18:28:24 +0100 Subject: New to list and question In-Reply-To: <1141317175.5514.83.camel@guru.puzzled.xs4all.nl> References: <1141313287.5514.67.camel@guru.puzzled.xs4all.nl> <20060302171230.7669dc41.anaconda@maisondubonheur.com> <1141317175.5514.83.camel@guru.puzzled.xs4all.nl> Message-ID: <20060302182824.1df2bbda.anaconda@maisondubonheur.com> On Thu, 02 Mar 2006 17:32:55 +0100 Patrick [Patrick] wrote: Patrick> On Thu, 2006-03-02 at 17:12 +0100, Alain TAUCH wrote: Patrick> > The kickstart %post section syntax is documented in your favorite Patrick> > redhat Patrick> > linux documentation. Patrick> Patrick> Well, man system-config-kickstart doesn't really provide much info. The Patrick> stuff in /usr/share/doc/system-config-kickstart-*/ was a little better Patrick> but still not much to go on. what about: Red Hat Linux 9: Red Hat Linux Customization Guide Chapter 7. Kickstart Installations 7.7. Post-installation Script or: Red Hat Enterprise Linux 3: System Administration Guide Chapter 9. Kickstart Installations 9.7. Post-installation Script or even: Red Hat Enterprise Linux 4: System Administration Guide Chapter 1. Kickstart Installations 1.7. Post-installation Script Don't get me wrong, I'm not used to throw some RTFM just for the fun. Regards -- Alain TAUCH From ks at puzzled.xs4all.nl Thu Mar 2 17:59:39 2006 From: ks at puzzled.xs4all.nl (Patrick) Date: Thu, 02 Mar 2006 18:59:39 +0100 Subject: New to list and question In-Reply-To: <20060302182824.1df2bbda.anaconda@maisondubonheur.com> References: <1141313287.5514.67.camel@guru.puzzled.xs4all.nl> <20060302171230.7669dc41.anaconda@maisondubonheur.com> <1141317175.5514.83.camel@guru.puzzled.xs4all.nl> <20060302182824.1df2bbda.anaconda@maisondubonheur.com> Message-ID: <1141322379.5514.96.camel@guru.puzzled.xs4all.nl> On Thu, 2006-03-02 at 18:28 +0100, Alain TAUCH wrote: [snip] > or even: > Red Hat Enterprise Linux 4: System Administration Guide > Chapter 1. Kickstart Installations > 1.7. Post-installation Script That's the one I should have found myself. Being accustomed to FC I usually just search for the docs on my system and if I can't find it (also with google) then I turn to the ML. That's prolly the reason I totally missed that nice admin guide (and all the other guides they offer). > Don't get me wrong, I'm not used to throw some RTFM just for the fun. Point taken, I should have looked harder for the info. Thanks for taking the time to look it up. Regards, Patrick From Chris.Pierzycki at ge.com Fri Mar 3 17:41:22 2006 From: Chris.Pierzycki at ge.com (Pierzycki, Chris (GE Healthcare, consultant)) Date: Fri, 3 Mar 2006 11:41:22 -0600 Subject: Pkgorder of x86_64 RHEL 4.2 on i386 RHEL 4.2 Message-ID: <65E50D2CD54E9649B6CAA522C3EF65910C9A0F81@MKEMLVEM08.e2k.ad.ge.com> Hi, I'm running into problems while trying to rebuild the package order of somewhat customized RHEL 4.2 x86_64 distribution. While it works fine on a RHEL 4.2 x86_64 installed system, when I try i386 installed system it won't work. I get the following error: Traceback (most recent call last): File "/usr/lib/anaconda-runtime/pkgorder", line 209, in ? pkgOrder.append(hdlist[package].nevra()) File "/usr/lib/anaconda/hdrlist.py", line 422, in __getitem__ raise KeyError, "No such package %s" %(item,) KeyError: 'No such package kernel-smp' The following is the script I am using export PYTHONPATH=/usr/lib/anaconda export PATH=$PATH:/usr/lib/anaconda-runtime genhdlist $1 # Full path required pkgorder $1 x86_64 > /tmp/pkgorder.txt genhdlist --fileorder /tmp/pkgorder.txt $DIST Any help greatly appreciated. From mshuler at rackspace.com Fri Mar 3 17:49:15 2006 From: mshuler at rackspace.com (Michael Shuler) Date: Fri, 03 Mar 2006 11:49:15 -0600 Subject: Pkgorder of x86_64 RHEL 4.2 on i386 RHEL 4.2 In-Reply-To: <65E50D2CD54E9649B6CAA522C3EF65910C9A0F81@MKEMLVEM08.e2k.ad.ge.com> References: <65E50D2CD54E9649B6CAA522C3EF65910C9A0F81@MKEMLVEM08.e2k.ad.ge.com> Message-ID: <4408819B.2090807@rackspace.com> Pierzycki, Chris (GE Healthcare, consultant) wrote: > Traceback (most recent call last): > File "/usr/lib/anaconda-runtime/pkgorder", line 209, in ? > pkgOrder.append(hdlist[package].nevra()) > File "/usr/lib/anaconda/hdrlist.py", line 422, in __getitem__ > raise KeyError, "No such package %s" %(item,) > KeyError: 'No such package kernel-smp' > > The following is the script I am using > > export PYTHONPATH=/usr/lib/anaconda > export PATH=$PATH:/usr/lib/anaconda-runtime > > genhdlist $1 # Full path required > pkgorder $1 x86_64 > /tmp/pkgorder.txt > genhdlist --fileorder /tmp/pkgorder.txt $DIST Our installer runs AS-3_i386 - you need to trick genhdlist ;) Kind Regards, Michael ===== michael at ares: ~ $ cat /mnt/kickstart/redhat/rhas-4-64/RedHat/base/make_comps.sh #!/bin/sh # This file should be put into /kickstart//RedHat/base # and be run on the kickstart server whenever we make a change # to the RPM pool name=$(cd ../../;pwd) # run genhdlist echo "Running genhdlist..." # to build x86_64 tree, we need to trick genhdlist and pkgorder echo x86_64-redhat-linux > /etc/rpm/platform /usr/lib/anaconda-runtime/genhdlist --withnumbers $name export PYTHONPATH=/usr/lib/anaconda /usr/lib/anaconda-runtime/pkgorder $name i386 | /exports/tools/pkgorder_fix.pl $name > $name/RedHat/base/filelist /usr/lib/anaconda-runtime/genhdlist --withnumbers --fileorder $name/RedHat/base/filelist $name # set back the correct /etc/rpm/platform echo i686-redhat-linux > /etc/rpm/platform echo "Done." From forltran at yahoo.com Fri Mar 3 21:37:34 2006 From: forltran at yahoo.com (Lambert Tran) Date: Fri, 3 Mar 2006 13:37:34 -0800 (PST) Subject: Pkgorder of x86_64 RHEL 4.2 on i386 RHEL 4.2 In-Reply-To: <4408819B.2090807@rackspace.com> Message-ID: <20060303213734.73994.qmail@web80231.mail.yahoo.com> It seems like you are missing some package. You try this and make sure all packages are in the RPMS: # cd to-your-RPMS-dir mkdir /tmp/testdb rpm --initdb --dbpath /tmp/testdb rpm --test --dbpath /tmp/testdb -Uvh *.rpm I would always do this first before I try the genhlist-pkgorder-genhlist stuffs. -lt --- Michael Shuler wrote: > Pierzycki, Chris (GE Healthcare, consultant) wrote: > > Traceback (most recent call last): > > File "/usr/lib/anaconda-runtime/pkgorder", line 209, in ? > > pkgOrder.append(hdlist[package].nevra()) > > File "/usr/lib/anaconda/hdrlist.py", line 422, in __getitem__ > > raise KeyError, "No such package %s" %(item,) > > KeyError: 'No such package kernel-smp' > > > > The following is the script I am using > > > > export PYTHONPATH=/usr/lib/anaconda > > export PATH=$PATH:/usr/lib/anaconda-runtime > > > > genhdlist $1 # Full path required > > pkgorder $1 x86_64 > /tmp/pkgorder.txt > > genhdlist --fileorder /tmp/pkgorder.txt $DIST > > Our installer runs AS-3_i386 - you need to trick genhdlist ;) > > Kind Regards, > Michael > > ===== > > michael at ares: ~ $ cat > /mnt/kickstart/redhat/rhas-4-64/RedHat/base/make_comps.sh > #!/bin/sh > > # This file should be put into /kickstart//RedHat/base > # and be run on the kickstart server whenever we make a change > # to the RPM pool > > name=$(cd ../../;pwd) > > # run genhdlist > echo "Running genhdlist..." > > # to build x86_64 tree, we need to trick genhdlist and pkgorder > echo x86_64-redhat-linux > /etc/rpm/platform > > /usr/lib/anaconda-runtime/genhdlist --withnumbers $name > export PYTHONPATH=/usr/lib/anaconda > /usr/lib/anaconda-runtime/pkgorder $name i386 | > /exports/tools/pkgorder_fix.pl $name > $name/RedHat/base/filelist > /usr/lib/anaconda-runtime/genhdlist --withnumbers --fileorder > $name/RedHat/base/filelist $name > > # set back the correct /etc/rpm/platform > echo i686-redhat-linux > /etc/rpm/platform > > echo "Done." > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From Elizabeth.Brosch at thomson.com Tue Mar 7 17:06:27 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Tue, 7 Mar 2006 12:06:27 -0500 Subject: TFTP Open Timeout Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E5614@tshuspaphimbx01.ERF.THOMSON.COM> I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. I see the DHCP server issues DHCPDISCOVER then DHCPOFFER then DHCPREQUEST then DHCPACK Then on the client, you see TFTP attempting to start but it dies with timeout. I have been doing this for quite a while with no issues. The only difference with my configuration now is we have a new subnet. I had issues with DHCP but worked through it when I found the correct configuration for dhcpd.conf. But now I get TFTP timeout. Has anyone encountered this that could share some information with me? Thanks in advance. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 -------------- next part -------------- An HTML attachment was scrubbed... URL: From kevin at ev1servers.net Tue Mar 7 17:09:55 2006 From: kevin at ev1servers.net (Kevin Landreth) Date: Tue, 07 Mar 2006 11:09:55 -0600 Subject: TFTP Open Timeout In-Reply-To: <2213CAA79DA9094AB52517FDADEF07FA021E5614@tshuspaphimbx01.ERF.THOMSON.COM> References: <2213CAA79DA9094AB52517FDADEF07FA021E5614@tshuspaphimbx01.ERF.THOMSON.COM> Message-ID: <440DBE63.5010909@ev1servers.net> TFTP uses UDP, so my first assumption would be a firewall, due to the timeout error. However, I just noticed when you do a getent services tftp it reports 69/tcp, which is wrong, its 69/udp - Kevin L. Elizabeth.Brosch at thomson.com wrote: > I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. > > > > I see the DHCP server issues > > DHCPDISCOVER > > then > > DHCPOFFER > > then > > DHCPREQUEST > > then > > DHCPACK > > > > Then on the client, you see TFTP attempting to start but it dies with > timeout. > > > > I have been doing this for quite a while with no issues. The only > difference with my configuration now is we have a new subnet. I had > issues with DHCP but worked through it when I found the correct > configuration for dhcpd.conf. > > > > But now I get TFTP timeout. > > > > Has anyone encountered this that could share some information with me? > > > > Thanks in advance. > > > > > > Liz Brosch > > TSH System Services > > Philadelphia, PA 19104 > > office: (215) 386-0100 x1144 > > cell: (267) 784-9166 > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > 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: smime.p7s Type: application/x-pkcs7-signature Size: 3141 bytes Desc: S/MIME Cryptographic Signature URL: From Chip.Shabazian at bankofamerica.com Tue Mar 7 18:22:07 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Tue, 07 Mar 2006 10:22:07 -0800 Subject: TFTP Open Timeout Message-ID: We have seen this issue on some switches due to the time it takes for the NIC to negotiate. What version are you using, and is portfast turned on at the switch? You can force the speed and duplex by appending the following to the kernel boot parameters: ksdevice=eth0 eth0_ethtool="autoneg=off speed=100 duplex=full" Adjust as appropriate. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Kevin Landreth Sent: Tuesday, March 07, 2006 9:10 AM To: Discussion list about Kickstart Subject: Re: TFTP Open Timeout TFTP uses UDP, so my first assumption would be a firewall, due to the timeout error. However, I just noticed when you do a getent services tftp it reports 69/tcp, which is wrong, its 69/udp - Kevin L. Elizabeth.Brosch at thomson.com wrote: > I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. > > > > I see the DHCP server issues > > DHCPDISCOVER > > then > > DHCPOFFER > > then > > DHCPREQUEST > > then > > DHCPACK > > > > Then on the client, you see TFTP attempting to start but it dies with > timeout. > > > > I have been doing this for quite a while with no issues. The only > difference with my configuration now is we have a new subnet. I had > issues with DHCP but worked through it when I found the correct > configuration for dhcpd.conf. > > > > But now I get TFTP timeout. > > > > Has anyone encountered this that could share some information with me? > > > > Thanks in advance. > > > > > > Liz Brosch > > TSH System Services > > Philadelphia, PA 19104 > > office: (215) 386-0100 x1144 > > cell: (267) 784-9166 > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From Joseph.Cipolla at hsn.net Tue Mar 7 18:28:14 2006 From: Joseph.Cipolla at hsn.net (Cipolla, Joseph) Date: Tue, 7 Mar 2006 13:28:14 -0500 Subject: TFTP Open Timeout Message-ID: Isn't there a timeout value that you can also set in the kernel boot params to give it more time to establish a connection and DHCP address? I remember needing it a while back, but unfortunately it eludes me. Does anyone happen to know what that is? Thanks. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Shabazian, Chip Sent: Tuesday, March 07, 2006 1:22 PM To: kevin at ev1servers.net; Discussion list about Kickstart Subject: RE: TFTP Open Timeout We have seen this issue on some switches due to the time it takes for the NIC to negotiate. What version are you using, and is portfast turned on at the switch? You can force the speed and duplex by appending the following to the kernel boot parameters: ksdevice=eth0 eth0_ethtool="autoneg=off speed=100 duplex=full" Adjust as appropriate. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Kevin Landreth Sent: Tuesday, March 07, 2006 9:10 AM To: Discussion list about Kickstart Subject: Re: TFTP Open Timeout TFTP uses UDP, so my first assumption would be a firewall, due to the timeout error. However, I just noticed when you do a getent services tftp it reports 69/tcp, which is wrong, its 69/udp - Kevin L. Elizabeth.Brosch at thomson.com wrote: > I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. > > > > I see the DHCP server issues > > DHCPDISCOVER > > then > > DHCPOFFER > > then > > DHCPREQUEST > > then > > DHCPACK > > > > Then on the client, you see TFTP attempting to start but it dies with > timeout. > > > > I have been doing this for quite a while with no issues. The only > difference with my configuration now is we have a new subnet. I had > issues with DHCP but worked through it when I found the correct > configuration for dhcpd.conf. > > > > But now I get TFTP timeout. > > > > Has anyone encountered this that could share some information with me? > > > > Thanks in advance. > > > > > > Liz Brosch > > TSH System Services > > Philadelphia, PA 19104 > > office: (215) 386-0100 x1144 > > cell: (267) 784-9166 > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > 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 Chip.Shabazian at bankofamerica.com Tue Mar 7 18:29:57 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Tue, 07 Mar 2006 10:29:57 -0800 Subject: TFTP Open Timeout Message-ID: If ANYONE knows about this switch, PLEASE forward to the list. I've been looking for that switch for a while, and have been told that it doesn't exist. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Cipolla, Joseph Sent: Tuesday, March 07, 2006 10:28 AM To: Discussion list about Kickstart Subject: RE: TFTP Open Timeout Isn't there a timeout value that you can also set in the kernel boot params to give it more time to establish a connection and DHCP address? I remember needing it a while back, but unfortunately it eludes me. Does anyone happen to know what that is? Thanks. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Shabazian, Chip Sent: Tuesday, March 07, 2006 1:22 PM To: kevin at ev1servers.net; Discussion list about Kickstart Subject: RE: TFTP Open Timeout We have seen this issue on some switches due to the time it takes for the NIC to negotiate. What version are you using, and is portfast turned on at the switch? You can force the speed and duplex by appending the following to the kernel boot parameters: ksdevice=eth0 eth0_ethtool="autoneg=off speed=100 duplex=full" Adjust as appropriate. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Kevin Landreth Sent: Tuesday, March 07, 2006 9:10 AM To: Discussion list about Kickstart Subject: Re: TFTP Open Timeout TFTP uses UDP, so my first assumption would be a firewall, due to the timeout error. However, I just noticed when you do a getent services tftp it reports 69/tcp, which is wrong, its 69/udp - Kevin L. Elizabeth.Brosch at thomson.com wrote: > I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. > > > > I see the DHCP server issues > > DHCPDISCOVER > > then > > DHCPOFFER > > then > > DHCPREQUEST > > then > > DHCPACK > > > > Then on the client, you see TFTP attempting to start but it dies with > timeout. > > > > I have been doing this for quite a while with no issues. The only > difference with my configuration now is we have a new subnet. I had > issues with DHCP but worked through it when I found the correct > configuration for dhcpd.conf. > > > > But now I get TFTP timeout. > > > > Has anyone encountered this that could share some information with me? > > > > Thanks in advance. > > > > > > Liz Brosch > > TSH System Services > > Philadelphia, PA 19104 > > office: (215) 386-0100 x1144 > > cell: (267) 784-9166 > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > 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 Elizabeth.Brosch at thomson.com Tue Mar 7 18:34:54 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Tue, 7 Mar 2006 13:34:54 -0500 Subject: TFTP Open Timeout Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E5616@tshuspaphimbx01.ERF.THOMSON.COM> I have sent a request to Network Services to see if portfast is enabled. I am almost positive that it is. I am using the following tftp: tftp-server-0.39-1 on RHEL AS v4 Update 1 I'll try the boot parameters you listed below. I was able to get slightly more information in my messages file. After the DHCPACK, I see the following: in.tftpd[26784]: RRQ from 10.xx.xx.xx filename pxelinux.0 in.tftpd[26784]: tftp: client does not accept options in.tftpd[26784]: RRQ from 10.xx.xx.xx filename pxelinux.0 Then I see on the client TFTP open timeout. Thanks for the response. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Shabazian, Chip Sent: Tuesday, March 07, 2006 1:22 PM To: kevin at ev1servers.net; Discussion list about Kickstart Subject: RE: TFTP Open Timeout We have seen this issue on some switches due to the time it takes for the NIC to negotiate. What version are you using, and is portfast turned on at the switch? You can force the speed and duplex by appending the following to the kernel boot parameters: ksdevice=eth0 eth0_ethtool="autoneg=off speed=100 duplex=full" Adjust as appropriate. From hsuttong at ariba.com Tue Mar 7 19:02:32 2006 From: hsuttong at ariba.com (Hugh Sutton-Gee) Date: Tue, 7 Mar 2006 11:02:32 -0800 Subject: TFTP Open Timeout References: Message-ID: <40D30D99949893468C5BD4686B1A6C1D01A30642@us-hqmail2.ariba.com> Which switch? The one with dhcpd delays due to port-fast issues? I had this problem on cisco c4003's. Once I enabled port-fast everything was okay. --Hugh --- --Original Message----- From: kickstart-list-bounces at redhat.com on behalf of Shabazian, Chip Sent: Tue 3/7/2006 10:29 AM To: Discussion list about Kickstart Subject: RE: TFTP Open Timeout If ANYONE knows about this switch, PLEASE forward to the list. I've been looking for that switch for a while, and have been told that it doesn't exist. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Cipolla, Joseph Sent: Tuesday, March 07, 2006 10:28 AM To: Discussion list about Kickstart Subject: RE: TFTP Open Timeout Isn't there a timeout value that you can also set in the kernel boot params to give it more time to establish a connection and DHCP address? I remember needing it a while back, but unfortunately it eludes me. Does anyone happen to know what that is? Thanks. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Shabazian, Chip Sent: Tuesday, March 07, 2006 1:22 PM To: kevin at ev1servers.net; Discussion list about Kickstart Subject: RE: TFTP Open Timeout We have seen this issue on some switches due to the time it takes for the NIC to negotiate. What version are you using, and is portfast turned on at the switch? You can force the speed and duplex by appending the following to the kernel boot parameters: ksdevice=eth0 eth0_ethtool="autoneg=off speed=100 duplex=full" Adjust as appropriate. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Kevin Landreth Sent: Tuesday, March 07, 2006 9:10 AM To: Discussion list about Kickstart Subject: Re: TFTP Open Timeout TFTP uses UDP, so my first assumption would be a firewall, due to the timeout error. However, I just noticed when you do a getent services tftp it reports 69/tcp, which is wrong, its 69/udp - Kevin L. Elizabeth.Brosch at thomson.com wrote: > I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. > > > > I see the DHCP server issues > > DHCPDISCOVER > > then > > DHCPOFFER > > then > > DHCPREQUEST > > then > > DHCPACK > > > > Then on the client, you see TFTP attempting to start but it dies with > timeout. > > > > I have been doing this for quite a while with no issues. The only > difference with my configuration now is we have a new subnet. I had > issues with DHCP but worked through it when I found the correct > configuration for dhcpd.conf. > > > > But now I get TFTP timeout. > > > > Has anyone encountered this that could share some information with me? > > > > Thanks in advance. > > > > > > Liz Brosch > > TSH System Services > > Philadelphia, PA 19104 > > office: (215) 386-0100 x1144 > > cell: (267) 784-9166 > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > 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 _______________________________________________ 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 Elizabeth.Brosch at thomson.com Tue Mar 7 19:08:48 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Tue, 7 Mar 2006 14:08:48 -0500 Subject: TFTP Open Timeout Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E5617@tshuspaphimbx01.ERF.THOMSON.COM> I am not sure what type of switch we have in this data center. But I am almost certain that port-fast is enabled. So that is why I am confused about the TCP timeout I am receiving. Thanks. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 ________________________________ From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Hugh Sutton-Gee Sent: Tuesday, March 07, 2006 2:03 PM To: Discussion list about Kickstart Subject: RE: TFTP Open Timeout Which switch? The one with dhcpd delays due to port-fast issues? I had this problem on cisco c4003's. Once I enabled port-fast everything was okay. --Hugh --- --Original Message----- From: kickstart-list-bounces at redhat.com on behalf of Shabazian, Chip Sent: Tue 3/7/2006 10:29 AM To: Discussion list about Kickstart Subject: RE: TFTP Open Timeout If ANYONE knows about this switch, PLEASE forward to the list. I've been looking for that switch for a while, and have been told that it doesn't exist. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Cipolla, Joseph Sent: Tuesday, March 07, 2006 10:28 AM To: Discussion list about Kickstart Subject: RE: TFTP Open Timeout Isn't there a timeout value that you can also set in the kernel boot params to give it more time to establish a connection and DHCP address? I remember needing it a while back, but unfortunately it eludes me. Does anyone happen to know what that is? Thanks. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Shabazian, Chip Sent: Tuesday, March 07, 2006 1:22 PM To: kevin at ev1servers.net; Discussion list about Kickstart Subject: RE: TFTP Open Timeout We have seen this issue on some switches due to the time it takes for the NIC to negotiate. What version are you using, and is portfast turned on at the switch? You can force the speed and duplex by appending the following to the kernel boot parameters: ksdevice=eth0 eth0_ethtool="autoneg=off speed=100 duplex=full" Adjust as appropriate. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Kevin Landreth Sent: Tuesday, March 07, 2006 9:10 AM To: Discussion list about Kickstart Subject: Re: TFTP Open Timeout TFTP uses UDP, so my first assumption would be a firewall, due to the timeout error. However, I just noticed when you do a getent services tftp it reports 69/tcp, which is wrong, its 69/udp - Kevin L. Elizabeth.Brosch at thomson.com wrote: > I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. > > > > I see the DHCP server issues > > DHCPDISCOVER > > then > > DHCPOFFER > > then > > DHCPREQUEST > > then > > DHCPACK > > > > Then on the client, you see TFTP attempting to start but it dies with > timeout. > > > > I have been doing this for quite a while with no issues. The only > difference with my configuration now is we have a new subnet. I had > issues with DHCP but worked through it when I found the correct > configuration for dhcpd.conf. > > > > But now I get TFTP timeout. > > > > Has anyone encountered this that could share some information with me? > > > > Thanks in advance. > > > > > > Liz Brosch > > TSH System Services > > Philadelphia, PA 19104 > > office: (215) 386-0100 x1144 > > cell: (267) 784-9166 > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > 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 _______________________________________________ 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 Elizabeth.Brosch at thomson.com Tue Mar 7 20:24:14 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Tue, 7 Mar 2006 15:24:14 -0500 Subject: TFTP Open Timeout Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E561F@tshuspaphimbx01.ERF.THOMSON.COM> Thanks for the suggestion Chip but this didn't make any difference. I see traffic between my dhcpd/Kickstart server and client by displaying the output from tcpdump. I can't seem to get around this issue. It appears it is some sort of timeout either with dhcp or tftp or possibly both. Like I stated earlier, I have this new subnet that I am attempting to Kickstart with no luck. Thanks anyway. