From brilong at cisco.com Fri Apr 1 13:29:10 2005 From: brilong at cisco.com (Brian Long) Date: Fri, 01 Apr 2005 08:29:10 -0500 Subject: Anaconda Bug in RHEL 3? Message-ID: <1112362150.993.21.camel@localhost.localdomain> I posted this in anaconda-devel-list and got zero bites. Is anyone else affected by this --nisserver server1,server2,server3 bug? https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=153080 /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From brilong at cisco.com Fri Apr 1 13:38:07 2005 From: brilong at cisco.com (Brian Long) Date: Fri, 01 Apr 2005 08:38:07 -0500 Subject: Kickstart and Authentication Proxy Message-ID: <1112362687.993.26.camel@localhost.localdomain> Does anyone support users trying to kickstart over a VPN connection when said connection requires authentication on a web site? When our newest VPN users try to connect to internal sites, they get redirected to a website which requires their authentication. I'm wondering how people can perform kickstarts when this type of authentication is required. The way a few folks have worked around this is to boot a Knoppix CD, authenticate on the web site, reboot and the kickstart succeeds. This is obviously not scalable. Since many of these VPN users are full-time telecommuters, I need to find a scalable way to support them. Any ideas would be greatly appreciated! /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From docmani at zahadum.net Fri Apr 1 19:42:30 2005 From: docmani at zahadum.net (John Sabean) Date: Fri, 01 Apr 2005 14:42:30 -0500 Subject: Trying to install GRUB but getting "L 99 99 99" error! Message-ID: I've been working at building a kickstart disk for a RedHat Enterprise 4 configuration that I have (using the demo/trial version). My ks.cfg file has the following bootloader command: bootloader --location=mbr --md5pass=$blahblahblah When the host reboots, it prints out "L 99 99 99 99....". Thats a LILO error, isn't it??? (and yes, this disk previously had LILO loaded on it) I've even gone in, in rescue mode, run chroot, and performed 'grub-install /dev/sda', and rebooting I get the same error. I'm at a total loss! Here are some goodies from my ks.cfg file: install cdrom lang en_US.UTF-8 langsupport --default=en_US.UTF-8 en_US.UTF-8 keyboard us xconfig --card "ATI Mach64" --videoram 8128 --hsync 30-61 --vsync 56-76 --resolution 800x600 --depth 16 network --network setup rootpw --iscrypted $blablahblah firewall --enabled --port=22:tcp selinux --enforcing authconfig --enableshadow --enablemd5 timezone America/New_York # The following is the partition information you requested # Note that any partitions you deleted are not expressed # here so unless you clear all partitions first, this is # not guaranteed to work clearpart --all zerombr yes part /boot --fstype "ext3" --ondisk sda --start 1 --end 13 part /usr --fstype "ext3" --ondisk sda --start 14 --end 2563 part /opt --fstype "ext3" --ondisk sda --start 2564 --end 3838 part /export --fstype "ext3" --ondisk sda --start 3839 --end 5113 part /tmp --fstype "ext3" --ondisk sda --start 5114 --end 6133 part /usr/local --fstype "ext3" --ondisk sda --start 6134 --end 6898 part /var --fstype "ext3" --ondisk sda --start 6899 --end 7535 part swap --ondisk sda --start 7536 --end 7599 part / --fstype "ext3" --ondisk sda --start 7600 --end 8924 part /junk --fstype "ext3" --ondisk sdb --start 1 --end 8924 bootloader --driveorder=sda,sdb --location=mbr --md5pass=$blahblahblah %packages ... etc Any help would be greatly appriciated. From ebrown at lanl.gov Fri Apr 1 19:48:00 2005 From: ebrown at lanl.gov (Ed Brown) Date: Fri, 01 Apr 2005 12:48:00 -0700 Subject: Trying to install GRUB but getting "L 99 99 99" error! In-Reply-To: References: Message-ID: <1112384879.16494.70.camel@edbrown.lanl.gov> Try checking the boot order in the system BIOS or the SCSI BIOS. You may be installing grub to sda, but still booting from sdb? -Ed On Fri, 2005-04-01 at 12:42, John Sabean wrote: > I've been working at building a kickstart disk for a RedHat Enterprise 4 > configuration that I have (using the demo/trial version). > > My ks.cfg file has the following bootloader command: > > bootloader --location=mbr --md5pass=$blahblahblah > > When the host reboots, it prints out "L 99 99 99 99....". Thats a LILO > error, isn't it??? (and yes, this disk previously had LILO loaded on it) > > I've even gone in, in rescue mode, run chroot, and performed 'grub-install > /dev/sda', and rebooting I get the same error. > > I'm at a total loss! > > Here are some goodies from my ks.cfg file: > > > install > cdrom > lang en_US.UTF-8 > langsupport --default=en_US.UTF-8 en_US.UTF-8 > keyboard us > xconfig --card "ATI Mach64" --videoram 8128 --hsync 30-61 --vsync 56-76 > --resolution 800x600 --depth 16 > network --network setup > rootpw --iscrypted $blablahblah > firewall --enabled --port=22:tcp > selinux --enforcing > authconfig --enableshadow --enablemd5 > timezone America/New_York > # The following is the partition information you requested > # Note that any partitions you deleted are not expressed > # here so unless you clear all partitions first, this is > # not guaranteed to work > clearpart --all > zerombr yes > part /boot --fstype "ext3" --ondisk sda --start 1 --end 13 > part /usr --fstype "ext3" --ondisk sda --start 14 --end 2563 > part /opt --fstype "ext3" --ondisk sda --start 2564 --end 3838 > part /export --fstype "ext3" --ondisk sda --start 3839 --end 5113 > part /tmp --fstype "ext3" --ondisk sda --start 5114 --end 6133 > part /usr/local --fstype "ext3" --ondisk sda --start 6134 --end 6898 > part /var --fstype "ext3" --ondisk sda --start 6899 --end 7535 > part swap --ondisk sda --start 7536 --end 7599 > part / --fstype "ext3" --ondisk sda --start 7600 --end 8924 > part /junk --fstype "ext3" --ondisk sdb --start 1 --end 8924 > bootloader --driveorder=sda,sdb --location=mbr --md5pass=$blahblahblah > > %packages > ... etc > > Any help would be greatly appriciated. > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From docmani at zahadum.net Fri Apr 1 19:57:10 2005 From: docmani at zahadum.net (John Sabean) Date: Fri, 01 Apr 2005 14:57:10 -0500 Subject: Trying to install GRUB but getting "L 99 99 99" error! In-Reply-To: <1112384879.16494.70.camel@edbrown.lanl.gov> References: <1112384879.16494.70.camel@edbrown.lanl.gov> Message-ID: I thought of that. I removed all the other drives from the system aside from /dev/sda and rebooted. Its definately booting from that drive. On Fri, 01 Apr 2005 12:48:00 -0700, Ed Brown wrote: > Try checking the boot order in the system BIOS or the SCSI BIOS. You > may be installing grub to sda, but still booting from sdb? > > -Ed > > > On Fri, 2005-04-01 at 12:42, John Sabean wrote: >> I've been working at building a kickstart disk for a RedHat Enterprise 4 >> configuration that I have (using the demo/trial version). >> >> My ks.cfg file has the following bootloader command: >> >> bootloader --location=mbr --md5pass=$blahblahblah >> >> When the host reboots, it prints out "L 99 99 99 99....". Thats a LILO >> error, isn't it??? (and yes, this disk previously had LILO loaded on it) >> >> I've even gone in, in rescue mode, run chroot, and performed >> 'grub-install >> /dev/sda', and rebooting I get the same error. >> >> I'm at a total loss! >> >> Here are some goodies from my ks.cfg file: >> >> >> install >> cdrom >> lang en_US.UTF-8 >> langsupport --default=en_US.UTF-8 en_US.UTF-8 >> keyboard us >> xconfig --card "ATI Mach64" --videoram 8128 --hsync 30-61 --vsync 56-76 >> --resolution 800x600 --depth 16 >> network --network setup >> rootpw --iscrypted $blablahblah >> firewall --enabled --port=22:tcp >> selinux --enforcing >> authconfig --enableshadow --enablemd5 >> timezone America/New_York >> # The following is the partition information you requested >> # Note that any partitions you deleted are not expressed >> # here so unless you clear all partitions first, this is >> # not guaranteed to work >> clearpart --all >> zerombr yes >> part /boot --fstype "ext3" --ondisk sda --start 1 --end 13 >> part /usr --fstype "ext3" --ondisk sda --start 14 --end 2563 >> part /opt --fstype "ext3" --ondisk sda --start 2564 --end 3838 >> part /export --fstype "ext3" --ondisk sda --start 3839 --end 5113 >> part /tmp --fstype "ext3" --ondisk sda --start 5114 --end 6133 >> part /usr/local --fstype "ext3" --ondisk sda --start 6134 --end 6898 >> part /var --fstype "ext3" --ondisk sda --start 6899 --end 7535 >> part swap --ondisk sda --start 7536 --end 7599 >> part / --fstype "ext3" --ondisk sda --start 7600 --end 8924 >> part /junk --fstype "ext3" --ondisk sdb --start 1 --end 8924 >> bootloader --driveorder=sda,sdb --location=mbr --md5pass=$blahblahblah >> >> %packages >> ... etc >> >> Any help would be greatly appriciated. >> >> >> _______________________________________________ >> 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 docmani at zahadum.net Fri Apr 1 20:23:44 2005 From: docmani at zahadum.net (John Sabean) Date: Fri, 01 Apr 2005 15:23:44 -0500 Subject: Trying to install GRUB but getting "L 99 99 99" error! In-Reply-To: <1112384879.16494.70.camel@edbrown.lanl.gov> References: <1112384879.16494.70.camel@edbrown.lanl.gov> Message-ID: Never mind, I'm stupid. :) I have a SAN attached to the system. When I install, it recognizes the SAN as sda. When the system boots, it recognizes my first HD as sda. Never mind... :) On Fri, 01 Apr 2005 12:48:00 -0700, Ed Brown wrote: > Try checking the boot order in the system BIOS or the SCSI BIOS. You > may be installing grub to sda, but still booting from sdb? > > -Ed > > > On Fri, 2005-04-01 at 12:42, John Sabean wrote: >> I've been working at building a kickstart disk for a RedHat Enterprise 4 >> configuration that I have (using the demo/trial version). >> >> My ks.cfg file has the following bootloader command: >> >> bootloader --location=mbr --md5pass=$blahblahblah >> >> When the host reboots, it prints out "L 99 99 99 99....". Thats a LILO >> error, isn't it??? (and yes, this disk previously had LILO loaded on it) >> >> I've even gone in, in rescue mode, run chroot, and performed >> 'grub-install >> /dev/sda', and rebooting I get the same error. >> >> I'm at a total loss! >> >> Here are some goodies from my ks.cfg file: >> >> >> install >> cdrom >> lang en_US.UTF-8 >> langsupport --default=en_US.UTF-8 en_US.UTF-8 >> keyboard us >> xconfig --card "ATI Mach64" --videoram 8128 --hsync 30-61 --vsync 56-76 >> --resolution 800x600 --depth 16 >> network --network setup >> rootpw --iscrypted $blablahblah >> firewall --enabled --port=22:tcp >> selinux --enforcing >> authconfig --enableshadow --enablemd5 >> timezone America/New_York >> # The following is the partition information you requested >> # Note that any partitions you deleted are not expressed >> # here so unless you clear all partitions first, this is >> # not guaranteed to work >> clearpart --all >> zerombr yes >> part /boot --fstype "ext3" --ondisk sda --start 1 --end 13 >> part /usr --fstype "ext3" --ondisk sda --start 14 --end 2563 >> part /opt --fstype "ext3" --ondisk sda --start 2564 --end 3838 >> part /export --fstype "ext3" --ondisk sda --start 3839 --end 5113 >> part /tmp --fstype "ext3" --ondisk sda --start 5114 --end 6133 >> part /usr/local --fstype "ext3" --ondisk sda --start 6134 --end 6898 >> part /var --fstype "ext3" --ondisk sda --start 6899 --end 7535 >> part swap --ondisk sda --start 7536 --end 7599 >> part / --fstype "ext3" --ondisk sda --start 7600 --end 8924 >> part /junk --fstype "ext3" --ondisk sdb --start 1 --end 8924 >> bootloader --driveorder=sda,sdb --location=mbr --md5pass=$blahblahblah >> >> %packages >> ... etc >> >> Any help would be greatly appriciated. >> >> >> _______________________________________________ >> 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 barjunk at attglobal.net Sat Apr 2 21:35:47 2005 From: barjunk at attglobal.net (Barsalou) Date: Sat, 02 Apr 2005 12:35:47 -0900 Subject: e1000 still In-Reply-To: <42413A32.8030005@science.uva.nl> References: <42413A32.8030005@science.uva.nl> Message-ID: <1112477747.8284.12.camel@localhost.localdomain> Although this is only a story with no technical backup, we have experienced issues using a Cisco 2950 with some ethernet devices not properly negotiating. We have spanning tree in portfast mode AND turned trunk mode to access. This is actually important with regard to OS X machines getting connected. The solution we have used is to put a hub/switch between the two devices to make things work. In our case, the device was a Tally printer. We didn't spend lots of time trying to determine the problem, but by changing the connection from a 2950 to an older 1900 we were able to get things to work. So it is definitely something with that particular ASIC or the way it was programmed. Hopefully someone with great technical expertise than myself can clarify for us. Mike B. On Wed, 2005-03-23 at 10:43 +0100, Jeroen Roodhart wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > |I don't believe there is a need to wait 30 or 45 seconds for spanning > |tree calculations on the switch. There is something much more subtle > |going on here. As I mentioned before, we've tried capturing traffic > |with a passive analyzer between the box being built and the switch, and > |just this change to the network connections allow the install to >proceed > |normally (as does inserting a small 10/100 hub inline, or replacing the > |gig switch with a 100Mb switch.) It isn't about spanning tree > |calculations on the switch, I think there is some sub-second change in > |the characteristics of the negotiation or anaconda's network setup, >when > |spanning tree is on or off, that make a difference. > | > |-Ed > > I just want to support Ed here, this is the _exact_ behaviour that we > experienced. > > I really don't see the issue here. I you retry try when it fails, you > just make the proces more fault tolerant (normal programming I think, > especially [as said before] when you use UDP). > > If it works in one go, you don't have any adverse effects (it's just as > fast) and if it doesn't, it works with a short delay. What's the > deal? > > In the mean time, when we experience this behaviour again (seems to > have dissapeared when going to FC3) I'll patch the code myself. > > With kind regards, > > Jeroen > - -- > Jeroen Roodhart University of Amsterdam > jeroen at science.uva.nl Faculty of Science / ICT-Group > Systeem- en netwerkbeheer Tel. 020 525 7203 / 06 51338165 > - -- > See http://www.science.uva.nl/~jeroen for openPGP public key > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.2.6 (GNU/Linux) > > iD8DBQFCQToy37AP1zFtDU0RAtnwAJ9RipKbZxzWWSRsbyVW/I39B7sH6wCeN2BS > VAu/Rj9aha/xl0m37IcUGSE= > =TY1P > -----END PGP SIGNATURE----- > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -- Barsalou From athos10 at altern.org Sun Apr 3 07:35:30 2005 From: athos10 at altern.org (athos10) Date: Sun, 03 Apr 2005 09:35:30 +0200 Subject: Kickstart and Authentication Proxy In-Reply-To: <1112362687.993.26.camel@localhost.localdomain> References: <1112362687.993.26.camel@localhost.localdomain> Message-ID: <1112513731.17300.5.camel@christelle> Brian, Do you mean something such as F5 Firepass ? You need to dld only the ks file or using a full network (https/vpn) install ? I didn't understand clearly why your work around is not acceptable. What do you mean in 'This >is obviously not scalable. Since many of these VPN users are full-time >telecommuters, I need to find a scalable way to support them.' ? I hope it'll help me to help you. conf t wr ;-). Athos10 aka michael.aubertin at free.fr Le vendredi 01 avril 2005 ? 