From tamas.1.simon at nokia.com Wed Feb 7 11:08:34 2007 From: tamas.1.simon at nokia.com (tamas.1.simon at nokia.com) Date: Wed, 7 Feb 2007 12:08:34 +0100 Subject: Install order In-Reply-To: <45BF53CA.4000805@rampaginggeek.com> References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be><20070129092012.GJ1426@thomson.net><45BE8B63.3050905@rampaginggeek.com> <45BF53CA.4000805@rampaginggeek.com> Message-ID: <4B5BDDC9AE76CA47AF37CE848F0DE46C033C69AB@buebe101.NOE.Nokia.com> Hi, I have a custom cd with 249 rpm packages. In the ks.cfg file the %packages section look like this (alphabet order): #Package install information %packages --resolvedeps @ everything anacron apr apr-util audit audit-libs authconfig basesystem .... But it starts to install the packages with the 'comps' package. How does it decide what package to start the install with? Thanks Tamas From email at jasonkohles.com Wed Feb 7 14:19:37 2007 From: email at jasonkohles.com (Jason Kohles) Date: Wed, 7 Feb 2007 09:19:37 -0500 Subject: Install order In-Reply-To: <4B5BDDC9AE76CA47AF37CE848F0DE46C033C69AB@buebe101.NOE.Nokia.com> References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be><20070129092012.GJ1426@thomson.net><45BE8B63.3050905@rampaginggeek.com> <45BF53CA.4000805@rampaginggeek.com> <4B5BDDC9AE76CA47AF37CE848F0DE46C033C69AB@buebe101.NOE.Nokia.com> Message-ID: <01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com> On Feb 7, 2007, at 6:08 AM, wrote: > Hi, > > I have a custom cd with 249 rpm packages. > In the ks.cfg file the %packages section look like this (alphabet > order): > > #Package install information > %packages --resolvedeps > @ everything > anacron > apr > apr-util > audit > audit-libs > authconfig > basesystem > .... > > But it starts to install the packages with the 'comps' package. > > How does it decide what package to start the install with? > Based on the dependencies of each package. It installs comps first because comps contains dependency information for the rest of the packages, then it determines what order to install based on that dependency information, so if package 'bar' requires package 'foo' be installed, then foo will be installed before bar, regardless of what order they are in in your ks.cfg. -- Jason Kohles email at jasonkohles.com http://www.jasonkohles.com/ "A witty saying proves nothing." -- Voltaire From tamas.1.simon at nokia.com Wed Feb 7 14:41:37 2007 From: tamas.1.simon at nokia.com (tamas.1.simon at nokia.com) Date: Wed, 7 Feb 2007 15:41:37 +0100 Subject: Install order In-Reply-To: <01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com> References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be><20070129092012.GJ1426@thomson.net><45BE8B63.3050905@rampaginggeek.com><45BF53CA.4000805@rampaginggeek.com><4B5BDDC9AE76CA47AF37CE848F0DE46C033C69AB@buebe101.NOE.Nokia.com> <01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com> Message-ID: <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA1AA@buebe101.NOE.Nokia.com> Thanks, I just have some packages which actually doesn't have dependency but in the pre and post install script of rpm they use eg. grep to check something, and at this point grep is not installed, so the package will not be installed with kickstart. Is there any option in kickstart to tell a package 'when' it should be installed? Tamas -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of ext Jason Kohles Sent: Wednesday, February 07, 2007 3:20 PM To: Discussion list about Kickstart Subject: Re: Install order On Feb 7, 2007, at 6:08 AM, wrote: > Hi, > > I have a custom cd with 249 rpm packages. > In the ks.cfg file the %packages section look like this (alphabet > order): > > #Package install information > %packages --resolvedeps > @ everything > anacron > apr > apr-util > audit > audit-libs > authconfig > basesystem > .... > > But it starts to install the packages with the 'comps' package. > > How does it decide what package to start the install with? > Based on the dependencies of each package. It installs comps first because comps contains dependency information for the rest of the packages, then it determines what order to install based on that dependency information, so if package 'bar' requires package 'foo' be installed, then foo will be installed before bar, regardless of what order they are in in your ks.cfg. -- Jason Kohles email at jasonkohles.com http://www.jasonkohles.com/ "A witty saying proves nothing." -- Voltaire _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From jkeating at redhat.com Wed Feb 7 14:46:04 2007 From: jkeating at redhat.com (Jesse Keating) Date: Wed, 7 Feb 2007 09:46:04 -0500 Subject: Install order In-Reply-To: <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA1AA@buebe101.NOE.Nokia.com> References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be> <01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com> <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA1AA@buebe101.NOE.Nokia.com> Message-ID: <200702070946.05044.jkeating@redhat.com> On Wednesday 07 February 2007 09:41, tamas.1.simon at nokia.com wrote: > I just have some packages which actually doesn't have dependency but in > the pre and post install script of rpm they use eg. grep to check > something, and at this point grep is not installed, so the package will > not be installed with kickstart. > > Is there any option in kickstart to tell a package 'when' it should be > installed? Use Requires(pre) or Requires(post) in your package if it uses grep. This will ensure ordering so that grep is installed before your package. -- Jesse Keating Release Engineer: Fedora -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: From tamas.1.simon at nokia.com Wed Feb 7 14:57:01 2007 From: tamas.1.simon at nokia.com (tamas.1.simon at nokia.com) Date: Wed, 7 Feb 2007 15:57:01 +0100 Subject: Install order In-Reply-To: <200702070946.05044.jkeating@redhat.com> References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be><01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com><4B5BDDC9AE76CA47AF37CE848F0DE46C033EA1AA@buebe101.NOE.Nokia.com> <200702070946.05044.jkeating@redhat.com> Message-ID: <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA201@buebe101.NOE.Nokia.com> Ok, thanks, but the packege is not mine but made by hp, so I don't want to modify it. Tamas -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of ext Jesse Keating Sent: Wednesday, February 07, 2007 3:46 PM To: kickstart-list at redhat.com Subject: Re: Install order On Wednesday 07 February 2007 09:41, tamas.1.simon at nokia.com wrote: > I just have some packages which actually doesn't have dependency but > in the pre and post install script of rpm they use eg. grep to check > something, and at this point grep is not installed, so the package > will not be installed with kickstart. > > Is there any option in kickstart to tell a package 'when' it should be > installed? Use Requires(pre) or Requires(post) in your package if it uses grep. This will ensure ordering so that grep is installed before your package. -- Jesse Keating Release Engineer: Fedora From drindt at mx-solutions.de Wed Feb 7 15:15:25 2007 From: drindt at mx-solutions.de (Daniel Rindt) Date: Wed, 07 Feb 2007 16:15:25 +0100 Subject: setting specific route by kickstart Message-ID: <1170861325.9486.26.camel@heron.mx-solutions.de> Hello, my Hoster normally uses DHCP to get servers the IP Adress. But in my case the dhcp seems not to work. I get from the provider the information that i must initiate a hostroute for eth0 over 10.255.255.1. The gateway has the same IP adress. Iam no network specialist, hope anyone can help me. How i can set the route by kickstarting? Thanks in advance Daniel Rindt -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Dies ist ein digital signierter Nachrichtenteil URL: From email at jasonkohles.com Wed Feb 7 15:50:15 2007 From: email at jasonkohles.com (Jason Kohles) Date: Wed, 7 Feb 2007 10:50:15 -0500 Subject: Install order In-Reply-To: <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA201@buebe101.NOE.Nokia.com> References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be><01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com><4B5BDDC9AE76CA47AF37CE848F0DE46C033EA1AA@buebe101.NOE.Nokia.com> <200702070946.05044.jkeating@redhat.com> <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA201@buebe101.NOE.Nokia.com> Message-ID: <9B6031BA-267A-4B25-932B-E7C44028A4AD@jasonkohles.com> On Feb 7, 2007, at 9:57 AM, wrote: > Ok, thanks, but the packege is not mine but made by hp, so I don't > want > to modify it. > If the package is wrong and can't be modified, the most reasonable solution is just to run rpm in a %post script to install it manually. > > Tamas > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of ext Jesse > Keating > Sent: Wednesday, February 07, 2007 3:46 PM > To: kickstart-list at redhat.com > Subject: Re: Install order > > On Wednesday 07 February 2007 09:41, tamas.1.simon at nokia.com wrote: >> I just have some packages which actually doesn't have dependency but >> in the pre and post install script of rpm they use eg. grep to check >> something, and at this point grep is not installed, so the package >> will not be installed with kickstart. >> >> Is there any option in kickstart to tell a package 'when' it >> should be > >> installed? > > Use Requires(pre) or Requires(post) in your package if it uses grep. > This will ensure ordering so that grep is installed before your > package. > > -- > Jesse Keating > Release Engineer: Fedora > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- Jason Kohles email at jasonkohles.com http://www.jasonkohles.com/ "A witty saying proves nothing." -- Voltaire -------------- next part -------------- An HTML attachment was scrubbed... URL: From teng at dataway.com Wed Feb 7 19:14:28 2007 From: teng at dataway.com (Tedman Eng) Date: Wed, 7 Feb 2007 11:14:28 -0800 Subject: setting specific route by kickstart Message-ID: A host route means you want to route TO a specific host VIA a different gateway than the default gateway. You have only specified one machine, and it's unclear whether you want to route TO or VIA the address you describe. In a general sense, For the kickstart environment: the default route specified in the kickstart IP configuration would be your host route. The kickstart environment only ever contacts one server (your kickstart server), so there is no difference between a default route and a host route. For the build target: use %post to add some routing statements to you startup files, for example to /etc/rc.local Usually "route add xxx.xxx.xxx.xxx gw xxx.xxx.xxx.xxx" > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Daniel Rindt > Sent: Wednesday, February 07, 2007 7:15 AM > To: kickstart-list at redhat.com > Subject: setting specific route by kickstart > > Hello, > > my Hoster normally uses DHCP to get servers the IP Adress. But in my > case the dhcp seems not to work. I get from the provider the > information > that i must initiate a hostroute for eth0 over 10.255.255.1. > The gateway > has the same IP adress. Iam no network specialist, hope > anyone can help > me. > How i can set the route by kickstarting? > > Thanks in advance > Daniel Rindt > From drindt at mx-solutions.de Wed Feb 7 20:28:36 2007 From: drindt at mx-solutions.de (Daniel Rindt) Date: Wed, 07 Feb 2007 21:28:36 +0100 Subject: setting specific route by kickstart In-Reply-To: References: Message-ID: <1170880116.3315.8.camel@luminary.homelinux.net> Am Mittwoch, den 07.02.2007, 11:14 -0800 schrieb Tedman Eng: > > For the build target: use %post to add some routing statements to you > startup files, for example to /etc/rc.local > Usually "route add xxx.xxx.xxx.xxx gw xxx.xxx.xxx.xxx" The kickstartfile will be loaded via http. means that the installer needs the route at first. here is the example: ifconfig eth0 87.106.18.81 netmask 255.255.255.255 route add -host 10.255.255.1 dev eth0 route add default gw 10.255.255.1 dev eth0 this isnt possible with kernel paramters? Thanks in advance Daniel -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Dies ist ein digital signierter Nachrichtenteil URL: From sjbrowne at bluebottle.com Wed Feb 7 21:56:46 2007 From: sjbrowne at bluebottle.com (Stuart J. Browne) Date: Thu, 8 Feb 2007 08:56:46 +1100 Subject: Install order In-Reply-To: <9B6031BA-267A-4B25-932B-E7C44028A4AD@jasonkohles.com> References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be><01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com><4B5BDDC9AE76CA47AF37CE848F0DE46C033EA1AA@buebe101.NOE.Nokia.com><200702070946.05044.jkeating@redhat.com><4B5BDDC9AE76CA47AF37CE848F0DE46C033EA201@buebe101.NOE.Nokia.com> <9B6031BA-267A-4B25-932B-E7C44028A4AD@jasonkohles.com> Message-ID: <01e801c74b02$de86ea30$4001a8c0@trusteddelivery.local> Also, jump