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Shabazian, Chip Sent: Tuesday, March 07, 2006 1:22 PM To: kevin at ev1servers.net; Discussion list about Kickstart Subject: RE: TFTP Open Timeout We have seen this issue on some switches due to the time it takes for the NIC to negotiate. What version are you using, and is portfast turned on at the switch? You can force the speed and duplex by appending the following to the kernel boot parameters: ksdevice=eth0 eth0_ethtool="autoneg=off speed=100 duplex=full" Adjust as appropriate. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Kevin Landreth Sent: Tuesday, March 07, 2006 9:10 AM To: Discussion list about Kickstart Subject: Re: TFTP Open Timeout TFTP uses UDP, so my first assumption would be a firewall, due to the timeout error. However, I just noticed when you do a getent services tftp it reports 69/tcp, which is wrong, its 69/udp - Kevin L. Elizabeth.Brosch at thomson.com wrote: > I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. > > > > I see the DHCP server issues > > DHCPDISCOVER > > then > > DHCPOFFER > > then > > DHCPREQUEST > > then > > DHCPACK > > > > Then on the client, you see TFTP attempting to start but it dies with > timeout. > > > > I have been doing this for quite a while with no issues. The only > difference with my configuration now is we have a new subnet. I had > issues with DHCP but worked through it when I found the correct > configuration for dhcpd.conf. > > > > But now I get TFTP timeout. > > > > Has anyone encountered this that could share some information with me? > > > > Thanks in advance. > > > > > > Liz Brosch > > TSH System Services > > Philadelphia, PA 19104 > > office: (215) 386-0100 x1144 > > cell: (267) 784-9166 > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > 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 _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From Chip.Shabazian at bankofamerica.com Tue Mar 7 20:44:46 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Tue, 07 Mar 2006 12:44:46 -0800 Subject: TFTP Open Timeout Message-ID: The key is WHEN do you see the traffic. The NIC gets an IP address for PXEBoot, gets it's info needed via tftp, then the NIC is RESTARTED, the driver loaded, the request sent again. It's this second request where we have the timeout issue. Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Elizabeth.Brosch at thomson.com Sent: Tuesday, March 07, 2006 12:24 PM To: kickstart-list at redhat.com Subject: RE: TFTP Open Timeout Thanks for the suggestion Chip but this didn't make any difference. I see traffic between my dhcpd/Kickstart server and client by displaying the output from tcpdump. I can't seem to get around this issue. It appears it is some sort of timeout either with dhcp or tftp or possibly both. Like I stated earlier, I have this new subnet that I am attempting to Kickstart with no luck. Thanks anyway. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Shabazian, Chip Sent: Tuesday, March 07, 2006 1:22 PM To: kevin at ev1servers.net; Discussion list about Kickstart Subject: RE: TFTP Open Timeout We have seen this issue on some switches due to the time it takes for the NIC to negotiate. What version are you using, and is portfast turned on at the switch? You can force the speed and duplex by appending the following to the kernel boot parameters: ksdevice=eth0 eth0_ethtool="autoneg=off speed=100 duplex=full" Adjust as appropriate. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Kevin Landreth Sent: Tuesday, March 07, 2006 9:10 AM To: Discussion list about Kickstart Subject: Re: TFTP Open Timeout TFTP uses UDP, so my first assumption would be a firewall, due to the timeout error. However, I just noticed when you do a getent services tftp it reports 69/tcp, which is wrong, its 69/udp - Kevin L. Elizabeth.Brosch at thomson.com wrote: > I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. > > > > I see the DHCP server issues > > DHCPDISCOVER > > then > > DHCPOFFER > > then > > DHCPREQUEST > > then > > DHCPACK > > > > Then on the client, you see TFTP attempting to start but it dies with > timeout. > > > > I have been doing this for quite a while with no issues. The only > difference with my configuration now is we have a new subnet. I had > issues with DHCP but worked through it when I found the correct > configuration for dhcpd.conf. > > > > But now I get TFTP timeout. > > > > Has anyone encountered this that could share some information with me? > > > > Thanks in advance. > > > > > > Liz Brosch > > TSH System Services > > Philadelphia, PA 19104 > > office: (215) 386-0100 x1144 > > cell: (267) 784-9166 > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > 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 _______________________________________________ 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 kevin at ev1servers.net Tue Mar 7 20:51:29 2006 From: kevin at ev1servers.net (Kevin Landreth) Date: Tue, 07 Mar 2006 14:51:29 -0600 Subject: TFTP Open Timeout In-Reply-To: <2213CAA79DA9094AB52517FDADEF07FA021E561F@tshuspaphimbx01.ERF.THOMSON.COM> References: <2213CAA79DA9094AB52517FDADEF07FA021E561F@tshuspaphimbx01.ERF.THOMSON.COM> Message-ID: <440DF251.1000108@ev1servers.net> With your logs, I see that the actual requests are getting to the server. Is TFTP configured to be in a chroot? Is the filename option in dhcpd.conf take this into account? It's like its requesting pxelinux.0 but not finding it. I did have a similar issue as chip, but it was after the RHE kernel was booting and it loaded the modules for the NIC. I have a very similiar setup using all native RHEL ########################################## [root at pxe ~]# cat /etc/redhat-release Red Hat Enterprise Linux ES release 4 (Nahant Update 2) [root at pxe ~]# rpm -q tftp-server tftp tftp-server-0.39-1 tftp-0.39-1 dhcpd.conf ######################################### deny unknown-clients; not authoritative; ddns-update-style ad-hoc; option domain-name "redhat.ev1servers.net"; option domain-name-servers 192.168.0.9, 192.168.0.10; option subnet-mask 255.255.255.0; allow bootp; allow booting; option ip-forwarding false; # No IP forwarding option mask-supplier false; # Don't respond to ICMP Mask req subnet 192.168.1.0 netmask 255.255.255.0 { option routers 192.168.1.1; option domain-name-servers 192.168.1.1; } group { next-server 192.168.1.1; # name of the TFTP server filename "/rhel4_u2/pxe/pxelinux.bin"; # name of the bootloader host pxe_bootp6 { hardware ethernet 00:c0:a8:fd:47:d5; fixed-address 192.168.1.6; } - Kevin L. Elizabeth.Brosch at thomson.com wrote: > Thanks for the suggestion Chip but this didn't make any difference. > > I see traffic between my dhcpd/Kickstart server and client by displaying > the output from tcpdump. > > I can't seem to get around this issue. It appears it is some sort of > timeout either with dhcp or tftp or possibly both. > > Like I stated earlier, I have this new subnet that I am attempting to > Kickstart with no luck. > > Thanks anyway. > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Shabazian, Chip > Sent: Tuesday, March 07, 2006 1:22 PM > To: kevin at ev1servers.net; Discussion list about Kickstart > Subject: RE: TFTP Open Timeout > > > We have seen this issue on some switches due to the time it takes for > the NIC to negotiate. What version are you using, and is portfast > turned on at the switch? > > You can force the speed and duplex by appending the following to the > kernel boot parameters: > > ksdevice=eth0 eth0_ethtool="autoneg=off speed=100 duplex=full" > > Adjust as appropriate. > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Kevin Landreth > Sent: Tuesday, March 07, 2006 9:10 AM > To: Discussion list about Kickstart > Subject: Re: TFTP Open Timeout > > TFTP uses UDP, so my first assumption would be a firewall, due to the > timeout > error. However, I just noticed when you do a getent services tftp it > reports > 69/tcp, which is wrong, its 69/udp > > - Kevin L. > > Elizabeth.Brosch at thomson.com wrote: > >>I am having problems with TFTP. I receive PXE-E32: TFTP open timeout. >> >> >> >>I see the DHCP server issues >> >>DHCPDISCOVER >> >>then >> >>DHCPOFFER >> >>then >> >>DHCPREQUEST >> >>then >> >>DHCPACK >> >> >> >>Then on the client, you see TFTP attempting to start but it dies with >>timeout. >> >> >> >>I have been doing this for quite a while with no issues. The only >>difference with my configuration now is we have a new subnet. I had >>issues with DHCP but worked through it when I found the correct >>configuration for dhcpd.conf. >> >> >> >>But now I get TFTP timeout. >> >> >> >>Has anyone encountered this that could share some information with me? >> >> >> >>Thanks in advance. >> >> >> >> >> >>Liz Brosch >> >>TSH System Services >> >>Philadelphia, PA 19104 >> >>office: (215) 386-0100 x1144 >> >>cell: (267) 784-9166 >> >> >> >> >> > > ------------------------------------------------------------------------ > >>_______________________________________________ >>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 > > _______________________________________________ > 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 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 3141 bytes Desc: S/MIME Cryptographic Signature URL: From Elizabeth.Brosch at thomson.com Tue Mar 7 20:52:10 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Tue, 7 Mar 2006 15:52:10 -0500 Subject: TFTP Open Timeout Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E5624@tshuspaphimbx01.ERF.THOMSON.COM> I see the traffic for the first DHCP DISCOVER/OFFER/REQUEST/ACK. Then I see traffic back and forth a few times after I see in.tftpd called. But then the traffic dies and the client shows TFTP open timeout. I have put in a call to the Network team. They need to double check the configuration on the switch. There has to be something going on there. Like I stated earlier, they have confirmed for me in the past that port-fast enabled was configured but maybe something changed! Thanks for the quick reply. I appreciate it. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 From Chris.Pierzycki at ge.com Tue Mar 7 22:20:11 2006 From: Chris.Pierzycki at ge.com (Pierzycki, Chris (GE Healthcare, consultant)) Date: Tue, 7 Mar 2006 16:20:11 -0600 Subject: Pkgorder of x86_64 RHEL 4.2 on i386 RHEL 4.2 Message-ID: <65E50D2CD54E9649B6CAA522C3EF65910CBCFEFF@MKEMLVEM08.e2k.ad.ge.com> I checked all the packages. The error complains about kernel-smp package ... which I know is there. It looks like Michael was right. But now I am trying to build a 32 bit distro on 64 bit platform and I am running to different problem. Everything goes well with the creation process but when the media loads, it says it doesn't support the long mode... please use 32 bit system. Joy. Chris -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Lambert Tran Sent: Friday, March 03, 2006 3:38 PM To: Discussion list about Kickstart Subject: Re: Pkgorder of x86_64 RHEL 4.2 on i386 RHEL 4.2 It seems like you are missing some package. You try this and make sure all packages are in the RPMS: # cd to-your-RPMS-dir mkdir /tmp/testdb rpm --initdb --dbpath /tmp/testdb rpm --test --dbpath /tmp/testdb -Uvh *.rpm I would always do this first before I try the genhlist-pkgorder-genhlist stuffs. -lt --- Michael Shuler wrote: > Pierzycki, Chris (GE Healthcare, consultant) wrote: > > Traceback (most recent call last): > > File "/usr/lib/anaconda-runtime/pkgorder", line 209, in ? > > pkgOrder.append(hdlist[package].nevra()) > > File "/usr/lib/anaconda/hdrlist.py", line 422, in __getitem__ > > raise KeyError, "No such package %s" %(item,) > > KeyError: 'No such package kernel-smp' > > > > The following is the script I am using > > > > export PYTHONPATH=/usr/lib/anaconda > > export PATH=$PATH:/usr/lib/anaconda-runtime > > > > genhdlist $1 # Full path required > > pkgorder $1 x86_64 > /tmp/pkgorder.txt > > genhdlist --fileorder /tmp/pkgorder.txt $DIST > > Our installer runs AS-3_i386 - you need to trick genhdlist ;) > > Kind Regards, > Michael > > ===== > > michael at ares: ~ $ cat > /mnt/kickstart/redhat/rhas-4-64/RedHat/base/make_comps.sh > #!/bin/sh > > # This file should be put into /kickstart//RedHat/base > # and be run on the kickstart server whenever we make a change > # to the RPM pool > > name=$(cd ../../;pwd) > > # run genhdlist > echo "Running genhdlist..." > > # to build x86_64 tree, we need to trick genhdlist and pkgorder > echo x86_64-redhat-linux > /etc/rpm/platform > > /usr/lib/anaconda-runtime/genhdlist --withnumbers $name > export PYTHONPATH=/usr/lib/anaconda > /usr/lib/anaconda-runtime/pkgorder $name i386 | > /exports/tools/pkgorder_fix.pl $name > $name/RedHat/base/filelist > /usr/lib/anaconda-runtime/genhdlist --withnumbers --fileorder > $name/RedHat/base/filelist $name > > # set back the correct /etc/rpm/platform > echo i686-redhat-linux > /etc/rpm/platform > > echo "Done." > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From secroft at micron.com Tue Mar 7 22:50:03 2006 From: secroft at micron.com (secroft at micron.com) Date: Tue, 7 Mar 2006 15:50:03 -0700 Subject: TFTP Open Timeout Message-ID: <029E352135E99D49B4B14E42F95D8C50021D3F61@ntxboimbx12.micron.com> I just went through this with setting up a new pxe/tftp server and it all came down to the location of the pxelinux.0 file, but if you've been using it for a while, your issue sounds more like a possible routing problem for pxe to the subnet possibly? Have you confirmed that your pxe/dhcp/tftp server is working on a known configuration that it previously worked on? It'd be nice to see a trace of the traffic you're seeing. Scott -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Elizabeth.Brosch at thomson.com Sent: Wednesday, March 08, 2006 4:52 AM To: kickstart-list at redhat.com Subject: RE: TFTP Open Timeout I see the traffic for the first DHCP DISCOVER/OFFER/REQUEST/ACK. Then I see traffic back and forth a few times after I see in.tftpd called. But then the traffic dies and the client shows TFTP open timeout. I have put in a call to the Network team. They need to double check the configuration on the switch. There has to be something going on there. Like I stated earlier, they have confirmed for me in the past that port-fast enabled was configured but maybe something changed! Thanks for the quick reply. I appreciate it. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From mshuler at rackspace.com Tue Mar 7 23:22:47 2006 From: mshuler at rackspace.com (Michael Shuler) Date: Tue, 07 Mar 2006 17:22:47 -0600 Subject: Pkgorder of x86_64 RHEL 4.2 on i386 RHEL 4.2 In-Reply-To: <65E50D2CD54E9649B6CAA522C3EF65910CBCFEFF@MKEMLVEM08.e2k.ad.ge.com> References: <65E50D2CD54E9649B6CAA522C3EF65910CBCFEFF@MKEMLVEM08.e2k.ad.ge.com> Message-ID: <440E15C7.5060602@rackspace.com> That error actually looks like you are serving up a 64-bit installer kernel/initrd to a 32-bit client. I have not tried running one of my installers from a 64-bit platform. Try replacing the installer kernel/initrd again with the correct 32-bit images that you are trying to install, just to be sure. Is this a PXE installer you are working on? Kind Regards, Michael Pierzycki, Chris (GE Healthcare, consultant) wrote: > I checked all the packages. The error complains about kernel-smp > package ... which I know is there. > > It looks like Michael was right. But now I am trying to build a 32 bit > distro on 64 bit platform and I am running to different problem. > Everything goes well with the creation process but when the media loads, > it says it doesn't support the long mode... please use 32 bit system. > Joy. > > Chris > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Lambert Tran > Sent: Friday, March 03, 2006 3:38 PM > To: Discussion list about Kickstart > Subject: Re: Pkgorder of x86_64 RHEL 4.2 on i386 RHEL 4.2 > > It seems like you are missing some package. > You try this and make sure all packages are in the RPMS: > > # cd to-your-RPMS-dir > mkdir /tmp/testdb > rpm --initdb --dbpath /tmp/testdb > rpm --test --dbpath /tmp/testdb -Uvh *.rpm > > I would always do this first before I try the genhlist-pkgorder-genhlist > stuffs. > > -lt > > --- Michael Shuler wrote: > > >>Pierzycki, Chris (GE Healthcare, consultant) wrote: >> >>>Traceback (most recent call last): >>> File "/usr/lib/anaconda-runtime/pkgorder", line 209, in ? >>> pkgOrder.append(hdlist[package].nevra()) >>> File "/usr/lib/anaconda/hdrlist.py", line 422, in __getitem__ >>> raise KeyError, "No such package %s" %(item,) >>>KeyError: 'No such package kernel-smp' >>> >>>The following is the script I am using >>> >>>export PYTHONPATH=/usr/lib/anaconda >>>export PATH=$PATH:/usr/lib/anaconda-runtime >>> >>>genhdlist $1 # Full path required >>>pkgorder $1 x86_64 > /tmp/pkgorder.txt >>>genhdlist --fileorder /tmp/pkgorder.txt $DIST >> >>Our installer runs AS-3_i386 - you need to trick genhdlist ;) >> >>Kind Regards, >>Michael >> >>===== >> >>michael at ares: ~ $ cat >>/mnt/kickstart/redhat/rhas-4-64/RedHat/base/make_comps.sh >>#!/bin/sh >> >># This file should be put into /kickstart//RedHat/base >># and be run on the kickstart server whenever we make a change >># to the RPM pool >> >>name=$(cd ../../;pwd) >> >># run genhdlist >>echo "Running genhdlist..." >> >># to build x86_64 tree, we need to trick genhdlist and pkgorder >>echo x86_64-redhat-linux > /etc/rpm/platform >> >>/usr/lib/anaconda-runtime/genhdlist --withnumbers $name >>export PYTHONPATH=/usr/lib/anaconda >>/usr/lib/anaconda-runtime/pkgorder $name i386 | >>/exports/tools/pkgorder_fix.pl $name > $name/RedHat/base/filelist >>/usr/lib/anaconda-runtime/genhdlist --withnumbers --fileorder >>$name/RedHat/base/filelist $name >> >># set back the correct /etc/rpm/platform >>echo i686-redhat-linux > /etc/rpm/platform >> >>echo "Done." >> >>_______________________________________________ >>Kickstart-list mailing list >>Kickstart-list at redhat.com >>https://www.redhat.com/mailman/listinfo/kickstart-list >> > > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > > _______________________________________________ > 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 Elizabeth.Brosch at thomson.com Wed Mar 8 03:13:39 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Tue, 7 Mar 2006 22:13:39 -0500 Subject: TFTP Open Timeout Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E5634@tshuspaphimbx01.ERF.THOMSON.COM> Thanks for reply. My pxe/dhcp/kickstart has been working and I was able to install 12 hosts. But to do sure everything still works as it should I am going to install to host on same subnet. If that works then I will continue to investigate problems by looking into switch, routes. Thanks -------------------------- Sent from my BlackBerry Wireless Handheld -----Original Message----- From: kickstart-list-bounces at redhat.com To: kickstart-list at redhat.com Sent: Tue Mar 07 17:50:03 2006 Subject: RE: TFTP Open Timeout I just went through this with setting up a new pxe/tftp server and it all came down to the location of the pxelinux.0 file, but if you've been using it for a while, your issue sounds more like a possible routing problem for pxe to the subnet possibly? Have you confirmed that your pxe/dhcp/tftp server is working on a known configuration that it previously worked on? It'd be nice to see a trace of the traffic you're seeing. Scott -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Elizabeth.Brosch at thomson.com Sent: Wednesday, March 08, 2006 4:52 AM To: kickstart-list at redhat.com Subject: RE: TFTP Open Timeout I see the traffic for the first DHCP DISCOVER/OFFER/REQUEST/ACK. Then I see traffic back and forth a few times after I see in.tftpd called. But then the traffic dies and the client shows TFTP open timeout. I have put in a call to the Network team. They need to double check the configuration on the switch. There has to be something going on there. Like I stated earlier, they have confirmed for me in the past that port-fast enabled was configured but maybe something changed! Thanks for the quick reply. I appreciate it. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 _______________________________________________ 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 phr at doc.ic.ac.uk Wed Mar 8 12:11:49 2006 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 8 Mar 2006 12:11:49 +0000 (GMT) Subject: TFTP Open Timeout In-Reply-To: References: Message-ID: On Tue, 7 Mar 2006, Shabazian, Chip wrote: > If ANYONE knows about this switch, PLEASE forward to the list. I've > been looking for that switch for a while, and have been told that it > doesn't exist. I think you mean "linksleep", which appeared around FC3 IIRC. See this previous list posting for more info: https://www.redhat.com/archives/kickstart-list/2004-June/msg00042.html [snip 132 lines] I checked the current FC4 packages, and it's still undocumented! Time for a patch, methinks. Cheers, Phil From ks at puzzled.xs4all.nl Wed Mar 8 17:00:14 2006 From: ks at puzzled.xs4all.nl (Patrick) Date: Wed, 08 Mar 2006 18:00:14 +0100 Subject: Preserve $HOME in %post? Message-ID: <1141837214.5985.7.camel@guru.puzzled.xs4all.nl> Hi all, I noticed that $HOME becomes /tmp when I use it in the %post section of my ks file. Is there a way to preserve "$HOME"? What happens now is something like this: cat <>/home/test/.rpmmacros %_topdir %(echo $HOME)/redhat EOF chown test.test /home/test/.rpmmacros After the install & reboot I see: $ cat /home/test/.rpmmacros %_topdir %(echo /tmp)/redhat And I would like it to say: %_topdir %(echo $HOME)/redhat Thanks for any suggestions. Regards, Patrick From Chip.Shabazian at bankofamerica.com Wed Mar 8 19:06:06 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 08 Mar 2006 11:06:06 -0800 Subject: Preserve $HOME in %post? Message-ID: Just escape out the $ cat <>/home/test/.rpmmacros %_topdir %(echo \$HOME)/redhat EOF chown test.test /home/test/.rpmmacros -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Patrick Sent: Wednesday, March 08, 2006 9:00 AM To: Discussion list about Kickstart Subject: Preserve $HOME in %post? Hi all, I noticed that $HOME becomes /tmp when I use it in the %post section of my ks file. Is there a way to preserve "$HOME"? What happens now is something like this: cat <>/home/test/.rpmmacros %_topdir %(echo $HOME)/redhat EOF chown test.test /home/test/.rpmmacros After the install & reboot I see: $ cat /home/test/.rpmmacros %_topdir %(echo /tmp)/redhat And I would like it to say: %_topdir %(echo $HOME)/redhat Thanks for any suggestions. Regards, Patrick _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From gkapitany at rogers.com Wed Mar 8 19:26:16 2006 From: gkapitany at rogers.com (GABRIEL KAPITANY) Date: Wed, 8 Mar 2006 14:26:16 -0500 (EST) Subject: pxe on IBM x366 Message-ID: <20060308192616.60524.qmail@web88102.mail.re2.yahoo.com> Hi guys, I have a problem trying to use pxe and nfs install on an IBM x366 box. I'd like to mention that I have this configuration working on a different hardware. The server boots using pxe but when in tries to do an nfs mount it times out. Here are the messages on the client (Alt+F3): - eth1 has link, using it - doing kickstart... setting it up - waiting for link... - 2 seconds - reverse name lookup failed - url is 192.168.1.10:/images/AS3.0U6-x366/kslapp.cfg - file location: nfs://192.168.1.10:/images/AS3.0U6-x366/kslapp.cfg - failed to mount nfs source - trying to mount device hda At this point I'm prompted for the type of install and if I choose nfs then it works fine but I'm not able to use kickstart.... On the pxe server in the logs I can see: tftp: client does not accept options The pxe configuration for this box is: default RHEL-AS3.0U6-x366 label RHEL-AS3.0U6-x366 kernel RHEL-AS3.0U6-x366/vmlinuz append initrd=RHEL-AS3.0U6-x366/initrd.img ramdisk_size=7000 ks=nfs:192.168.1.10:/images/AS3.0U6-x366/kslapp.cfg ksdevice =link ipappend 1 Any help would be appreciated, Thx, Gabriel From dan at id-confirm.com Wed Mar 8 20:11:00 2006 From: dan at id-confirm.com (Dan Trainor) Date: Wed, 08 Mar 2006 13:11:00 -0700 Subject: Small image based on comps.xml alone? Message-ID: <440F3A54.7040906@id-confirm.com> Hello, all - I've been working on a few different items lately, one of which I'm working on today. I'd like to build images which contain packages which are either directly or indirectly referenced via comps.xml, being dependencies of, or explicitly named packages, in comps.xml. The reason I'm doing this is to first allow myself some degree of flexibility when adding my own packages to comps.xml and having them being installed, and also to slim down an install to under 1 complete CD. I know that we've talked about comps.xml on the list in the past, but I don't think we've quite covered this aspect of things. Thanks again for the time - it is greatly appreciated. Thanks -dant From Chip.Shabazian at bankofamerica.com Wed Mar 8 20:36:03 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 08 Mar 2006 12:36:03 -0800 Subject: Small image based on comps.xml alone? Message-ID: What kind of information are you looking for? I haven't had time to work on this yet myself, but I did take the standard comps.xml file, remove all of the language entries, and it suddenly made a LOT more sense to me. Here are some of the links I've been saving for when I have time to do this: http://people.redhat.com/rkeech/custom-distro.txt http://fedora.redhat.com/projects/anaconda-installer/comps.html http://www.linuxjournal.com/article/6473 http://www.linuxdevcenter.com/pub/a/linux/2005/02/17/kickstart_updates.h tml?page=2 http://www.linuxdevcenter.com/pub/a/linux/2004/11/04/advanced_kickstart. html?page=2 I would like to create a comps.xml which TRULY builds a minimal system (I have a build down to about 100 packages), then allow me to add onto that as appropriate. Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Dan Trainor Sent: Wednesday, March 08, 2006 12:11 PM To: kickstart-list at redhat.com Subject: Small image based on comps.xml alone? Hello, all - I've been working on a few different items lately, one of which I'm working on today. I'd like to build images which contain packages which are either directly or indirectly referenced via comps.xml, being dependencies of, or explicitly named packages, in comps.xml. The reason I'm doing this is to first allow myself some degree of flexibility when adding my own packages to comps.xml and having them being installed, and also to slim down an install to under 1 complete CD. I know that we've talked about comps.xml on the list in the past, but I don't think we've quite covered this aspect of things. Thanks again for the time - it is greatly appreciated. Thanks -dant _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From dan at id-confirm.com Wed Mar 8 20:41:27 2006 From: dan at id-confirm.com (Dan Trainor) Date: Wed, 08 Mar 2006 13:41:27 -0700 Subject: Small image based on comps.xml alone? In-Reply-To: References: Message-ID: <440F4177.1040906@id-confirm.com> Shabazian, Chip wrote: > What kind of information are you looking for? > > I haven't had time to work on this yet myself, but I did take the > standard comps.xml file, remove all of the language entries, and it > suddenly made a LOT more sense to me. > > Here are some of the links I've been saving for when I have time to do > this: > http://people.redhat.com/rkeech/custom-distro.txt > http://fedora.redhat.com/projects/anaconda-installer/comps.html > http://www.linuxjournal.com/article/6473 > http://www.linuxdevcenter.com/pub/a/linux/2005/02/17/kickstart_updates.h > tml?page=2 > http://www.linuxdevcenter.com/pub/a/linux/2004/11/04/advanced_kickstart. > html?page=2 > > I would like to create a comps.xml which TRULY builds a minimal system > (I have a build down to about 100 packages), then allow me to add onto > that as appropriate. > > Chip > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Dan Trainor > Sent: Wednesday, March 08, 2006 12:11 PM > To: kickstart-list at redhat.com > Subject: Small image based on comps.xml alone? > > Hello, all - > > I've been working on a few different items lately, one of which I'm > working on today. I'd like to build images which contain packages which > are either directly or indirectly referenced via comps.xml, being > dependencies of, or explicitly named packages, in comps.xml. > > The reason I'm doing this is to first allow myself some degree of > flexibility when adding my own packages to comps.xml and having them > being installed, and also to slim down an install to under 1 complete > CD. > > I know that we've talked about comps.xml on the list in the past, but I > don't think we've quite covered this aspect of things. > > Thanks again for the time - it is greatly appreciated. > > Thanks > -dant > Thanks for the reply, Chip - I was thinking, based on what I've read of your posts, "if anyone would respond to this, it would be Chip". haha. Anyway - yes, I'm trying to do the same sort of whing which you had mentioned. I've got it down pretty well, however, I don't know how to make my medium include ONLY those packages specified in comps.xml. The best I've been able to do is an HTTP boot, and record the packages downloaded by the installing client, based on Apache's logs. I'll then take this list of packages and throw them all on one bootable CD, and corss my fingers. However, there *has* to be a more elegant way of doing this. Thanks -- - Dan Trainor - id-Confirm, Inc. - Direct: 720.241.5580 From skvidal at linux.duke.edu Wed Mar 8 20:49:47 2006 From: skvidal at linux.duke.edu (seth vidal) Date: Wed, 08 Mar 2006 15:49:47 -0500 Subject: Small image based on comps.xml alone? In-Reply-To: <440F4177.1040906@id-confirm.com> References: <440F4177.1040906@id-confirm.com> Message-ID: <1141850987.10548.22.camel@cutter> > Thanks for the reply, Chip - > > I was thinking, based on what I've read of your posts, "if anyone would > respond to this, it would be Chip". haha. > > Anyway - yes, I'm trying to do the same sort of whing which you had > mentioned. I've got it down pretty well, however, I don't know how to > make my medium include ONLY those packages specified in comps.xml. > > The best I've been able to do is an HTTP boot, and record the packages > downloaded by the installing client, based on Apache's logs. I'll then > take this list of packages and throw them all on one bootable CD, and > corss my fingers. However, there *has* to be a more elegant way of > doing this. take your comps.xml - put it in as the groups file in a yum repo made by createrepo. and yum should be able to mark-for-install and depsolve everything in that group. then you know what pkgs you need. you could make it even easier than that with a little bit of coding. -sv From dan at id-confirm.com Wed Mar 8 21:06:09 2006 From: dan at id-confirm.com (Dan Trainor) Date: Wed, 08 Mar 2006 14:06:09 -0700 Subject: Small image based on comps.xml alone? In-Reply-To: <1141850987.10548.22.camel@cutter> References: <440F4177.1040906@id-confirm.com> <1141850987.10548.22.camel@cutter> Message-ID: <440F4741.9090201@id-confirm.com> seth vidal wrote: >> Thanks for the reply, Chip - >> >> I was thinking, based on what I've read of your posts, "if anyone would >> respond to this, it would be Chip". haha. >> >> Anyway - yes, I'm trying to do the same sort of whing which you had >> mentioned. I've got it down pretty well, however, I don't know how to >> make my medium include ONLY those packages specified in comps.xml. >> >> The best I've been able to do is an HTTP boot, and record the packages >> downloaded by the installing client, based on Apache's logs. I'll then >> take this list of packages and throw them all on one bootable CD, and >> corss my fingers. However, there *has* to be a more elegant way of >> doing this. > > take your comps.xml - put it in as the groups file in a yum repo made by > createrepo. > > and yum should be able to mark-for-install and depsolve everything in > that group. > > then you know what pkgs you need. > > you could make it even easier than that with a little bit of coding. > > -sv > > > Thanks for the reply, Seth - Interesting solution, and I think that solves part of the deal. Some of those articles that Chip had mentioned are a great reference. I'm going to toy around with a few of them right now. Thanks -- - Dan Trainor - id-Confirm, Inc. - Direct: 720.241.5580 From egan at sense.net Wed Mar 8 21:02:40 2006 From: egan at sense.net (Egan Ford) Date: Wed, 8 Mar 2006 14:02:40 -0700 Subject: pxe on IBM x366 In-Reply-To: <20060308192616.60524.qmail@web88102.mail.re2.yahoo.com> Message-ID: <042301c642f3$a5235480$8df00c09@oberon> What OS? (Version and Update) > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of > GABRIEL KAPITANY > Sent: Wednesday, March 08, 2006 12:26 PM > To: kickstart-list at redhat.com > Subject: pxe on IBM x366 > > > Hi guys, > > I have a problem trying to use pxe and nfs install on > an IBM x366 box. I'd like to mention that I have this > configuration working on a different hardware. > > The server boots using pxe but when in tries to do an > nfs mount it times out. > > Here are the messages on the client (Alt+F3): > > - eth1 has link, using it > - doing kickstart... setting it up > - waiting for link... > - 2 seconds > - reverse name lookup failed > - url is 192.168.1.10:/images/AS3.0U6-x366/kslapp.cfg > - file location: > nfs://192.168.1.10:/images/AS3.0U6-x366/kslapp.cfg > - failed to mount nfs source > - trying to mount device hda > > At this point I'm prompted for the type of install and > if I choose nfs then it works fine but I'm not able to > use kickstart.... > > On the pxe server in the logs I can see: > > tftp: client does not accept options > > The pxe configuration for this box is: > > default RHEL-AS3.0U6-x366 > > label RHEL-AS3.0U6-x366 > kernel RHEL-AS3.0U6-x366/vmlinuz > append initrd=RHEL-AS3.0U6-x366/initrd.img > ramdisk_size=7000 > ks=nfs:192.168.1.10:/images/AS3.0U6-x366/kslapp.cfg > ksdevice > =link > ipappend 1 > > Any help would be appreciated, > Thx, > > Gabriel > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From ks at puzzled.xs4all.nl Wed Mar 8 23:44:08 2006 From: ks at puzzled.xs4all.nl (Patrick) Date: Thu, 09 Mar 2006 00:44:08 +0100 Subject: Preserve $HOME in %post? In-Reply-To: References: Message-ID: <1141861448.8822.1.camel@guru.puzzled.xs4all.nl> On Wed, 2006-03-08 at 11:06 -0800, Shabazian, Chip wrote: > Just escape out the $ > > cat <>/home/test/.rpmmacros > %_topdir %(echo \$HOME)/redhat > EOF > chown test.test /home/test/.rpmmacros Thanks, will give it a try. Regards, Patrck From gkapitany at rogers.com Thu Mar 9 13:41:37 2006 From: gkapitany at rogers.com (GABRIEL KAPITANY) Date: Thu, 9 Mar 2006 08:41:37 -0500 (EST) Subject: pxe on IBM x366 In-Reply-To: <042301c642f3$a5235480$8df00c09@oberon> Message-ID: <20060309134137.48936.qmail@web88110.mail.re2.yahoo.com> Hi, I'm using a linux RHEL 3.0 AS Update6 for the image. In the mean time I figured that the timeout might came from the network card settings so instead of connecting through a switch I've used a crossover cable an the install worked fine. I'll try to play with the options: line_speed=100 and full_duplex=1, this might fix the problem. Thanks, Gabriel --- Egan Ford wrote: > What OS? (Version and Update) > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com > > [mailto:kickstart-list-bounces at redhat.com] On > Behalf Of > > GABRIEL KAPITANY > > Sent: Wednesday, March 08, 2006 12:26 PM > > To: kickstart-list at redhat.com > > Subject: pxe on IBM x366 > > > > > > Hi guys, > > > > I have a problem trying to use pxe and nfs install > on > > an IBM x366 box. I'd like to mention that I have > this > > configuration working on a different hardware. > > > > The server boots using pxe but when in tries to do > an > > nfs mount it times out. > > > > Here are the messages on the client (Alt+F3): > > > > - eth1 has link, using it > > - doing kickstart... setting it up > > - waiting for link... > > - 2 seconds > > - reverse name lookup failed > > - url is > 192.168.1.10:/images/AS3.0U6-x366/kslapp.cfg > > - file location: > > nfs://192.168.1.10:/images/AS3.0U6-x366/kslapp.cfg > > - failed to mount nfs source > > - trying to mount device hda > > > > At this point I'm prompted for the type of install > and > > if I choose nfs then it works fine but I'm not > able to > > use kickstart.... > > > > On the pxe server in the logs I can see: > > > > tftp: client does not accept options > > > > The pxe configuration for this box is: > > > > default RHEL-AS3.0U6-x366 > > > > label RHEL-AS3.0U6-x366 > > kernel RHEL-AS3.0U6-x366/vmlinuz > > append initrd=RHEL-AS3.0U6-x366/initrd.img > > ramdisk_size=7000 > > > ks=nfs:192.168.1.10:/images/AS3.0U6-x366/kslapp.cfg > > ksdevice > > =link > > ipappend 1 > > > > Any help would be appreciated, > > Thx, > > > > Gabriel > > > > _______________________________________________ > > 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 gkapitany at rogers.com Fri Mar 10 19:32:09 2006 From: gkapitany at rogers.com (GABRIEL KAPITANY) Date: Fri, 10 Mar 2006 14:32:09 -0500 (EST) Subject: modified netboot.img for ppc Message-ID: <20060310193209.89356.qmail@web88111.mail.re2.yahoo.com> Hi, I'd like to modify the default netboot.img for PPC and I'd like to append some parameters to the kernel: ks=.. ksdevice=.. What I'd like to achieve is an unattended install similar to pxe, using a kickstart file. The tool used is mkzimage(ld) but I don't know how can I do that, hardcode in the kernel? Thanks, Gabriel From dan at id-confirm.com Wed Mar 15 00:38:00 2006 From: dan at id-confirm.com (Dan Trainor) Date: Tue, 14 Mar 2006 17:38:00 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line Message-ID: <441761E8.8090705@id-confirm.com> Hello, all - I'm trying to take a few command-line arguments during Anaconda's boot sequence for a new ISO which I am remastering, and I can't quite seem to extract them into ks.cfg itself. I'm using something similar to this to extract the values, in %pre: if grep -i -q "ipaddr=[a-zA-Z0-9]" /proc/cmdline then IPADDR=`cat /proc/cmdline | sed 's/.*ipaddr=\([^ ]*\).*/\1/'` else IPADDR=1.2.3.4 fi I then can extract ${IPADDR} from most anywhere - except my ks.cfg. I'm using something similar to this, to define network arguments: network --device=eth0 \ --bootproto=static \ --ip=`/usr/bin/echo ${IPADDR}` \ --netmask=`/usr/bin/echo ${NMADDR}` \ --gateway=`/usr/bin/echo ${GWADDR}` \ --nameserver=`/usr/bin/echo ${NSADDR}` ...to no avail. I see the literal contents (IPADDR=`/usr/bin/echo ) inside my system config scripts, instead of the extracted value being written to these files. I've even gone so far as to get rid of my echo statement in an attempt to get it to work, as well as just placing ${VARNAME}, to get more familiar with how Anaconda extracts these values trying to find my mistake. I'm sure what I'm doing is not overly complicated - but for now, I'm stuck. If anyone has any feedback, I'd be very eager to hear back from you. Thanks! -dant From clumens at redhat.com Wed Mar 15 00:57:54 2006 From: clumens at redhat.com (Chris Lumens) Date: Tue, 14 Mar 2006 19:57:54 -0500 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <441761E8.8090705@id-confirm.com> References: <441761E8.8090705@id-confirm.com> Message-ID: <20060315005754.GA11357@exeter.boston.redhat.com> > I'm using something similar to this, to define network arguments: > > network --device=eth0 \ > --bootproto=static \ > --ip=`/usr/bin/echo ${IPADDR}` \ > --netmask=`/usr/bin/echo ${NMADDR}` \ > --gateway=`/usr/bin/echo ${GWADDR}` \ > --nameserver=`/usr/bin/echo ${NSADDR}` > > ...to no avail. I see the literal contents (IPADDR=`/usr/bin/echo ) > inside my system config scripts, instead of the extracted value being > written to these files. I've even gone so far as to get rid of my echo > statement in an attempt to get it to work, as well as just placing > ${VARNAME}, to get more familiar with how Anaconda extracts these values > trying to find my mistake. > > I'm sure what I'm doing is not overly complicated - but for now, I'm > stuck. If anyone has any feedback, I'd be very eager to hear back from you. Kickstart doesn't have any provisions for reading environment variables or capturing the output of programs like you're trying to do. However, you can do this stuff in a %pre script, as that just gets fed to the shell (or whatever you set the interpreter to). So you could write a pre script that generates the dynamic chunks and writes it to a file, then %include that file up in the command section. That should work. - Chris From dan at id-confirm.com Wed Mar 15 00:58:49 2006 From: dan at id-confirm.com (Dan Trainor) Date: Tue, 14 Mar 2006 17:58:49 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <20060315005754.GA11357@exeter.boston.redhat.com> References: <441761E8.8090705@id-confirm.com> <20060315005754.GA11357@exeter.boston.redhat.com> Message-ID: <441766C9.9060206@id-confirm.com> Chris Lumens wrote: >> I'm using something similar to this, to define network arguments: >> >> network --device=eth0 \ >> --bootproto=static \ >> --ip=`/usr/bin/echo ${IPADDR}` \ >> --netmask=`/usr/bin/echo ${NMADDR}` \ >> --gateway=`/usr/bin/echo ${GWADDR}` \ >> --nameserver=`/usr/bin/echo ${NSADDR}` >> >> ...to no avail. I see the literal contents (IPADDR=`/usr/bin/echo ) >> inside my system config scripts, instead of the extracted value being >> written to these files. I've even gone so far as to get rid of my echo >> statement in an attempt to get it to work, as well as just placing >> ${VARNAME}, to get more familiar with how Anaconda extracts these values >> trying to find my mistake. >> >> I'm sure what I'm doing is not overly complicated - but for now, I'm >> stuck. If anyone has any feedback, I'd be very eager to hear back from you. > > Kickstart doesn't have any provisions for reading environment variables > or capturing the output of programs like you're trying to do. However, > you can do this stuff in a %pre script, as that just gets fed to the > shell (or whatever you set the interpreter to). So you could write a > pre script that generates the dynamic chunks and writes it to a file, > then %include that file up in the command section. > > That should work. > > - Chris > Thanks for the response, Chris - Interesting... I'll give it a shot here and let you know. Thanks! -dant From dan at id-confirm.com Wed Mar 15 01:45:19 2006 From: dan at id-confirm.com (Dan Trainor) Date: Tue, 14 Mar 2006 18:45:19 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <20060315005754.GA11357@exeter.boston.redhat.com> References: <441761E8.8090705@id-confirm.com> <20060315005754.GA11357@exeter.boston.redhat.com> Message-ID: <441771AF.1000609@id-confirm.com> Chris Lumens wrote: >> I'm using something similar to this, to define network arguments: >> >> network --device=eth0 \ >> --bootproto=static \ >> --ip=`/usr/bin/echo ${IPADDR}` \ >> --netmask=`/usr/bin/echo ${NMADDR}` \ >> --gateway=`/usr/bin/echo ${GWADDR}` \ >> --nameserver=`/usr/bin/echo ${NSADDR}` >> >> ...to no avail. I see the literal contents (IPADDR=`/usr/bin/echo ) >> inside my system config scripts, instead of the extracted value being >> written to these files. I've even gone so far as to get rid of my echo >> statement in an attempt to get it to work, as well as just placing >> ${VARNAME}, to get more familiar with how Anaconda extracts these values >> trying to find my mistake. >> >> I'm sure what I'm doing is not overly complicated - but for now, I'm >> stuck. If anyone has any feedback, I'd be very eager to hear back from you. > > Kickstart doesn't have any provisions for reading environment variables > or capturing the output of programs like you're trying to do. However, > you can do this stuff in a %pre script, as that just gets fed to the > shell (or whatever you set the interpreter to). So you could write a > pre script that generates the dynamic chunks and writes it to a file, > then %include that file up in the command section. > > That should work. > > - Chris > Hello again, Chris - I'm using %pre to write to a file called /tmp/netsettings. I'm then trying to %include it from within my commands section. Every time, I'm given an error saying that /tmp/netsettings cannot be found. I cannot verify this, or it's location, because Anaconda bombs out and locks up vt2. I think I'm lost in chroot land. Do I need to mount a disk in %pre in order to write my temp file? What would you suggest? Thanks! -dant From sjbrowne at bluebottle.com Wed Mar 15 01:58:06 2006 From: sjbrowne at bluebottle.com (Stuart J. Browne) Date: Wed, 15 Mar 2006 12:58:06 +1100 Subject: Extracting variables from within ks.cfg, interpreted from command-line References: <441761E8.8090705@id-confirm.com><20060315005754.GA11357@exeter.boston.redhat.com> <441771AF.1000609@id-confirm.com> Message-ID: <003f01c647d4$00a9ba80$4201a8c0@bluebottle.com> You shouldn't need to, the /tmp/ should be the same at that point for the %include (It was when I was doing this). You should be able to add a simple 'read JUNK' into your %pre to pause it there so you can check VT2 and see if the file was created properly, or just cat it at that point to another of the VT's. ----- Original Message ----- From: "Dan Trainor" To: "Discussion list about Kickstart" Sent: Wednesday, March 15, 2006 12:45 PM Subject: Re: Extracting variables from within ks.cfg,interpreted from command-line > Chris Lumens wrote: >>> I'm using something similar to this, to define network arguments: >>> >>> network --device=eth0 \ >>> --bootproto=static \ >>> --ip=`/usr/bin/echo ${IPADDR}` \ >>> --netmask=`/usr/bin/echo ${NMADDR}` \ >>> --gateway=`/usr/bin/echo ${GWADDR}` \ >>> --nameserver=`/usr/bin/echo ${NSADDR}` >>> >>> ...to no avail. I see the literal contents (IPADDR=`/usr/bin/echo ) >>> inside my system config scripts, instead of the extracted value being >>> written to these files. I've even gone so far as to get rid of my echo >>> statement in an attempt to get it to work, as well as just placing >>> ${VARNAME}, to get more familiar with how Anaconda extracts these values >>> trying to find my mistake. >>> >>> I'm sure what I'm doing is not overly complicated - but for now, I'm >>> stuck. If anyone has any feedback, I'd be very eager to hear back from >>> you. >> >> Kickstart doesn't have any provisions for reading environment variables >> or capturing the output of programs like you're trying to do. However, >> you can do this stuff in a %pre script, as that just gets fed to the >> shell (or whatever you set the interpreter to). So you could write a >> pre script that generates the dynamic chunks and writes it to a file, >> then %include that file up in the command section. >> >> That should work. >> >> - Chris >> > > Hello again, Chris - > > I'm using %pre to write to a file called /tmp/netsettings. I'm then > trying to %include it from within my commands section. > > Every time, I'm given an error saying that /tmp/netsettings cannot be > found. I cannot verify this, or it's location, because Anaconda bombs out > and locks up vt2. > > I think I'm lost in chroot land. Do I need to mount a disk in %pre in > order to write my temp file? What would you suggest? > > Thanks! > -dant > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From dan at id-confirm.com Wed Mar 15 06:59:19 2006 From: dan at id-confirm.com (Dan Trainor) Date: Tue, 14 Mar 2006 23:59:19 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <003f01c647d4$00a9ba80$4201a8c0@bluebottle.com> References: <441761E8.8090705@id-confirm.com><20060315005754.GA11357@exeter.boston.redhat.com> <441771AF.1000609@id-confirm.com> <003f01c647d4$00a9ba80$4201a8c0@bluebottle.com> Message-ID: <4417BB47.5030801@id-confirm.com> Stuart J. Browne wrote: > You shouldn't need to, the /tmp/ should be the same at that point for > the %include (It was when I was doing this). > > You should be able to add a simple 'read JUNK' into your %pre to pause > it there so you can check VT2 and see if the file was created properly, > or just cat it at that point to another of the VT's. > ----- Original Message ----- From: "Dan Trainor" > To: "Discussion list about Kickstart" > Sent: Wednesday, March 15, 2006 12:45 PM > Subject: Re: Extracting variables from within ks.cfg,interpreted from > command-line > > >> Chris Lumens wrote: >>>> I'm using something similar to this, to define network arguments: >>>> >>>> network --device=eth0 \ >>>> --bootproto=static \ >>>> --ip=`/usr/bin/echo ${IPADDR}` \ >>>> --netmask=`/usr/bin/echo ${NMADDR}` \ >>>> --gateway=`/usr/bin/echo ${GWADDR}` \ >>>> --nameserver=`/usr/bin/echo ${NSADDR}` >>>> >>>> ...to no avail. I see the literal contents (IPADDR=`/usr/bin/echo ) >>>> inside my system config scripts, instead of the extracted value >>>> being written to these files. I've even gone so far as to get rid >>>> of my echo statement in an attempt to get it to work, as well as >>>> just placing ${VARNAME}, to get more familiar with how Anaconda >>>> extracts these values trying to find my mistake. >>>> >>>> I'm sure what I'm doing is not overly complicated - but for now, I'm >>>> stuck. If anyone has any feedback, I'd be very eager to hear back >>>> from you. >>> >>> Kickstart doesn't have any provisions for reading environment variables >>> or capturing the output of programs like you're trying to do. However, >>> you can do this stuff in a %pre script, as that just gets fed to the >>> shell (or whatever you set the interpreter to). So you could write a >>> pre script that generates the dynamic chunks and writes it to a file, >>> then %include that file up in the command section. >>> >>> That should work. >>> >>> - Chris >>> >> >> Hello again, Chris - >> >> I'm using %pre to write to a file called /tmp/netsettings. I'm then >> trying to %include it from within my commands section. >> >> Every time, I'm given an error saying that /tmp/netsettings cannot be >> found. I cannot verify this, or it's location, because Anaconda bombs >> out and locks up vt2. >> >> I think I'm lost in chroot land. Do I need to mount a disk in %pre in >> order to write my temp file? What would you suggest? >> >> Thanks! >> -dant >> hi - Alright, so, I think something wierd is going on. I'm still seeing Anaconda bomb out saying that /tmp/netsettings does not exist. I can verify that /tmp/netsettings does not get created. However, I cannot figure out why. vty3 tells me something similar to the following: : command not founde 1: : command not founde 3: : command not founde 6: : command not founde 9: : command not founde 11: /tmp/ks-script: line 44: syntax error: unexpected end of file * WARNING - error code 512 encountered running a kickstart %pre%post script ...and it repeats one more time. It's interesting because those numbers do not correspond with anything in my kickstart script, nor do I have the phrase "founde" anywhere in my kickstart. In fact, I can't find any references to that phrase in my