08:38 -0500, Brian Long a ?crit : >Does anyone support users trying to kickstart over a VPN connection when >said connection requires authentication on a web site? When our newest >VPN users try to connect to internal sites, they get redirected to a >website which requires their authentication. I'm wondering how people >can perform kickstarts when this type of authentication is required. > >The way a few folks have worked around this is to boot a Knoppix CD, >authenticate on the web site, reboot and the kickstart succeeds. This >is obviously not scalable. Since many of these VPN users are full-time >telecommuters, I need to find a scalable way to support them. > >Any ideas would be greatly appreciated! > >/Brian/ From jean-paul at leclere.org Sun Apr 3 09:12:34 2005 From: jean-paul at leclere.org (=?ISO-8859-1?Q?Jean-PaulLecl=E8re?=) Date: Sun, 03 Apr 2005 11:12:34 +0200 Subject: Installing one more rpm in %post Message-ID: <424FB382.7060501@leclere.org> I have a problem with my distribution. I need to install to different rpm versions with the same name : libstdc++-2.96-113.i386.rpm and libstdc++-3.2-1.i386.rpm. The two versions do not conflicts and can be manually installed together with rpm -i. My installer is a RH 7.3 installer and libstdc++ is set in comps file. Only libstdc++-2.96-113.i386.rpm get installed... My question is : can I install the second rpm in %post section of install.cfg and what is the command to do it : rpm -i libstdc++-3.2-1.i386.rpm (if I do that, will the rpm command find the package) rpm -i /mnt/source/RedHat/RPMS/libstdc++-3.2-1.i386.rpm (and %post --nochroot) (if I do that will the command install the package in the right place ?) Or is there a better way to get the two packages installed ? Thanks for helping me. Jean-Paul Lecl?re -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.308 / Virus Database: 266.9.1 - Release Date: 01/04/2005 From azeem81 at msn.com Mon Apr 4 09:31:09 2005 From: azeem81 at msn.com (azeem ahmad) Date: Mon, 04 Apr 2005 14:31:09 +0500 Subject: isolinux Message-ID: hi list how can i make a hard disk bootable with isolinux Regards Azeem _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ From dan at half-asleep.com Mon Apr 4 12:03:53 2005 From: dan at half-asleep.com (Daniel Segall) Date: Mon, 4 Apr 2005 08:03:53 -0400 (EDT) Subject: Installing one more rpm in %post In-Reply-To: <424FB382.7060501@leclere.org> References: <424FB382.7060501@leclere.org> Message-ID: <41277.192.80.55.74.1112616233.squirrel@webmail.half-asleep.com> It really depends on your environment. The way I do it, is just have any additional packages sitting on an NFS share which gets mounted at %post, and whatever is needed for that install (based on variables in my post install script) gets copied to /root, then installed. This way you are still in chroot. -Dan > I have a problem with my distribution. I need to install to different > rpm versions with the same name : libstdc++-2.96-113.i386.rpm and > libstdc++-3.2-1.i386.rpm. The two versions do not conflicts and can be > manually installed together with rpm -i. > > My installer is a RH 7.3 installer and libstdc++ is set in comps file. > > Only libstdc++-2.96-113.i386.rpm get installed... > > My question is : can I install the second rpm in %post section of > install.cfg and what is the command to do it : > > rpm -i libstdc++-3.2-1.i386.rpm (if I do that, will the rpm command find > the package) > > rpm -i /mnt/source/RedHat/RPMS/libstdc++-3.2-1.i386.rpm (and %post > --nochroot) (if I do that will the command install the package in the > right place ?) > > Or is there a better way to get the two packages installed ? > > Thanks for helping me. > > Jean-Paul Lecl?re > > > > -- > No virus found in this outgoing message. > Checked by AVG Anti-Virus. > Version: 7.0.308 / Virus Database: 266.9.1 - Release Date: 01/04/2005 > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > From ajaymulwani at gmail.com Mon Apr 4 13:48:35 2005 From: ajaymulwani at gmail.com (Ajay Mulwani) Date: Mon, 4 Apr 2005 19:18:35 +0530 Subject: Installing one more rpm in %post In-Reply-To: <424FB382.7060501@leclere.org> References: <424FB382.7060501@leclere.org> Message-ID: <5520b27905040406481217363b@mail.gmail.com> > > My question is : can I install the second rpm in %post section of > install.cfg and what is the command to do it : > Yes, you can. > rpm -i libstdc++-3.2-1.i386.rpm (if I do that, will the rpm command find > the package) > No. > rpm -i /mnt/source/RedHat/RPMS/libstdc++-3.2-1.i386.rpm (and %post > --nochroot) (if I do that will the command install the package in the > right place ?) > No, try the following in your ks.cfg (or install.cfg) : ***** ks.cfg OR install.cfg *** %post --nochroot cp /mnt/source/RedHat/RPMS/libstdc++-3.2-1.i386.rpm /mnt/sysimage/tmp %post rpm -i /tmp/libstdc++-3.2-1.i386.rpm **** END *** Ajay From ajaymulwani at gmail.com Mon Apr 4 13:48:35 2005 From: ajaymulwani at gmail.com (Ajay Mulwani) Date: Mon, 4 Apr 2005 19:18:35 +0530 Subject: Installing one more rpm in %post In-Reply-To: <424FB382.7060501@leclere.org> References: <424FB382.7060501@leclere.org> Message-ID: <5520b27905040406481217363b@mail.gmail.com> > > My question is : can I install the second rpm in %post section of > install.cfg and what is the command to do it : > Yes, you can. > rpm -i libstdc++-3.2-1.i386.rpm (if I do that, will the rpm command find > the package) > No. > rpm -i /mnt/source/RedHat/RPMS/libstdc++-3.2-1.i386.rpm (and %post > --nochroot) (if I do that will the command install the package in the > right place ?) > No, try the following in your ks.cfg (or install.cfg) : ***** ks.cfg OR install.cfg *** %post --nochroot cp /mnt/source/RedHat/RPMS/libstdc++-3.2-1.i386.rpm /mnt/sysimage/tmp %post rpm -i /tmp/libstdc++-3.2-1.i386.rpm **** END *** Ajay From lists at xoul.com Tue Apr 5 16:12:25 2005 From: lists at xoul.com (AGHEMO Danilo) Date: Tue, 05 Apr 2005 18:12:25 +0200 Subject: Ethernet probe order (again) Message-ID: <1112717545.5074.103.camel@localhost.localdomain> I'm experiencing some problems trying to install Fedora Core 2 via PXE/Kickstart on some HP ProLiant DL3x0 G3 servers. These server have two Broadcom BCM5703X Gigabit NICs on board and two Intel 82546EB Gigabit NICs inside one PCI slot. The first two NICs are connected to a network switch, while the other two PCI NICs are cross- connected to another server. It happens that the Intel NICs are always probed and found before the Broadcom NICs, so that e1000 is used for eth0 and eth1 while tg3 is used for eth2 and eth3. Actually, I would prefer to use eth0 and eth1 with tg3 and eth2 and eth3 with e1000. Now I'm getting to the point: I would like to load a kickstart file via ksdevice=eth0 and have eth0 refer to my first tg3 card after reboot. How can I "reorder" my network cards at kickstart time? Maybe I'm getting fool, but I'm sure I had done this before: after a manual installation, using lspci I can see that I/O ports for the Intel cards starts at 4000 and 4040 each, with a size of 64 each, so I used a reserve=0x4000,128 along with other append parameters... But this is not working on these server! Maybe I should use some pci=whatever? Any idea? I can't use ksdevice=eth0 to load my kickstart file, but if I use ksdevice=eth2 I can manage to install my server. But after reboot I can't reach my server, since eth0 is my cross-connected e1000 card. I've even tried to place a "device eth tg3" inside my kickstart file, but with no success. What kind of param should I pass to my kernel to have ksdevice=eth0 refer to my first tg3 card? What can I put inside my kickstart file to "swap" eth0/1 with eth2/3 so that, again, eth0 is my first tg3 card? I know that I can just modify /etc/modprobe.conf after install, but this seems not to be the smartest choice... I've read of a patch by Patrick Device, but I would like not to modify any package... Can you help? From aaron at activehotels.com Wed Apr 6 10:11:57 2005 From: aaron at activehotels.com (Aaron Walker) Date: Wed, 06 Apr 2005 11:11:57 +0100 Subject: adding packages to RHEL-3up4 for a kickstart install Message-ID: <4253B5ED.8030604@activehotels.com> I currently have my RHEL3_x86_64 distro copied to /exports/rhel-3. I am using sun v20z's. Dual opterons with 4GB RAM. I can successfully install other boxes using pxe/kickstart. What I now want to do is to install customised RPMS when doing a kiskstart install. This is what I have done. Created my RPM's and copied them over to /exports/rhel-3/RedHat/RPMS run the genhdlist command /usr/lib/anaconda-runtime/genhdlist /exports/rhel-3 I am then running a kickstart install without any changes to the kickstart config file.( this was to see if the changes so far, would impact the install) The install appears to complete successfully. However on the first reboot of the new server (aka the client). Grub fails to find the kernel image. So I have booted the client up using knoppix . It appears the installation has been completed with the exception of the kernel. There are no kernel modules in /lib and /boot is empty except for the grub directory. Any help would be much appreciated. From redhat at iamafreeman.com Wed Apr 6 16:24:08 2005 From: redhat at iamafreeman.com (redhat at iamafreeman.com) Date: Wed, 6 Apr 2005 17:24:08 +0100 Subject: adding packages to RHEL-3up4 for a kickstart install In-Reply-To: <4253B5ED.8030604@activehotels.com> References: <4253B5ED.8030604@activehotels.com> Message-ID: <200504061724.08506.redhat@iamafreeman.com> merging you own rpms into redhats is complicated by the need to get anaconda to accept the package signature, you will have to sign them and make your own anaconda (unless there is an easier way anyone?) this is difficult enough that I suspect most (including me) just rpm -Uvh path in the post section running one of the open up2date server (or shelling out for rhn proxy) will simplify that even if they are redhats rpms (say updates) you'll need to create a pkgorder file (with pkgorder) and pass that to genhdlist Neil On Wednesday 06 Apr 2005 11:11, Aaron Walker wrote: > I currently have my RHEL3_x86_64 distro copied to /exports/rhel-3. > I am using sun v20z's. Dual opterons with 4GB RAM. > > I can successfully install other boxes using pxe/kickstart. > > What I now want to do is to install customised RPMS when doing a > kiskstart install. > > This is what I have done. > Created my RPM's and copied them over to /exports/rhel-3/RedHat/RPMS > run the genhdlist command > /usr/lib/anaconda-runtime/genhdlist /exports/rhel-3 > > I am then running a kickstart install without any changes to the > kickstart config file.( this was to see if the changes so far, would > impact the install) > > The install appears to complete successfully. However on the first > reboot of the new server (aka the client). Grub fails to find the > kernel image. > So I have booted the client up using knoppix . It appears the > installation has been completed with the exception of the kernel. There > are no kernel modules in /lib and > /boot is empty except for the grub directory. > > Any help would be much appreciated. > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From steve.mah at oracle.com Wed Apr 6 16:53:15 2005 From: steve.mah at oracle.com (Stephen Mah) Date: Wed, 06 Apr 2005 09:53:15 -0700 Subject: how to prevent Mozilla from being installed? Message-ID: <425413FB.7030103@oracle.com> For some reason Mozilla is being install during a kickstart installation. I've check the comps.xml file to make sure it's marked as optional or removed completely. I do see that the rpm "devhelp" requires mozilla. It's listed under x-software-dev. Can I remove this? I'll try for my next build. thanks steve From christopher.hotchkiss at gmail.com Thu Apr 7 10:32:50 2005 From: christopher.hotchkiss at gmail.com (Christopher Hotchkiss) Date: Thu, 7 Apr 2005 06:32:50 -0400 Subject: how to prevent Mozilla from being installed? In-Reply-To: <425413FB.7030103@oracle.com> References: <425413FB.7030103@oracle.com> Message-ID: <7f48492a05040703322c021c0e@mail.gmail.com> On Apr 6, 2005 12:53 PM, Stephen Mah wrote: > > For some reason Mozilla is being install during a kickstart installation. > > I've check the comps.xml file to make sure it's marked as optional or > removed completely. > > I do see that the rpm "devhelp" requires mozilla. It's listed under > x-software-dev. Can I remove this? If you aren't going to need the help packages I don't see why not. BTW:Last time I checked Mozilla was a dependancy of Gnome as well. -- Christopher Hotchkiss (813)960-9273 http://www.post227.org -------------- next part -------------- An HTML attachment was scrubbed... URL: From drew at sundawg.org Thu Apr 7 13:19:41 2005 From: drew at sundawg.org (drew at sundawg.org) Date: Thu, 7 Apr 2005 08:19:41 -0500 Subject: Physical Volume names Message-ID: <20050407131941.GB18040@pooh.internal.org> Hi, I'm new to the list, but not so new to linux. I'm working on a project, and I've hit a bit of a stumbling block. I'm trying to write a script to basically generate a "how things look now" kickstart file. I've had good success, except for determining the pv numbers that are used to build a volume group. Unfortunately, I only have one test system, but when installing this particular system, I created three partitions. One was /boot, the other two I designated as LVM PVs. These were numbered pv.8 and pv.10. After searching through /proc and all the LVM commands I can think of, I can't seem to correlate pv.8 and pv.10 to anything (in fact, I can find no mention of those "names" outside kickstart at all). Any suggestions? drew -- Never offend people with style