on the HP ITRC forums and tell them that it's screwed up. They're pretty good at fixing their screwup's. ________________________________ From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Jason Kohles Sent: Thursday, February 08, 2007 2:50 AM To: Discussion list about Kickstart Subject: Re: Install order On Feb 7, 2007, at 9:57 AM, wrote: Ok, thanks, but the packege is not mine but made by hp, so I don't want to modify it. If the package is wrong and can't be modified, the most reasonable solution is just to run rpm in a %post script to install it manually. Tamas -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of ext Jesse Keating Sent: Wednesday, February 07, 2007 3:46 PM To: kickstart-list at redhat.com Subject: Re: Install order On Wednesday 07 February 2007 09:41, tamas.1.simon at nokia.com wrote: I just have some packages which actually doesn't have dependency but in the pre and post install script of rpm they use eg. grep to check something, and at this point grep is not installed, so the package will not be installed with kickstart. Is there any option in kickstart to tell a package 'when' it should be installed? Use Requires(pre) or Requires(post) in your package if it uses grep. This will ensure ordering so that grep is installed before your package. From binand at gmail.com Thu Feb 8 01:27:48 2007 From: binand at gmail.com (Binand Sethumadhavan) Date: Thu, 8 Feb 2007 02:27:48 +0100 Subject: Install order In-Reply-To: <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA201@buebe101.NOE.Nokia.com> References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be> <01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com> <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA1AA@buebe101.NOE.Nokia.com> <200702070946.05044.jkeating@redhat.com> <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA201@buebe101.NOE.Nokia.com> Message-ID: On 07/02/07, tamas.1.simon at nokia.com wrote: > Ok, thanks, but the packege is not mine but made by hp, so I don't want > to modify it. Are you talking about the HP server management tools RPMs (hpasm, hprsm etc.)? I install them post-install, post-reboot. They are quite haphazardly packaged - I found my /etc/shadow gone when I included these RPMs in the install tree. They have other problems too, viz: - Some of them are packaged with rpm v3, so if I sign them with rpm v4 (to provision via my RHN satellite), they become corrupted. I need to run up2date --nosig only for them. - They include files which are deleted as part of post-installation configuration - a one-time configuration utility is installed in /sbin which will delete itself at the end of its own first run. Messes up rpm -V. - They create symlinks in %post (and delete them in %preun) instead of packaging those in. - Some of them run gcc in %post to compile modules for the running kernel. To be fair, I don't know how this can be avoided. Binand From tamas.1.simon at nokia.com Thu Feb 8 09:33:02 2007 From: tamas.1.simon at nokia.com (tamas.1.simon at nokia.com) Date: Thu, 8 Feb 2007 10:33:02 +0100 Subject: Install order In-Reply-To: References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be><01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com><4B5BDDC9AE76CA47AF37CE848F0DE46C033EA1AA@buebe101.NOE.Nokia.com><200702070946.05044.jkeating@redhat.com><4B5BDDC9AE76CA47AF37CE848F0DE46C033EA201@buebe101.NOE.Nokia.com> Message-ID: <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA7F7@buebe101.NOE.Nokia.com> Hi, No, I am not talking about those 'hpasm, hprsm etc' packages but vlan_init-1.0.1e-1.linux.rpm, which only contains the script, config files and text file to configure VLAN's on HP server. Thank you for the help, actually to put this package to the %post section worked. Tamas -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of ext Binand Sethumadhavan Sent: Thursday, February 08, 2007 2:28 AM To: Discussion list about Kickstart Subject: Re: Install order On 07/02/07, tamas.1.simon at nokia.com wrote: > Ok, thanks, but the packege is not mine but made by hp, so I don't > want to modify it. Are you talking about the HP server management tools RPMs (hpasm, hprsm etc.)? I install them post-install, post-reboot. They are quite haphazardly packaged - I found my /etc/shadow gone when I included these RPMs in the install tree. They have other problems too, viz: - Some of them are packaged with rpm v3, so if I sign them with rpm v4 (to provision via my RHN satellite), they become corrupted. I need to run up2date --nosig only for them. - They include files which are deleted as part of post-installation configuration - a one-time configuration utility is installed in /sbin which will delete itself at the end of its own first run. Messes up rpm -V. - They create symlinks in %post (and delete them in %preun) instead of packaging those in. - Some of them run gcc in %post to compile modules for the running kernel. To be fair, I don't know how this can be avoided. Binand _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From debian at herakles.homelinux.org Thu Feb 8 22:11:39 2007 From: debian at herakles.homelinux.org (John Summerfield) Date: Fri, 09 Feb 2007 07:11:39 +0900 Subject: Install order In-Reply-To: <9B6031BA-267A-4B25-932B-E7C44028A4AD@jasonkohles.com> References: <2F4F96E90349F241809293630A893B3D0212C051@MAIL03.rvponp.fgov.be><01A04652-E820-4C47-B9D5-41C69043DD37@jasonkohles.com><4B5BDDC9AE76CA47AF37CE848F0DE46C033EA1AA@buebe101.NOE.Nokia.com> <200702070946.05044.jkeating@redhat.com> <4B5BDDC9AE76CA47AF37CE848F0DE46C033EA201@buebe101.NOE.Nokia.com> <9B6031BA-267A-4B25-932B-E7C44028A4AD@jasonkohles.com> Message-ID: <45CBA01B.6080703@herakles.homelinux.org> Jason Kohles wrote: > On Feb 7, 2007, at 9:57 AM, > wrote: > >> Ok, thanks, but the packege is not mine but made by hp, so I don't want >> to modify it. Or create your own package that requires HP's package and grep appropriately. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu Please do not reply off-list From debian at herakles.homelinux.org Thu Feb 8 22:29:18 2007 From: debian at herakles.homelinux.org (John Summerfield) Date: Fri, 09 Feb 2007 07:29:18 +0900 Subject: Server Provisioning - Slightly OT In-Reply-To: <8552c3a30701282103v76f95d97x7c92e291c6da669a@mail.gmail.com> References: <8552c3a30701282103v76f95d97x7c92e291c6da669a@mail.gmail.com> Message-ID: <45CBA43E.4010102@herakles.homelinux.org> Mike M wrote: > Hello, > > Since this list deals with OS provisioning (albeit Red Hat-focused), I > figured this would be the right group of people to get some opinions > from. > > I work in a data center with >2000 servers, about 90% of which are > running some variation of Red Hat Linux, from 7.3 to ES 4. The > remainder are Windows Server 2003 and Solaris 10 x86. As you can > tell, OS provisioning is very important for a shop of this size. > > Using kickstart, we have been able to make our larger Linux > deployments very quick, because there is usually a day or two in the > project plan that allows us to get the legwork done in advance. > However, we run into pretty severe problems when we only have one or > two servers that need to be provisioned in one shot. This is because > there is a lot of stuff that needs to be done in advance in the > standard kickstart process - you need to harvest MAC addresses, enable > portfast on the switch, create DNS entries, create a kickstart config, > etc. While we could certainly install from CD, there is the > possibility of non-standard package selections, and a by-hand %post of > all non-Red Hat packages and tweaks. > > Now, as I see it, there are two options in the provisioning world - > image-based (ghost, partimage), or native installation (kickstart, > RIS). Each option seems to have pros and cons. > > So here are my questions to the group: > > If you run a multi-platform server farm, how do you provision the OS? > Have you purchased a product, used one specific package, or cobbled > your own from free tools? Do you use image-based technology, and if > so, how hard do you find image management? If you use images, how do > you tackle "uniqueness" (hostnames, IP addresses, etc.) of > newly-provisioned machines? I only do tens of machines. I use dhcp on Linux (Debian) for Windows and Linux. I've not tried chaining from pxelinux to the MS boot code, but I suspect it can be done. What gets installed here depends, in part, which network the machine boots on. I don't understand why people need to harvest the MAC addresses; I choose from the pxelinux menu. I _do_ configure dhcp to give different replies (including IP address, filename etc) depending what's asking: the BIOS gets a different answer from what I tell Linux. You can also configure it to give different answers for Mac's boot code, and I expect, for *sparc*. Probably for NIC vendor too - an expression involving the hardware address should work. Here's a fragment to recognise an Intelish box: class "pxeclients" { match if substring (option vendor-class-identifier, 0, 9) = "PXEClient"; option vendor-class-identifier "PXEClient"; # At least one of the vendor-specific option must be set. We set # the MCAST IP address to 0.0.0.0 to tell the bootrom to still use # TFTP (address 0.0.0.0 is considered as "no address") # option PXE.mtftp-ip 192.168.9.1; vendor-option-space PXE; option PXE.mtftp-ip 0.0.0.0; vendor-option-space PXE; } Here are some more:-) class "anaconda" { match if substring (option vendor-class-identifier, 0, 8) = "anaconda"; option vendor-class-identifier "anaconda"; } class "debian-installer" { match if substring (option vendor-class-identifier, 0, 3) = "d-i"; option root-path "192.168.9.4:/tftpboot/PXE/linux-rootfs/kubuntu/edgy"; option interface-mtu 1500; } Etherboot: # Filched from http://clic.mandrakesoft.com/documentation/pxe/ch04.html class "Etherboot" { match if substring (option vendor-class-identifier, 0, 13) = "Etherboot-5.0"; option vendor-encapsulated-options 3c:09:45:74:68:65:72:62:6f:6f:74:ff; option vendor-class-identifier "Etherboot-5.0"; filename "tftpboot.vc"; } I configured this in hopes: pool { allow members of "anaconda"; deny members of "pxeclients"; default-lease-time 900; filename "http://Fedora.demo.lan/5/i386/os/Fedora/"; max-lease-time 1800; range 192.168.9.170 192.168.9.179; option log-servers 192.168.9.4; } Red Hat's objection was the ill-founded notion they'd need to register something with IANA. Not so. > > Any advice you can give will be much appreciated. I'd love to hear > from anyone, especially those with large mixed environments. As you > can imagine, we have to use different methods for each OS (Kickstart, > Jumpstart, and RIS/Ghost), and I would love to see something that can > handle them all, or at least simplify and unify the three methods. > > Mike > > P.S. And yes, as much as it may seem otherwise, I really don't have > rose-colored glasses on...I understand that there's probably no *easy* > way to do it, but I figured it couldn't hurt to ask. ;) > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu Please do not reply off-list From binand at gmail.com Mon Feb 19 02:39:16 2007 From: binand at gmail.com (Binand Sethumadhavan) Date: Mon, 19 Feb 2007 03:39:16 +0100 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: References: <45B8CC1A.3010305@rampaginggeek.com> Message-ID: On 26/01/07, Panu Matilainen wrote: > - if you run 32bit kernel on x86_64 it wont report itself as x86_64 but > i686 or athlon, this will cause various places (including anaconda Newbie-ish question... what exactly is the difference between i686 and x86_64? I imagine athlon is specifically for AMD 64-bit processors? Is there a weblink that explains what each of the flags in /proc/cpuinfo stands for? Thanks, Binand From klaus.steden at thomson.net Mon Feb 19 02:59:59 2007 From: klaus.steden at thomson.net (Klaus Steden) Date: Sun, 18 Feb 2007 21:59:59 -0500 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: References: <45B8CC1A.3010305@rampaginggeek.com> Message-ID: <20070219025959.GF1253@thomson.net> > >- if you run 32bit kernel on x86_64 it wont report itself as x86_64 but > > i686 or athlon, this will cause various places (including anaconda > > Newbie-ish question... what exactly is the difference between i686 and > x86_64? I imagine athlon is specifically for AMD 64-bit processors? > > Is there a weblink that explains what each of the flags in /proc/cpuinfo > stands for? > One short answer is that a 32-bit CPU (i386 - Pentium, Athlon, etc.) uses 32-bit addressing, while a 64-bit CPU (x86_64 - Opteron, Itanium, etc.) uses 64-bit addressing -- meaning your OS and applications can address more than 4 GB of memory at one time. hth, Klaus From binand at gmail.com Mon Feb 19 03:24:55 2007 From: binand at gmail.com (Binand Sethumadhavan) Date: Mon, 19 Feb 