Anaconda source trees, either. I know someone has to have done this and had it working successfully before. I'm just looking for some insight here. Are you guys sure I'm not just lost in a chroot jungle? Is placement of %pre in relation to the networking directive relative to the situation? Anything like that? Thanks again, guys! -dant From phr at doc.ic.ac.uk Wed Mar 15 07:20:38 2006 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 15 Mar 2006 07:20:38 +0000 (GMT) Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <4417BB47.5030801@id-confirm.com> References: <441761E8.8090705@id-confirm.com><20060315005754.GA11357@exeter.boston.redhat.com> <441771AF.1000609@id-confirm.com> <003f01c647d4$00a9ba80$4201a8c0@bluebottle.com> <4417BB47.5030801@id-confirm.com> Message-ID: On Tue, 14 Mar 2006, Dan Trainor wrote: > Alright, so, I think something wierd is going on. > > I'm still seeing Anaconda bomb out saying that /tmp/netsettings does > not exist. I can verify that /tmp/netsettings does not get created. > However, I cannot figure out why. > > vty3 tells me something similar to the following: > > : command not founde 1: > : command not founde 3: > : command not founde 6: [snip] This doesn't quite feel like CRLF-line-endings, but worth a check. Can you post your kickstart file? Cheers, Phil From dan at id-confirm.com Wed Mar 15 07:24:40 2006 From: dan at id-confirm.com (Dan Trainor) Date: Wed, 15 Mar 2006 00:24:40 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: References: <441761E8.8090705@id-confirm.com><20060315005754.GA11357@exeter.boston.redhat.com> <441771AF.1000609@id-confirm.com> <003f01c647d4$00a9ba80$4201a8c0@bluebottle.com> <4417BB47.5030801@id-confirm.com> Message-ID: <4417C138.2070909@id-confirm.com> Philip Rowlands wrote: > On Tue, 14 Mar 2006, Dan Trainor wrote: > >> Alright, so, I think something wierd is going on. >> >> I'm still seeing Anaconda bomb out saying that /tmp/netsettings does >> not exist. I can verify that /tmp/netsettings does not get created. >> However, I cannot figure out why. >> >> vty3 tells me something similar to the following: >> >> : command not founde 1: >> : command not founde 3: >> : command not founde 6: > [snip] > > This doesn't quite feel like CRLF-line-endings, but worth a check. Can > you post your kickstart file? > > > Cheers, > Phil > Thanks for the response, Phil - I've pasted it at http://pastebin.com/603104 Thanks -dant From jkeating at j2solutions.net Wed Mar 15 12:10:39 2006 From: jkeating at j2solutions.net (Jesse Keating) Date: Wed, 15 Mar 2006 07:10:39 -0500 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <4417C138.2070909@id-confirm.com> References: <441761E8.8090705@id-confirm.com><20060315005754.GA11357@exeter.boston.redhat.com> <441771AF.1000609@id-confirm.com> <003f01c647d4$00a9ba80$4201a8c0@bluebottle.com> <4417BB47.5030801@id-confirm.com> <4417C138.2070909@id-confirm.com> Message-ID: <4418043F.5060108@j2solutions.net> On 03/15/2006 Dan Trainor wrote: > > I've pasted it at http://pastebin.com/603104 I've had bad luck putting a bash script in %pre. Instead I wget a script in %pre and then execute it. The command not founde is probably a typo, and it is trying ot run your script but is not able to, some part of it is missing in the anaconda env (like bash?). The line numbers are really hard to grok when dealing w/ ks script errors. -- Jesse Keating RHCE (geek.j2solutions.net) Fedora Legacy Team (www.fedoralegacy.org) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 191 bytes Desc: OpenPGP digital signature URL: From rwcitek at alum.calberkeley.org Wed Mar 15 15:49:06 2006 From: rwcitek at alum.calberkeley.org (Robert Citek) Date: Wed, 15 Mar 2006 09:49:06 -0600 Subject: swapfile instead of swap partition? Message-ID: <2FCC8C7D-1338-45CF-9F89-139E3651B17D@alum.calberkeley.org> Is there a way to specify that the kickstart script should create a swapfile instead of a swap partition? My goal is to partition the disk with just one partition formatted as ext3 and that contains a swapfile. Searched the list archives with google[1] but only found hits for "swap" but not "swapfile". Also had a look through the docs[2] but again only "swap" no "swapfile". Thanks in advance for any pointers. [1] http://www.google.com/search?q=site:www.redhat.com/archives/ kickstart-list+swapfile [2] http://www.redhat.com/docs/manuals/linux/RHL-7.3-Manual/custom- guide/ch-kickstart2.html Regards, - Robert http://www.cwelug.org/downloads Help others get OpenSource software. Distribute FLOSS for Windows, Linux, *BSD, and MacOS X with BitTorrent From Chip.Shabazian at bankofamerica.com Wed Mar 15 17:22:30 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 15 Mar 2006 09:22:30 -0800 Subject: Extracting variables from within ks.cfg, interpreted from command-line Message-ID: This is what I use in my command section: %include /tmp/partinfo And this is in my %pre: ETH=`grep DEVICE /tmp/netinfo | cut -d = -f 2` IP=`ifconfig $ETH | grep inet | cut -d : -f 2 | cut -d " " -f 1` NETMASK=`ifconfig $ETH | grep inet | cut -d : -f 4` GATEWAY=`route -n| grep "^0.0.0.0" | cut -b 17-32 | cut -d " " -f 1` HOSTNAME=`grep HOSTNAME /tmp/netinfo | cut -d = -f 2 | cut -d . -f 1` cat << EOF > /tmp/buildnet network --device $ETH --bootproto static --ip=$IP --netmask=$NETMASK --gateway=$ GATEWAY --hostname=$HOSTNAME EOF This works every time, whether you get your IP address from DHCP, or specify it on the command line. Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Philip Rowlands Sent: Tuesday, March 14, 2006 11:21 PM To: Discussion list about Kickstart Subject: Re: Extracting variables from within ks.cfg, interpreted from command-line On Tue, 14 Mar 2006, Dan Trainor wrote: > Alright, so, I think something wierd is going on. > > I'm still seeing Anaconda bomb out saying that /tmp/netsettings does > not exist. I can verify that /tmp/netsettings does not get created. > However, I cannot figure out why. > > vty3 tells me something similar to the following: > > : command not founde 1: > : command not founde 3: > : command not founde 6: [snip] This doesn't quite feel like CRLF-line-endings, but worth a check. Can you post your kickstart file? Cheers, Phil _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From Chip.Shabazian at bankofamerica.com Wed Mar 15 17:30:19 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 15 Mar 2006 09:30:19 -0800 Subject: swapfile instead of swap partition? Message-ID: Just build the system with no swap disk, then create your swap file in the %post -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Robert Citek Sent: Wednesday, March 15, 2006 7:49 AM To: Kickstart-list at redhat.com Subject: swapfile instead of swap partition? Is there a way to specify that the kickstart script should create a swapfile instead of a swap partition? My goal is to partition the disk with just one partition formatted as ext3 and that contains a swapfile. Searched the list archives with google[1] but only found hits for "swap" but not "swapfile". Also had a look through the docs[2] but again only "swap" no "swapfile". Thanks in advance for any pointers. [1] http://www.google.com/search?q=site:www.redhat.com/archives/ kickstart-list+swapfile [2] http://www.redhat.com/docs/manuals/linux/RHL-7.3-Manual/custom- guide/ch-kickstart2.html Regards, - Robert http://www.cwelug.org/downloads Help others get OpenSource software. Distribute FLOSS for Windows, Linux, *BSD, and MacOS X with BitTorrent _______________________________________________ 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 Mar 15 18:21:52 2006 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 15 Mar 2006 18:21:52 +0000 (GMT) Subject: swapfile instead of swap partition? In-Reply-To: <2FCC8C7D-1338-45CF-9F89-139E3651B17D@alum.calberkeley.org> References: <2FCC8C7D-1338-45CF-9F89-139E3651B17D@alum.calberkeley.org> Message-ID: On Wed, 15 Mar 2006, Robert Citek wrote: > Is there a way to specify that the kickstart script should create a > swapfile instead of a swap partition? I don't believe anaconda (kickstart) has a way of doing this, but it would certainly be possible in a %post script. Cheers, Phil From rwcitek at alum.calberkeley.org Wed Mar 15 19:45:27 2006 From: rwcitek at alum.calberkeley.org (Robert Citek) Date: Wed, 15 Mar 2006 13:45:27 -0600 Subject: %post script for swapfile (was: swapfile instead of swap partition?) In-Reply-To: References: Message-ID: <51FDB933-3078-45F5-8D18-8584BA7F2E59@alum.calberkeley.org> On Mar 15, 2006, at 11:30 AM, Shabazian, Chip wrote: > Just build the system with no swap disk, then create your swap file in > the %post On Mar 15, 2006, at 12:21 PM, Philip Rowlands wrote: > On Wed, 15 Mar 2006, Robert Citek wrote: > >> Is there a way to specify that the kickstart script should create >> a swapfile instead of a swap partition? > > I don't believe anaconda (kickstart) has a way of doing this, but > it would certainly be possible in a %post script. Thanks, Chip and Phil. I'm not that familiar with putting a script in %post. In a "normal" environment, the script would look something like this: dd if=/dev/zero of=/swapfile bs=1M size=100 mkswap /swapfile echo -e "/swapfile\tswap\tswap\tdefaults\t0\t0" >> /etc/fstab From the docs[1] it looks as though this script will work as-is. If that's not the case, how would I need to modify that script so that it will work in a %post section? Also, what default interpreter is being used to run the script: bash, sh, other? The docs mention that I can change it, but not what the default is, unless I overlooked it. Again, thanks in advance for any pointers. [1] http://www.redhat.com/docs/manuals/linux/RHL-7.3-Manual/custom- guide/s1-kickstart2-options.html#S2-KICKSTART2-POSTINSTALLCONFIG Regards, - Robert http://www.cwelug.org/downloads Help others get OpenSource software. Distribute FLOSS for Windows, Linux, *BSD, and MacOS X with BitTorrent From Chip.Shabazian at bankofamerica.com Wed Mar 15 20:05:40 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 15 Mar 2006 12:05:40 -0800 Subject: %post script for swapfile (was: swapfile instead of swap partition?) Message-ID: That should work just fine as is -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Robert Citek Sent: Wednesday, March 15, 2006 11:45 AM To: Discussion list about Kickstart Subject: %post script for swapfile (was: swapfile instead of swap partition?) On Mar 15, 2006, at 11:30 AM, Shabazian, Chip wrote: > Just build the system with no swap disk, then create your swap file in > the %post On Mar 15, 2006, at 12:21 PM, Philip Rowlands wrote: > On Wed, 15 Mar 2006, Robert Citek wrote: > >> Is there a way to specify that the kickstart script should create a >> swapfile instead of a swap partition? > > I don't believe anaconda (kickstart) has a way of doing this, but it > would certainly be possible in a %post script. Thanks, Chip and Phil. I'm not that familiar with putting a script in %post. In a "normal" environment, the script would look something like this: dd if=/dev/zero of=/swapfile bs=1M size=100 mkswap /swapfile echo -e "/swapfile\tswap\tswap\tdefaults\t0\t0" >> /etc/fstab From the docs[1] it looks as though this script will work as-is. If that's not the case, how would I need to modify that script so that it will work in a %post section? Also, what default interpreter is being used to run the script: bash, sh, other? The docs mention that I can change it, but not what the default is, unless I overlooked it. Again, thanks in advance for any pointers. [1] http://www.redhat.com/docs/manuals/linux/RHL-7.3-Manual/custom- guide/s1-kickstart2-options.html#S2-KICKSTART2-POSTINSTALLCONFIG Regards, - Robert http://www.cwelug.org/downloads Help others get OpenSource software. Distribute FLOSS for Windows, Linux, *BSD, and MacOS X with BitTorrent _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From rwcitek at alum.calberkeley.org Wed Mar 15 21:00:25 2006 From: rwcitek at alum.calberkeley.org (Robert Citek) Date: Wed, 15 Mar 2006 15:00:25 -0600 Subject: %post script for swapfile In-Reply-To: References: Message-ID: On Mar 15, 2006, at 2:05 PM, Shabazian, Chip wrote: > That should work just fine as is Thanks, Chip. The reason I asked is because there doesn't seem to be any easy way to test/debug a %post script. That is, the only way to test/debug is to run the entire install process again. So, it's nice to know in advance that a post script should work. Regards, - Robert http://www.cwelug.org/downloads Help others get OpenSource software. Distribute FLOSS for Windows, Linux, *BSD, and MacOS X with BitTorrent From Chip.Shabazian at bankofamerica.com Wed Mar 15 21:02:31 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 15 Mar 2006 13:02:31 -0800 Subject: %post script for swapfile Message-ID: When testing/debugging, I put a sleep 99999999999 in my %pre and %post. You can then go into the Alt-F2 shell and test whatever you want. When it works the way you want, just kill the sleep process, and the build will continue. Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Robert Citek Sent: Wednesday, March 15, 2006 1:00 PM To: Discussion list about Kickstart Subject: Re: %post script for swapfile On Mar 15, 2006, at 2:05 PM, Shabazian, Chip wrote: > That should work just fine as is Thanks, Chip. The reason I asked is because there doesn't seem to be any easy way to test/debug a %post script. That is, the only way to test/debug is to run the entire install process again. So, it's nice to know in advance that a post script should work. Regards, - Robert http://www.cwelug.org/downloads Help others get OpenSource software. Distribute FLOSS for Windows, Linux, *BSD, and MacOS X with BitTorrent _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From lists.james at gmail.com Wed Mar 15 21:32:19 2006 From: lists.james at gmail.com (James P) Date: Wed, 15 Mar 2006 15:32:19 -0600 Subject: %post script for swapfile In-Reply-To: References: Message-ID: <441887E3.7090708@gmail.com> I'd also recommend using /bin/ash if you want to test drive any wierd or convoluted logic you're going to put into the %pre section. /bin/bash works for testing %post stuff, but the %pre interpreter isn't a full-fledged bash shell, so you can't do fancy stuff like ${parameter//pattern/string} like you can do in bash. I'm not 100% sure if ash is even the interpreter used in %pre, but it's been a perfect simulator for me so far. Also, if you don't want to sleep for an eternity, you could put something like this in %pre or %post: exec < /dev/tty3 > /dev/tty3 chvt 3 echo "Press to continue with installation." read foo chvt 1 ...... Then you could go mess around in Alt-F2 for as long as you like; and just Alt-F3, when you're ready to continue the build. Shabazian, Chip wrote: >When testing/debugging, I put a sleep 99999999999 in my %pre and %post. >You can then go into the Alt-F2 shell and test whatever you want. When >it works the way you want, just kill the sleep process, and the build >will continue. > >Chip > >-----Original Message----- >From: kickstart-list-bounces at redhat.com >[mailto:kickstart-list-bounces at redhat.com] On Behalf Of Robert Citek >Sent: Wednesday, March 15, 2006 1:00 PM >To: Discussion list about Kickstart >Subject: Re: %post script for swapfile > > >On Mar 15, 2006, at 2:05 PM, Shabazian, Chip wrote: > > >>That should work just fine as is >> >> > >Thanks, Chip. > >The reason I asked is because there doesn't seem to be any easy way to >test/debug a %post script. That is, the only way to test/debug is to >run the entire install process again. So, it's nice to know in advance >that a post script should work. > >Regards, >- Robert >http://www.cwelug.org/downloads >Help others get OpenSource software. Distribute FLOSS for Windows, >Linux, *BSD, and MacOS X with BitTorrent > >_______________________________________________ >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 smah at vmware.com Wed Mar 15 22:32:46 2006 From: smah at vmware.com (Stephen Mah) Date: Wed, 15 Mar 2006 14:32:46 -0800 Subject: kickstart from a HD image Message-ID: <4418960E.7090403@vmware.com> Has anyone created a Hard drive image that runs a linux kernel or syslinux to kickstart Red Hat via network install? What I'm looking for is a way to send a hard drive image using something like Ghost. The image would contain the vmlinuz and initrd.img file to start the installer. The image would also need to work on multiple machine types, ie: different disk geometry, hardware, etc... -steve From jim at rossberry.com Wed Mar 15 22:47:21 2006 From: jim at rossberry.com (Jim Wildman) Date: Wed, 15 Mar 2006 17:47:21 -0500 (EST) Subject: kickstart from a HD image In-Reply-To: <4418960E.7090403@vmware.com> References: <4418960E.7090403@vmware.com> Message-ID: On Wed, 15 Mar 2006, Stephen Mah wrote: > Has anyone created a Hard drive image that runs a linux kernel or > syslinux to kickstart Red Hat via network install? > Just drop the initrd and vmlinuz from the appropriate boot.iso into /boot/grub, edit grub.conf and off you go. ------------------------------------------------------------------------ Jim Wildman, CISSP, RHCE jim at rossberry.com http://www.rossberry.com "Society in every state is a blessing, but Government, even in its best state, is a necessary evil; in its worst state, an intolerable one." Thomas Paine From Chip.Shabazian at bankofamerica.com Wed Mar 15 22:39:52 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 15 Mar 2006 14:39:52 -0800 Subject: kickstart from a HD image Message-ID: Sounds like you may want something like systemimager, not kickstart: http://www.systemimager.org/ -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Stephen Mah Sent: Wednesday, March 15, 2006 2:33 PM To: kickstart-list at redhat.com Subject: kickstart from a HD image Has anyone created a Hard drive image that runs a linux kernel or syslinux to kickstart Red Hat via network install? What I'm looking for is a way to send a hard drive image using something like Ghost. The image would contain the vmlinuz and initrd.img file to start the installer. The image would also need to work on multiple machine types, ie: different disk geometry, hardware, etc... -steve _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From ebrown at lanl.gov Thu Mar 16 01:16:00 2006 From: ebrown at lanl.gov (Ed Brown) Date: Wed, 15 Mar 2006 18:16:00 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <441761E8.8090705@id-confirm.com> References: <441761E8.8090705@id-confirm.com> Message-ID: <1142471760.25366.321.camel@edbrown.lanl.gov> On Tue, 2006-03-14 at 17:38 -0700, Dan Trainor wrote: > I'm using something similar to this to extract the values, in %pre: > > if grep -i -q "ipaddr=[a-zA-Z0-9]" /proc/cmdline > then > IPADDR=`cat /proc/cmdline | sed 's/.*ipaddr=\([^ ]*\).*/\1/'` > else > IPADDR=1.2.3.4 > fi > > I then can extract ${IPADDR} from most anywhere - except my ks.cfg. > > I'm using something similar to this, to define network arguments: > > network --device=eth0 \ > --bootproto=static \ > --ip=`/usr/bin/echo ${IPADDR}` \ > --netmask=`/usr/bin/echo ${NMADDR}` \ > --gateway=`/usr/bin/echo ${GWADDR}` \ > --nameserver=`/usr/bin/echo ${NSADDR}` > Quite possibly I don't understand what you are trying to do, but if you are providing static ip network info at the anaconda 'boot:' prompt, why would you need to write this information into the kickstart file on the network line? You can provide these parameters like: boot: linux ks=... method=... ip=... gateway=... dns=... netmask=... and have this in your ks.cfg: network --device=eth0 --bootproto=static and there is no need to edit ks.cfg. Or, if you are creating a custom syslinux.cfg, your stanzas could list all the parameters except perhaps the ip, or ip and gateway: syslinux.cfg: ------------ ... label mylabel kernel vmlinuz append initrd=initrd.img dns=our.dns.server netmask=255.255.255.0 ks=URL_of_ks.cfg method=URL_of_distro at anaconda 'boot:' prompt: boot: mylabel ip=... gateway=... Again, no need to edit ks.cfg hope this helps, Ed From smah at vmware.com Thu Mar 16 02:50:11 2006 From: smah at vmware.com (Stephen Mah) Date: Wed, 15 Mar 2006 18:50:11 -0800 Subject: running loadlin with ks=nfs Message-ID: <4418D263.8010207@vmware.com> Has anyone started the Red Hat installer from loadlin passing the ks=nfs option? Mine got to stage1 of the installer and can't find the kickstart file. steve From dan at id-confirm.com Thu Mar 16 08:24:14 2006 From: dan at id-confirm.com (Dan Trainor) Date: Thu, 16 Mar 2006 01:24:14 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <1142471760.25366.321.camel@edbrown.lanl.gov> References: <441761E8.8090705@id-confirm.com> <1142471760.25366.321.camel@edbrown.lanl.gov> Message-ID: <441920AE.10200@id-confirm.com> Ed Brown wrote: > On Tue, 2006-03-14 at 17:38 -0700, Dan Trainor wrote: > >> I'm using something similar to this to extract the values, in %pre: >> >> if grep -i -q "ipaddr=[a-zA-Z0-9]" /proc/cmdline >> then >> IPADDR=`cat /proc/cmdline | sed 's/.*ipaddr=\([^ ]*\).*/\1/'` >> else >> IPADDR=1.2.3.4 >> fi >> >> I then can extract ${IPADDR} from most anywhere - except my ks.cfg. >> >> I'm using something similar to this, to define network arguments: >> >> network --device=eth0 \ >> --bootproto=static \ >> --ip=`/usr/bin/echo ${IPADDR}` \ >> --netmask=`/usr/bin/echo ${NMADDR}` \ >> --gateway=`/usr/bin/echo ${GWADDR}` \ >> --nameserver=`/usr/bin/echo ${NSADDR}` >> > > Quite possibly I don't understand what you are trying to do, but if you > are providing static ip network info at the anaconda 'boot:' prompt, why > would you need to write this information into the kickstart file on the > network line? > > You can provide these parameters like: > > boot: linux ks=... method=... ip=... gateway=... dns=... netmask=... > > and have this in your ks.cfg: > > network --device=eth0 --bootproto=static > > and there is no need to edit ks.cfg. > > > Or, if you are creating a custom syslinux.cfg, your stanzas could list > all the parameters except perhaps the ip, or ip and gateway: > > syslinux.cfg: > ------------ > ... > label mylabel > kernel vmlinuz > append initrd=initrd.img dns=our.dns.server netmask=255.255.255.0 > ks=URL_of_ks.cfg method=URL_of_distro > > at anaconda 'boot:' prompt: > boot: mylabel ip=... gateway=... > > Again, no need to edit ks.cfg > > hope this helps, > Ed > > > > Hello, Ed - What I'm wanting to do is to let the user supply boot:-time arguments for a few settings, but if they do not specify these settings, use a default. I think that Chip had responded earlier with a solution, which I have yet to try. Regardless, I'll report back. Thanks! -- - Dan Trainor - id-Confirm, Inc. - Direct: 720.241.5580 From jdixson at metreos.com Thu Mar 16 20:15:55 2006 From: jdixson at metreos.com (James Dixson) Date: Thu, 16 Mar 2006 14:15:55 -0600 Subject: Running anaconda twice Message-ID: <6EE053013F00084096AF0F02CFDB88B8254CD1@securemail.metreos.com> I am building a network appliance based on linux. I want this appliance to have three bootable partitions, a recovery partition and two runtime partitions (designated 0 and 1). To simply manufacturing as well as to facilitate functional testing in virtual hardware, I want to build a bootable ISO that I can use to either initialize new hardware or to initialize virtual hardware with all three partitions. I have decided to use kickstart/anaconda to automate the install process. I have succeeded so far in constructing a netstg2.img that uses a "shim" anaconda to invoke the "real" anaconda twice in succession, but with different ks.cfg files, and then invoke partimage to duplicate image0 to create image1. My new netstg2.img works perfectly, but success with stage2.img and ISO/CD installs (as opposed to network-installs) is eluding me. My shim anaconda is included below (the original 10.2.1.5-2 anaconda has been copied to /usr/bin/anaconda.orig). My problem is the following... When running under netstg2.img the 'reset_mounts' function correctly umounts the ks.cfg defined partitions, and removes all of the anaconda-created device nodes allowing anaconda to be run a second time with a different ks.cfg. However when running under stage2.img, the 'reset_mounts' functions fails to umount the "/mnt/sysimage" partition with a "Device Busy" error. Furthermore switching to the alternative tty and issuing 'umount /mnt/sysimage' also results in a Device Busy error. I am sure the problem is a subtle different in the way stage2 operates from netstg2, but I at a loss as to what it could be. -- James -------------------------------------------- #!