when you can offend them with substance. - Sam Brown, Washington Post, 1977 From SSeremeth at anacomp.com Thu Apr 7 14:28:48 2005 From: SSeremeth at anacomp.com (Seremeth, Stephen) Date: Thu, 7 Apr 2005 10:28:48 -0400 Subject: Physical Volume names Message-ID: <66911C8E1F2DA24EB29DE289297A135E4855F3@usrd104> > Unfortunately, I only have one test system, but when installing this > particular system, I created three partitions. One was > /boot, the other > two I designated as LVM PVs. These were numbered pv.8 and > pv.10. After > searching through /proc and all the LVM commands I can think > of, I can't > seem to correlate pv.8 and pv.10 to anything (in fact, I can find no > mention of those "names" outside kickstart at all). > I believe you're looking for: lvdisplay -m From drew at sundawg.org Thu Apr 7 22:20:00 2005 From: drew at sundawg.org (drew) Date: Thu, 7 Apr 2005 17:20:00 -0500 Subject: Physical Volume names In-Reply-To: <66911C8E1F2DA24EB29DE289297A135E4855F3@usrd104> References: <66911C8E1F2DA24EB29DE289297A135E4855F3@usrd104> Message-ID: <20050407222000.GA5127@pooh.internal.org> * Seremeth, Stephen (SSeremeth at anacomp.com) wrote: > > Unfortunately, I only have one test system, but when installing this > > particular system, I created three partitions. One was > > /boot, the other > > two I designated as LVM PVs. These were numbered pv.8 and > > pv.10. After > > searching through /proc and all the LVM commands I can think > > of, I can't > > seem to correlate pv.8 and pv.10 to anything (in fact, I can find no > > mention of those "names" outside kickstart at all). > > I believe you're looking for: > > lvdisplay -m Thanks for the reply Stephen. lvdisplay -m on my system throws an error. I guess I neglected to mention, this is on RHEL 3, which I guess uses LVM1. My RHEL4 systems have LVM2 and that command works fine. Let me clarify my question a bit. Are the pv numbers arbitrary? Say my kickstart file created during installation contains the following: part /boot --fstype "ext3" --onpart hda1 part pv.8 --noformat --onpart hda2 part pv.10 --noformat --onpart hda3 volgroup vg00 --pesize=4096 pv.8 pv.10 If I wanted to write a new, but equivalent kickstart file, could I use the following: part /boot --fstype "ext3" --onpart hda1 part pv.1 --noformat --onpart hda2 part pv.2 --noformat --onpart hda3 volgroup vg00 --pesize=4096 pv.1 pv.2 Or do the pv.#s actually correspond to a partition? If so, how do I derive that relationship from an existing installation? Thanks, drew -- Don't break your shin on a stool that is not in your way. - Irish Proverb From phr at doc.ic.ac.uk Sat Apr 9 12:50:49 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Sat, 9 Apr 2005 13:50:49 +0100 (BST) Subject: Kickstart and Authentication Proxy In-Reply-To: <1112362687.993.26.camel@localhost.localdomain> References: <1112362687.993.26.camel@localhost.localdomain> Message-ID: On Fri, 1 Apr 2005, Brian Long wrote: >Does anyone support users trying to kickstart over a VPN connection when >said connection requires authentication on a web site? When our newest >VPN users try to connect to internal sites, they get redirected to a >website which requires their authentication. I can't quite picture the scenario. External users needing kickstart installs are hitting (say) kickstart.private.example.com, which redirects off to a login page? Which network resources are being accessed (kickstart file, 2nd-stage loader, install dir, %post stuff), and how are the users being asked to boot the install? To generalize, if everything works *except* the awkward login/redirect, you might be able to do something with %pre, dialog(1) and wget/curl to prompt and login at the beginning of kickstarting. Cheers, Phil From phr at doc.ic.ac.uk Sat Apr 9 13:20:58 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Sat, 9 Apr 2005 14:20:58 +0100 (BST) Subject: adding packages to RHEL-3up4 for a kickstart install In-Reply-To: <4253B5ED.8030604@activehotels.com> References: <4253B5ED.8030604@activehotels.com> Message-ID: On Wed, 6 Apr 2005, Aaron Walker wrote: >This is what I have done. > Created my RPM's and copied them over to /exports/rhel-3/RedHat/RPMS > run the genhdlist command > /usr/lib/anaconda-runtime/genhdlist /exports/rhel-3 I've not been paying close attention, but there is/was a bug regarding package ordering which made this sort of rebuild more complicated. Try searching for genhdlist and pkgorder for the exact incantation. Cheers, Phil From timm at fnal.gov Wed Apr 13 14:38:02 2005 From: timm at fnal.gov (Steven Timm) Date: Wed, 13 Apr 2005 09:38:02 -0500 (CDT) Subject: cups Message-ID: I am wondering if it is possible to veto certain packages from a group, either at kickstart time or by actually modifying the comps.xml file. For instance: I want to include the X Window System group in my install, so I could say @X Window System in kickstart. But X Windows System includes printing tools as a subcategory and none of my nodes are ever going to print. So can I at kickstart time veto an rpm? In other words, install all of the groups above, but not the cups rpm? Or can I do it in the comps.xml by defining a new group which includes the X Windows Systems group, but vetoes the one rpm? Steve Timm ------------------------------------------------------------------ Steven C. Timm, Ph.D (630) 840-8525 timm at fnal.gov http://home.fnal.gov/~timm/ Fermilab Computing Div/Core Support Services Dept./Scientific Computing Section Assistant Group Leader, Farms and Clustered Systems Group Lead of Computing Farms Team From phr at doc.ic.ac.uk Wed Apr 13 15:02:59 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 13 Apr 2005 16:02:59 +0100 (BST) Subject: cups In-Reply-To: References: Message-ID: On Wed, 13 Apr 2005, Steven Timm wrote: >I am wondering if it is possible to veto certain packages >from a group, either at kickstart time or by actually modifying >the comps.xml file. You can write stuff like this in the kickstart file: @ Base @ X Window System -cups -nc -bzip2-devel but bear in mind that a dependency check *might* re-include the package automatically. Cheers, Phil From cmadams at hiwaay.net Wed Apr 13 15:26:24 2005 From: cmadams at hiwaay.net (Chris Adams) Date: Wed, 13 Apr 2005 10:26:24 -0500 Subject: cups In-Reply-To: References: Message-ID: <20050413152624.GF1206118@hiwaay.net> Once upon a time, Steven Timm said: > I am wondering if it is possible to veto certain packages > from a group, either at kickstart time or by actually modifying > the comps.xml file. Read the kickstart docs. > For instance: > I want to include the X Window System group in my install, so > I could say > > @X Window System > > in kickstart. But X Windows System includes printing tools as a > subcategory and none of my nodes are ever going to print. @ X Window System -cups Make sure nothing else selected depends on cups though (or it'll get pulled right back in). -- Chris Adams Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble. From jerome.tournier at idealx.com Thu Apr 14 15:47:16 2005 From: jerome.tournier at idealx.com (Jerome Tournier) Date: Thu, 14 Apr 2005 17:47:16 +0200 Subject: hostname don't asked when not set in ks.cfg Message-ID: <20050414174716.A30261@sheffield.idealx.com> Hello every body, i set up a kickstart file to do everything automatically, exept the hostname and network configuration. I whant those parameters to be asked at the begining of the installation. So nothing about that appear in ks.cfg. The network configuration is automatically asked, but not the hostname one. Anyone have an idea about how to force anaconda to ask me this ? Thanks ! -- Jerome From Eric.Doutreleau at int-evry.fr Fri Apr 15 08:07:31 2005 From: Eric.Doutreleau at int-evry.fr (Eric Doutreleau) Date: Fri, 15 Apr 2005 10:07:31 +0200 Subject: cfengine for maintaining kickstart files Message-ID: <1113552451.5155.10.camel@rezo.int-evry.fr> Hi we re maitaining hundreds of linux machine with a lot of different profiles. for that we use kickstart for installation and yum and cfengine for maintaining the machine. but right now mainting our kickstart files is becoming really a mess. we have a lot of different kickstart files and it s really difficult to maintain all these files. for exemple when we want to change the root password for a small subset of our machines it becomes a pain in all the kickstart files. Has someone already succeeded in maintaining all its kickstart file up to date? i'm considering using cfengine for that in the "kickstart" server. any ideas about that? From jss at ast.cam.ac.uk Fri Apr 15 16:52:59 2005 From: jss at ast.cam.ac.uk (Jeremy Sanders) Date: Fri, 15 Apr 2005 17:52:59 +0100 (BST) Subject: Kickstart root on raid problems Message-ID: Hi - I'm trying to use kickstart to install root on raid using preexisting partitions. THis used to work with the kickstart in RH 7.3, but seems to fail with anaconda-10.1.1.13. The partitioning information is as follows in the kickstart file: part raid.01 --onpart hda1 part raid.02 --onpart hda2 part raid.03 --onpart hda3 --noformat part raid.04 --onpart hdb1 part raid.05 --onpart hdb2 part raid.06 --onpart hdb3 --noformat raid / --level=1 --device=md0 --fstype ext3 raid.01 raid.04 raid swap --level=1 --device=md1 --fstype swap raid.02 raid.05 raid /soft3 --level=1 --device=md2 --fstype ext3 --noformat raid.03 raid.06 I've tried removing everything except the root partitioning but it still fails in partsize = req.getActualSize(partitions, diskset) AttributeError: 'NoneType' object has no attribute 'getActualSize' You can see the full anaconda dump at http://www-xray.ast.cam.ac.uk/~jss/data/anacdump.txt I've tried adding --size 1 to each of the part options with no effect. I've also tried using --noformat on each of the part options. Each of the partition types is set to RAID autodetect, and they contain preexisting raid devices (with the same partitions as the install is using). Any ideas?? Thanks Jeremy -- Jeremy Sanders http://www-xray.ast.cam.ac.uk/~jss/ X-Ray Group, Institute of Astronomy, University of Cambridge, UK. Public Key Server PGP Key ID: E1AAE053 From ebrown at lanl.gov Fri Apr 15 17:06:46 2005 From: ebrown at lanl.gov (Ed Brown) Date: Fri, 15 Apr 2005 11:06:46 -0600 Subject: cfengine for maintaining kickstart files In-Reply-To: <1113552451.5155.10.camel@rezo.int-evry.fr> References: <1113552451.5155.10.camel@rezo.int-evry.fr> Message-ID: <1113584806.24763.153.camel@edbrown.lanl.gov> You might consider looking at how much of what distinguishes your kickstart files could actually be done with cfengine instead. Your example of root passwords is exactly the kind of thing that could be in your cfengine configuration files, rather than in kickstart. Even packages, above and beyond the base and core groups, can be installed via cfengine, rather than the packages list in the kickstart file. And adding a little runtime flexibility to kickstart can take care of the fringe cases, avoiding the need for a special kickstart file for one or two machines. I've posted before about how to allow for runtime changes, to an otherwise default partitioning scheme for example. -Ed On Fri, 2005-04-15 at 02:07, Eric Doutreleau wrote: > Hi > > we re maitaining hundreds of linux machine with a lot of different > profiles. > for that we use kickstart for installation and yum and cfengine for > maintaining the machine. > > but right now mainting our kickstart files is becoming really a mess. > we have a lot of different kickstart files and it s really difficult to > maintain all these files. > > for exemple when we want to change the root password for a small subset > of our machines it becomes a pain in all the kickstart files. > > Has someone already succeeded in maintaining all its kickstart file up > to date? > > i'm considering using cfengine for that in the "kickstart" server. > any ideas about that? > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From drkludge at cox.net Sat Apr 16 00:11:53 2005 From: drkludge at cox.net (Greg Morgan) Date: Fri, 15 Apr 2005 17:11:53 -0700 Subject: Rau Wiki Message-ID: <42605849.5060408@cox.net> Does anyone know if the anaconda wiki hosted by rau.homedns.org is still available or is the server down right now? There was some talk in February 2005 that it was moribund. Paul didn't take a gun and shot it because it was almost dead did he? ;-) Thanks, Greg From alexander_rau at yahoo.com Sat Apr 16 00:37:49 2005 From: alexander_rau at yahoo.com (Alexander Rau) Date: Fri, 15 Apr 2005 17:37:49 -0700 (PDT) Subject: Rau Wiki In-Reply-To: 6667 Message-ID: <20050416003750.5079.qmail@web52106.mail.yahoo.com> Unfortunately the wiki was down due to the fact that my provider closed shop without letting us know. I moved the site and it should be up and running fairly shortly. Let me know should you still have issues. I apologize for any inconvenience. Thanks AR --- Greg Morgan wrote: > Does anyone know if the anaconda wiki hosted by > rau.homedns.org is still > available or is the server down right now? There > was some talk in > February 2005 that it was moribund. Paul didn't > take a gun and shot it > because it was almost dead did he? ;-) > > Thanks, > Greg > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From gwaugh at frontstreetnetworks.com Sat Apr 16 15:22:02 2005 From: gwaugh at frontstreetnetworks.com (Gerald Waugh) Date: Sat, 16 Apr 2005 11:22:02 -0400 Subject: Kivckstart vis USB cdrom Message-ID: <1113664922.1684.100.camel@soyo.raqware.com> Hi, First post; We are installing CentOS4 via a USB cdrom, and have a problem. The server boots just fine from the USB cdrom. At the boot prompt we enter "linux ks=cdrom". If we do nothing else, the install drops into 'text mode' If we pull the USB cable and reinsert when the console reports "running /sbin/loader" the kickstart script is run and everything is fine. Any ideas as to how to fix this, much appreciated. TIA Gerald From phr at doc.ic.ac.uk Sat Apr 16 16:02:44 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Sat, 16 Apr 2005 17:02:44 +0100 (BST) Subject: Kivckstart vis USB cdrom In-Reply-To: <1113664922.1684.100.camel@soyo.raqware.com> References: <1113664922.1684.100.camel@soyo.raqware.com> Message-ID: On Sat, 16 Apr 2005, Gerald Waugh wrote: >First post; Welcome. >We are installing CentOS4 via a USB cdrom, and have a problem. >The server boots just fine from the USB cdrom. >At the boot prompt we enter "linux ks=cdrom". >If we do nothing else, the install drops into 'text mode' With a language prompt? That happens when the kickstart file can't be found (for whatever reason). >If we pull the USB cable and reinsert when the console