2007 04:24:55 +0100 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: <20070219025959.GF1253@thomson.net> References: <45B8CC1A.3010305@rampaginggeek.com> <20070219025959.GF1253@thomson.net> Message-ID: On 19/02/07, Klaus Steden wrote: > > Newbie-ish question... what exactly is the difference between i686 and > > x86_64? I imagine athlon is specifically for AMD 64-bit processors? > > > One short answer is that a 32-bit CPU (i386 - Pentium, Athlon, etc.) uses > 32-bit addressing, while a 64-bit CPU (x86_64 - Opteron, Itanium, etc.) uses > 64-bit addressing -- meaning your OS and applications can address more than 4 > GB of memory at one time. I know that much... my question was specifically about x86_64 and i686 (since I was under the impression that i686 also means a 64-bit processor). I imagine i686 covers a much wider range of CPUs, not all of which are 64-bit? Binand From jkeating at redhat.com Mon Feb 19 13:14:37 2007 From: jkeating at redhat.com (Jesse Keating) Date: Mon, 19 Feb 2007 08:14:37 -0500 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: References: <45B8CC1A.3010305@rampaginggeek.com> <20070219025959.GF1253@thomson.net> Message-ID: <200702190814.37552.jkeating@redhat.com> On Sunday 18 February 2007 22:24, Binand Sethumadhavan wrote: > I know that much... my question was specifically about x86_64 and i686 > (since I was under the impression that i686 also means a 64-bit > processor). > > I imagine i686 covers a much wider range of CPUs, not all of which are > 64-bit? i686 does not mean 64-bit. i686 is still 32bit. It's just a newer family of 32bit processors. (newer than i586/i486/i386) -- Jesse Keating Release Engineer: Fedora -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: From wanderso at csl.co.uk Mon Feb 19 15:29:52 2007 From: wanderso at csl.co.uk (William Anderson) Date: Mon, 19 Feb 2007 15:29:52 +0000 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: <200702190814.37552.jkeating@redhat.com> References: <45B8CC1A.3010305@rampaginggeek.com> <20070219025959.GF1253@thomson.net> <200702190814.37552.jkeating@redhat.com> Message-ID: <45D9C270.8030401@csl.co.uk> Jesse Keating wrote: > On Sunday 18 February 2007 22:24, Binand Sethumadhavan wrote: >> I know that much... my question was specifically about x86_64 and i686 >> (since I was under the impression that i686 also means a 64-bit >> processor). >> >> I imagine i686 covers a much wider range of CPUs, not all of which are >> 64-bit? > > i686 does not mean 64-bit. i686 is still 32bit. It's just a newer family of > 32bit processors. (newer than i586/i486/i386) To clarify: i286 = 80286 i386 = 80386 i486 = 80486 i586 = Pentium i686 = Pentium Pro, Pentium {II|III|4|M|D} / Celeron {M|D} / Xeon / Core x86_64 = AMD {Sempron|Athlon64|Opteron} / Pentium {4|D} / Xeon / Core running in 64-bit mode Binand, when you ask "detect" 64-bitness in a kickstart, do you mean in the initial options to determine a repo to use, or in %pre / %post? If the latter, `uname -i` should suffice. -- William Anderson, Systems Administrator Concept Systems Ltd. | +44 (0)131 200 4789 wanderso at csl.co.uk | http://www.csl.co.uk/ This email and any files transmitted with it are confidential and are intended solely for the use of the individual or entity to whom they are addressed. If you are not the original recipient or the person responsible for delivering the email to the intended recipient, be advised that you have received this email in error, and that any use, dissemination, forwarding, printing, or copying of this email is strictly prohibited. If you received this email in error, please immediately notify the sender and delete the original. From klaus.steden at thomson.net Mon Feb 19 15:43:06 2007 From: klaus.steden at thomson.net (Klaus Steden) Date: Mon, 19 Feb 2007 10:43:06 -0500 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: <45D9C270.8030401@csl.co.uk> References: <45B8CC1A.3010305@rampaginggeek.com> <20070219025959.GF1253@thomson.net> <200702190814.37552.jkeating@redhat.com> <45D9C270.8030401@csl.co.uk> Message-ID: <20070219154306.GH1253@thomson.net> > Binand, when you ask "detect" 64-bitness in a kickstart, do you mean in > the initial options to determine a repo to use, or in %pre / %post? If > the latter, `uname -i` should suffice. > Sadly, that won't work, since 'uname -i' will only report on what the architecture the kernel is ... you'd have to examine /proc/cpuinfo for the complete story (which is what started this thread I think). Klaus From binand at gmail.com Mon Feb 19 16:59:34 2007 From: binand at gmail.com (Binand Sethumadhavan) Date: Mon, 19 Feb 2007 17:59:34 +0100 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: <45D9C270.8030401@csl.co.uk> References: <45B8CC1A.3010305@rampaginggeek.com> <20070219025959.GF1253@thomson.net> <200702190814.37552.jkeating@redhat.com> <45D9C270.8030401@csl.co.uk> Message-ID: On 19/02/07, William Anderson wrote: > Binand, when you ask "detect" 64-bitness in a kickstart, do you mean in > the initial options to determine a repo to use, or in %pre / %post? If > the latter, `uname -i` should suffice. I didn't ask this question - it was Jason Edgecombe who started this thread. I just wanted to extricate myself from the various alphanumero-underscoric acronyms when it came to processor architectures :) Binand From jason at rampaginggeek.com Mon Feb 19 21:15:18 2007 From: jason at rampaginggeek.com (Jason Edgecombe) Date: Mon, 19 Feb 2007 16:15:18 -0500 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: References: <45B8CC1A.3010305@rampaginggeek.com> <20070219025959.GF1253@thomson.net> <200702190814.37552.jkeating@redhat.com> <45D9C270.8030401@csl.co.uk> Message-ID: <45DA1366.60002@rampaginggeek.com> Binand Sethumadhavan wrote: > On 19/02/07, William Anderson wrote: >> Binand, when you ask "detect" 64-bitness in a kickstart, do you mean in >> the initial options to determine a repo to use, or in %pre / %post? If >> the latter, `uname -i` should suffice. > > I didn't ask this question - it was Jason Edgecombe who started this > thread. I just wanted to extricate myself from the various > alphanumero-underscoric acronyms when it came to processor > architectures :) > > Binand To summarize the thread, the final analysis says that the i386/ia_64 logic cannot be done in kickstart and must be done earlier in the boot process. There was a feature suggestion on the syslinux mailing list