/usr/bin/python import os import sys # If this is a syslogd request, handle it if sys.argv[1] == '--syslogd': os.execv('/usr/bin/anaconda.orig', ('syslogd',sys.argv[1],sys.argv[2],sys.argv[3])) method = 'cdrom' def _invoke(cmdpath, args): pid = os.fork() if not pid: os.execv(cmdpath, args) os.waitpid(pid, 0) def do_anaconda(imageName, args): nargs = args[:] # .. other stuff to change arg list to load a dynamically generated ks.cfg _invoke('/usr/bin/anaconda.orig',['anaconda',]+list(nargs)) def reset_mounts(zapdevices=True): # if CD install create CDROM mount point if method == "cdrom": os.mkdir("/mnt/cdrom") # Unmount sysimage mounts ('maint','data','image0','image1' are all partitions created in ks.cfg) sysMounts = ('dev','selinux','sys','proc','boot','maint','data','image0','image1') for m in sysMounts: _invoke('/usr/bin/umount',('umount','/mnt/sysimage/%s'%m)) _invoke('/usr/bin/umount',('umount','/mnt/sysimage')) # /dev/hdc7 is provisioned as the swap partition in ks.cfg _invoke('/usr/bin/swapoff',('swapoff','/dev/hdc7')) # Drop device nodes deviceGlobs = ('/dev/hdc*','/dev/md*','dev/sda*','dev/initctl','/dev/mapper','/dev/log ','/dev/hdc*') if zapdevices: map(lambda x: _invoke('/usr/bin/rm',('rm','-rf',x)), deviceGlobs) if method == 'cdrom': _invoke('/usr/bin/mount',('mount','/tmp/cdrom','/mnt/source')) os.unlink('/tmp/cdrom') try: args = sys.argv[1:] # Run anaconda on maint do_anaconda('maint',args) # Run anaconda on image0 reset_mounts() do_anaconda('image0',args) # .. other stuff to canonicalize maint and image0, update grub # Capture image0 and duplicate to image1 reset_mounts(zapdevices=False) # .. other stuff to use partimage to copy image0 to image1 and canonicalize image1 except: print "%s %s\n%s" % (sys.exc_info()[0],sys.exc_info()[1],repr(sys.exc_info()[2])) # Done print "Press any key to reboot" sys.stdin.readline() sys.exit(0) ---------------------------------------------- -------------- next part -------------- An HTML attachment was scrubbed... URL: From Elizabeth.Brosch at thomson.com Fri Mar 17 21:49:41 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Fri, 17 Mar 2006 16:49:41 -0500 Subject: No network drivers for doing kickstart Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E5698@tshuspaphimbx01.ERF.THOMSON.COM> All: I am attempting to Kickstart some HP DL 585 servers but it fails with "no network drivers for doing Kickstart" It has Dual Port NC7782 embedded NIC. I am attempting to install RHEL AS v4 Update 2 - kernel 2.6.9-22.ELsmp Has anyone encountered this problem before? Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 -------------- next part -------------- An HTML attachment was scrubbed... URL: From jim at rossberry.com Sat Mar 18 04:33:09 2006 From: jim at rossberry.com (Jim Wildman) Date: Fri, 17 Mar 2006 23:33:09 -0500 (EST) Subject: No network drivers for doing kickstart In-Reply-To: <2213CAA79DA9094AB52517FDADEF07FA021E5698@tshuspaphimbx01.ERF.THOMSON.COM> References: <2213CAA79DA9094AB52517FDADEF07FA021E5698@tshuspaphimbx01.ERF.THOMSON.COM> Message-ID: On Fri, 17 Mar 2006 Elizabeth.Brosch at thomson.com wrote: > I am attempting to Kickstart some HP DL 585 servers but it fails with > "no network drivers for doing Kickstart" > > It has Dual Port NC7782 embedded NIC. I am attempting to install RHEL > AS v4 Update 2 - kernel 2.6.9-22.ELsmp > > Has anyone encountered this problem before? We have not seen that error in the course of installing dozens of rh3 and rh4 instances on both dl380 and dl585 machines. We have had trouble with the DHCP/PXE/tftp exchange failing due to too much lag time in the repsonse. ------------------------------------------------------------------------ Jim Wildman, CISSP, RHCE jim at rossberry.com http://www.rossberry.com "Society in every state is a blessing, but Government, even in its best state, is a necessary evil; in its worst state, an intolerable one." Thomas Paine From abelanger at bankofny.com Sat Mar 18 17:01:58 2006 From: abelanger at bankofny.com (abelanger at bankofny.com) Date: Sat, 18 Mar 2006 12:01:58 -0500 Subject: Alan Belanger/NY/DOMESTIC/BNY is out of the office. Message-ID: I will be out of the office starting 03/18/2006 and will not return until 04/03/2006. I will respond to your message when I return. Please contact Tony Totera for any USS related questions or issues ________________________________________________________________________ The information in this e-mail, and any attachment therein, is confidential and for use by the addressee only. If you are not the intended recipient, please return the e-mail to the sender and delete it from your computer. Although The Bank of New York attempts to sweep e-mail and attachments for viruses, it does not guarantee that either are virus-free and accepts no liability for any damage sustained as a result of viruses. From Trevor.Lauder at telus.com Sun Mar 19 00:10:50 2006 From: Trevor.Lauder at telus.com (Trevor Lauder) Date: Sat, 18 Mar 2006 17:10:50 -0700 Subject: Kickstart initialization over authenticated FTP Message-ID: We have been trying to find out if there is a way to pull the kickstart config file down through authenticated FTP. Running something like "linux text ks=ftp://username:password at somehost.domain/kickstart.cfg......." on the bootloader prompt doesn't work. Is there a way to pull that file down with FTP using a username / password? Anonymous FTP is not an option. Thanks, Trevor Lauder Senior Windows / Linux SA-CFT3 TELUS ITI Western Region - CIDC p 403.503.3799 c 403.816.1508 e Trevor.Lauder at TELUS.com TELUS(r) the future is friendly www.TELUS.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From phr at doc.ic.ac.uk Sun Mar 19 03:36:34 2006 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Sun, 19 Mar 2006 03:36:34 +0000 (GMT) Subject: Kickstart initialization over authenticated FTP In-Reply-To: References: Message-ID: On Sat, 18 Mar 2006, Trevor Lauder wrote: > We have been trying to find out if there is a way to pull the > kickstart config file down through authenticated FTP. Running > something like "linux text > ks=ftp://username:password at somehost.domain/kickstart.cfg......." on > the bootloader prompt doesn't work. Is there a way to pull that file > down with FTP using a username / password? Anonymous FTP is not an > option. You don't say which version of Redhat/Fedora this is, but there's certainly FTP username/password support in the current version of anaconda. See ftp.c and urls.c in the source for the gory details. Cheers, Phil From Trevor.Lauder at telus.com Sun Mar 19 03:57:29 2006 From: Trevor.Lauder at telus.com (Trevor Lauder) Date: Sat, 18 Mar 2006 20:57:29 -0700 Subject: Kickstart initialization over authenticated FTP Message-ID: Ok thank you for the information, I'll take a look at the source and see what I can get working. We are trying this on RHEL3. Cheers, Trevor Lauder Senior Windows / Linux SA-CFT3 TELUS ITI Western Region - CIDC p 403.503.3799 c 403.816.1508 e Trevor.Lauder at TELUS.com TELUS(r) the future is friendly www.TELUS.com -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Philip Rowlands Sent: Saturday, March 18, 2006 8:37 PM To: Discussion list about Kickstart Subject: Re: Kickstart initialization over authenticated FTP On Sat, 18 Mar 2006, Trevor Lauder wrote: > We have been trying to find out if there is a way to pull the > kickstart config file down through authenticated FTP. Running > something like "linux text > ks=ftp://username:password at somehost.domain/kickstart.cfg......." on > the bootloader prompt doesn't work. Is there a way to pull that file > down with FTP using a username / password? Anonymous FTP is not an > option. You don't say which version of Redhat/Fedora this is, but there's certainly FTP username/password support in the current version of anaconda. See ftp.c and urls.c in the source for the gory details. Cheers, Phil _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From abelanger at bankofny.com Sun Mar 19 17:28:57 2006 From: abelanger at bankofny.com (abelanger at bankofny.com) Date: Sun, 19 Mar 2006 12:28:57 -0500 Subject: Alan Belanger/NY/DOMESTIC/BNY is out of the office. Message-ID: I will be out of the office starting 03/18/2006 and will not return until 04/03/2006. I will respond to your message when I return. Please contact Tony Totera for any USS related questions or issues ________________________________________________________________________ The information in this e-mail, and any attachment therein, is confidential and for use by the addressee only. If you are not the intended recipient, please return the e-mail to the sender and delete it from your computer. Although The Bank of New York attempts to sweep e-mail and attachments for viruses, it does not guarantee that either are virus-free and accepts no liability for any damage sustained as a result of viruses. From cleve.cornelius at kroger.com Sun Mar 19 18:18:15 2006 From: cleve.cornelius at kroger.com (cleve.cornelius at kroger.com) Date: Sun, 19 Mar 2006 13:18:15 -0500 Subject: Cleve Cornelius/MIS/CORP/KrogerCo is out of the office. Message-ID: I will be out of the office starting Sun 03/19/2006 and will not return until Mon 03/27/2006. If this is an urgent matter please contact the NOC and they will page the on call engineer to assist you. From Elizabeth.Brosch at thomson.com Mon Mar 20 15:46:31 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Mon, 20 Mar 2006 10:46:31 -0500 Subject: No network drivers for doing kickstart Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E569C@tshuspaphimbx01.ERF.THOMSON.COM> Thanks for reply. I don't know what my issue is but I am unable to get past it. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 -----Original Message----- We have not seen that error in the course of installing dozens of rh3 and rh4 instances on both dl380 and dl585 machines. We have had trouble with the DHCP/PXE/tftp exchange failing due to too much lag time in the repsonse. ------------------------------------------------------------------------ Jim Wildman, CISSP, RHCE jim at rossberry.com http://www.rossberry.com "Society in every state is a blessing, but Government, even in its best state, is a necessary evil; in its worst state, an intolerable one." Thomas Paine _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From dxh at yahoo.com Mon Mar 20 16:28:06 2006 From: dxh at yahoo.com (Don Hoover) Date: Mon, 20 Mar 2006 08:28:06 -0800 (PST) Subject: Multi-version boot CD possible? Message-ID: <20060320162806.90625.qmail@web60724.mail.yahoo.com> Is it even possible to have a boot CD that will allow network kickstarts for multiple versions? It get really tired of having to generate new boot cd's for the admins everything we get a new update version from redhat. It would be nice if one version could be used to kick off installs for say both Update 2 and Update 3 etc. I always get the "install tree does not seem to match your boot media" error when I try do attempt it in various ways. Anyone gotten around this limitation? From tibbs at math.uh.edu Mon Mar 20 16:41:34 2006 From: tibbs at math.uh.edu (Jason L Tibbitts III) Date: Mon, 20 Mar 2006 10:41:34 -0600 Subject: Multi-version boot CD possible? In-Reply-To: <20060320162806.90625.qmail@web60724.mail.yahoo.com> (Don Hoover's message of "Mon, 20 Mar 2006 08:28:06 -0800 (PST)") References: <20060320162806.90625.qmail@web60724.mail.yahoo.com> Message-ID: >>>>> "DH" == Don Hoover writes: DH> Is it even possible to have a boot CD that will allow network DH> kickstarts for multiple versions? You can put as many boot images as you want on a CD (or USB drive). Just name the images appropriately (8.3 filenames) and configure syslinux appropriately: label ks-4 kernel vml4 append initrd=ird4.img ramdisk_size=8192 ksdevice=link ks=whatever label ks6-4 kernel vm64 append initrd=ir64.img ramdisk_size=10000 ksdevice=link ks=whatever label ks-3 kernel vml3 append initrd=ird3.img ramdisk_size=8192 ksdevice=link ks=whatever label ks6-c4 kernel vm6c4 append initrd=ir6c4.img ramdisk_size=10000 ksdevice=link ks=whatever label ks-c4 kernel vmlc4 append initrd=irdc4.img ramdisk_size=10000 ksdevice=link ks=whatever label memtest86 kernel memtest append - Of course you have to type the right thing and make sure the appropriate kickstart files are on the network. - J< From dan at id-confirm.com Mon Mar 20 17:28:39 2006 From: dan at id-confirm.com (Dan Trainor) Date: Mon, 20 Mar 2006 10:28:39 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <441761E8.8090705@id-confirm.com> References: <441761E8.8090705@id-confirm.com> Message-ID: <441EE647.1070409@id-confirm.com> Dan Trainor wrote: > Hello, all - > > I'm trying to take a few command-line arguments during Anaconda's boot > sequence for a new ISO which I am remastering, and I can't quite seem to > extract them into ks.cfg itself. > > I'm using something similar to this to extract the values, in %pre: > > if grep -i -q "ipaddr=[a-zA-Z0-9]" /proc/cmdline > then > IPADDR=`cat /proc/cmdline | sed 's/.*ipaddr=\([^ ]*\).*/\1/'` > else > IPADDR=1.2.3.4 > fi > > I then can extract ${IPADDR} from most anywhere - except my ks.cfg. > > I'm using something similar to this, to define network arguments: > > network --device=eth0 \ > --bootproto=static \ > --ip=`/usr/bin/echo ${IPADDR}` \ > --netmask=`/usr/bin/echo ${NMADDR}` \ > --gateway=`/usr/bin/echo ${GWADDR}` \ > --nameserver=`/usr/bin/echo ${NSADDR}` > > ...to no avail. I see the literal contents (IPADDR=`/usr/bin/echo ) > inside my system config scripts, instead of the extracted value being > written to these files. I've even gone so far as to get rid of my echo > statement in an attempt to get it to work, as well as just placing > ${VARNAME}, to get more familiar with how Anaconda extracts these values > trying to find my mistake. > > I'm sure what I'm doing is not overly complicated - but for now, I'm > stuck. If anyone has any feedback, I'd be very eager to hear back from > you. > > Thanks! > -dant > Hello again, everyone - I must be doing something wrong. My %pre script has the exact same layout (less the same commands) as the one provided in the example: http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-guide/s1-kickstart2-preinstallconfig.html but my kickstart is still bombing out. I am replying to my own post, in the hopes that others may see what I'm doing, and might be able to provide some feedback (moreso than you all have already posted (which I greatly appreciate, btw)). I use an include to include a file which I generate in %pre, which is used to establish some network settings. When booting Anaconda, I get a: IOError: [Error 2] No such file or directory: '/tmp/netsettings' I understand the error perfectly, but can't figure out why this file cannot be found. Perhaps I have my %pre in the wrong place, but according to the manual, anywhere after the commands section should suffice. I've tried to issue sleep commands at the end of %pre hoping to catch more of the error and to explore a bit more, but it seems that Anaconda bombs out prior to to completing my %Pre script - which leads me to believe that %pre is not parsed prior to the kickstart's initiation - not in this instance. Thanks! -dant From Chip.Shabazian at bankofamerica.com Mon Mar 20 19:15:54 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Mon, 20 Mar 2006 11:15:54 -0800 Subject: Extracting variables from within ks.cfg, interpreted from command-line Message-ID: Sounds like there are problems somewhere in the %pre that is causing it to not execute properly. I have had to debug those also, and just put my sleep commands in various places until I was able to narrow down where it was failing. Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Dan Trainor Sent: Monday, March 20, 2006 9:29 AM To: Discussion list about Kickstart Subject: Re: Extracting variables from within ks.cfg, interpreted from command-line Dan Trainor wrote: > Hello, all - > > I'm trying to take a few command-line arguments during Anaconda's boot > sequence for a new ISO which I am remastering, and I can't quite seem > to extract them into ks.cfg itself. > > I'm using something similar to this to extract the values, in %pre: > > if grep -i -q "ipaddr=[a-zA-Z0-9]" /proc/cmdline then > IPADDR=`cat /proc/cmdline | sed 's/.*ipaddr=\([^ ]*\).*/\1/'` else > IPADDR=1.2.3.4 > fi > > I then can extract ${IPADDR} from most anywhere - except my ks.cfg. > > I'm using something similar to this, to define network arguments: > > network --device=eth0 \ > --bootproto=static \ > --ip=`/usr/bin/echo ${IPADDR}` \ > --netmask=`/usr/bin/echo ${NMADDR}` \ > --gateway=`/usr/bin/echo ${GWADDR}` \ > --nameserver=`/usr/bin/echo ${NSADDR}` > > ...to no avail. I see the literal contents (IPADDR=`/usr/bin/echo ) > inside my system config scripts, instead of the extracted value being > written to these files. I've even gone so far as to get rid of my > echo statement in an attempt to get it to work, as well as just > placing ${VARNAME}, to get more familiar with how Anaconda extracts > these values trying to find my mistake. > > I'm sure what I'm doing is not overly complicated - but for now, I'm > stuck. If anyone has any feedback, I'd be very eager to hear back > from you. > > Thanks! > -dant > Hello again, everyone - I must be doing something wrong. My %pre script has the exact same layout (less the same commands) as the one provided in the example: http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-gui de/s1-kickstart2-preinstallconfig.html but my kickstart is still bombing out. I am replying to my own post, in the hopes that others may see what I'm doing, and might be able to provide some feedback (moreso than you all have already posted (which I greatly appreciate, btw)). I use an include to include a file which I generate in %pre, which is used to establish some network settings. When booting Anaconda, I get a: IOError: [Error 2] No such file or directory: '/tmp/netsettings' I understand the error perfectly, but can't figure out why this file cannot be found. Perhaps I have my %pre in the wrong place, but according to the manual, anywhere after the commands section should suffice. I've tried to issue sleep commands at the end of %pre hoping to catch more of the error and to explore a bit more, but it seems that Anaconda bombs out prior to to completing my %Pre script - which leads me to believe that %pre is not parsed prior to the kickstart's initiation - not in this instance. Thanks! -dant _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From dan at id-confirm.com Mon Mar 20 23:05:28 2006 From: dan at id-confirm.com (Dan Trainor) Date: Mon, 20 Mar 2006 16:05:28 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <4418043F.5060108@j2solutions.net> References: <441761E8.8090705@id-confirm.com><20060315005754.GA11357@exeter.boston.redhat.com> <441771AF.1000609@id-confirm.com> <003f01c647d4$00a9ba80$4201a8c0@bluebottle.com> <4417BB47.5030801@id-confirm.com> <4417C138.2070909@id-confirm.com> <4418043F.5060108@j2solutions.net> Message-ID: <441F3538.701@id-confirm.com> Jesse Keating wrote: > On 03/15/2006 Dan Trainor wrote: > >> I've pasted it at http://pastebin.com/603104 > > > I've had bad luck putting a bash script in %pre. Instead I wget a > script in %pre and then execute it. The command not founde is probably > a typo, and it is trying ot run your script but is not able to, some > part of it is missing in the anaconda env (like bash?). The line > numbers are really hard to grok when dealing w/ ks script errors. > Hello, Jesse - What kind of "bad luck" have you found? I'm still hot after this one, and it's really bothering me. Bash is available to me in %pre, yes? This stuff would even work in sh, as far as my testing shows. %pre does get parsed before my network info is used with the 'network' directive in my kickstart file, right? Thanks! -dant From dan at id-confirm.com Mon Mar 20 23:06:39 2006 From: dan at id-confirm.com (Dan Trainor) Date: Mon, 20 Mar 2006 16:06:39 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: References: <441761E8.8090705@id-confirm.com><20060315005754.GA11357@exeter.boston.redhat.com> <441771AF.1000609@id-confirm.com> <003f01c647d4$00a9ba80$4201a8c0@bluebottle.com> <4417BB47.5030801@id-confirm.com> Message-ID: <441F357F.5030704@id-confirm.com> Philip Rowlands wrote: > On Tue, 14 Mar 2006, Dan Trainor wrote: > >> Alright, so, I think something wierd is going on. >> >> I'm still seeing Anaconda bomb out saying that /tmp/netsettings does >> not exist. I can verify that /tmp/netsettings does not get created. >> However, I cannot figure out why. >> >> vty3 tells me something similar to the following: >> >> : command not founde 1: >> : command not founde 3: >> : command not founde 6: > [snip] > > This doesn't quite feel like CRLF-line-endings, but worth a check. Can > you post your kickstart file? > > > Cheers, > Phil > Hello, Phil - Were you able to see the link to the kickstart which I was using? I'm interested in hearing about your CRLF-line-ending concerns, as it seem that both my ks.cfg and ks-script files in /tmp carry a ^M after every line. It's been my experience that control characters where they do not belong, make really funny things happen. Thanks! -dant From jkeating at j2solutions.net Mon Mar 20 23:28:24 2006 From: jkeating at j2solutions.net (Jesse Keating) Date: Mon, 20 Mar 2006 15:28:24 -0800 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <441F3538.701@id-confirm.com> References: <441761E8.8090705@id-confirm.com> <4418043F.5060108@j2solutions.net> <441F3538.701@id-confirm.com> Message-ID: <200603201528.24532.jkeating@j2solutions.net> On Monday 20 March 2006 15:05, Dan Trainor wrote: > What kind of "bad luck" have you found? ?I'm still hot after this one, > and it's really bothering me. > > Bash is available to me in %pre, yes? ?This stuff would even work in sh, > as far as my testing shows. > > %pre does get parsed before my network info is used with the 'network' > directive in my kickstart file, right? So what I ended up doing was using python. Python is there because anaconda uses it. %pre can use the network as supplied on how to get the ks file or the content itself. Not necessarily the same network info supplied in the 'network' section as that is used for the installed system, not for the installer. I personally do PXE booting and get the ks config file from a webserver, so the environment already has network going. I pull down a pre script from a webserver (use IP as DNS does not work in install env), then execute the python prescript to parse /proc/cmdline for arguments that will lead to stub files getting created. Later in the kickstart script those stub files are included for the various things. Here is a really ugly early version of this script. I can't find the latest one I was responsible for, but here ya go: http://geek.j2solutions.net/stuff/autokickrhel4.py -- Jesse Keating RHCE (geek.j2solutions.net) Fedora Legacy Team (www.fedoralegacy.org) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 191 bytes Desc: not available URL: From dan at id-confirm.com Mon Mar 20 23:36:08 2006 From: dan at id-confirm.com (Dan Trainor) Date: Mon, 20 Mar 2006 16:36:08 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <200603201528.24532.jkeating@j2solutions.net> References: <441761E8.8090705@id-confirm.com> <4418043F.5060108@j2solutions.net> <441F3538.701@id-confirm.com> <200603201528.24532.jkeating@j2solutions.net> Message-ID: <441F3C68.8070108@id-confirm.com> Jesse Keating wrote: > On Monday 20 March 2006 15:05, Dan Trainor wrote: >> What kind of "bad luck" have you found? I'm still hot after this one, >> and it's really bothering me. >> >> Bash is available to me in %pre, yes? This stuff would even work in sh, >> as far as my testing shows. >> >> %pre does get parsed before my network info is used with the 'network' >> directive in my kickstart file, right? > > So what I ended up doing was using python. Python is there because anaconda > uses it. %pre can use the network as supplied on how to get the ks file or > the content itself. Not necessarily the same network info supplied in the > 'network' section as that is used for the installed system, not for the > installer. I personally do PXE booting and get the ks config file from a > webserver, so the environment already has network going. I pull down a pre > script from a webserver (use IP as DNS does not work in install env), then > execute the python prescript to parse /proc/cmdline for arguments that will > lead to stub files getting created. Later in the kickstart script those stub > files are included for the various things. > > Here is a really ugly early version of this script. I can't find the latest > one I was responsible for, but here ya go: > > http://geek.j2solutions.net/stuff/autokickrhel4.py > Hello, Jesse - Alright, you're going about the same route as I am - sorta. As much as I'd like to boot off of the network, that's not an option in this particular case. Since this is not working, I'm just going to forget the idea and make a really, really ugly hack to establish all of this in %post, and make it work as the machine boots the next time. Ultimately that's all I need. I really appreciate all the input that you all have given. Although I've not figured out how to do what I originally wanted, I learned a lot of other things. Thanks again for all your time -dant From phb at llgc.org.uk Tue Mar 21 10:21:58 2006 From: phb at llgc.org.uk (Phil Bettinson) Date: Tue, 21 Mar 2006 10:21:58 +0000 Subject: No network drivers for doing kickstart In-Reply-To: <2213CAA79DA9094AB52517FDADEF07FA021E569C@tshuspaphimbx01.ERF.THOMSON.COM> References: <2213CAA79DA9094AB52517FDADEF07FA021E569C@tshuspaphimbx01.ERF.THOMSON.COM> Message-ID: <441FD3C6.3020606@llgc.org.uk> Morning, Your issue might possibly be because of the switch that the machines are connected too. Some (newer?) switches do some negotiations and speed tests before clearing the port for use, this can cause the DHCP to time out (as the network request doesn't get further than the port). I've had this problem on a few machines here, but our network engineer simply removed or decreased the timing for the port (sorry for the vague answer, but I havn't seen the operation myself yet), and that seemed to solve the problem Hope that Helps, Phil Elizabeth.Brosch at thomson.com wrote: > Thanks for reply. I don't know what my issue is but I am unable to get > past it. > > Liz Brosch > TSH System Services > Philadelphia, PA 