reports "running >/sbin/loader" the kickstart script is run and everything is fine. Sounds like: - The BIOS is using the CD drive correctly as a boot device. - ISOLINUX is discovering the kernel and initrd properly - The kernel doesn't realise, unless triggered by the hot-plug, that the cdrom is what you mean by ks=cdrom At a guess, either try ks=/dev/scd?? if the drive shows up as a SCSI device, or see if there's are kernel USB/CD/SCSI options which can be passed in with ks=cdrom to influence the probe. Cheers, Phil From gwaugh at frontstreetnetworks.com Sat Apr 16 16:37:33 2005 From: gwaugh at frontstreetnetworks.com (Gerald Waugh) Date: Sat, 16 Apr 2005 12:37:33 -0400 Subject: Kivckstart vis USB cdrom In-Reply-To: References: <1113664922.1684.100.camel@soyo.raqware.com> Message-ID: <1113669453.1684.110.camel@soyo.raqware.com> On Sat, 2005-04-16 at 17:02 +0100, Philip Rowlands wrote: > Sounds like: > - The BIOS is using the CD drive correctly as a boot device. > - ISOLINUX is discovering the kernel and initrd properly > - The kernel doesn't realise, unless triggered by the hot-plug, that the > cdrom is what you mean by ks=cdrom > > At a guess, either try ks=/dev/scd?? if the drive shows up as a SCSI > device, or see if there's are kernel USB/CD/SCSI options which can be > passed in with ks=cdrom to influence the probe. Thanks for the reply; I booted the server (running with a previous install) with a regular CD in the usb cdrom. the OS recognizes the cdrom as /dev/scd0. It is in /etc/fstab as /dev/scd0 /media/cdrecorder auto pamconsole,exec,noauto,managed 0 0 Yet if I use "ks=/dev/scd0" I get; "loading USB storage driver" So it got that part correct... Then I get "Error opening kickstart file (null): Bad Address" It seems as though the BIOS is possibly not releasing the connection to the usb cdrom, so it is busy? Which might explain why pulling the plug and reinserting it works? Gerald From phr at doc.ic.ac.uk Sat Apr 16 20:59:30 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Sat, 16 Apr 2005 21:59:30 +0100 (BST) Subject: Kivckstart vis USB cdrom In-Reply-To: <1113669453.1684.110.camel@soyo.raqware.com> References: <1113664922.1684.100.camel@soyo.raqware.com> <1113669453.1684.110.camel@soyo.raqware.com> Message-ID: On Sat, 16 Apr 2005, Gerald Waugh wrote: >I booted the server (running with a previous install) with a regular CD >in the usb cdrom. the OS recognizes the cdrom as /dev/scd0. >It is in /etc/fstab as >/dev/scd0 /media/cdrecorder auto pamconsole,exec,noauto,managed 0 0 > >Yet if I use "ks=/dev/scd0" I get; >"loading USB storage driver" So it got that part correct... >Then I get "Error opening kickstart file (null): Bad Address" Erm, that won't work... The syntax would be something like ks=file:???, but I don't know if/where it would be mounted. Do any of these documented options help?? nousb Do not load USB support (helps if install hangs early sometimes). nousbstorage Do not load usbstorage module in loader. May help with device ordering on SCSI systems. (taken from /usr/share/doc/anaconda-*/command-line.txt) Cheers, Phil From nate at math.umn.edu Tue Apr 19 22:11:35 2005 From: nate at math.umn.edu (Nate Steffan) Date: Tue, 19 Apr 2005 17:11:35 -0500 Subject: Options for grabbing kickstart Message-ID: <42658217.40405@math.umn.edu> I have a question regarding what are the different ways a kickstart can be called using the fedora core rescue cd. We currently are using nfs, however due to recent nfs dameon issues with our file servers we are looking to switch over to either afs or samba. More likely afs. My question what are the different ways the kickstart can be called from the fedora core. Is there a way to call it from a http source or any other networked source. For example: ks=http://pathtodomain/kickdirectory/ks.cfg Something like that. We are basically looking for any alternatives to nfs. Thanks, Nate Steffan -------------- next part -------------- A non-text attachment was scrubbed... Name: nate.vcf Type: text/x-vcard Size: 338 bytes Desc: not available URL: From tdiehl at rogueind.com Tue Apr 19 22:31:07 2005 From: tdiehl at rogueind.com (Tom Diehl) Date: Tue, 19 Apr 2005 18:31:07 -0400 (EDT) Subject: Options for grabbing kickstart In-Reply-To: <42658217.40405@math.umn.edu> References: <42658217.40405@math.umn.edu> Message-ID: On Tue, 19 Apr 2005, Nate Steffan wrote: > I have a question regarding what are the different ways a kickstart can > be called using the fedora core rescue cd. > > We currently are using nfs, however due to recent nfs dameon issues with > our file servers we are looking to switch over to either afs or samba. > More likely afs. My question what are the different ways the kickstart > can be called from the fedora core. Is there a way to call it from a > http source or any other networked source. For example: > ks=http://pathtodomain/kickdirectory/ks.cfg That is what I do. Look in the anaconda docs. Something like /usr/share/doc/anaconda* should get the info you want. Make sure you have the anaconda package installed of course. Regards, Tom Diehl tdiehl at rogueind.com Spamtrap address mtd123 at rogueind.com From samuli.jarvinen at hp.com Wed Apr 20 07:48:03 2005 From: samuli.jarvinen at hp.com (Samuli Jarvinen) Date: Wed, 20 Apr 2005 10:48:03 +0300 Subject: Kivckstart vis USB cdrom In-Reply-To: <1113664922.1684.100.camel@soyo.raqware.com> References: <1113664922.1684.100.camel@soyo.raqware.com> Message-ID: <1113983283.8907.8.camel@sajarnb> You can place the kickstart file ks.cfg inside the initrd.img located in the cdrom. Then boot up the installer with command linux ks=file:/ks.cfg That's how I'm going around that problem when I'm not able to get the kickstart via http. -Samuli On Sat, 2005-04-16 at 11:22 -0400, Gerald Waugh wrote: > Hi, > > First post; > We are installing CentOS4 via a USB cdrom, and have a problem. > The server boots just fine from the USB cdrom. > At the boot prompt we enter "linux ks=cdrom". > If we do nothing else, the install drops into 'text mode' > > If we pull the USB cable and reinsert when the console reports > "running /sbin/loader" the kickstart script is run and everything is > fine. > > Any ideas as to how to fix this, much appreciated. > > TIA > Gerald > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From Christian.Rohrmeier at SCHERING.DE Thu Apr 21 14:26:36 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Thu, 21 Apr 2005 16:26:36 +0200 Subject: Kickstart %include directive and %pre script issues Message-ID: Hello Dear Kickstarters, I'm running RHEL 2.1, 3, and 4, all of which need to be served by the kickstart script in question. I have found a problem with the "%include" directive when including a file that is created by the %pre script. What I want to do: dynamic disk partitioning. A great example of this can be found on page 56 of the RHEL 3 System Administration Guide. The idea is, in your %pre script you check out the disk situation, or ask the user something, and write the apropriate kickstart partitioning commands (as they would normally be found in the commands section of the kickstart script) into a file. This file was already included (via the %include directive) in the commands section of the kickstart file. And thats the catch: while anaconda parses the kickstart script, it encounters the %include before it executes the the %pre script, and it dies a horrible death, spewing python errors, because it can't get to those precious "part" commands. Alternatively, if one places the "%include" line alllllllll the way at the end of the command section, as the very last item before the %pre script starts, then Anaconda doesn't die. Its runs the %pre script no problem, which in this case write the file that is included via %include. However, it then starts asking about how to partition the drive. This means that anaconda failed to include the %include. Position of the %include directive in the command section of the kickstart script is therefore obviously important. Unfortunately it is not documented where, relative to other commands and the entire kickstart script in general, the %include directive must be placed in order to successfully use it to include partitioning information gleaned from a %pre script. Your help would be greatly appreciated. Sincerely, Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 From danny280279 at hotmail.com Thu Apr 21 14:29:31 2005 From: danny280279 at hotmail.com (DANIEL hoggan) Date: Thu, 21 Apr 2005 14:29:31 +0000 Subject: Kickstart and anaconda Message-ID: I don't know if this is the right question to ask, but if I use anaconda to install my distro what does kickstart do? _________________________________________________________________ Want to block unwanted pop-ups? Download the free MSN Toolbar now! http://toolbar.msn.co.uk/ From marhoy at gmail.com Thu Apr 21 14:35:23 2005 From: marhoy at gmail.com (=?ISO-8859-1?Q?Martin_H=F8y?=) Date: Thu, 21 Apr 2005 16:35:23 +0200 Subject: Kickstart %include directive and %pre script issues In-Reply-To: References: Message-ID: <9307250504210735749b9c11@mail.gmail.com> On 4/21/05, Christian.Rohrmeier at schering.de wrote: > Hello Dear Kickstarters, > > I'm running RHEL 2.1, 3, and 4, all of which need to be served by the > kickstart script in question. > > I have found a problem with the "%include" directive when including a file > that is created by the %pre script. > > What I want to do: dynamic disk partitioning. A great example of this can > be found on page 56 of the RHEL 3 System Administration Guide. The idea is, > in your %pre script you check out the disk situation, or ask the user > something, and write the apropriate kickstart partitioning commands (as > they would normally be found in the commands section of the kickstart > script) into a file. This file was already included (via the %include > directive) in the commands section of the kickstart file. > > And thats the catch: while anaconda parses the kickstart script, it > encounters the %include before it executes the the %pre script, and it dies > a horrible death, spewing python errors, because it can't get to those > precious "part" commands. I'm doing exactly that, and it works. The %pre-script checks the present partition layout, and decides wether it will re-use the existing partitions or create new ones. See the attached ks.cfg. Sorry about the mixed english/norwegian in the script. We're just - norwegian. Regards, Martin -------------- next part -------------- A non-text attachment was scrubbed... Name: ks.cfg Type: application/octet-stream Size: 5913 bytes Desc: not available URL: From phr at doc.ic.ac.uk Thu Apr 21 14:50:18 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 21 Apr 2005 15:50:18 +0100 (BST) Subject: Kickstart and anaconda In-Reply-To: References: Message-ID: On Thu, 21 Apr 2005, DANIEL hoggan wrote: >I don't know if this is the right question to ask, but if I use >anaconda to install my distro what does kickstart do? In a nutshell... It allows you to give pre-defined answers to the questions which the installer (anaconda) asks regarding basic system setup, disk partitioning, package selection etc. Also, there is provision for some post-installation scripting to perform further customisation. Hope this helps, Phil From Christian.Rohrmeier at SCHERING.DE Thu Apr 21 16:09:39 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Thu, 21 Apr 2005 18:09:39 +0200 Subject: Kickstart %include directive and %pre script issues Message-ID: Hi Martin, Well, that went quick! Thanks for your reply. I found the problem. I notice that your example is on RHEL4, and the the other example I have is in RHEL 3 documentation. Because of its small size, I develop with RHEL2.1 (it installs so quickly!). And thats the problem: RHEL 2.1 does not support the "%include" directive. I tested my Kickstart script with %include on RHEL 3 and 4, and there it works fine. Does anyone have a solution for dynamic partitioning under RHEL 2.1??? Mit freundlichen Gr??en, Christian Rohrmeier _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 Martin H?y Sent by: To: Discussion list about Kickstart kickstart-list-bounces cc: @redhat.com Subject: Re: Kickstart %include directive and %pre script issues 21.04.2005 16:35 Please respond to martin.hoy; Please respond to Discussion list about Kickstart On 4/21/05, Christian.Rohrmeier at schering.de wrote: > Hello Dear Kickstarters, > > I'm running RHEL 2.1, 3, and 4, all of which need to be served by the > kickstart script in question. > > I have found a problem with the "%include" directive when including a file > that is created by the %pre script. > > What I want to do: dynamic disk partitioning. A great example of this can > be found on page 56 of the RHEL 3 System Administration Guide. The idea is, > in your %pre script you check out the disk situation, or ask the user > something, and write the apropriate kickstart partitioning commands (as > they would normally be found in the commands section of the kickstart > script) into a file. This file was already included (via the %include > directive) in the commands section of the kickstart file. > > And thats the catch: while anaconda parses the kickstart script, it > encounters the %include before it executes the the %pre script, and it dies > a horrible death, spewing python errors, because it can't get to those > precious "part" commands. I'm doing exactly that, and it works. The %pre-script checks the present partition layout, and decides wether it will re-use the existing partitions or create new ones. See the attached ks.cfg. Sorry about the mixed english/norwegian in the script. We're just - norwegian. Regards, Martin (See attached file: ks.cfg)_______________________________________________ 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: ks.cfg Type: application/octet-stream Size: 5913 bytes Desc: not available URL: From mikem.rtp at gmail.com Thu Apr 21 16:15:15 2005 From: mikem.rtp at gmail.com (Mike McLean) Date: Thu, 21 Apr 2005 12:15:15 -0400 Subject: Kickstart %include directive and %pre script issues In-Reply-To: References: Message-ID: <4f50e06805042109153de4e6ae@mail.gmail.com> On 4/21/05, Christian.Rohrmeier at schering.de wrote: > Does anyone have a solution for dynamic partitioning under RHEL 2.1??? Dynamically generate the kickstart externally (or before booting the installer). From email at jasonkohles.com Thu Apr 21 16:24:36 2005 From: email at jasonkohles.com (Jason Kohles) Date: Thu, 21 Apr 2005 12:24:36 -0400 Subject: Kickstart %include directive and %pre script issues In-Reply-To: References: Message-ID: <20050421162436.GA19868@mail.jasonkohles.com> On