about having a 32bit default vs a 64bit default. Branching in syslinux looks to be the best option. Sincerely, Jason Edgecombe From jkeating at j2solutions.net Mon Feb 19 21:51:52 2007 From: jkeating at j2solutions.net (Jesse Keating) Date: Mon, 19 Feb 2007 16:51:52 -0500 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: <45DA1366.60002@rampaginggeek.com> References: <45B8CC1A.3010305@rampaginggeek.com> <45DA1366.60002@rampaginggeek.com> Message-ID: <200702191651.52959.jkeating@j2solutions.net> On Monday 19 February 2007 16:15, Jason Edgecombe wrote: > To summarize the thread, the final analysis says that the i386/ia_64 > logic cannot be done in kickstart and must be done earlier in the boot > process. There was a feature suggestion on the syslinux mailing list > about having a 32bit default vs a 64bit default. Branching in syslinux > looks to be the best option. Perhaps you need to explain a bit more about what you wanted to _do_ with that detection. There are many things you could do in %pre based on what kernel is loaded. It would require conscious thought on what kernel to boot at syslinux or pxelinux, but that isn't very hard. -- Jesse Keating RHCE (geek.j2solutions.net) Fedora Legacy Team (www.fedoralegacy.org) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: From jason at rampaginggeek.com Mon Feb 19 23:28:55 2007 From: jason at rampaginggeek.com (Jason Edgecombe) Date: Mon, 19 Feb 2007 18:28:55 -0500 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: <200702191651.52959.jkeating@j2solutions.net> References: <45B8CC1A.3010305@rampaginggeek.com> <45DA1366.60002@rampaginggeek.com> <200702191651.52959.jkeating@j2solutions.net> Message-ID: <45DA32B7.4040804@rampaginggeek.com> Jesse Keating wrote: > On Monday 19 February 2007 16:15, Jason Edgecombe wrote: > >> To summarize the thread, the final analysis says that the i386/ia_64 >> logic cannot be done in kickstart and must be done earlier in the boot >> process. There was a feature suggestion on the syslinux mailing list >> about having a 32bit default vs a 64bit default. Branching in syslinux >> looks to be the best option. >> > > Perhaps you need to explain a bit more about what you wanted to _do_ with that > detection. There are many things you could do in %pre based on what kernel > is loaded. It would require conscious thought on what kernel to boot at > syslinux or pxelinux, but that isn't very hard. > Basically, I wanted to autodetect a 32/64 bit processor and install the proper 32/64 bit RedHat based on the processor type. I've found out since that my users aren't clamoring for 64bit and I'll probably want to consciously choose 32/64bit rather than autodetect. Sincerely, Jason From debian at herakles.homelinux.org Mon Feb 19 23:26:12 2007 From: debian at herakles.homelinux.org (John Summerfield) Date: Tue, 20 Feb 2007 08:26:12 +0900 Subject: detecting i386 or x86_64 in kickstart In-Reply-To: References: <45B8CC1A.3010305@rampaginggeek.com> <20070219025959.GF1253@thomson.net> Message-ID: <45DA3214.2090309@herakles.homelinux.org> Binand Sethumadhavan wrote: > On 19/02/07, Klaus Steden wrote: > >> > Newbie-ish question... what exactly is the difference between i686 and >> > x86_64? I imagine athlon is specifically for AMD 64-bit processors? >> > >> One short answer is that a 32-bit CPU (i386 - Pentium, Athlon, etc.) uses >> 32-bit addressing, while a 64-bit CPU (x86_64 - Opteron, Itanium, >> etc.) uses >> 64-bit addressing -- meaning your OS and applications can address more >> than 4 >> GB of memory at one time. > > > I know that much... my question was specifically about x86_64 and i686 > (since I was under the impression that i686 also means a 64-bit > processor). > > I imagine i686 covers a much wider range of CPUs, not all of which are > 64-bit? Trye, As does "athlon." Similar on other architectures too, not all Sparcs, Powers, S/360 family CPUs do 64-bits. Doubtless the flags depend on atchitecture too: the flags register is an 8086 thing, S/370 uses control registers for this (and more). -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu Please do not reply off-list From tbergema at eso.org Tue Feb 20 14:37:35 2007 From: tbergema at eso.org (Timo Bergemann) Date: Tue, 20 Feb 2007 15:37:35 +0100 Subject: [Fwd: yum updates] Message-ID: <45DB07AF.6080806@eso.org> Hello all, I was working earlier with kickstart and I was used to make first a basic installation via kickstart and then to upgrade to the latest packages via yum in the %POST. unfortunately this I have seen the last time working in FC3. when I try to update all packages in %post I get a lot like followed: Traceback (most recent call last): File "/mnt/sysimage/usr/bin/yum", line 28, in ? import yummain ImportError: No module named yummain even if I do not believe that i should do a chroot to /mnt/sysimage, if I try the same but before "chroot /mnt/sysimage" I come further but not far enough. I get a list of all packages which will get upgraded but then finally a crash with a lot of errors like this which I will append later if needed. But the fix should be anyway NOT to chroot Another problem is if i try to run a perl script. it also complains about missing libs which are not in / but already in /mnt/sysimage But if I try to chroot to /mnt/sysimage before executing the perl script it works I believe that this is the same error: unreached / unknown libs. can anybody tell me what has been changed or even better - how to I get yum running during the installation again? Or/ AND is there a way to install from updated repositories without creating my own updated "base" repo, if I give all repositories to the kickstart script: nfs --server oscrepo.hq.eso.org --dir /repo/mirrors/fedora-core-6-base/i386/os/ repo --name=updates --baseurl=ftp://oscrepo.hq.eso.org/linuxrepos/hosts/nb012931/eso/formats/yum/ repo --name=eso --baseurl=ftp://oscrepo.hq.eso.org/linuxrepos/hosts/nb012931/eso/formats/yum/ %packages --resolvedeps Thank you for your help Timo -- \|/ (o o) _______________________oOO__(.)__OOo_________________________ ______ ______ _______ / ____/ / ____/ / ___ / Timo Bergemann / /___ / /___ / / / / European Southern Observatory / ____/ /___ / / / / / Karl-Schwarzschildstr. 