19104 > office: (215) 386-0100 x1144 > cell: (267) 784-9166 > -----Original Message----- > > > We have not seen that error in the course of installing dozens of rh3 > and > rh4 instances on both dl380 and dl585 machines. We have had trouble > with the DHCP/PXE/tftp exchange failing due to too much lag time in the > repsonse. > > ------------------------------------------------------------------------ > Jim Wildman, CISSP, RHCE jim at rossberry.com > http://www.rossberry.com > "Society in every state is a blessing, but Government, even in its best > state, is a necessary evil; in its worst state, an intolerable one." > Thomas Paine > > _______________________________________________ > 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 -- ----------------------------------------------------------------------- Phil Bettinson Swyddog Gweinyddu Systemau Systems Administrator Isadran Cyfrifiaduron Computer Section Llyfrgell Genedlaethol Cymru National Library of Wales Aberystwyth, Aberystwyth Ceredigion, Ceredigion SY23 3BU SY23 3BU Ff?n / Phone: 01970 632867 Ffacs / Fax: 01970 615709 EBost / Email: phb at llgc.org.uk We / Web: http://www.llgc.org.uk ----------------------------------------------------------------------- Yn siarad drosof fy hun, nid LlGC - Speaking personally, not for NLW From briandlong at gmail.com Tue Mar 21 13:49:26 2006 From: briandlong at gmail.com (Brian Long) Date: Tue, 21 Mar 2006 08:49:26 -0500 Subject: RHEL 3 and "nostorage"? Message-ID: <1c59d60b0603210549g37668a0euda267644fb7a05d9@mail.gmail.com> On my RHEL 4 kickstarts, I specify nostorage on the command line and "device scsi cciss" in ks.cfg. This ensures no SAN-connected devices will be seen by Anaconda and get accidentally overwritten. Has anyone duplicated this behavior on RHEL 3? I looked at Anaconda source for U5 and U6 and nostorage is not present. I believe the only similar option would be "noprobe" and then specify device eth and scsi in ks.cfg. I'm not very familiar with noprobe; will the kickstart go interactive for other options as well? We already specify the monitor in ks.cfg, so that should be fine. Thanks for any pointers on how to emulate nostorage in RHEL 3. /Brian/ -------------- next part -------------- An HTML attachment was scrubbed... URL: From phr at doc.ic.ac.uk Tue Mar 21 15:02:03 2006 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 21 Mar 2006 15:02:03 +0000 (GMT) Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: <441F357F.5030704@id-confirm.com> References: <441761E8.8090705@id-confirm.com><20060315005754.GA11357@exeter.boston.redhat.com> <441771AF.1000609@id-confirm.com> <003f01c647d4$00a9ba80$4201a8c0@bluebottle.com> <4417BB47.5030801@id-confirm.com> <441F357F.5030704@id-confirm.com> Message-ID: On Mon, 20 Mar 2006, Dan Trainor wrote: > Were you able to see the link to the kickstart which I was using? Yes, but I wouldn't be surprised if the pastebin site auto-corrects line endings to make text files readable. > I'm interested in hearing about your CRLF-line-ending concerns, as it seem > that both my ks.cfg and ks-script files in /tmp carry a ^M after every line. > It's been my experience that control characters where they do not belong, make > really funny things happen. Shell scripts do not like CRLF line endings. Are you using a Windows editor? Did you try running everything through dos2unix? Cheers, Phil From phb at llgc.org.uk Tue Mar 21 15:04:13 2006 From: phb at llgc.org.uk (Phil Bettinson) Date: Tue, 21 Mar 2006 15:04:13 +0000 Subject: No network drivers for doing kickstart In-Reply-To: <441FD3C6.3020606@llgc.org.uk> References: <2213CAA79DA9094AB52517FDADEF07FA021E569C@tshuspaphimbx01.ERF.THOMSON.COM> <441FD3C6.3020606@llgc.org.uk> Message-ID: <442015ED.9090709@llgc.org.uk> Hello again, I have also experienced a similar problem, this seems to be with a problem (for me at least) with the tg3 driver with Redhat As 3 update 6, a problem we didn't have (and don't have) if we just use Redhat AS 3 (with no updates). This seems to be more of a problem with the vmlinuz and the initrd for this release. As we didn't want ot re-build these, we went back to using RedHat AS 3, which works. I note, however that in RedHat AS 3 update 7, there has been an update to this driver. I'm busy downloading the Isos so I can give it a go with that, and see if it works. Phil (again) Phil Bettinson wrote: > Morning, > > Your issue might possibly be because of the switch that the machines are > connected too. Some (newer?) switches do some negotiations and speed > tests before clearing the port for use, this can cause the DHCP to time > out (as the network request doesn't get further than the port). I've had > this problem on a few machines here, but our network engineer simply > removed or decreased the timing for the port (sorry for the vague > answer, but I havn't seen the operation myself yet), and that seemed to > solve the problem > > Hope that Helps, > > Phil > > Elizabeth.Brosch at thomson.com wrote: > >> Thanks for reply. I don't know what my issue is but I am unable to get >> past it. >> Liz Brosch >> TSH System Services >> Philadelphia, PA 19104 >> office: (215) 386-0100 x1144 >> cell: (267) 784-9166 >> -----Original Message----- >> >> >> We have not seen that error in the course of installing dozens of rh3 >> and rh4 instances on both dl380 and dl585 machines. We have had trouble >> with the DHCP/PXE/tftp exchange failing due to too much lag time in the >> repsonse. >> >> ------------------------------------------------------------------------ >> Jim Wildman, CISSP, RHCE jim at rossberry.com >> http://www.rossberry.com >> "Society in every state is a blessing, but Government, even in its best >> state, is a necessary evil; in its worst state, an intolerable one." >> Thomas Paine >> >> _______________________________________________ >> 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 > > -- ----------------------------------------------------------------------- Phil Bettinson Swyddog Gweinyddu Systemau Systems Administrator Isadran Cyfrifiaduron Computer Section Llyfrgell Genedlaethol Cymru National Library of Wales Aberystwyth, Aberystwyth Ceredigion, Ceredigion SY23 3BU SY23 3BU Ff?n / Phone: 01970 632867 Ffacs / Fax: 01970 615709 EBost / Email: phb at llgc.org.uk We / Web: http://www.llgc.org.uk ----------------------------------------------------------------------- Yn siarad drosof fy hun, nid LlGC - Speaking personally, not for NLW From demel at zid.tuwien.ac.at Tue Mar 21 15:19:25 2006 From: demel at zid.tuwien.ac.at (Johannes Demel) Date: Tue, 21 Mar 2006 16:19:25 +0100 Subject: xconfig statement with FC5 Message-ID: <010601c64cfa$da0bcfe0$0e22a8c0@kom.tuwien.ac.at> During Kickstart installation with FC4, I used the following Line to configure X-window and start it up after booting (on a DL385 system). xconfig --videocard "ATI Mach64" --resolution 1024x768 --depth 16 --startxonboot --defaultdesktop kde with Fedore Core 5 this allways bomps during the installation (somethins like rphlx.videocard missing). Without the xconfig statement all wents fine, but X does not come up after the boot (only runlevel 3). I tried xconfig --startxonboot --defaultdesktop=KDE What I want is: Configure X to use KDE by default and it should come up automatically after a reboot. -- Johannes Demel demel at zid.tuwien.ac.at Johannes.Demel at tuwien.ac.at Information Technology Services, Head of Communication Group Vienna University of Technology, Austria Wiedner Hauptstrasse 8-10/020, A 1040 Wien Tel: ++43(1)58801-42040 Fax: ++43(1)58801-42099 From Elizabeth.Brosch at thomson.com Tue Mar 21 15:27:06 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Tue, 21 Mar 2006 10:27:06 -0500 Subject: No network drivers for doing kickstart Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E56B7@tshuspaphimbx01.ERF.THOMSON.COM> Thanks Phil for the reply. I am using RHEL 4 Update 2. I was thinking about going to either Update 3 or reverting back to RHEL 4 w/out updates. I began to think that the vmlinuz and the initrd for this release was missing something that is causing my problems. I don't suspect it has anything to do with timing (switches, etc) but I could be wrong. I am going to attempt Kickstart with another release and see if that solves my problem. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Phil Bettinson Sent: Tuesday, March 21, 2006 10:04 AM To: Discussion list about Kickstart Subject: Re: No network drivers for doing kickstart Hello again, I have also experienced a similar problem, this seems to be with a problem (for me at least) with the tg3 driver with Redhat As 3 update 6, a problem we didn't have (and don't have) if we just use Redhat AS 3 (with no updates). This seems to be more of a problem with the vmlinuz and the initrd for this release. As we didn't want ot re-build these, we went back to using RedHat AS 3, which works. I note, however that in RedHat AS 3 update 7, there has been an update to this driver. I'm busy downloading the Isos so I can give it a go with that, and see if it works. Phil (again) Phil Bettinson wrote: > Morning, > > Your issue might possibly be because of the switch that the machines are > connected too. Some (newer?) switches do some negotiations and speed > tests before clearing the port for use, this can cause the DHCP to time > out (as the network request doesn't get further than the port). I've had > this problem on a few machines here, but our network engineer simply > removed or decreased the timing for the port (sorry for the vague > answer, but I havn't seen the operation myself yet), and that seemed to > solve the problem > > Hope that Helps, > > Phil > > Elizabeth.Brosch at thomson.com wrote: > >> Thanks for reply. I don't know what my issue is but I am unable to get >> past it. >> Liz Brosch >> TSH System Services >> Philadelphia, PA 19104 >> office: (215) 386-0100 x1144 >> cell: (267) 784-9166 >> -----Original Message----- >> >> >> We have not seen that error in the course of installing dozens of rh3 >> and rh4 instances on both dl380 and dl585 machines. We have had trouble >> with the DHCP/PXE/tftp exchange failing due to too much lag time in the >> repsonse. >> >> ------------------------------------------------------------------------ >> Jim Wildman, CISSP, RHCE jim at rossberry.com >> http://www.rossberry.com >> "Society in every state is a blessing, but Government, even in its best >> state, is a necessary evil; in its worst state, an intolerable one." >> Thomas Paine >> >> _______________________________________________ >> 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 > > -- ----------------------------------------------------------------------- Phil Bettinson Swyddog Gweinyddu Systemau Systems Administrator Isadran Cyfrifiaduron Computer Section Llyfrgell Genedlaethol Cymru National Library of Wales Aberystwyth, Aberystwyth Ceredigion, Ceredigion SY23 3BU SY23 3BU Ff?n / Phone: 01970 632867 Ffacs / Fax: 01970 615709 EBost / Email: phb at llgc.org.uk We / Web: http://www.llgc.org.uk ----------------------------------------------------------------------- Yn siarad drosof fy hun, nid LlGC - Speaking personally, not for NLW _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From clumens at redhat.com Tue Mar 21 16:08:38 2006 From: clumens at redhat.com (Chris Lumens) Date: Tue, 21 Mar 2006 11:08:38 -0500 Subject: xconfig statement with FC5 In-Reply-To: <010601c64cfa$da0bcfe0$0e22a8c0@kom.tuwien.ac.at> References: <010601c64cfa$da0bcfe0$0e22a8c0@kom.tuwien.ac.at> Message-ID: <20060321160838.GD3137@exeter.boston.redhat.com> > During Kickstart installation with FC4, I used the following Line to > configure X-window and start it up after booting (on a DL385 system). > > xconfig --videocard "ATI Mach64" --resolution 1024x768 --depth 16 > --startxonboot --defaultdesktop kde > > with Fedore Core 5 this allways bomps during the installation (somethins > like rphlx.videocard missing). > > Without the xconfig statement all wents fine, but X does not come up after > the boot > (only runlevel 3). > I tried > xconfig --startxonboot --defaultdesktop=KDE > > What I want is: > Configure X to use KDE by default and it should come up automatically after > a reboot. Is this a text kickstart install? - Chris From demel at zid.tuwien.ac.at Tue Mar 21 17:10:22 2006 From: demel at zid.tuwien.ac.at (Johannes Demel) Date: Tue, 21 Mar 2006 18:10:22 +0100 Subject: xconfig statement with FC5 References: <010601c64cfa$da0bcfe0$0e22a8c0@kom.tuwien.ac.at> Message-ID: <013e01c64d0a$5a287390$0e22a8c0@kom.tuwien.ac.at> yes, it is a text kickstart here the main part of my kickstart file: =============== lang en_US keyboard us xconfig --videocard "ATI Mach64" --resolution 1024x768 --depth 16 --startxonboot --defaultdesktop kde install url --url http://pxe.kom.tuwien.ac.at/tftpboot/fedora/core/5/x86_64/os/ timezone Europe/Vienna network --device=eth0 --bootproto=static --ip=128.131.34.31 --netmask=255.255.255.0 --gateway=128.131.34.1 --nameserver=128.13 0.2.3 --hostname=frisko.kom.tuwien.ac.at %include common/70GB firewall --enabled --ssh --http --port=5901:tcp --port=10000:tcp selinux --permissive auth --enablemd5 --useshadow #skipx text %packages %include common/standard-packages5 %include common/additional-rpms5 %include common/standard-removes5 ========== - Johannes ---- Original Message ---- Date: Tue, 21 Mar 2006 11:08:38 -0500 From: Chris Lumens Subject: Re: xconfig statement with FC5 To: kickstart-list at redhat.com Message-ID: <20060321160838.GD3137 at exeter.boston.redhat.com> Content-Type: text/plain; charset=us-ascii > During Kickstart installation with FC4, I used the following Line to > configure X-window and start it up after booting (on a DL385 system). > > xconfig --videocard "ATI Mach64" --resolution 1024x768 --depth 16 --startxonboot --defaultdesktop kde > > with Fedore Core 5 this allways bomps during the installation (somethins like rphlx.videocard missing). > > Without the xconfig statement all wents fine, but X does not come up after the boot > (only runlevel 3). I tried xconfig --startxonboot --defaultdesktop=KDE > > What I want is: > Configure X to use KDE by default and it should come up automatically after a reboot. Is this a text kickstart install? - Chris ------------------ ---- Johannes Demel demel at zid.tuwien.ac.at Johannes.Demel at tuwien.ac.at From Chip.Shabazian at bankofamerica.com Tue Mar 21 18:26:35 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Tue, 21 Mar 2006 10:26:35 -0800 Subject: No network drivers for doing kickstart Message-ID: I have had issues in the past where a newer initrd worked, but for many reasons, we had to stay with the build based off an older version. In these cases, I found that you could replace the loader in the older initrd with the loader from the newer one, and all was fine. If you still want to build with the latest to minimize packages that need to be upgraded, you could give this a shot. Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Elizabeth.Brosch at thomson.com Sent: Tuesday, March 21, 2006 7:27 AM To: kickstart-list at redhat.com Subject: RE: No network drivers for doing kickstart Thanks Phil for the reply. I am using RHEL 4 Update 2. I was thinking about going to either Update 3 or reverting back to RHEL 4 w/out updates. I began to think that the vmlinuz and the initrd for this release was missing something that is causing my problems. I don't suspect it has anything to do with timing (switches, etc) but I could be wrong. I am going to attempt Kickstart with another release and see if that solves my problem. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Phil Bettinson Sent: Tuesday, March 21, 2006 10:04 AM To: Discussion list about Kickstart Subject: Re: No network drivers for doing kickstart Hello again, I have also experienced a similar problem, this seems to be with a problem (for me at least) with the tg3 driver with Redhat As 3 update 6, a problem we didn't have (and don't have) if we just use Redhat AS 3 (with no updates). This seems to be more of a problem with the vmlinuz and the initrd for this release. As we didn't want ot re-build these, we went back to using RedHat AS 3, which works. I note, however that in RedHat AS 3 update 7, there has been an update to this driver. I'm busy downloading the Isos so I can give it a go with that, and see if it works. Phil (again) Phil Bettinson wrote: > Morning, > > Your issue might possibly be because of the switch that the machines > are connected too. Some (newer?) switches do some negotiations and > speed tests before clearing the port for use, this can cause the DHCP > to time out (as the network request doesn't get further than the > port). I've had this problem on a few machines here, but our network > engineer simply removed or decreased the timing for the port (sorry > for the vague answer, but I havn't seen the operation myself yet), and > that seemed to solve the problem > > Hope that Helps, > > Phil > > Elizabeth.Brosch at thomson.com wrote: > >> Thanks for reply. I don't know what my issue is but I am unable to >> get past it. >> Liz Brosch >> TSH System Services >> Philadelphia, PA 19104 >> office: (215) 386-0100 x1144 >> cell: (267) 784-9166 >> -----Original Message----- >> >> >> We have not seen that error in the course of installing dozens of rh3 >> and rh4 instances on both dl380 and dl585 machines. We have had >> trouble with the DHCP/PXE/tftp exchange failing due to too much lag >> time in the repsonse. >> >> ------------------------------------------------------------------------ >> Jim Wildman, CISSP, RHCE jim at rossberry.com >> http://www.rossberry.com >> "Society in every state is a blessing, but Government, even in its >> best state, is a necessary evil; in its worst state, an intolerable one." >> Thomas Paine >> >> _______________________________________________ >> 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 > > -- ----------------------------------------------------------------------- Phil Bettinson Swyddog Gweinyddu Systemau Systems Administrator Isadran Cyfrifiaduron Computer Section Llyfrgell Genedlaethol Cymru National Library of Wales Aberystwyth, Aberystwyth Ceredigion, Ceredigion SY23 3BU SY23 3BU Ff?n / Phone: 01970 632867 Ffacs / Fax: 01970 615709 EBost / Email: phb at llgc.org.uk We / Web: http://www.llgc.org.uk ----------------------------------------------------------------------- Yn siarad drosof fy hun, nid LlGC - Speaking personally, not for NLW _______________________________________________ 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 Chip.Shabazian at bankofamerica.com Tue Mar 21 18:47:39 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Tue, 21 Mar 2006 10:47:39 -0800 Subject: RHEL 3 and "nostorage"? Message-ID: Have you tried using latefcload? This first appeared in RHEL3 U6 ________________________________ From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Brian Long Sent: Tuesday, March 21, 2006 5:49 AM To: kickstart-list at redhat.com Subject: RHEL 3 and "nostorage"? On my RHEL 4 kickstarts, I specify nostorage on the command line and "device scsi cciss" in ks.cfg. This ensures no SAN-connected devices will be seen by Anaconda and get accidentally overwritten. Has anyone duplicated this behavior on RHEL 3? I looked at Anaconda source for U5 and U6 and nostorage is not present. I believe the only similar option would be "noprobe" and then specify device eth and scsi in ks.cfg. I'm not very familiar with noprobe; will the kickstart go interactive for other options as well? We already specify the monitor in ks.cfg, so that should be fine. Thanks for any pointers on how to emulate nostorage in RHEL 3. /Brian/ -------------- next part -------------- An HTML attachment was scrubbed... URL: From Elizabeth.Brosch at thomson.com Tue Mar 21 19:42:29 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Tue, 21 Mar 2006 14:42:29 -0500 Subject: No network drivers for doing kickstart Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E56C6@tshuspaphimbx01.ERF.THOMSON.COM> Thanks very much. I will try that. Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Shabazian, Chip Sent: Tuesday, March 21, 2006 1:27 PM To: Discussion list about Kickstart Subject: RE: No network drivers for doing kickstart I have had issues in the past where a newer initrd worked, but for many reasons, we had to stay with the build based off an older version. In these cases, I found that you could replace the loader in the older initrd with the loader from the newer one, and all was fine. If you still want to build with the latest to minimize packages that need to be upgraded, you could give this a shot. Chip From Gordon.Keegan at FMR.COM Tue Mar 21 22:03:40 2006 From: Gordon.Keegan at FMR.COM (Keegan, Gordon) Date: Tue, 21 Mar 2006 17:03:40 -0500 Subject: RHEL 3 and "nostorage"? Message-ID: <942A778F73A72141889C1F53A8EAA1C53A6061@MSGMMKCLM2WIN.DMN1.FMR.COM> I don't know if it's exactly what you're looking for, but another RHEL 4 install directive is "latefcload" which tells anaconda to scan any Fibre devices last, which insures that any local scsi disks are the first in the list (ie. /dev/sda is a local scsi device, not a San lun, if both are present.) It's been back-ported to RHEL 3 and is included in U6. If you need to make San storage completely invisible during the install, one option is to unroll the install media and in the initrd.img file in /modules/pcitable file, change any occurrences of "qla2xxx" to "ignore" (ie. qla2100, qla2200, qla2300...) or whatever the device is for your HBA's. It's kind of a pain, but should prevent them from being seen at all during the initial boot/install, but will load the appropriate drivers for the running OS during the install. Gordon -----Original Message----- From: Brian Long [mailto:briandlong at gmail.com] Sent: Tuesday, March 21, 2006 8:49 AM To: kickstart-list at redhat.com Subject: RHEL 3 and "nostorage"? On my RHEL 4 kickstarts, I specify nostorage on the command line and "device scsi cciss" in ks.cfg. This ensures no SAN-connected devices will be seen by Anaconda and get accidentally overwritten. Has anyone duplicated this behavior on RHEL 3? I looked at Anaconda source for U5 and U6 and nostorage is not present. I believe the only similar option would be "noprobe" and then specify device eth and scsi in ks.cfg. I'm not very familiar with noprobe; will the kickstart go interactive for other options as well? We already specify the monitor in ks.cfg, so that should be fine. Thanks for any pointers on how to emulate nostorage in RHEL 3. /Brian/ From error27 at gmail.com Tue Mar 21 22:07:03 2006 From: error27 at gmail.com (Dan Carpenter) Date: Tue, 21 Mar 2006 14:07:03 -0800 Subject: No network drivers for doing kickstart In-Reply-To: <2213CAA79DA9094AB52517FDADEF07FA021E56B7@tshuspaphimbx01.ERF.THOMSON.COM> References: <2213CAA79DA9094AB52517FDADEF07FA021E56B7@tshuspaphimbx01.ERF.THOMSON.COM> Message-ID: On 3/21/06, Elizabeth.Brosch at thomson.com wrote: > Thanks Phil for the reply. I am using RHEL 4 Update 2. I was thinking about going to either Update 3 or reverting back to RHEL 4 w/out updates. I began to think that the vmlinuz and the initrd for this release was missing something that is causing my problems. I don't suspect it has anything to do with timing (switches, etc) but I could be wrong. I am going to attempt Kickstart with another release and see if that solves my problem. > 1) If you install using CDROM does the network work by default? 2) Which driver does it use? 3) When you boot over the network does it say that it is loading the correct driver on tty3? 