Thu, Apr 21, 2005 at 04:26:36PM +0200, Christian.Rohrmeier at SCHERING.DE wrote: > Hello Dear Kickstarters, > > I'm running RHEL 2.1, 3, and 4, all of which need to be served by the > kickstart script in question. > > I have found a problem with the "%include" directive when including a file > that is created by the %pre script. > > What I want to do: dynamic disk partitioning. A great example of this can > be found on page 56 of the RHEL 3 System Administration Guide. The idea is, > in your %pre script you check out the disk situation, or ask the user > something, and write the apropriate kickstart partitioning commands (as > they would normally be found in the commands section of the kickstart > script) into a file. This file was already included (via the %include > directive) in the commands section of the kickstart file. > > And thats the catch: while anaconda parses the kickstart script, it > encounters the %include before it executes the the %pre script, and it dies > a horrible death, spewing python errors, because it can't get to those > precious "part" commands. > It depends on the version of anaconda, the way it worked prior to AS2.1 was that the %pre sections were all run first, then the whole kickstart file was re-read, ignoring the %pre sections. Red Hat decided to change the way this works, and the approach they took was to remove the old way, release AS2.1, and then start working on adding in the new way. Unfortunately what this means is that this really won't work with 2.1, as the code needed to support it doesn't even exist in anaconda for that version. What I ended up doing was backporting the support for this to the AS2.1 installer. You can find a working copy of kickstart.py on my website at http://www.jasonkohles.com/snippets/Red-Hat-Linux/as21-pre-fix.html, drop it your RHupdates folder, and you should get this functionality back. -- Jason Kohles A witty saying proves nothing. email at jasonkohles.com -- Voltaire (1694 - 1778) http://www.jasonkohles.com/ From Christian.Rohrmeier at SCHERING.DE Thu Apr 21 17:00:39 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Thu, 21 Apr 2005 19:00:39 +0200 Subject: Kickstart %include directive and %pre script issues Message-ID: Hi Jason, Thanks for the excellent advice. Rolling my own install CD is a horrifying thought. I so much abhor this idea, that I install all RPMs that aren't on the default install media via APT in the %post script. There is a .../RedHat/instimage/usr/lib/anaconda/kickstart.py in the install media directory on my kickstart server (which I get to via http). I was hoping I could just plop the new kickstart.py in there, but even though its there, its not used it seems. The CD needs to be modified. Thanks for the idea though. If my users complain too much maybe I'll make my own install CD... -Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 Jason Kohles To: Discussion list about Kickstart Sent by: cc: Christian.Rohrmeier at schering.de kickstart-list-bounces Subject: Re: Kickstart %include directive and %pre script issues @redhat.com 21.04.2005 18:24 Please respond to Discussion list about Kickstart On Thu, Apr 21, 2005 at 04:26:36PM +0200, Christian.Rohrmeier at SCHERING.DE wrote: > Hello Dear Kickstarters, > > I'm running RHEL 2.1, 3, and 4, all of which need to be served by the > kickstart script in question. > > I have found a problem with the "%include" directive when including a file > that is created by the %pre script. > > What I want to do: dynamic disk partitioning. A great example of this can > be found on page 56 of the RHEL 3 System Administration Guide. The idea is, > in your %pre script you check out the disk situation, or ask the user > something, and write the apropriate kickstart partitioning commands (as > they would normally be found in the commands section of the kickstart > script) into a file. This file was already included (via the %include > directive) in the commands section of the kickstart file. > > And thats the catch: while anaconda parses the kickstart script, it > encounters the %include before it executes the the %pre script, and it dies > a horrible death, spewing python errors, because it can't get to those > precious "part" commands. > It depends on the version of anaconda, the way it worked prior to AS2.1 was that the %pre sections were all run first, then the whole kickstart file was re-read, ignoring the %pre sections. Red Hat decided to change the way this works, and the approach they took was to remove the old way, release AS2.1, and then start working on adding in the new way. Unfortunately what this means is that this really won't work with 2.1, as the code needed to support it doesn't even exist in anaconda for that version. What I ended up doing was backporting the support for this to the AS2.1 installer. You can find a working copy of kickstart.py on my website at http://www.jasonkohles.com/snippets/Red-Hat-Linux/as21-pre-fix.html, drop it your RHupdates folder, and you should get this functionality back. -- Jason Kohles A witty saying proves nothing. email at jasonkohles.com -- Voltaire (1694 - 1778) http://www.jasonkohles.com/ _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From error27 at gmail.com Thu Apr 21 23:24:49 2005 From: error27 at gmail.com (Dan Carpenter) Date: Thu, 21 Apr 2005 16:24:49 -0700 Subject: sample driver disk for fc3_i386 Message-ID: Does anyone have a sample driver disk for FC3 (32 bit Intel)? I'm trying to create a bcm5700 driver but it says: * module(s) bcm5700 not found I figure that if I could look at someone elses driver disk I could probably tell what I'm doing wrong. My modules.cgz tree looks like this: 2.6.9-1.667/i586/bcm5700.ko 2.6.9-1.667/i686/bcm5700.ko 2.6.9-1.667smp/i586/bcm5700.ko 2.6.9-1.667smp/i686/bcm5700.ko regards, dan carpenter From marhoy at gmail.com Fri Apr 22 07:37:05 2005 From: marhoy at gmail.com (=?ISO-8859-1?Q?Martin_H=F8y?=) Date: Fri, 22 Apr 2005 09:37:05 +0200 Subject: Kickstart %include directive and %pre script issues In-Reply-To: References: Message-ID: <93072505042200377b1bc54c@mail.gmail.com> On 4/21/05, Christian.Rohrmeier at schering.de wrote: > > Hi Jason, > > Thanks for the excellent advice. Rolling my own install CD is a horrifying > thought. I so much abhor this idea, that I install all RPMs that aren't on > the default install media via APT in the %post script. It's actually quite easy (at least on newer RH/FC, don't know about AS2.1). I do this regularily because I want to include all the released updates in the install-tree. That way, I save install-time because I don't have to upgrade a lot of packages afterwards (which on FC3 is quite a few). See the previous postings on including updates (or other .rpms) in the install-tree. Regards, Martin From email at jasonkohles.com Fri Apr 22 17:41:25 2005 From: email at jasonkohles.com (Jason Kohles) Date: Fri, 22 Apr 2005 13:41:25 -0400 Subject: Kickstart %include directive and %pre script issues In-Reply-To: References: Message-ID: <20050422174125.GA20860@mail.jasonkohles.com> On Thu, Apr 21, 2005 at 07:00:39PM +0200, Christian.Rohrmeier at SCHERING.DE wrote: > > Hi Jason, > > Thanks for the excellent advice. Rolling my own install CD is a horrifying > thought. I so much abhor this idea, that I install all RPMs that aren't on > the default install media via APT in the %post script. > > There is a .../RedHat/instimage/usr/lib/anaconda/kickstart.py in the > install media directory on my kickstart server (which I get to via http). I > was hoping I could just plop the new kickstart.py in there, but even though > its there, its not used it seems. The CD needs to be modified. > It isn't as hard as you think, you don't have to replace the existing copy of kickstart,py, if you make a directory on the CD called RHupdates, anaconda will look for updated files there automatically, just drop kickstart.py in there and burn the cd... Check the documentation for more details on the RHupdates directory... -- Jason Kohles A witty saying proves nothing. email at jasonkohles.com -- Voltaire (1694 - 1778) http://www.jasonkohles.com/ From info at hostinthebox.net Tue Apr 26 21:57:07 2005 From: info at hostinthebox.net (dan) Date: Tue, 26 Apr 2005 14:57:07 -0700 Subject: Minimal Install Guidelines Message-ID: <426EB933.1040307@hostinthebox.net> Hello, all - I've been working with a few documents that describe how to go about making a minimalistic install, but they all seem kinda hoakey. I know you guys probably have a lot of experience doing this sort of thing, so I'm hoping that I might be able to get some answers from you all. Is there some sort of template that you guys use to create and install a minimalistic install set, that is small but functional? I'm about to start going through the comps.xml template and picking and choosing what I want/need, but I'd rather not re-invent the wheel if necessary. Thanks! -dant From phr at doc.ic.ac.uk Tue Apr 26 22:14:50 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 26 Apr 2005 23:14:50 +0100 (BST) Subject: Minimal Install Guidelines In-Reply-To: <426EB933.1040307@hostinthebox.net> References: <426EB933.1040307@hostinthebox.net> Message-ID: On Tue, 26 Apr 2005, dan wrote: >Is there some sort of template that you guys use to create and install >a minimalistic install set, that is small but functional? I'm about to >start going through the comps.xml template and picking and choosing >what I want/need, but I'd rather not re-invent the wheel if necessary. Minimal as in lightweight, or minimal as in must-fit-on-a-64MB-flash-card? If the latter, then a stock Fedora install won't cut it. If you merely wish to improve build times and decrease disk usage, however, try specifying just the @Base group and any other individual packages you need. IIRC this gives something in the low-hundreds MB. Cheers, Phil From info at hostinthebox.net Tue Apr 26 22:20:44 2005 From: info at hostinthebox.net (dan) Date: Tue, 26 Apr 2005 15:20:44 -0700 Subject: Minimal Install Guidelines In-Reply-To: References: <426EB933.1040307@hostinthebox.net> Message-ID: <426EBEBC.5020508@hostinthebox.net> Philip Rowlands wrote: > On Tue, 26 Apr 2005, dan wrote: > > >>Is there some sort of template that you guys use to create and install >>a minimalistic install set, that is small but functional? I'm about to >>start going through the comps.xml template and picking and choosing >>what I want/need, but I'd rather not re-invent the wheel if necessary. > > > Minimal as in lightweight, or minimal as in > must-fit-on-a-64MB-flash-card? If the latter, then a stock Fedora > install won't cut it. > > If you merely wish to improve build times and decrease disk usage, > however, try specifying just the @Base group and any other individual > packages you need. IIRC this gives something in the low-hundreds MB. > > > Cheers, > Phil Phil - Like I said, I've accomplished a small install aroudn the neighborhood of 480M, which works, but I'd like to go smaller. The target platform is not an embedded one, I just want to cut down on the whole system footprint, and would like to take a proactive approach to security by installing only the most required of packages. From there, I can add to the system what I see fit, and when i see fit. I don't care about flash cards (but that'd be pretty cool, huh??), I want "minimal as in lightweight". IIRC, the manual states that @Base and @Core groups are installed by default, however, I believe specific packages can be "removed" from those groups, or not installed, by prefixing each package name with a "-" under %packagse. I'm going to try to give this a shot. Regardless, I'll report my findings. Thanks! -dant From phr at doc.ic.ac.uk Tue Apr 26 22:33:50 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 26 Apr 2005 23:33:50 +0100 (BST) Subject: Minimal Install Guidelines In-Reply-To: <426EBEBC.5020508@hostinthebox.net> References: <426EB933.1040307@hostinthebox.net> <426EBEBC.5020508@hostinthebox.net> Message-ID: On Tue, 26 Apr 2005, dan wrote: >IIRC, the manual states that @Base and @Core groups are installed by >default, however, I believe specific packages can be "removed" from >those groups, or not installed, by prefixing each package name with a >"-" under %packagse. I'm going to try to give this a shot. Packages can be excluded, but be aware of the --ignoredeps/--resolvedeps/--ignoremissing flags to %packages, which decide how anaconda will react to missing dependencies. It *is* possible to remove some leaf-node packages from @Base and @Core, but I don't know of a quick'n'easy way to suggest for pruning. (At one point I had a notion to write a GUI tool for this, but I was befuddled by the rpm Python libs. Perhaps someone will chime in to point out the myriad dependency-tracking package management GUIs which already exist.) Cheers, Phil From info at hostinthebox.net Tue Apr 26 22:37:57 2005 From: info at hostinthebox.net (dan) Date: Tue, 26 Apr 2005 15:37:57 -0700 Subject: Minimal Install Guidelines In-Reply-To: References: <426EB933.1040307@hostinthebox.net> <426EBEBC.5020508@hostinthebox.net> Message-ID: <426EC2C5.5040100@hostinthebox.net> Philip Rowlands wrote: > On Tue, 26 Apr 2005, dan wrote: > > >>IIRC, the manual states that @Base and @Core groups are installed by >>default, however, I believe specific packages can be "removed" from >>those groups, or not installed, by prefixing each package name with a >>"-" under %packagse. I'm going to try to give this a shot. > > > Packages can be excluded, but be aware of the > --ignoredeps/--resolvedeps/--ignoremissing flags to %packages, which > decide how anaconda will react to missing dependencies. > > It *is* possible to remove some leaf-node packages from @Base and @Core, > but I don't know of a quick'n'easy way to suggest for pruning. (At one > point I had a notion to write a GUI tool for this, but I was befuddled > by the rpm Python libs. Perhaps someone will chime in to point out the > myriad dependency-tracking package management GUIs which already exist.) > > Yes, it'd be nice to see something like that, a dependency list of some sort, in the same manner in which we have a comps.xml, haha. Well, I appreciate your time. I'm attempting a kickstart with @Base and @Core, with a few packages missing. We'll see how it goes. I'll report back later. Thanks -dant From info at hostinthebox.net Tue Apr 26 23:11:30 2005 From: info at hostinthebox.net (dan) Date: Tue, 26 Apr 2005 16:11:30 -0700 Subject: Minimal Install Guidelines In-Reply-To: <426EC2C5.5040100@hostinthebox.net> References: <426EB933.1040307@hostinthebox.net> <426EBEBC.5020508@hostinthebox.net> <426EC2C5.5040100@hostinthebox.net> Message-ID: <426ECAA2.3070303@hostinthebox.net> dan wrote: > Philip Rowlands wrote: > >> On Tue, 26 Apr 2005, dan wrote: >> >> >>> IIRC, the manual states that @Base and @Core groups are installed by >>> default, however, I believe specific packages can be "removed" from >>> those groups, or not installed, by prefixing each package name with a >>> "-" under %packagse. I'm going to try to give this a shot. >> >> >> >> Packages can be excluded, but be aware of the >> --ignoredeps/--resolvedeps/--ignoremissing flags to %packages, which >> decide how anaconda will react to missing dependencies. >> >> It *is* possible to remove some leaf-node packages from @Base and @Core, >> but I don't know of a quick'n'easy way to suggest for pruning. (At one >> point I had a notion to write a GUI tool for this, but I was befuddled >> by the rpm Python libs. Perhaps someone will chime in to point out the >> myriad dependency-tracking package management GUIs which already exist.) >> >> > Yes, it'd be nice to see something like that, a dependency list of some > sort, in the same manner in which we have a comps.xml, haha. > > Well, I appreciate your time. I'm attempting a kickstart with @Base and > @Core, with a few packages missing. We'll see how it goes. I'll report > back later. > > Thanks > -dant > Well, I got it down to 479M. Yay for me. The one problem I see is glibc-common is over 200M in itself. I've debated in the past about compiling every package, in source form, against a smaller libc, such as uClibc, but thought that this would be a HUGE pain in the ass just to save 200M. However, uClibc's library, in it's fullest form, is around 50M. However cool it would be, I think I'd just be wasting my time. Thanks again for the help -dant From drkludge at cox.net Wed Apr 27 02:40:18 2005 From: drkludge at cox.net (Greg Morgan) Date: Tue, 26 Apr 2005 19:40:18 -0700 Subject: Minimal Install Guidelines In-Reply-To: <426ECAA2.3070303@hostinthebox.net> References: <426EB933.1040307@hostinthebox.net> <426EBEBC.5020508@hostinthebox.net> <426EC2C5.5040100@hostinthebox.net> <426ECAA2.3070303@hostinthebox.net> Message-ID: <426EFB92.8050800@cox.net> dan wrote: >> -dant >> > > Well, I got it down to 479M. Yay for me. The one problem I see is > glibc-common is over 200M in itself. > > I've debated in the past about compiling every package, in source form, > against a smaller libc, such as uClibc, but thought that this would be a > HUGE pain in the ass just to save 200M. However, uClibc's library, in > it's fullest form, is around 50M. > > However cool it would be, I think I'd just be wasting my time. > Dan, There is a practical side in what your are trying to do. The smaller the binaries and install base, then the lower the chance for a potential buffer overflow. I don't know of anyone that has done uClibc on the scale that your are talking about. However, take a look at the leaf project at the links below. The bering-uclibc is a Linux firewall that runs off a 1.44Meg floppy disk. Browse the cvs repository. I believe there is a buildtool that automates part of what the team has done with uClibc. That tool may be of interest to you if you really have the time to shrink Fedora. I think you may have some challenges, however. Just like Anaconda, all leaf firewalls use busybox. SF Project Page http://sourceforge.net/projects/leaf There is a picture at the bottom of the description page. http://leaf-project.org/index.php?module=pagemaster&PAGE_user_op=view_page&PAGE_id=2&MMN_position=16:16 Bering-uclibc CVS. http://cvs.sourceforge.net/viewcvs.py/leaf/src/bering-uclibc/ Greg From pmatilai at welho.com Wed Apr 27 13:46:37 2005 From: pmatilai at welho.com (Panu Matilainen) Date: Wed, 27 Apr 2005 16:46:37 +0300 (EEST) Subject: Minimal Install Guidelines In-Reply-To: References: <426EB933.1040307@hostinthebox.net> <426EBEBC.5020508@hostinthebox.net> Message-ID: On Tue, 26 Apr 2005, Philip Rowlands wrote: > On Tue, 26 Apr 2005, dan wrote: > >> IIRC, the manual states that @Base and @Core groups are installed by >> default, however, I believe specific packages can be "removed" from >> those groups, or not installed, by prefixing each package name with a >> "-" under %packagse. I'm going to try to give this a shot. > > Packages can be excluded, but be aware of the > --ignoredeps/--resolvedeps/--ignoremissing flags to %packages, which > decide how anaconda will react to missing dependencies. > > It *is* possible to remove some leaf-node packages from @Base and @Core, > but I don't know of a quick'n'easy way to suggest for pruning. (At one > point I had a notion to write a GUI tool for this, but I was befuddled > by the rpm Python libs. Perhaps someone will chime in to point out the > myriad dependency-tracking package management GUIs which already exist.) You can look for leafnodes with this script: http://laiskiainen.org/rpm/scripts/rpmorphan.py By default it only looks for leafnodes which look like libraries, for real pruning run with -a option to show all leafnodes. Beware though, there are things like 'grub' which nothing depends on :) - Panu - From bvitiello at egenera.com Wed Apr 27 17:05:10 2005 From: bvitiello at egenera.com (Brian Vitiello) Date: Wed, 27 Apr 2005 13:05:10 -0400 Subject: Kickstart for rescue mode? Message-ID: Does anyone know if it is possible to use a kickstart file in Rescue Mode. If so, do you know how to do this or where I can look to find this out? Basically, I need to boot up in rescue mode and get through all those screens to select rescue image locations, NFS server name, DHCP or static config. I need to be able to boot into rescue mode and just run a few commands but it needs to be entirely automated/without any user interaction. Thanks! -------------- next part -------------- An HTML attachment was scrubbed... URL: From rreed at ops.sgp.arm.gov Thu Apr 28 16:57:52 2005 From: rreed at ops.sgp.arm.gov (Ronald Reed) Date: Thu, 28 Apr 2005 11:57:52 -0500 Subject: Device Parameter in Kickstart File Message-ID: <1114707471.2619.23.camel@ronnote> I have a kickstart file that has 5 device parameters in the file. They look like this: device scsi aacraid device scsi megaraid device scsi aic7xxx device eth e100 device eth tg3 This is done with a kernel boot time parameter of 'expert noprobe', so that the aacraid module will be loaded first. This works, except it appears that it only pays attention to the first device parameter. The megaraid and aic7xxx module are never loaded. Has anyone else used the device parameter and got multiple ones working? -- =========================== Ron Reed RedHat Certified Engineer SGP Computer Department Manager Unix Systems Administrator ARM SGP CART Site (580)388-4053 ron.reed at arm.gov From info at hostinthebox.net Thu Apr 28 17:59:14 2005 From: info at hostinthebox.net (dan) Date: Thu, 28 Apr 2005 10:59:14 -0700 Subject: Device Parameter in Kickstart File In-Reply-To: <1114707471.2619.23.camel@ronnote> References: <1114707471.2619.23.camel@ronnote> Message-ID: <42712472.8080802@hostinthebox.net> Ronald Reed wrote: > I have a kickstart file that has 5 device parameters in the file. They > look like this: > > device scsi aacraid > device scsi megaraid > device scsi aic7xxx > device eth e100 > device eth tg3 > > This is done with a kernel boot time parameter of 'expert noprobe', so > that the aacraid module will be loaded first. This works, except it > appears that it only pays attention to the first device parameter. The > megaraid and aic7xxx module are never loaded. Has anyone else used the > device parameter and got multiple ones working? > Ron - I'm taking a shot in the dark here, but what if you were to try to specify them all on one line, or on one line specified by commas, or some such. I'm not entirely sure. Thanks -dant From rreed at ops.sgp.arm.gov Thu Apr 28 18:09:31 2005 From: rreed at ops.sgp.arm.gov (Ronald Reed) Date: Thu, 28 Apr 2005 13:09:31 -0500 Subject: Device Parameter in Kickstart File In-Reply-To: <42712472.8080802@hostinthebox.net> References: <1114707471.2619.23.camel@ronnote> <42712472.8080802@hostinthebox.net> Message-ID: <1114711771.2619.35.camel@ronnote> I thought of that right after I posted, so I have burned a new CD and I am trying it now. Well, that didn't work. If I change the device lines to read: device scsi aacraid,megaraid,aic7xxx device eth e100,tg3 It doesn't load any modules and tells me that it can't find a hard drive to install on. Anyone know what section of the anaconda code read the device parameter so that I can find it in the source code? On Thu, 2005-04-28 at 12:59, dan wrote: > Ronald Reed wrote: > > I have a kickstart file that has 5 device parameters in the file. They > > look like this: > > > > device scsi aacraid > > device scsi megaraid > > device scsi aic7xxx > > device eth e100 > > device eth tg3 > > > > This is done with a kernel boot time parameter of 'expert noprobe', so > > that the aacraid module will be loaded first. This works, except it > > appears that it only pays attention to the first device parameter. The > > megaraid and aic7xxx module are never loaded. Has anyone else used the > > device parameter and got multiple ones working? > > > > Ron - > > I'm taking a shot in the dark here, but what if you were to try to > specify them all on one line, or on one line specified by commas, or > some such. I'm not entirely sure. > > Thanks > -dant > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -- =========================== Ron Reed RedHat Certified Engineer SGP Computer Department Manager Unix Systems Administrator ARM SGP CART Site (580)388-4053 ron.reed at arm.gov From info at hostinthebox.net Thu Apr 28 18:18:40 2005 From: info at hostinthebox.net (dan) Date: Thu, 28 Apr 2005 11:18:40 -0700 Subject: Interpreting Anaconda, distributions Message-ID: <42712900.70309@hostinthebox.net> Hello, all - While I've been reading the list over the past few weeks, I've noticed that most everyone pretty much knows Anaconda like the back of their hand. Is there some sort of clearinghouse that details the inner-workings of Anaconda? If I wanted to dive into it and try to figure some stuff out, is there a specific version I should look for? I just really don't know where to get started on this, please pardon my ignorance. Is it safe to say that the Kickstart process is the same throught RedHat 9.0+, Fedora, CentOS, RHEL, and all other RedHat-based distributions? It's been my experience that this is true, however, I don't want to start working on something someday that looks like it might work, and have it totally blow up in my face. I hope these questions aren't too dull for you all to answer. I've had varying success with Kickstart-based installs until recently, and I hope that I can even improve on that, with a bit of your help. Thanks! -dant From bedouglas at earthlink.net Thu Apr 28 18:28:50 2005 From: bedouglas at earthlink.net (bruce) Date: Thu, 28 Apr 2005 11:28:50 -0700 Subject: trying to setup for a remote installation Message-ID: <09f401c54c20$2180f0d0$0301a8c0@Mesa.com> hi... new to setting up kickstart for remote installation. here's what i'd like to do. i'd like to be able to remotely access a server, and use VNC to graphically go through the fedora installation process. the server currently has RH8 on it. as we understand it, you can place the FC3 isos on a HD on another system, and access the isos via the kickstart file. however, our questions have to do with how/what to do to implement what we're trying to do. so.. our questions... -can we do what we want to do, allowing us to use VNC -can the kickstart file be created, and placed on the targeted system's HD. -if the kickstart file can be placed on the targeted system's HD, where should it go?? (what dir should it be placed in...) -would the targeted kickstart file have the VNC 'hooks' in the kickstart file -would this allow us to use VNC to then graphically install FC3 on the targeted system any help/ideas/opinions/comments would be helpful... we don't want to somehow screw this up!! -bruce bedouglas at earthlink.net From phr at doc.ic.ac.uk Thu Apr 28 19:53:18 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 28 Apr 2005 20:53:18 +0100 (BST) Subject: trying to setup for a remote installation In-Reply-To: <09f401c54c20$2180f0d0$0301a8c0@Mesa.com> References: <09f401c54c20$2180f0d0$0301a8c0@Mesa.com> Message-ID: On Thu, 28 Apr 2005, bruce wrote: >here's what i'd like to do. i'd like to be able to remotely access a server, >and use VNC to graphically go through the fedora installation process. the >server currently has RH8 on it. > >as we understand it, you can place the FC3 isos on a HD on another system, >and access the isos via the kickstart file. however, our questions have to >do with how/what to do to implement what we're trying to do. "Kickstart" as a general concept provides full automation to a Fedora/RedHat installation. If you wish to proceed through the install GUI via VNC, kickstart ain't necessary. Suggested reading includes the anaconda documentation (/usr/share/doc/anaconda*), particularly the vnc arguments to bootup. (I realiase this isn't a super-helpful answer, but I think all your questions are covered in the docs. Feel free to reply if things are still unclear.) Cheers, Phil From phr at doc.ic.ac.uk Thu Apr 28 20:23:41 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 28 Apr 2005 21:23:41 +0100 (BST) Subject: trying to setup for a remote installation In-Reply-To: <0a1f01c54c2e$149c95a0$0301a8c0@Mesa.com> References: <0a1f01c54c2e$149c95a0$0301a8c0@Mesa.com> Message-ID: Please do not reply off-list; postings archived for future reference help reduce repeated questions. On Thu, 28 Apr 2005, bruce wrote: >still foggy! > >in looking over sites from google... ii'm not sure how you can simply modify >the grub.conf (i'm assuming that's what you meant by bootup) to do the >remote upgrade via VNC. /usr/share/doc/anaconda-10.*/command-line.txt says: vnc Enable vnc-based installation. You will need to connect to the machine using a vnc client application. anaconda arguments go on the kernel command line; try GRUB's command-line editing while testing. >if i place the fedora isos on a httpd server, then what? /usr/share/doc/anaconda-10.