2 / /___ ___ / / / /__/ / D-85748 Garching bei Muenchen /_____/ /_____/ /______/ Germany fax : +49 89 3200 6380 e-mail : tbergema at eso.org voice : +49 89 3200 6760 www : http://www.eso.org _____________________________________________________________ From skvidal at linux.duke.edu Tue Feb 20 16:11:52 2007 From: skvidal at linux.duke.edu (seth vidal) Date: Tue, 20 Feb 2007 11:11:52 -0500 Subject: [Fwd: yum updates] In-Reply-To: <45DB07AF.6080806@eso.org> References: <45DB07AF.6080806@eso.org> Message-ID: <1171987912.28200.14.camel@cutter> On Tue, 2007-02-20 at 15:37 +0100, Timo Bergemann wrote: > Hello all, > > I was working earlier with kickstart and I was used to make first a > basic installation via kickstart and then to upgrade to the latest > packages via yum in the %POST. > unfortunately this I have seen the last time working in FC3. > > when I try to update all packages in %post I get a lot like followed: > > Traceback (most recent call last): > File "/mnt/sysimage/usr/bin/yum", line 28, in ? > import yummain > ImportError: No module named yummain > > even if I do not believe that i should do a chroot to /mnt/sysimage, > if I try the same but before "chroot /mnt/sysimage" I come further but > not far enough. I get a list of all packages which will get upgraded but > then finally a crash with a lot of errors like this which I will append > later if needed. > But the fix should be anyway NOT to chroot > 1. you do need to be in the chroot 2. if you can post the errors you're seeing on update that would be appreciated. -sv From terryb at gg.specsavers.com Tue Feb 20 16:53:40 2007 From: terryb at gg.specsavers.com (Terry Black) Date: Tue, 20 Feb 2007 16:53:40 +0000 Subject: [Fwd: yum updates] Message-ID: A problem I had was that I ran out of space in /var/spool/up2date. If there are a lot of packages you would be surprised hoe much space is needed. Terry On Tue, 2007-02-20 at 15:37 +0100, Timo Bergemann wrote: > Hello all, > > I was working earlier with kickstart and I was used to make first a > basic installation via kickstart and then to upgrade to the latest > packages via yum in the %POST. > unfortunately this I have seen the last time working in FC3. > > when I try to update all packages in %post I get a lot like followed: > > Traceback (most recent call last): > File "/mnt/sysimage/usr/bin/yum", line 28, in ? > import yummain > ImportError: No module named yummain > > even if I do not believe that i should do a chroot to /mnt/sysimage, > if I try the same but before "chroot /mnt/sysimage" I come further but > not far enough. I get a list of all packages which will get upgraded but > then finally a crash with a lot of errors like this which I will append > later if needed. > But the fix should be anyway NOT to chroot > 1. you do need to be in the chroot 2. if you can post the errors you're seeing on update that would be appreciated. -sv _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list ----------------------------------------- NOTICE: This message contains privileged and confidential information intended only for the addressee. If you have received this message in error you must not disseminate, copy or take action on it; please notify postmaster at specsavers.com Opinions expressed in this message are those of the sender and do not necessarily represent those of Specsavers. Although this e-mail and any attachments are believed to be virus free, e-mail communications are not 100% secure and Specsavers makes no warranty that this message is secure or virus free. All references to Specsavers mean Specsavers Optical Group Limited or any member of the Specsavers Group of Companies. Nothing in this transmission shall or shall be deemed to constitute an offer or acceptance of an offer or otherwise have the effect of forming a contract by electronic communication. Your name and address may be stored to facilitate communications. From tbergema at eso.org Wed Feb 21 14:13:55 2007 From: tbergema at eso.org (Timo Bergemann) Date: Wed, 21 Feb 2007 15:13:55 +0100 Subject: [Fwd: yum updates] In-Reply-To: References: Message-ID: <45DC53A3.30307@eso.org> Hello, thank for your reply. just to make it complete... I assumed that something changed, that the installer does not anymore a chroot by default. But it should, as I just found options to disables this like %post --nochroot. anyway, yum is running again. I simply had some errors in my post installations which confused yum that much that it was not working anymore like written below. Thanks Seth, your hind was directing me in the right direction. THX Timo Terry Black schrieb: > A problem I had was that I ran out of space in /var/spool/up2date. If there are a lot of packages you would be surprised hoe much space is needed. > Terry > > > On Tue, 2007-02-20 at 15:37 +0100, Timo Bergemann wrote: > >> Hello all, >> >> I was working earlier with kickstart and I was used to make first a >> basic installation via kickstart and then to upgrade to the latest >> packages via yum in the %POST. >> unfortunately this I have seen the last time working in FC3. >> >> when I try to update all packages in %post I get a lot like followed: >> >> Traceback (most recent call last): >> File "/mnt/sysimage/usr/bin/yum", line 28, in ? >> import yummain >> ImportError: No module named yummain >> >> even if I do not believe that i should do a chroot to /mnt/sysimage, >> if I try the same but before "chroot /mnt/sysimage" I come further but >> not far enough. I get a list of all packages which will get upgraded but >> then finally a crash with a lot of errors like this which I will append >> later if needed. >> But the fix should be anyway NOT to chroot >> >> > > 1. you do need to be in the chroot > 2. if you can post the errors you're seeing on update that would be > appreciated. > > > -sv > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > ----------------------------------------- > NOTICE: This message contains privileged and confidential > information intended only for the addressee. If you have received > this message in error you must not disseminate, copy or take action > on it; please notify postmaster at specsavers.com Opinions expressed > in this message are those of the sender and do not necessarily > represent those of Specsavers. Although this e-mail and any > attachments are believed to be virus free, e-mail communications > are not 100% secure and Specsavers makes no warranty that this > message is secure or virus free. All references to Specsavers mean > Specsavers Optical Group Limited or any member of the Specsavers > Group of Companies. Nothing in this transmission shall or shall be > deemed to constitute an offer or acceptance of an offer or > otherwise have the effect of forming a contract by electronic > communication. Your name and address may be stored to facilitate > communications. > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- \|/ (o o) _______________________oOO__(.)