4) When you boot over the network does the kernel print any error messages on tty4? regards, dan carpenter From dan at id-confirm.com Wed Mar 22 07:04:51 2006 From: dan at id-confirm.com (Dan Trainor) Date: Wed, 22 Mar 2006 00:04:51 -0700 Subject: Execution of commands in %post Message-ID: <4420F713.2060406@id-confirm.com> Hello, all - In %post, I have a long command which is run. I'm trying to use env vars which are set previously in %post to use as arguments to this command. However, the variables are not expanded, i.e. I see literal values of ${MYVARIABLE} in my debugging, instead of the variable value. I've tried to even go as far as using an `echo ${MYVARIABLE}` in place of just ${MYVARIABLE} to see if that yields any other results, to no avail. I see something interesting before the command is run. Something similar to this: CHDIR path /var/lib/mysql/config CHDIR path /var/lib/mysql/config/myfile Am I being told that Anaconda is literally changing directories to the one specified? The command *is* a bit odd, in the fact that I specify alsbolute paths to each file involved. It'd be the equiv of: /bin/grep string /path/to/my/file I don't see how this would be any different, unless Anaconda treats the third argument as an actual command. I think, actually, this might be what is happening. Can anyone identify this? Thanks! -- - Dan Trainor - id-Confirm, Inc. - Direct: 720.241.5580 From dan at id-confirm.com Wed Mar 22 07:41:31 2006 From: dan at id-confirm.com (Dan Trainor) Date: Wed, 22 Mar 2006 00:41:31 -0700 Subject: Execution of commands in %post In-Reply-To: <4420F713.2060406@id-confirm.com> References: <4420F713.2060406@id-confirm.com> Message-ID: <4420FFAB.4040109@id-confirm.com> Dan Trainor wrote: > Hello, all - > > In %post, I have a long command which is run. I'm trying to use env > vars which are set previously in %post to use as arguments to this > command. However, the variables are not expanded, i.e. I see literal > values of ${MYVARIABLE} in my debugging, instead of the variable value. > > I've tried to even go as far as using an `echo ${MYVARIABLE}` in place > of just ${MYVARIABLE} to see if that yields any other results, to no avail. > > I see something interesting before the command is run. Something > similar to this: > > CHDIR path /var/lib/mysql/config > CHDIR path /var/lib/mysql/config/myfile > > Am I being told that Anaconda is literally changing directories to the > one specified? The command *is* a bit odd, in the fact that I specify > alsbolute paths to each file involved. It'd be the equiv of: > > /bin/grep string /path/to/my/file > > I don't see how this would be any different, unless Anaconda treats the > third argument as an actual command. I think, actually, this might be > what is happening. > > Can anyone identify this? > > Thanks! > Hi - It's late. How about I also provide an example of the kind of command which I'm trying to execute: /path/to/command write /path/to/destination/file 3301 100 $SLSQLUID STRING1 $SQLUSER $SQLPASS 0 3306 127.0.0.1 theDatabase I've tried most every variation of a one-liner which I can think of. Calling the variables ${VARIABLE}, doing `echo ${VARIABLE}`, and so on. I can't get them to print their values. I can only get them to print their literal values. Thanks! -- - Dan Trainor - id-Confirm, Inc. - Direct: 720.241.5580 From jkeating at j2solutions.net Wed Mar 22 08:00:28 2006 From: jkeating at j2solutions.net (Jesse Keating) Date: Wed, 22 Mar 2006 00:00:28 -0800 Subject: Execution of commands in %post In-Reply-To: <4420FFAB.4040109@id-confirm.com> References: <4420F713.2060406@id-confirm.com> <4420FFAB.4040109@id-confirm.com> Message-ID: <200603220000.33308.jkeating@j2solutions.net> On Tuesday 21 March 2006 23:41, Dan Trainor wrote: > Hi - > > It's late. ?How about I also provide an example of the kind of command > which I'm trying to execute: > > /path/to/command write /path/to/destination/file 3301 100 $SLSQLUID > STRING1 $SQLUSER $SQLPASS 0 3306 127.0.0.1 theDatabase > > I've tried most every variation of a one-liner which I can think of. > Calling the variables ${VARIABLE}, doing `echo ${VARIABLE}`, and so on. > ? I can't get them to print their values. ?I can only get them to print > their literal values. It would be best to put your commands into a bash script and call this script from %post. wget it from a web address (no DNS use IP) and then execute it. -- Jesse Keating RHCE (geek.j2solutions.net) Fedora Legacy Team (www.fedoralegacy.org) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 191 bytes Desc: not available URL: From dan at id-confirm.com Wed Mar 22 08:05:01 2006 From: dan at id-confirm.com (Dan Trainor) Date: Wed, 22 Mar 2006 01:05:01 -0700 Subject: Execution of commands in %post In-Reply-To: <200603220000.33308.jkeating@j2solutions.net> References: <4420F713.2060406@id-confirm.com> <4420FFAB.4040109@id-confirm.com> <200603220000.33308.jkeating@j2solutions.net> Message-ID: <4421052D.1080807@id-confirm.com> Jesse Keating wrote: > On Tuesday 21 March 2006 23:41, Dan Trainor wrote: >> Hi - >> >> It's late. How about I also provide an example of the kind of command >> which I'm trying to execute: >> >> /path/to/command write /path/to/destination/file 3301 100 $SLSQLUID >> STRING1 $SQLUSER $SQLPASS 0 3306 127.0.0.1 theDatabase >> >> I've tried most every variation of a one-liner which I can think of. >> Calling the variables ${VARIABLE}, doing `echo ${VARIABLE}`, and so on. >> I can't get them to print their values. I can only get them to print >> their literal values. > > It would be best to put your commands into a bash script and call this script > from %post. wget it from a web address (no DNS use IP) and then execute it. > Good morning, Jesse - I completely agree. However, for these installs, I cannot assume that this install will have connectivity to grab such a script. That's why I'd like to do it all from %post in this manner. Like I said, everything else works, except this "special" line. Thanks -- - Dan Trainor - id-Confirm, Inc. - Direct: 720.241.5580 From jkeating at j2solutions.net Wed Mar 22 08:15:39 2006 From: jkeating at j2solutions.net (Jesse Keating) Date: Wed, 22 Mar 2006 00:15:39 -0800 Subject: Execution of commands in %post In-Reply-To: <4421052D.1080807@id-confirm.com> References: <4420F713.2060406@id-confirm.com> <200603220000.33308.jkeating@j2solutions.net> <4421052D.1080807@id-confirm.com> Message-ID: <200603220015.40080.jkeating@j2solutions.net> On Wednesday 22 March 2006 00:05, Dan Trainor wrote: > I completely agree. ?However, for these installs, I cannot assume that > this install will have connectivity to grab such a script. ?That's why > I'd like to do it all from %post in this manner. > > Like I said, everything else works, except this "special" line. Are you specifying the parser for the %post section? You can specify bash, or python or whathaveyou. It might be best to specify such a parser so that the content gets used in the correct way. -- Jesse Keating RHCE (geek.j2solutions.net) Fedora Legacy Team (www.fedoralegacy.org) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 191 bytes Desc: not available URL: From dan at id-confirm.com Wed Mar 22 08:16:39 2006 From: dan at id-confirm.com (Dan Trainor) Date: Wed, 22 Mar 2006 01:16:39 -0700 Subject: Execution of commands in %post In-Reply-To: <200603220015.40080.jkeating@j2solutions.net> References: <4420F713.2060406@id-confirm.com> <200603220000.33308.jkeating@j2solutions.net> <4421052D.1080807@id-confirm.com> <200603220015.40080.jkeating@j2solutions.net> Message-ID: <442107E7.1040709@id-confirm.com> Jesse Keating wrote: > On Wednesday 22 March 2006 00:05, Dan Trainor wrote: >> I completely agree. However, for these installs, I cannot assume that >> this install will have connectivity to grab such a script. That's why >> I'd like to do it all from %post in this manner. >> >> Like I said, everything else works, except this "special" line. > > Are you specifying the parser for the %post section? You can specify bash, or > python or whathaveyou. It might be best to specify such a parser so that the > content gets used in the correct way. > Hello again, Jesse - I am not specifying a parser. As I've said, I've had my entire kickstart work up until this point. I don't think I'll plan on switching parsers at this point. I'm too stubborn, and have high hopes to a solution to this kind of problem, haha. Thanks -- - Dan Trainor - id-Confirm, Inc. - Direct: 720.241.5580 From vlad.agranovsky at gmail.com Wed Mar 22 08:19:09 2006 From: vlad.agranovsky at gmail.com (vlad.agranovsky at gmail.com) Date: Wed, 22 Mar 2006 10:19:09 +0200 Subject: Execution of commands in %post In-Reply-To: <4421052D.1080807@id-confirm.com> Message-ID: <44210887.754f1c76.6d14.ffffd77b@mx.gmail.com> Dan, Please post your ks.cfg and, I believe, we will instantly find the bugs :) If it's large, you may post it to http://pastebin.com and give us the URL. Thanks. Vlad From dan at id-confirm.com Wed Mar 22 08:22:59 2006 From: dan at id-confirm.com (Dan Trainor) Date: Wed, 22 Mar 2006 01:22:59 -0700 Subject: Execution of commands in %post In-Reply-To: <44210887.754f1c76.6d14.ffffd77b@mx.gmail.com> References: <44210887.754f1c76.6d14.ffffd77b@mx.gmail.com> Message-ID: <44210963.5030803@id-confirm.com> vlad.agranovsky at gmail.com wrote: > Dan, > > Please post your ks.cfg and, I believe, we will instantly find the bugs :) > > If it's large, you may post it to http://pastebin.com and give us the URL. > > Thanks. > Vlad Good morning, Vlad - I just resolved across my bug. I was expecting to see some different data than what I was presented with. It turns out that this application required the presence of a directory, which was not there. This caused erroneous results. Kinda sent me off looking for the wrong problem. So I've been chasing the wrong bug - however, I greatly appreciate the help of all of you. I learn something every time. Thanks -- - Dan Trainor - id-Confirm, Inc. - Direct: 720.241.5580 From clumens at redhat.com Wed Mar 22 15:31:19 2006 From: clumens at redhat.com (Chris Lumens) Date: Wed, 22 Mar 2006 10:31:19 -0500 Subject: xconfig statement with FC5 In-Reply-To: <013e01c64d0a$5a287390$0e22a8c0@kom.tuwien.ac.at> References: <010601c64cfa$da0bcfe0$0e22a8c0@kom.tuwien.ac.at> <013e01c64d0a$5a287390$0e22a8c0@kom.tuwien.ac.at> Message-ID: <20060322153118.GE3137@exeter.boston.redhat.com> > yes, it is a text kickstart Okay, this is known: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=185840 I apparently broke this a while back, but text kickstart installs never got tested. I find that weird since I seem to remember testing them at least once. - Chris From Elizabeth.Brosch at thomson.com Wed Mar 22 19:56:21 2006 From: Elizabeth.Brosch at thomson.com (Elizabeth.Brosch at thomson.com) Date: Wed, 22 Mar 2006 14:56:21 -0500 Subject: No network drivers for doing kickstart Message-ID: <2213CAA79DA9094AB52517FDADEF07FA021E56DA@tshuspaphimbx01.ERF.THOMSON.COM> I am able to use RHEL v4 Update 1 with no problems now. It has the appropriate network drivers. But I am now having an issue where kickstart fails finding the network interface. I have had this problem in the past where I was assuming that it was eth0 interface when Anaconda seen the interface as eth3. But I am unable to get around it this time on my DL585 servers. These servers have two NICS, eth0 and eth1. I have tried both with no success. I have booted from CD #1 and went into linux rescue. DHCP will assign IP addresses to eth0 and eth1 but kickstart will still fail. Any suggestions? Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Shabazian, Chip Sent: Tuesday, March 21, 2006 1:27 PM To: Discussion list about Kickstart Subject: RE: No network drivers for doing kickstart I have had issues in the past where a newer initrd worked, but for many reasons, we had to stay with the build based off an older version. In these cases, I found that you could replace the loader in the older initrd with the loader from the newer one, and all was fine. If you still want to build with the latest to minimize packages that need to be upgraded, you could give this a shot. Chip From Chip.Shabazian at bankofamerica.com Wed Mar 22 21:15:16 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 22 Mar 2006 13:15:16 -0800 Subject: No network drivers for doing kickstart Message-ID: Are you trying to pxeboot? Are both interfaces connected? Can you build manually? (if so, this may be a link delay issue). When you boot into rescue, take a look at the MAC addresses so you know which is eth0 and which is eth1, and ensure that they ARE in fact eth0 and 1. Do you have access to the dhcp server? If so, what are you seeing there? -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Elizabeth.Brosch at thomson.com Sent: Wednesday, March 22, 2006 11:56 AM To: kickstart-list at redhat.com Subject: RE: No network drivers for doing kickstart I am able to use RHEL v4 Update 1 with no problems now. It has the appropriate network drivers. But I am now having an issue where kickstart fails finding the network interface. I have had this problem in the past where I was assuming that it was eth0 interface when Anaconda seen the interface as eth3. But I am unable to get around it this time on my DL585 servers. These servers have two NICS, eth0 and eth1. I have tried both with no success. I have booted from CD #1 and went into linux rescue. DHCP will assign IP addresses to eth0 and eth1 but kickstart will still fail. Any suggestions? Liz Brosch TSH System Services Philadelphia, PA 19104 office: (215) 386-0100 x1144 cell: (267) 784-9166 -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Shabazian, Chip Sent: Tuesday, March 21, 2006 1:27 PM To: Discussion list about Kickstart Subject: RE: No network drivers for doing kickstart I have had issues in the past where a newer initrd worked, but for many reasons, we had to stay with the build based off an older version. In these cases, I found that you could replace the loader in the older initrd with the loader from the newer one, and all was fine. If you still want to build with the latest to minimize packages that need to be upgraded, you could give this a shot. Chip _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From joevandyk at gmail.com Thu Mar 23 08:25:53 2006 From: joevandyk at gmail.com (Joe Van Dyk) Date: Thu, 23 Mar 2006 00:25:53 -0800 Subject: Problem with kickstart installation if no dhcp Message-ID: Hi, I can start a kickstart installation fine with kickstart config file being on a nfs shared drive if the machine can get networking information via DHCP. But if I have to enter in static IP network information, then it doesn't seem to be able to find the kickstart file via NFS. Any ideas why this might be? I have the ISOs shared via NFS and I can use those as the install media with both configurations, but I can't access the kickstart file. For now, I can put the kickstart file on the CD disk, but it would be nice to be able to have the config be in some central location. Thanks, Joe From joevandyk at gmail.com Thu Mar 23 08:26:55 2006 From: joevandyk at gmail.com (Joe Van Dyk) Date: Thu, 23 Mar 2006 00:26:55 -0800 Subject: running loadlin with ks=nfs In-Reply-To: <4418D263.8010207@vmware.com> References: <4418D263.8010207@vmware.com> Message-ID: Are you using DHCP, or are you entering in static network info? It works for me via DHCP, but not with static network info. On 3/15/06, Stephen Mah wrote: > Has anyone started the Red Hat installer from loadlin passing the ks=nfs > option? > > Mine got to stage1 of the installer and can't find the kickstart file. > > steve > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From LLavender at kmg.com Thu Mar 23 15:29:35 2006 From: LLavender at kmg.com (Lavender, Leo) Date: Thu, 23 Mar 2006 09:29:35 -0600 Subject: Problem with kickstart installation if no dhcp Message-ID: <044BBA9ACDF3D948BC41F0DD8DFADF314F269A@houexg3.kmg.com> I don't know if I'm understanding the question, but is there a dhcp server, or an arp server running to get an IP to talk to NFS, to get your static IP from the kickstart file? Leo -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Joe Van Dyk Sent: Thursday, March 23, 2006 2:26 AM To: Discussion list about Kickstart Subject: Problem with kickstart installation if no dhcp Hi, I can start a kickstart installation fine with kickstart config file being on a nfs shared drive if the machine can get networking information via DHCP. But if I have to enter in static IP network information, then it doesn't seem to be able to find the kickstart file via NFS. Any ideas why this might be? I have the ISOs shared via NFS and I can use those as the install media with both configurations, but I can't access the kickstart file. For now, I can put the kickstart file on the CD disk, but it would be nice to be able to have the config be in some central location. Thanks, Joe _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From smah at vmware.com Thu Mar 23 17:54:16 2006 From: smah at vmware.com (Stephen Mah) Date: Thu, 23 Mar 2006 09:54:16 -0800 Subject: running loadlin with ks=nfs In-Reply-To: References: <4418D263.8010207@vmware.com> Message-ID: <4422E0C8.30504@vmware.com> I got it working, apparently there's a 64 character limitation on the nfs server:/path my nfs exports paths were pretty long. thanks Joe Van Dyk wrote: > Are you using DHCP, or are you entering in static network info? It > works for me via DHCP, but not with static network info. > > On 3/15/06, Stephen Mah wrote: > >> Has anyone started the Red Hat installer from loadlin passing the ks=nfs >> option? >> >> Mine got to stage1 of the installer and can't find the kickstart file. >> >> steve >> >> _______________________________________________ >> 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 Chip.Shabazian at bankofamerica.com Thu Mar 23 18:00:31 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Thu, 23 Mar 2006 10:00:31 -0800 Subject: Problem with kickstart installation if no dhcp Message-ID: What are you using as boot options to assign the IP address? -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Joe Van Dyk Sent: Thursday, March 23, 2006 12:26 AM To: Discussion list about Kickstart Subject: Problem with kickstart installation if no dhcp Hi, I can start a kickstart installation fine with kickstart config file being on a nfs shared drive if the machine can get networking information via DHCP. But if I have to enter in static IP network information, then it doesn't seem to be able to find the kickstart file via NFS. Any ideas why this might be? I have the ISOs shared via NFS and I can use those as the install media with both configurations, but I can't access the kickstart file. For now, I can put the kickstart file on the CD disk, but it would be nice to be able to have the config be in some central location. Thanks, Joe _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From joevandyk at gmail.com Thu Mar 23 20:14:30 2006 From: joevandyk at gmail.com (Joe Van Dyk) Date: Thu, 23 Mar 2006 12:14:30 -0800 Subject: Problem with kickstart installation if no dhcp In-Reply-To: <044BBA9ACDF3D948BC41F0DD8DFADF314F269A@houexg3.kmg.com> References: <044BBA9ACDF3D948BC41F0DD8DFADF314F269A@houexg3.kmg.com> Message-ID: No, I'm entering in the static IP information via the installer. The static IP info doesn't come from the kickstart file. On 3/23/06, Lavender, Leo wrote: > I don't know if I'm understanding the question, but is there a dhcp > server, or an arp server running to get an IP to talk to NFS, to get > your static IP from the kickstart file? > > Leo > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Joe Van Dyk > Sent: Thursday, March 23, 2006 2:26 AM > To: Discussion list about Kickstart > Subject: Problem with kickstart installation if no dhcp > > > Hi, > > I can start a kickstart installation fine with kickstart config file > being on a nfs shared drive if the machine can get networking > information via DHCP. But if I have to enter in static IP network > information, then it doesn't seem to be able to find the kickstart file > via NFS. Any ideas why this might be? I have the ISOs shared via NFS > and I can use those as the install media with both configurations, but I > can't access the kickstart file. > > For now, I can put the kickstart file on the CD disk, but it would be > nice to be able to have the config be in some central location. > > Thanks, > Joe > > _______________________________________________ > 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 gkapitany at rogers.com Fri Mar 24 13:52:44 2006 From: gkapitany at rogers.com (GABRIEL KAPITANY) Date: Fri, 24 Mar 2006 08:52:44 -0500 (EST) Subject: Problem with kickstart installation if no dhcp In-Reply-To: Message-ID: <20060324135244.10196.qmail@web88108.mail.re2.yahoo.com> Hi, What is the device you specify for ksdevice? Do you have multiple nics? If you set the ip/netmask via installer try to add ksdevice=link. Gabriel --- Joe Van Dyk wrote: > No, I'm entering in the static IP information via > the installer. The > static IP info doesn't come from the kickstart file. > > On 3/23/06, Lavender, Leo wrote: > > I don't know if I'm understanding the question, > but is there a dhcp > > server, or an arp server running to get an IP to > talk to NFS, to get > > your static IP from the kickstart file? > > > > Leo > > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com > > [mailto:kickstart-list-bounces at redhat.com] On > Behalf Of Joe Van Dyk > > Sent: Thursday, March 23, 2006 2:26 AM > > To: Discussion list about Kickstart > > Subject: Problem with kickstart installation if no > dhcp > > > > > > Hi, > > > > I can start a kickstart installation fine with > kickstart config file > > being on a nfs shared drive if the machine can get > networking > > information via DHCP. But if I have to enter in > static IP network > > information, then it doesn't seem to be able to > find the kickstart file > > via NFS. Any ideas why this might be? I have the > ISOs shared via NFS > > and I can use those as the install media with both > configurations, but I > > can't access the kickstart file. > > > > For now, I can put the kickstart file on the CD > disk, but it would be > > nice to be able to have the config be in some > central location. > > > > Thanks, > > Joe > > > > _______________________________________________ > > 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 Chip.Shabazian at bankofamerica.com Fri Mar 24 16:55:27 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Fri, 24 Mar 2006 08:55:27 -0800 Subject: Problem with kickstart installation if no dhcp Message-ID: What version are you installing? Enter your IP info at the boot: prompt like this Boot: ks=nfs:ip_of_server:/path/to/ks.cfg ksdevice=eth0 ip=123.123.123.123 netmask=255.255.255.0 gateway=123.123.123.1 And make sure you also have the network information in the kickstart file on the network line (the first is used to GET the ks.cfg, the second is used to BUILD the server). It would help if you could post at least your command section of your kickstart file. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Joe Van Dyk Sent: Thursday, March 23, 2006 12:15 PM To: Discussion list about Kickstart Subject: Re: Problem with kickstart installation if no dhcp No, I'm entering in the static IP information via the installer. The static IP info doesn't come from the kickstart file. On 3/23/06, Lavender, Leo wrote: > I don't know if I'm understanding the question, but is there a dhcp > server, or an arp server running to get an IP to talk to NFS, to get > your static IP from the kickstart file? > > Leo > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Joe Van Dyk > Sent: Thursday, March 23, 2006 2:26 AM > To: Discussion list about Kickstart > Subject: Problem with kickstart installation if no dhcp > > > Hi, > > I can start a kickstart installation fine with kickstart config file > being on a nfs shared drive if the machine can get networking > information via DHCP. But if I have to enter in static IP network > information, then it doesn't seem to be able to find the kickstart > file via NFS. Any ideas why this might be? I have the ISOs shared > via NFS and I can use those as the install media with both > configurations, but I can't access the kickstart file. > > For now, I can put the kickstart file on the CD disk, but it would be > nice to be able to have the config be in some central location. > > Thanks, > Joe > > _______________________________________________ > 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 jgao at optonline.net Sun Mar 26 22:30:29 2006 From: jgao at optonline.net (jgao at optonline.net) Date: Sun, 26 Mar 2006 17:30:29 -0500 Subject: How to Install all packages in FC5 Kickstart Message-ID: Hi, I can install all packages in FC4 by specify: %packages @everything But when I tried to do the same thing in FC5 kickstart, I can?t find some packages, such as kernel-devel, after the successful kickstart installation. Is there any changes in this area in FC5? How could I install all FC5 extra packages if I want? Thanks, Jian From imarks at comcast.net Sun Mar 26 22:56:40 2006 From: imarks at comcast.net (Ian Marks) Date: Sun, 26 Mar 2006 17:56:40 -0500 Subject: ACL Support Message-ID: <44271C28.2050302@comcast.net> Is there any easy way to add ACL support to specific partitions in kickstart? On my current kickstart image I leave the partitioning section commented out so the user will be prompted to manually partition the system. I would like something in the post section to automatically add ACL support if possible. Thanks, Ian From richzendy at gmail.com Mon Mar 27 02:59:08 2006 From: richzendy at gmail.com (Edwind Richzendy Contreras Soto) Date: Sun, 26 Mar 2006 22:59:08 -0400 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: References: Message-ID: <1143428348.22250.6.camel@localhost.localdomain> in this line: IP=ifconfig $ETH | grep inet | cut -d : -f 2 | cut -d " "\ -f 1 change for: IP=ifconfig $ETH | grep "inet addr" | awk '1{print $2,$5 }'\ | awk -F : '1{print $2 }' the result is = in both lines :D El mi?, 15-03-2006 a las 09:22 -0800, Shabazian, Chip escribi?: > This is what I use in my command section: > %include /tmp/partinfo > > And this is in my %pre: > ETH=`grep DEVICE /tmp/netinfo | cut -d = -f 2` > IP=`ifconfig $ETH | grep inet | cut -d : -f 2 | cut -d " " -f 1` > NETMASK=`ifconfig $ETH | grep inet | cut -d : -f 4` > GATEWAY=`route -n| grep "^0.0.0.0" | cut -b 17-32 | cut -d " " -f 1` > HOSTNAME=`grep HOSTNAME /tmp/netinfo | cut -d = -f 2 | cut -d . -f 1` > > cat << EOF > /tmp/buildnet > network --device $ETH --bootproto static --ip=$IP --netmask=$NETMASK > --gateway=$ > GATEWAY --hostname=$HOSTNAME > EOF > > This works every time, whether you get your IP address from DHCP, or > specify it on the command line. > > Chip > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Philip Rowlands > Sent: Tuesday, March 14, 2006 11:21 PM > To: Discussion list about Kickstart > Subject: Re: Extracting variables from within ks.cfg, interpreted from > command-line > > On Tue, 14 Mar 2006, Dan Trainor wrote: > > > Alright, so, I think something wierd is going on. > > > > I'm still seeing Anaconda bomb out saying that /tmp/netsettings does > > not exist. I can verify that /tmp/netsettings does not get created. > > However, I cannot figure out why. > > > > vty3 tells me something similar to the following: > > > > : command not founde 1: > > : command not founde 3: > > : command not founde 6: > [snip] > > This doesn't quite feel like CRLF-line-endings, but worth a check. Can > you post your kickstart file? > > > Cheers, > Phil > > _______________________________________________ > 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 -- Fingerprint: 43C5 2F43 1238 D72C BAA1 F9D7 CECE E171 FBF8 5909 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 191 bytes Desc: This is a digitally signed message part URL: From Chip.Shabazian at bankofamerica.com Mon Mar 27 17:47:16 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Mon, 27 Mar 2006 09:47:16 -0800 Subject: Extracting variables from within ks.cfg, interpreted from command-line Message-ID: That works just as well, but I'm not sure why you would be having problems with my original line. It works perfect for me, and hasn't failed in over 1000 builds. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Edwind Richzendy Contreras Soto Sent: Sunday, March 26, 2006 6:59 PM To: Discussion list about Kickstart Subject: RE: Extracting variables from within ks.cfg, interpreted from command-line in this line: IP=ifconfig $ETH | grep inet | cut -d : -f 2 | cut -d " "\ -f 1 change for: IP=ifconfig $ETH | grep "inet addr" | awk '1{print $2,$5 }'\ | awk -F : '1{print $2 }' the result is = in both lines :D El mi?, 15-03-2006 a las 09:22 -0800, Shabazian, Chip escribi?: > This is what I use in my command section: > %include /tmp/partinfo > > And this is in my %pre: > ETH=`grep DEVICE /tmp/netinfo | cut -d = -f 2` IP=`ifconfig $ETH | > grep inet | cut -d : -f 2 | cut -d " " -f 1` NETMASK=`ifconfig $ETH | > grep inet | cut -d : -f 4` GATEWAY=`route -n| grep "^0.0.0.0" | cut -b > 17-32 | cut -d " " -f 1` HOSTNAME=`grep HOSTNAME /tmp/netinfo | cut -d > = -f 2 | cut -d . -f 1` > > cat << EOF > /tmp/buildnet > network --device $ETH --bootproto static --ip=$IP --netmask=$NETMASK > --gateway=$ GATEWAY --hostname=$HOSTNAME EOF > > This works every time, whether you get your IP address from DHCP, or > specify it on the command line. > > Chip > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Philip > Rowlands > Sent: Tuesday, March 14, 2006 11:21 PM > To: Discussion list about Kickstart > Subject: Re: Extracting variables from within ks.cfg, interpreted from > command-line > > On Tue, 14 Mar 2006, Dan Trainor wrote: > > > Alright, so, I think something wierd is going on. > > > > I'm still seeing Anaconda bomb out saying that /tmp/netsettings does > > not exist. I can verify that /tmp/netsettings does not get created. > > However, I cannot figure out why. > > > > vty3 tells me something similar to the following: > > > > : command not founde 1: > > : command not founde 3: > > : command not founde 6: > [snip] > > This doesn't quite feel like CRLF-line-endings, but worth a check. Can > you post your kickstart file? > > > Cheers, > Phil > > _______________________________________________ > 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 -- Fingerprint: 43C5 2F43 1238 D72C BAA1 F9D7 CECE E171 FBF8 5909 From dan at id-confirm.com Mon Mar 27 18:19:21 2006 From: dan at id-confirm.com (Dan Trainor) Date: Mon, 27 Mar 2006 11:19:21 -0700 Subject: Extracting variables from within ks.cfg, interpreted from command-line In-Reply-To: References: Message-ID: <44282CA9.1070608@id-confirm.com> Hi, Chip - Because I'm anal. I don't want to use Anaconda's naming conventions for the components. I wanted to use something that comforms closer to our system as a whole, i.e. ip= would become MYAPPIPADDR=, stuff like that. Just wanted to let you know that, so you don't think I'm nuts, and you knew where I was going with that haha. Thanks! -dant Shabazian, Chip wrote: > That works just as well, but I'm not sure why you would be having problems with my original line. It works perfect for me, and hasn't failed in over 1000 builds. > > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Edwind Richzendy Contreras Soto > Sent: Sunday, March 26, 2006 6:59 PM > To: Discussion list about Kickstart > Subject: RE: Extracting variables from within ks.cfg, interpreted from command-line > > in this line: IP=ifconfig $ETH | grep inet | cut -d : -f 2 | cut -d " "\ -f 1 > > change for: IP=ifconfig $ETH | grep "inet addr" | awk '1{print $2,$5 }'\ > | awk -F : '1{print $2 }' > > the result is = in both lines :D > > El mi?, 15-03-2006 a las 09:22 -0800, Shabazian, Chip escribi?: >> This is what I use in my command section: >> %include /tmp/partinfo >> >> And this is in my %pre: >> ETH=`grep DEVICE /tmp/netinfo | cut -d = -f 2` IP=`ifconfig $ETH | >> grep inet | cut -d : -f 2 | cut -d " " -f 1` NETMASK=`ifconfig $ETH | >> grep inet | cut -d : -f 4` GATEWAY=`route -n| grep "^0.0.0.0" | cut -b >> 17-32 | cut -d " " -f 1` HOSTNAME=`grep HOSTNAME /tmp/netinfo | cut -d >> = -f 2 | cut -d . -f 1` >> >> cat << EOF > /tmp/buildnet >> network --device $ETH --bootproto static --ip=$IP --netmask=$NETMASK >> --gateway=$ GATEWAY --hostname=$HOSTNAME EOF >> >> This works every time, whether you get your IP address from DHCP, or >> specify it on the command line. >> >> Chip >> >> -----Original Message----- >> From: kickstart-list-bounces at redhat.com >> [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Philip >> Rowlands >> Sent: Tuesday, March 14, 2006 11:21 PM >> To: Discussion list about Kickstart >> Subject: Re: Extracting variables from within ks.cfg, interpreted from >> command-line >> >> On Tue, 14 Mar 2006, Dan Trainor wrote: >> >>> Alright, so, I think something wierd is going on. >>> >>> I'm still seeing Anaconda bomb out saying that /tmp/netsettings does >>> not exist. I can verify that /tmp/netsettings does not get created. >>> However, I cannot figure out why. >>> >>> vty3 tells me something similar to the following: >>> >>> : command not founde 1: >>> : command not founde 3: >>> : command not founde 6: >> [snip] >> >> This doesn't quite feel like CRLF-line-endings, but worth a check. Can >> you post your kickstart file? >> >> >> Cheers, >> Phil >> >> _______________________________________________ >> 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 > -- > Fingerprint: 43C5 2F43 1238 D72C BAA1 F9D7 CECE E171 FBF8 5909 > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- - Dan Trainor - id-Confirm, Inc. - Direct: 720.241.5580 From dxh at yahoo.com Mon Mar 27 21:58:29 2006 From: dxh at yahoo.com (Don Hoover) Date: Mon, 27 Mar 2006 13:58:29 -0800 (PST) Subject: Anyone Built a LIVECD + Kickstart Message-ID: <20060327215829.46139.qmail@web60718.mail.yahoo.com> We have kickstart CD's we use that allow us to boot and install several versions of RHEL from the one CD using CGI's etc. Bascially the idea is to have a single CD the admins can carry with them, boot off of to install any system using any of our currently supported versions of RHEL. I was thinking it would be cool to add something like DamnSmall etc that they could use as a rescue toolkit to have handy too on that CD. After all I am using like 40MB right now just to basically have a isolinux directory. Just curious if anyone has done that yet. --- Don Hoover dxh at yahoo.com From Curtis at GreenKey.net Tue Mar 28 05:20:21 2006 From: Curtis at GreenKey.net (Curtis Doty) Date: Mon, 27 Mar 2006 21:20:21 -0800 Subject: How to Install all packages in FC5 Kickstart In-Reply-To: References: Message-ID: <4428C795.5080601@GreenKey.net> jgao at optonline.net wrote: > I can install all packages in FC4 by specify: > > %packages > @everything > > But when I tried to do the same thing in FC5 kickstart, I can?t find some packages, such as kernel-devel, after the successful kickstart installation. Is there any changes in this area in FC5? > > How could I install all FC5 extra packages if I want? > Just for giggles, I did a: %packages --resolvedeps * Which seemed to work fine even though I think resolvedeps is supposedly deprecated too. This was over http. Maybe you are using cd iso's with this bug? https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=186202 ../C From dag at wieers.com Tue Mar 28 05:43:46 2006 From: dag at wieers.com (Dag Wieers) Date: Tue, 28 Mar 2006 07:43:46 +0200 (CEST) Subject: Anyone Built a LIVECD + Kickstart In-Reply-To: <20060327215829.46139.qmail@web60718.mail.yahoo.com> References: <20060327215829.46139.qmail@web60718.mail.yahoo.com> Message-ID: On Mon, 27 Mar 2006, Don Hoover wrote: > We have kickstart CD's we use that allow us to boot > and install several versions of RHEL from the one CD > using CGI's etc. Bascially the idea is to have a > single CD the admins can carry with them, boot off of > to install any system using any of our currently > supported versions of RHEL. > > I was thinking it would be cool to add something like > DamnSmall etc that they could use as a rescue toolkit > to have handy too on that CD. > > After all I am using like 40MB right now just to > basically have a isolinux directory. > > > Just curious if anyone has done that yet. We have a customized boot.iso with our own splash screen (text-only), be-latin1 keyboard mappings, serial console and with many isolinux labels (rescue images and kickstart commandline parameters for every system we have). Since we're not allowed to do DHCP/PXE installations, such an ISO allows us to have a single image that we can boot from as a "virtual" cdrom and either test the network/hardware or do the installation. Unfortunately I haven't find a small rescue image that includes a recent 2.6 kernel and wide network support. RIP with network is already 30MB (takes a long time to load over the network) and none of them have tcptraceroute functionality (which is important to be able to test and pinpoint firewall configuration problems). tcptraceroute in busybox is an essential feature for us, I'd donate money to have someone write it. Kind regards, -- dag wieers, dag at wieers.com, http://dag.wieers.com/ -- [all I want is a warm bed and a kind word and unlimited power] From Chip.Shabazian at bankofamerica.com Tue Mar 28 19:07:42 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Tue, 28 Mar 2006 11:07:42 -0800 Subject: Anyone Built a LIVECD + Kickstart Message-ID: I've used the instructions found here to add binaries to my netstg2.img. You should be able to modify it to your liking: http://linuxgazette.net/105/nielsen.html#install Not a "live cd" exactly, but it should give you what you are looking for. Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Dag Wieers Sent: Monday, March 27, 2006 9:44 PM To: Discussion list about Kickstart Subject: Re: Anyone Built a LIVECD + Kickstart On Mon, 27 Mar 2006, Don Hoover wrote: > We have kickstart CD's we use that allow us to boot and install > several versions of RHEL from the one CD > using CGI's etc. Bascially the idea is to have a > single CD the admins can carry with them, boot off of to install any > system using any of our currently supported versions of RHEL. > > I was thinking it would be cool to add something like DamnSmall etc > that they could use as a rescue toolkit to have handy too on that CD. > > After all I am using like 40MB right now just to basically have a > isolinux directory. > > > Just curious if anyone has done that yet. We have a customized boot.iso with our own splash screen (text-only), be-latin1 keyboard mappings, serial console and with many isolinux labels (rescue images and kickstart commandline parameters for every system we have). Since we're not allowed to do DHCP/PXE installations, such an ISO allows us to have a single image that we can boot from as a "virtual" cdrom and either test the network/hardware or do the installation. Unfortunately I haven't find a small rescue image that includes a recent 2.6 kernel and wide network support. RIP with network is already 30MB (takes a long time to load over the network) and none of them have tcptraceroute functionality (which is important to be able to test and pinpoint firewall configuration problems). tcptraceroute in busybox is an essential feature for us, I'd donate money to have someone write it. Kind regards, -- dag wieers, dag at wieers.com, http://dag.wieers.com/ -- [all I want is a warm bed and a kind word and unlimited power] _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From phb at llgc.org.uk Wed Mar 29 09:18:17 2006 From: phb at llgc.org.uk (Phil Bettinson) Date: Wed, 29 Mar 2006 10:18:17 +0100 Subject: Redhat RHN (re)-joining in Kickstart In-Reply-To: References: Message-ID: <442A50D9.7090009@llgc.org.uk> Hello All, I'm currently arguing with a small part of a kickstart script that joins, and updates the computers to the kickstart network. It works, appart from the fact that it registers them as "localhost.localhost", as opposed to their machine names. The machines have the correct names when they reboot (that is, the pick up the correct hostnames, and have the right ip addresses etc.), but it means that I have to re-register them with RHN in order to give them a sane name. I'm aware that there have been request for features put in for this, but I was wondering if anyone had found a way round it? or if there is a way of setting the hostname of the computer before registring to RHN so that it registers with the correct name. Thanks in advance, Phil Bettinson. -- ----------------------------------------------------------------------- Phil Bettinson Swyddog Gweinyddu Systemau Systems Administrator Isadran Cyfrifiaduron Computer Section Llyfrgell Genedlaethol Cymru National Library of Wales Aberystwyth, Aberystwyth Ceredigion, Ceredigion SY23 3BU SY23 3BU Ff?n / Phone: 01970 632867 Ffacs / Fax: 01970 615709 EBost / Email: phb AT.NOSPAM llgc.org.uk We / Web: http://www.llgc.org.uk ----------------------------------------------------------------------- Yn siarad drosof fy hun, nid LlGC - Speaking personally, not for NLW From phr at doc.ic.ac.uk Wed Mar 29 14:47:55 2006 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 29 Mar 2006 15:47:55 +0100 (BST) Subject: Redhat RHN (re)-joining in Kickstart In-Reply-To: <442A50D9.7090009@llgc.org.uk> References: <442A50D9.7090009@llgc.org.uk> Message-ID: On Wed, 29 Mar 2006, Phil Bettinson wrote: > I'm currently arguing with a small part of a kickstart script that joins, and > updates the computers to the kickstart network. It works, appart from the fact > that it registers them as "localhost.localhost", as opposed to their machine > names. > > The machines have the correct names when they reboot (that is, the pick up the > correct hostnames, and have the right ip addresses etc.), but it means that I > have to re-register them with RHN in order to give them a sane name. I had this as a %post script: #!/bin/sh - . /etc/sysconfig/network hostname -v -- "$HOSTNAME" which might do the trick. Cheers, Phil From lists.james at gmail.com Wed Mar 29 15:48:25 2006 From: lists.james at gmail.com (James P) Date: Wed, 29 Mar 2006 09:48:25 -0600 Subject: Redhat RHN (re)-joining in Kickstart In-Reply-To: References: <442A50D9.7090009@llgc.org.uk> Message-ID: <442AAC49.5030309@gmail.com> Kinda depends how you're setting your hostname and how your doing your registration during the kickstart, but it should be fairly easy no matter how you approach it. In my own builds, I set the hostname in the %pre section and then register in the %post like this: rhnreg_ks --profilename $HOSTNAME p.s. Not sure if this is what you meant, but you don't really have to re-register the systems just to change their profile names. You can edit that part of the RHN website manually. Philip Rowlands wrote: > On Wed, 29 Mar 2006, Phil Bettinson wrote: > >> I'm currently arguing with a small part of a kickstart script that >> joins, and updates the computers to the kickstart network. It works, >> appart from the fact that it registers them as "localhost.localhost", >> as opposed to their machine names. >> >> The machines have the correct names when they reboot (that is, the >> pick up the correct hostnames, and have the right ip addresses etc.), >> but it means that I have to re-register them with RHN in order to >> give them a sane name. > > > I had this as a %post script: > > #!/bin/sh - > . /etc/sysconfig/network > hostname -v -- "$HOSTNAME" > > > which might do the trick. > > > Cheers, > Phil > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From Chip.Shabazian at bankofamerica.com Wed Mar 29 17:00:38 2006 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 29 Mar 2006 09:00:38 -0800 Subject: Redhat RHN (re)-joining in Kickstart Message-ID: I set it in my post like this: export HOSTNAME=`grep HOSTNAME /etc/sysconfig/network | awk -F= '{print $2}'` hostname $HOSTNAME -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of James P Sent: Wednesday, March 29, 2006 7:48 AM To: Discussion list about Kickstart Subject: Re: Redhat RHN (re)-joining in Kickstart Kinda depends how you're setting your hostname and how your doing your registration during the kickstart, but it should be fairly easy no matter how you approach it. In my own builds, I set the hostname in the %pre section and then register in the %post like this: rhnreg_ks --profilename $HOSTNAME p.s. Not sure if this is what you meant, but you don't really have to re-register the systems just to change their profile names. You can edit that part of the RHN website manually. Philip Rowlands wrote: > On Wed, 29 Mar 2006, Phil Bettinson wrote: > >> I'm currently arguing with a small part of a kickstart script that >> joins, and updates the computers to the kickstart network. It works, >> appart from the fact that it registers them as "localhost.localhost", >> as opposed to their machine names. >> >> The machines have the correct names when they reboot (that is, the >> pick up the correct hostnames, and have the right ip addresses etc.), >> but it means that I have to re-register them with RHN in order to >> give them a sane name. > > > I had this as a %post script: > > #!/bin/sh - > . /etc/sysconfig/network > hostname -v -- "$HOSTNAME" > > > which might do the trick. > > > Cheers, > Phil > > _______________________________________________ > 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 Wed Mar 29 18:00:57 2006 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Wed, 29 Mar 2006 13:00:57 -0500 Subject: Redhat RHN (re)-joining in Kickstart In-Reply-To: Message-ID: The only way I was able to fix this (and it occurs for me after the install if I tried to register) was this in %post: echo "192.168.10.100 myhostname" >> /etc/hosts of course make the necessary substitutions, and perform this before you do the rhnreg_ks 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 03/29/2006 12:00:38 PM: > I set it in my post like this: > > export HOSTNAME=`grep HOSTNAME /etc/sysconfig/network | awk -F= '{print > $2}'` > hostname $HOSTNAME > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of James P > Sent: Wednesday, March 29, 2006 7:48 AM > To: Discussion list about Kickstart > Subject: Re: Redhat RHN (re)-joining in Kickstart > > Kinda depends how you're setting your hostname and how your doing your > registration during the kickstart, but it should be fairly easy no > matter how you approach it. In my own builds, I set the hostname in the > %pre section and then register in the %post like this: > > rhnreg_ks --profilename $HOSTNAME > > p.s. Not sure if this is what you meant, but you don't really have to > re-register the systems just to change their profile names. You can edit > that part of the RHN website manually. > > Philip Rowlands wrote: > > > On Wed, 29 Mar 2006, Phil Bettinson wrote: > > > >> I'm currently arguing with a small part of a kickstart script that > >> joins, and updates the computers to the kickstart network. It works, > >> appart from the fact that it registers them as "localhost.localhost", > > >> as opposed to their machine names. > >> > >> The machines have the correct names when they reboot (that is, the > >> pick up the correct hostnames, and have the right ip addresses etc.), > > >> but it means that I have to re-register them with RHN in order to > >> give them a sane name. > > > > > > I had this as a %post script: > > > > #!/bin/sh - > > . /etc/sysconfig/network > > hostname -v -- "$HOSTNAME" > > > > > > which might do the trick. > > > > > > Cheers, > > Phil > > > > _______________________________________________ > > 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 phb at llgc.org.uk Thu Mar 30 14:58:51 2006 From: phb at llgc.org.uk (Phil Bettinson) Date: Thu, 30 Mar 2006 15:58:51 +0100 Subject: Redhat RHN (re)-joining in Kickstart In-Reply-To: References: Message-ID: <442BF22B.7010009@llgc.org.uk> Thanks for the help guys. It seems the final solution looks a little something like this.... HOSTNAME=`grep $IP /kickstart/hosts | awk '{print $3}'` IP=`ifconfig eth0 | grep inet | awk '{print $2}' | awk -F: '{print $2}'` HOST=`grep $IP /kickstart/hosts | awk '{print $2}'` echo "$IP $HOST $HOSTNAME" >> /etc/hosts hostname -v -- "$HOSTNAME" /kickstart is a directory previously mounted from the NFS server, where I have a "host" file, that is a file that is laid out like a hostfile (ish), with no comments in. eg (sample line) 192.168.0.1 mymachine mymachine.mynetwork.co.uk 10.0.0.1 The last IP address allows me to place the machine on two different networks. Here's the kickstart Stuff echo "Registring with RHN" /usr/sbin/rhnreg_ks --activationkey= echo " " echo "Running up2date" /usr/sbin/up2date up2date /usr/sbin/up2date -u The reason for using an activation key means that they go back into the right group at the end of the kickstart (that bit was the important bit). the reason for not wanting to use --profile is that that implies that there is a profile there already (in our case there might not be) You may notice that I first up2date up2date, before updating the system, that's because in the version of RHAS3 (rh3 no updates) I'm kickstarting, it has problems with dependencies on the x64_86 architecture. I also don't up2date the kernel here because of various (other) system dependencies that break on a kernel upgrade without modification/recompile/reinstall. This ensure that the system comes up working, and as secure as possible to ensure that we can get working straight away. Admitedly I don't need all this detail here, but it's here for those that happen to google this problem in the future. Thanks for the help solving this problem Phil Bettinson James_Martin at ao.uscourts.gov wrote: > The only way I was able to fix this (and it occurs for me after the > install if I tried to register) was this in %post: > > echo "192.168.10.100 myhostname" >> /etc/hosts > > of course make the necessary substitutions, and perform this before you do > the rhnreg_ks > > 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 03/29/2006 12:00:38 PM: > > >>I set it in my post like this: >> >>export HOSTNAME=`grep HOSTNAME /etc/sysconfig/network | awk -F= '{print >>$2}'` >>hostname $HOSTNAME >> >> >>-----Original Message----- >>From: kickstart-list-bounces at redhat.com >>[mailto:kickstart-list-bounces at redhat.com] On Behalf Of James P >>Sent: Wednesday, March 29, 2006 7:48 AM >>To: Discussion list about Kickstart >>Subject: Re: Redhat RHN (re)-joining in Kickstart >> >>Kinda depends how you're setting your hostname and how your doing your >>registration during the kickstart, but it should be fairly easy no >>matter how you approach it. In my own builds, I set the hostname in the >>%pre section and then register in the %post like this: >> >>rhnreg_ks --profilename $HOSTNAME >> >>p.s. Not sure if this is what you meant, but you don't really have to >>re-register the systems just to change their profile names. You can edit >>that part of the RHN website manually. >> >>Philip Rowlands wrote: >> >> >>>On Wed, 29 Mar 2006, Phil Bettinson wrote: >>> >>> >>>>I'm currently arguing with a small part of a kickstart script that >>>>joins, and updates the computers to the kickstart network. It works, >>>>appart from the fact that it registers them as "localhost.localhost", >> >>>>as opposed to their machine names. >>>> >>>>The machines have the correct names when they reboot (that is, the >>>>pick up the correct hostnames, and have the right ip addresses etc.), >> >>>>but it means that I have to re-register them with RHN in order to >>>>give them a sane name. >>> >>> >>>I had this as a %post script: >>> >>>#!/bin/sh - >>>. /etc/sysconfig/network >>>hostname -v -- "$HOSTNAME" >>> >>> >>>which might do the trick. >>> >>> >>>Cheers, >>>Phil >>> >>>_______________________________________________ >>>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 > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -- ----------------------------------------------------------------------- Phil Bettinson Swyddog Gweinyddu Systemau Systems Administrator Isadran Cyfrifiaduron Computer Section Llyfrgell Genedlaethol Cymru National Library of Wales Aberystwyth, Aberystwyth Ceredigion, Ceredigion SY23 3BU SY23 3BU Ff?n / Phone: 01970 632867 Ffacs / Fax: 01970 615709 EBost / Email: phb AT NOSPAM llgc.org.uk We / Web: http://www.llgc.org.uk ----------------------------------------------------------------------- Yn siarad drosof fy hun, nid LlGC - Speaking personally, not for NLW From darian.lyons at elefire.com Sat Mar 11 09:50:31 2006 From: darian.lyons at elefire.com (Darian Lyons) Date: Sat, 11 Mar 2006 09:50:31 -0000 Subject: subscribe Message-ID: <44129D61.8050008@elefire.com> -- Kind Regards, Darian Lyons http://www.elefire.com 0800.316.8099 Elefire, Ltd. London