*/install-methods.txt says: - FTP/HTTP (from a directory of loopback-mounted ISOs) ------------------------------------------------------ Summary: Pulls files from tree via FTP. Looks in 'disc1/' directory to contain files from CD #1, 'disc2/' for CD #2, etc. These can be created on the server by loopback mounting the ISO images into these directories under the directory made available to ftp. so it's slightly more complicated that serving the plain ISOs, as they have to be loopback-mounted. (NFS-exported ISOs don't, if that's easier.) Cheers, Phil From bedouglas at earthlink.net Thu Apr 28 20:32:01 2005 From: bedouglas at earthlink.net (bruce) Date: Thu, 28 Apr 2005 13:32:01 -0700 Subject: trying to setup for a remote installation In-Reply-To: Message-ID: <0a3a01c54c31$560eb4c0$0301a8c0@Mesa.com> still confudnfusing... but thanks... -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Philip Rowlands Sent: Thursday, April 28, 2005 1:24 PM To: Discussion list about Kickstart Subject: RE: trying to setup for a remote installation Please do not reply off-list; postings archived for future reference help reduce repeated questions. On Thu, 28 Apr 2005, bruce wrote: >still foggy! > >in looking over sites from google... ii'm not sure how you can simply modify >the grub.conf (i'm assuming that's what you meant by bootup) to do the >remote upgrade via VNC. /usr/share/doc/anaconda-10.*/command-line.txt says: vnc Enable vnc-based installation. You will need to connect to the machine using a vnc client application. anaconda arguments go on the kernel command line; try GRUB's command-line editing while testing. >if i place the fedora isos on a httpd server, then what? /usr/share/doc/anaconda-10.*/install-methods.txt says: - FTP/HTTP (from a directory of loopback-mounted ISOs) ------------------------------------------------------ Summary: Pulls files from tree via FTP. Looks in 'disc1/' directory to contain files from CD #1, 'disc2/' for CD #2, etc. These can be created on the server by loopback mounting the ISO images into these directories under the directory made available to ftp. so it's slightly more complicated that serving the plain ISOs, as they have to be loopback-mounted. (NFS-exported ISOs don't, if that's easier.) Cheers, Phil _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From bedouglas at earthlink.net Thu Apr 28 21:32:25 2005 From: bedouglas at earthlink.net (bruce) Date: Thu, 28 Apr 2005 14:32:25 -0700 Subject: trying to setup for a remote installation Message-ID: <0a5101c54c39$c67d1460$0301a8c0@Mesa.com> ok... in order to do a remote FC3 installation, using vnc, it appears that i need to create a floppy bootdisk, copy a kickstart file to it, place the FC3 ISOs on a HTTP server, and add the required VNC commands to the bootloader of the kickstart file. is this anywhere close to being correct... 1) create the floppy bootdisk (use rawwrite, and copy the boot.img file to the a: floppy) 2) edit the ks.cfg (kickstart file) with the correct parameters 3) copy this file to the top/root of the floppy drive.. 4) copy the FC3 ISO files to the correct location of the http server (http:/foo.com/FC3/) the FC3 dir contains the FC3 ISO files... i assume i should allow the files/index to be displayed from the browser... 5) within the ks.cfg file, i have a line for the 'bootloader' bootloader --location=mbr --append="fff" <<<<<<<<<<< (should i add the 'vnc' information here, to initiate the vnc connection on the client server... remember, i want to use vnc on my server to graphically install FC3 on the remote server... 6) or, should i modify the existing/targeted server's grub file, and add the VNC attributes to the end of the 'kernel /vmlinuz... ' vnc vncconnect=192.168.1.4:5500 am i getting close to understanding what has to be done.... also, if i wanted to simply do all this remotely, and have the kickstart (ks.cfg) file on the targeted harddrive, where would i copy it, and how would i know what to use as the device 'linux=hd::file'? would i get the device by doing a 'df' to see what the device names are, using the actual device name, and not the mounted name.... -bruce bedouglas at earthlink.net -----Original Message----- From: bruce [mailto:bedouglas at earthlink.net] Sent: Thursday, April 28, 2005 11:29 AM To: 'Discussion list about Kickstart'; 'fedora-list at redhat.com' Subject: trying to setup for a remote installation hi... new to setting up kickstart for remote installation. here's what i'd like to do. i'd like to be able to remotely access a server, and use VNC to graphically go through the fedora installation process. the server currently has RH8 on it. as we understand it, you can place the FC3 isos on a HD on another system, and access the isos via the kickstart file. however, our questions have to do with how/what to do to implement what we're trying to do. so.. our questions... -can we do what we want to do, allowing us to use VNC -can the kickstart file be created, and placed on the targeted system's HD. -if the kickstart file can be placed on the targeted system's HD, where should it go?? (what dir should it be placed in...) -would the targeted kickstart file have the VNC 'hooks' in the kickstart file -would this allow us to use VNC to then graphically install FC3 on the targeted system any help/ideas/opinions/comments would be helpful... we don't want to somehow screw this up!! -bruce bedouglas at earthlink.net From bedouglas at earthlink.net Thu Apr 28 23:27:52 2005 From: bedouglas at earthlink.net (bruce) Date: Thu, 28 Apr 2005 16:27:52 -0700 Subject: problems creating/saving a kickstart file ... Message-ID: <0a9901c54c49$e8080350$0301a8c0@Mesa.com> hi... when i use VNS to remotely access a server, i appear to have a problem with Kickstart Configurator. Basically, I can go through the process, to specify the kickstart options, but when i try to save the file, nothing happens. The system effectively sits there with no action. I've also tried to open the original kickstart.cfg file in the '/rott' dir, and again, nothing happens. The Kickstart app displays the open file dialog, but does nothing when i select the kickstart.cfg file... has anyone else seen/experienced this kind of problem before... searching google didn't produce any solution, and i don't yet want to file a RH/FC bug on it if there's a solution... thanks bruce bedouglas at earthlink.net From phr at doc.ic.ac.uk Fri Apr 29 00:27:10 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Fri, 29 Apr 2005 01:27:10 +0100 (BST) Subject: trying to setup for a remote installation In-Reply-To: <0a5101c54c39$c67d1460$0301a8c0@Mesa.com> References: <0a5101c54c39$c67d1460$0301a8c0@Mesa.com> Message-ID: On Thu, 28 Apr 2005, bruce wrote: >in order to do a remote FC3 installation, using vnc, it appears that i need >to create a floppy bootdisk, copy a kickstart file to it, place the FC3 ISOs >on a HTTP server, and add the required VNC commands to the bootloader of the >kickstart file. Hi Bruce, I think I may have clouded rather than clarified the kickstart process: - Do you want to automate the installation (i.e. launch it and walk away), or go through the prompts (disk layout, package selection) manually? Kickstart files are necessary only in the first case; kickstart is not needed simply to perform a remote install. - How will the machine be booted up? Floppy images are severely lacking when it comes to network installation; you'd have to mess around with driver disks, which isn't much fun. CD would be simplest; PXE (i.e. boot entirely from the network) is another option, but requires DHCP, TFTP, HTTP/NFS servers. - What's the purpose of the VNC invocation? Is the machine in an inaccessible rack/office somewhere? Does the bootup to install have to work entirely "hands-off", or is there someone to insert a floppy/CD? To put a kickstart file on the harddisk, this is the syntax (from kickstart-docs.txt): ks=hd::/ e.g. ks=hd:sda3:/mydir/ks.cfg Cheers, Phil From erik.niessen at gmail.com Fri Apr 29 13:03:55 2005 From: erik.niessen at gmail.com (Erik Niessen) Date: Fri, 29 Apr 2005 15:03:55 +0200 Subject: kickstart and ide-scsi Message-ID: Hi all, I am trying to use kickstart from a ide cdrom with scsi emulation enabled. Here is a piece of my isolinux.cfg file default kickstart label kickstart kernel vmlinuz append initrd=initrd.img ramdisk_size=8192 hdc=scsi ks=cdrom:/ks.cfg This doesn't work it cannot find my ks file. When I delete the hdc=scsi it works Any ideas? Cheers, Erik Niessen From erik.niessen at gmail.com Fri Apr 29 13:35:24 2005 From: erik.niessen at gmail.com (Erik Niessen) Date: Fri, 29 Apr 2005 15:35:24 +0200 Subject: kickstart and ide-scsi In-Reply-To: <4272361A.7020600@uk.colt.net> References: <4272361A.7020600@uk.colt.net> Message-ID: I tried this but this also doesn't work I had a look at the messages reported on f3 and it tries to mount the cdrom under hdc. Failed to mount /dev/hdc: No such device or address On 4/29/05, Iain Conochie wrote: > Erik Niessen wrote: > > Hi all, > > > > I am trying to use kickstart from a ide cdrom with scsi emulation enabled. > > Here is a piece of my isolinux.cfg file > > default kickstart > > label kickstart > > kernel vmlinuz > > append initrd=initrd.img ramdisk_size=8192 hdc=scsi ks=cdrom:/ks.cfg > > > > This doesn't work it cannot find my ks file. When I delete the hdc=scsi it works > > > > Any ideas? > > Try: > > label kickstart > > kernel vmlinuz > > append initrd=initrd.img ramdisk_size=8192 hdc=ide-scsi \ > ks=cdrom:/ks.cfg > > ide-scsi is the scsi emulation module > > Cheers > > Iain > > > > > Cheers, > > > > Erik Niessen > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > From iain at uk.colt.net Fri Apr 29 13:26:50 2005 From: iain at uk.colt.net (Iain Conochie) Date: Fri, 29 Apr 2005 14:26:50 +0100 Subject: kickstart and ide-scsi In-Reply-To: References: Message-ID: <4272361A.7020600@uk.colt.net> Erik Niessen wrote: > Hi all, > > I am trying to use kickstart from a ide cdrom with scsi emulation enabled. > Here is a piece of my isolinux.cfg file > default kickstart > label kickstart > kernel vmlinuz > append initrd=initrd.img ramdisk_size=8192 hdc=scsi ks=cdrom:/ks.cfg > > This doesn't work it cannot find my ks file. When I delete the hdc=scsi it works > > Any ideas? Try: label kickstart > kernel vmlinuz > append initrd=initrd.img ramdisk_size=8192 hdc=ide-scsi \ ks=cdrom:/ks.cfg ide-scsi is the scsi emulation module Cheers Iain > > Cheers, > > Erik Niessen > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From drkludge at cox.net Fri Apr 29 14:03:36 2005 From: drkludge at cox.net (Greg Morgan) Date: Fri, 29 Apr 2005 07:03:36 -0700 Subject: Interpreting Anaconda, distributions In-Reply-To: <42712900.70309@hostinthebox.net> References: <42712900.70309@hostinthebox.net> Message-ID: <42723EB8.1010804@cox.net> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 dan wrote: > Hello, all - > > While I've been reading the list over the past few weeks, I've noticed > that most everyone pretty much knows Anaconda like the back of their hand. > > Is there some sort of clearinghouse that details the inner-workings of > Anaconda? If I wanted to dive into it and try to figure some stuff out, > is there a specific version I should look for? I just really don't know > where to get started on this, please pardon my ignorance. Clearinghouse http://rau.homedns.org/twiki/bin/view/Anaconda/AnacondaDocumentationProject Conversion of Clearinghouse to fedora wiki http://fedoraproject.org/wiki/Anaconda http://fedoraproject.org/wiki/GregMorgan = conversion status page I see I made one mistake on the organization of the pages. I was thinking media wiki when moin wiki uses pagename/pagename organization. I _hope_ to fix those problems this weekend. Honey dos from wife may rule the day. ;-) Let's not forget the official sites http://fedora.redhat.com/projects/anaconda-installer/ http://www.redhat.com/docs/manuals/linux/RHL-9-Manual/install-guide/ http://www.redhat.com/docs/manuals/linux/RHL-9-Manual/custom-guide/part-install-info.html The official sites focus more on kickstart info. The rau wiki shows other parts of anaconda that are used to build the release CDs. LOL you can put your very own _spin_ on your Red Hat/Fedora CDs. > > Is it safe to say that the Kickstart process is the same throught RedHat > 9.0+, Fedora, CentOS, RHEL, and all other RedHat-based distributions? Yes. Different versions same concepts. There's an anaconda rpm that can be installed or you can get the most recent version from CVS as described on this page http://fedora.redhat.com/projects/anaconda-installer/. Let's not forget to add that Ubuntu linux is considering adding the snake to humanity http://www.ubuntulinux.org/wiki/GraphicalInstaller. Progeny is _rattling_ it's saber err... tail at Ubuntu because they already have the snake installer http://lxer.com/module/newswire/view/26797/. LOL I am just full of bad puns today. > It's been my experience that this is true, however, I don't want to > start working on something someday that looks like it might work, and > have it totally blow up in my face. I borrowed the format from the Stateless linux page. I jotted down some ideas of what I see people doing on the list http://fedoraproject.org/wiki/AnacondaVision. The same thing works for each version/product. It's getting better with time. There are many ways Anaconda can be extended. Note that any good program has that same potential of blowing up in your face under the right circumstances. I believe you need some more information to make you comfortable when it does happen. > > I hope these questions aren't too dull for you all to answer. Nope. > I've had > varying success with Kickstart-based installs until recently, and I hope > that I can even improve on that, with a bit of your help. I am fascinated by Anaconda. I think what it is missing some more howto documentation. The key is to know what you want on your target system when using ks files. If you are going to start a service, then you'll want to use chkconfig or know how to make the link files by hand in the /etc/rc.d/ System V styled directories. Typical of computer documentation some of Anaconda/KS documentation is just a listing of options with a few brief code examples. What is needed in my opinion is a good sample ks howto using a well developed target installation guide. These examples would a help a new person see how to glue all the options together. One, idea is a remaster and kickstart howto but show how to add files from say, mythtv, to the mix. There are several well defined procedures here http://www.mythtv.org/docs/mythtv-HOWTO.html and fedora specific notes here http://wilsonet.com/mythtv/fcmyth.php. Another example may be a straight ks howto without the remaster. For example, I am also working on this wiki http://phpwebsite-comm.sourceforge.net/wiki/index.php?title=Generic_Install_Introduction. I have seen all the problems in the user forums. So I took my forum posts along with other's and created the guide. Most of the forum problems could be solved with a ks file that sets php.ini parms, and starts both an apache service plus a database. These kinds of examples that are talked and walked through would be great for someone like you that's just trying to get your arms around kickstart files and other features that Anaconda has. The last tip I'd give you is remember that Linux is running. You have a bash shell environment with a few limitations because it is running all in memory. Your ks file is just a bash script that you could develop in a terminal window before running it as a kickstart file. The only difference is that it runs in an installer and is defined in a %post section of the kick start file. That's it. Greg -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFCcj64xyxe5L6mr7IRApVXAJ0VpH151S32BqJHNRUiFMQsSrRBawCgiDZ1 yCGTmdfFoJJnrVSxyx+OIIk= =keJX -----END PGP SIGNATURE----- From drkludge at cox.net Fri Apr 29 14:38:34 2005 From: drkludge at cox.net (Greg Morgan) Date: Fri, 29 Apr 2005 07:38:34 -0700 Subject: kickstart and ide-scsi In-Reply-To: References: <4272361A.7020600@uk.colt.net> Message-ID: <427246EA.5020406@cox.net> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Erik Niessen wrote: > I tried this but this also doesn't work > > I had a look at the messages reported on f3 and it tries to mount the cdrom > under hdc. Failed to mount /dev/hdc: No such device or address Do you really need scsi emulation anymore? That could be the source of your error. That may be the answer to this statement of yours, "When I delete the hdc=scsi it works." The other reason for the error is that your CD drive may not be on /dev/hdc in the Anaconda runtime environment. Remember that it is a minimal Linux environment. You may have other temporary devices or special devices without drivers throwing off your device count. Start the installer without a kickstart file. Use the ALT F1, F2, F3, and F4 keys to take a look around. Then tune your install script for the right /dev/hdx. Greg > > On 4/29/05, Iain Conochie wrote: > >>Erik Niessen wrote: >> >>>Hi all, >>> >>>I am trying to use kickstart from a ide cdrom with scsi emulation enabled. >>>Here is a piece of my isolinux.cfg file >>>default kickstart >>>label kickstart >>> kernel vmlinuz >>> append initrd=initrd.img ramdisk_size=8192 hdc=scsi ks=cdrom:/ks.cfg >>> >>>This doesn't work it cannot find my ks file. When I delete the hdc=scsi it works >>> >>>Any ideas? >> >>Try: >> >>label kickstart >> >>> kernel vmlinuz >>> append initrd=initrd.img ramdisk_size=8192 hdc=ide-scsi \ >> >>ks=cdrom:/ks.cfg >> >>ide-scsi is the scsi emulation module >> >>Cheers >> >>Iain >> >> >>>Cheers, >>> >>> Erik Niessen -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFCckbqxyxe5L6mr7IRApg4AJ9HHTgSeoocdm9a0FCRuLjJVrkYigCePxmk ZrPtSYpMJ04C0VU3QpoLnfw= =U/za -----END PGP SIGNATURE----- From bedouglas at earthlink.net Fri Apr 29 16:04:47 2005 From: bedouglas at earthlink.net (bruce) Date: Fri, 29 Apr 2005 09:04:47 -0700 Subject: trying to setup for a remote installation In-Reply-To: Message-ID: <0bb701c54cd5$2b9817a0$0301a8c0@Mesa.com> phill.... thanks for the followup/being patient... initially, i'd like to be able to walk through the entire process from a remote system. thus the need for/use of vnc. i was hoping that i could somehow reboot the remote server, and have vnc pop up on my server/vnc window, allowing me to select the various attributes of the remote Fedora install on the remote machine... this would allow me to be able to do a manual remote install of Fedora, using VNc.. the reason i wanted to do this, is to allow me to check out the drive configuration, and to determine what apps i wanted to select at the beginning... if this isn't possible, then i guess the next step would be to create a kickstart file, and somehow do the install that way. the machine wil be booting off a harddrive. once i get the machine up, i could then always use yum to keep it updated... thanks -bruce -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Philip Rowlands Sent: Thursday, April 28, 2005 5:27 PM To: Discussion list about Kickstart Subject: RE: trying to setup for a remote installation On Thu, 28 Apr 2005, bruce wrote: >in order to do a remote FC3 installation, using vnc, it appears that i need >to create a floppy bootdisk, copy a kickstart file to it, place the FC3 ISOs >on a HTTP server, and add the required VNC commands to the bootloader of the >kickstart file. Hi Bruce, I think I may have clouded rather than clarified the kickstart process: - Do you want to automate the installation (i.e. launch it and walk away), or go through the prompts (disk layout, package selection) manually? Kickstart files are necessary only in the first case; kickstart is not needed simply to perform a remote install. - How will the machine be booted up? Floppy images are severely lacking when it comes to network installation; you'd have to mess around with driver disks, which isn't much fun. CD would be simplest; PXE (i.e. boot entirely from the network) is another option, but requires DHCP, TFTP, HTTP/NFS servers. - What's the purpose of the VNC invocation? Is the machine in an inaccessible rack/office somewhere? Does the bootup to install have to work entirely "hands-off", or is there someone to insert a floppy/CD? To put a kickstart file on the harddisk, this is the syntax (from kickstart-docs.txt): ks=hd::/ e.g. ks=hd:sda3:/mydir/ks.cfg Cheers, Phil _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From info at hostinthebox.net Fri Apr 29 16:14:18 2005 From: info at hostinthebox.net (dan) Date: Fri, 29 Apr 2005 09:14:18 -0700 Subject: trying to setup for a remote installation In-Reply-To: <0bb701c54cd5$2b9817a0$0301a8c0@Mesa.com> References: <0bb701c54cd5$2b9817a0$0301a8c0@Mesa.com> Message-ID: <42725D5A.8000902@hostinthebox.net> bruce wrote: > phill.... > > thanks for the followup/being patient... initially, i'd like to be able to > walk through the entire process from a remote system. thus the need for/use > of vnc. > > i was hoping that i could somehow reboot the remote server, and have vnc pop > up on my server/vnc window, allowing me to select the various attributes of > the remote Fedora install on the remote machine... > > this would allow me to be able to do a manual remote install of Fedora, > using VNc.. > > the reason i wanted to do this, is to allow me to check out the drive > configuration, and to determine what apps i wanted to select at the > beginning... > > if this isn't possible, then i guess the next step would be to create a > kickstart file, and somehow do the install that way. > > the machine wil be booting off a harddrive. once i get the machine up, i > could then always use yum to keep it updated... > > > thanks > > -bruce > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Philip Rowlands > Sent: Thursday, April 28, 2005 5:27 PM > To: Discussion list about Kickstart > Subject: RE: trying to setup for a remote installation > > > On Thu, 28 Apr 2005, bruce wrote: > > >>in order to do a remote FC3 installation, using vnc, it appears that i need >>to create a floppy bootdisk, copy a kickstart file to it, place the FC3 > > ISOs > >>on a HTTP server, and add the required VNC commands to the bootloader of > > the > >>kickstart file. > > > Hi Bruce, > > I think I may have clouded rather than clarified the kickstart process: > > - Do you want to automate the installation (i.e. launch it and > walk away), or go through the prompts (disk layout, package > selection) manually? > > Kickstart files are necessary only in the first case; kickstart is not > needed simply to perform a remote install. > > - How will the machine be booted up? > > Floppy images are severely lacking when it comes to network > installation; you'd have to mess around with driver disks, which isn't > much fun. CD would be simplest; PXE (i.e. boot entirely from the > network) is another option, but requires DHCP, TFTP, HTTP/NFS servers. > > - What's the purpose of the VNC invocation? > > Is the machine in an inaccessible rack/office somewhere? Does the bootup > to install have to work entirely "hands-off", or is there someone to > insert a floppy/CD? > > To put a kickstart file on the harddisk, this is the syntax (from > kickstart-docs.txt): > ks=hd::/ > e.g. ks=hd:sda3:/mydir/ks.cfg > > > Cheers, > Phil Bruce - Are you able to get your hands on another machine that you can test with, or even VMWare, from which you can load a virtual machine for testing? This is a step I always take. In fact, I've been doing kickstarts with varying success for ~2 years, and I've still not been able to muster the manhood to attempt a total remote installation such as the one that you're asking. I'd highly suggest testing the remote install method on a machine which you can just fool around with, before going about the real thing. Thanks -dant From info at hostinthebox.net Fri Apr 29 17:25:19 2005 From: info at hostinthebox.net (dan) Date: Fri, 29 Apr 2005 10:25:19 -0700 Subject: Setting root password w/ rootpw Message-ID: <42726DFF.1050303@hostinthebox.net> Hello, all - I was wondering what method you guys use for setting a root password during a kickstart install. I understand that it can be set by the 'rootpw' directive, and optionally, I can throw it the --isencrypted flag. I'd like to assign an encrypted password, but I was curious as to which method you guys use in an effort to simply convert a text password into the encrypted type asked for. I've tried copying sniplets from /etc/passwd from user accounts for which I set the password, and I knew what the password was, stuff like that - to no avail. So... I thought I'd ask here. Thanks! -dant From steve.mah at oracle.com Fri Apr 29 17:42:42 2005 From: steve.mah at oracle.com (Stephen Mah) Date: Fri, 29 Apr 2005 10:42:42 -0700 Subject: Setting root password w/ rootpw In-Reply-To: <42726DFF.1050303@hostinthebox.net> References: <42726DFF.1050303@hostinthebox.net> Message-ID: <42727212.3030109@oracle.com> It will take clear text or encrypted passwords. use the grub-md5-crypt command to convert the password. eg: "welcome" would be rootpw --iscrypted $1$6AVJt0$jzcyYkTBUq3kYItxttWlm0 dan wrote: > Hello, all - > > I was wondering what method you guys use for setting a root password > during a kickstart install. I understand that it can be set by the > 'rootpw' directive, and optionally, I can throw it the --isencrypted > flag. I'd like to assign an encrypted password, but I was curious as > to which method you guys use in an effort to simply convert a text > password into the encrypted type asked for. > > I've tried copying sniplets from /etc/passwd from user accounts for > which I set the password, and I knew what the password was, stuff like > that - to no avail. So... I thought I'd ask here. > > Thanks! > -dant > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From tibbs at math.uh.edu Fri Apr 29 17:50:07 2005 From: tibbs at math.uh.edu (Jason L Tibbitts III) Date: Fri, 29 Apr 2005 12:50:07 -0500 Subject: Setting root password w/ rootpw In-Reply-To: <42726DFF.1050303@hostinthebox.net> (dan's message of "Fri, 29 Apr 2005 10:25:19 -0700") References: <42726DFF.1050303@hostinthebox.net> Message-ID: >>>>> "d" == dan writes: d> I'd like to assign an encrypted password, d> but I was curious as to which method you guys use in an effort to d> simply convert a text password into the encrypted type asked for. perl -we 'print crypt("plaintext", "salt"). "\n"' will do. Replace "plaintext" by the plaintext password you want to encrypt. How you replace "salt" depends on what kind of passwords you want. If you want the cheezy old-style crypted passwords, use two randomly generated characters. If you want new-style MD5 passwords, use "$1$" followed by up to eight random characters. (The random characters should be letters, digits, "." or "/".) - J< From phr at doc.ic.ac.uk Fri Apr 29 17:52:35 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Fri, 29 Apr 2005 18:52:35 +0100 (BST) Subject: Setting root password w/ rootpw In-Reply-To: <42726DFF.1050303@hostinthebox.net> References: <42726DFF.1050303@hostinthebox.net> Message-ID: On Fri, 29 Apr 2005, dan wrote: >I've tried copying sniplets from /etc/passwd from user accounts for >which I set the password, and I knew what the password was, stuff like >that - to no avail. So... I thought I'd ask here. Ensure that you have USEMD5=yes in /etc/sysconfig/authconfig; this is set by kickstart's auth --enablemd5 line. MD5-based passwords look like this: $1$h6SOogDr$jZPK48gOLAgayPYLN4gSr0 whereas classic "crypt" passwords like this: 9tJNsoIV/pjgQ Both can be generated by openssl's passwd applet; see "openssl passwd --help" for more info. Cheers, Phil From info at hostinthebox.net Fri Apr 29 17:55:35 2005 From: info at hostinthebox.net (dan) Date: Fri, 29 Apr 2005 10:55:35 -0700 Subject: Setting root password w/ rootpw In-Reply-To: References: <42726DFF.1050303@hostinthebox.net> Message-ID: <42727517.6030305@hostinthebox.net> Philip Rowlands wrote: > On Fri, 29 Apr 2005, dan wrote: > > >>I've tried copying sniplets from /etc/passwd from user accounts for >>which I set the password, and I knew what the password was, stuff like >>that - to no avail. So... I thought I'd ask here. > > > Ensure that you have USEMD5=yes in /etc/sysconfig/authconfig; this is > set by kickstart's auth --enablemd5 line. > > MD5-based passwords look like this: > $1$h6SOogDr$jZPK48gOLAgayPYLN4gSr0 > > whereas classic "crypt" passwords like this: > 9tJNsoIV/pjgQ > > Both can be generated by openssl's passwd applet; > see "openssl passwd --help" for more info. > > > Cheers, > Phil > All - I know the difference between the passwords, but I think that Stephen Mah got me on the right track. He suggested using the grub-md5-crypt command, which works perfectly. Exactly what I was looking for. Thanks -dant From error27 at gmail.com Sat Apr 30 01:58:04 2005 From: error27 at gmail.com (Dan Carpenter) Date: Fri, 29 Apr 2005 21:58:04 -0400 Subject: Device Parameter in Kickstart File In-Reply-To: <1114711771.2619.35.camel@ronnote> References: <1114707471.2619.23.camel@ronnote> <42712472.8080802@hostinthebox.net> <1114711771.2619.35.camel@ronnote> Message-ID: On 4/28/05, Ronald Reed wrote: > Anyone know what section of the anaconda code read the device parameter > so that I can find it in the source code? I'm new to this as well. Check out anaconda CVS. cd anaconda/loader2/ look for loadKickstartModule in kickstart.c and modules.c I don't completely understand you're setup, but maybe you could try getting access to the install media and doing all the rest in the %pre and %post scripts. I've never had to do this myself... regards, dan