__OOo_________________________ ______ ______ _______ / ____/ / ____/ / ___ / Timo Bergemann / /___ / /___ / / / / European Southern Observatory / ____/ /___ / / / / / Karl-Schwarzschildstr. 2 / /___ ___ / / / /__/ / D-85748 Garching bei Muenchen /_____/ /_____/ /______/ Germany fax : +49 89 3200 6380 e-mail : tbergema at eso.org voice : +49 89 3200 6760 www : http://www.eso.org _____________________________________________________________ From hawke at hawkesnest.net Thu Feb 22 00:19:21 2007 From: hawke at hawkesnest.net (Alex Mauer) Date: Wed, 21 Feb 2007 18:19:21 -0600 Subject: kickstart install on "low" memory machine Message-ID: I've been trying to do a kickstart install on some machines with a "low" amount of memory (128MB). It gets through the initial setup, formatting and partitioning the hard drive. But then as it's installing packages, it will hang (generally on glibc-common). If I watch free on the second vt, I can see that it's swapping, and I of course expect it to be slow when it has to do that. But eventually it will just hang, and the HDD activity light stops and the free memory amount stops changing. It's not out of memory (there's generally about 16-32MB of memory free). It hasn't locked up (the system still accepts input, I can switch VTs and run programs (and kill anaconda too)). But I can leave it sitting there for hours (I think 16 was the longest I waited) and it won't ever get past whichever package it hangs on -- it does get to 100% on that package though. If I upgrade the system to 256MB of RAM, it runs through the install in about 45 min. Does anyone have any suggestions for additional debugging/diagnosis I can do? -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 252 bytes Desc: OpenPGP digital signature URL: From skvidal at linux.duke.edu Thu Feb 22 04:28:18 2007 From: skvidal at linux.duke.edu (seth vidal) Date: Wed, 21 Feb 2007 23:28:18 -0500 Subject: [Fwd: yum updates] In-Reply-To: <45DC53A3.30307@eso.org> References: <45DC53A3.30307@eso.org> Message-ID: <1172118498.8256.31.camel@cutter> On Wed, 2007-02-21 at 15:13 +0100, Timo Bergemann wrote: > Hello, > thank for your reply. > just to make it complete... > I assumed that something changed, that the installer does not anymore a > chroot by default. > But it should, as I just found options to disables this like %post > --nochroot. > anyway, yum is running again. > I simply had some errors in my post installations which confused yum > that much that it was not working anymore like written below. > Timo, I'm glad I could help. However, I have one comment to make: > your hind was directing me in the right direction. THX the typo in the above sentence is the most troubling thing I've read in quite some time. :-D I don't want my hind directing anyone! :) Good luck with your installs. -sv From Chip.Shabazian at bankofamerica.com Thu Feb 22 04:37:57 2007 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 21 Feb 2007 20:37:57 -0800 Subject: kickstart install on "low" memory machine In-Reply-To: Message-ID: Check out rule-project.org -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Alex Mauer Sent: Wednesday, February 21, 2007 4:19 PM To: Kickstart-list at redhat.com Subject: kickstart install on "low" memory machine I've been trying to do a kickstart install on some machines with a "low" amount of memory (128MB). It gets through the initial setup, formatting and partitioning the hard drive. But then as it's installing packages, it will hang (generally on glibc-common). If I watch free on the second vt, I can see that it's swapping, and I of course expect it to be slow when it has to do that. But eventually it will just hang, and the HDD activity light stops and the free memory amount stops changing. It's not out of memory (there's generally about 16-32MB of memory free). It hasn't locked up (the system still accepts input, I can switch VTs and run programs (and kill anaconda too)). But I can leave it sitting there for hours (I think 16 was the longest I waited) and it won't ever get past whichever package it hangs on -- it does get to 100% on that package though. If I upgrade the system to 256MB of RAM, it runs through the install in about 45 min. Does anyone have any suggestions for additional debugging/diagnosis I can do? From sdalton1 at gmail.com Thu Feb 22 15:49:55 2007 From: sdalton1 at gmail.com (Steven Dalton) Date: Thu, 22 Feb 2007 10:49:55 -0500 Subject: kickstart installation hardware support Message-ID: <7b5a52eb0702220749o4d014bc7oa8124ed5887f4ebe@mail.gmail.com> I am having some problems installing Fedora core 5 on my system. When I use the FC5 boot disk with the kickstart directive linux ks=http://somedomain/ks.cfg The installation ignores the ks configuration completely and continues on to the graphical installation. I do not know if it is relevent but I do not have dhcp on my network where the machine so I need to setup the ip address statically. As a note, when I try to boot using the FC6 boot disk with the same ks option with no DHCP everything works perfectly. I want to know what the problem might be and how to add driver support for my kickstart installation. Here is my system hardware description. - Intel Dual-Core 64-bit Xeon processor - Intel E7520 Chipset - Intel 82546GB dual-port Gigabit Ethernet controller -------------- next part -------------- An HTML attachment was scrubbed... URL: From matsugae.kiyoaki at gmail.com Thu Feb 22 16:18:49 2007 From: matsugae.kiyoaki at gmail.com (Kiyoaki Matsugae) Date: Thu, 22 Feb 2007 17:18:49 +0100 Subject: kickstart installation hardware support In-Reply-To: <7b5a52eb0702220749o4d014bc7oa8124ed5887f4ebe@mail.gmail.com> References: <7b5a52eb0702220749o4d014bc7oa8124ed5887f4ebe@mail.gmail.com> Message-ID: <31d9a9570702220818y29181f23ke3340479258398ea@mail.gmail.com> Steven, can you try to replace somedomain by the ip address of your server ? if dhcp does not give you a valid DNS server, that might be the reason why your install fails Kiyo On 2/22/07, Steven Dalton wrote: > > I am having some problems installing Fedora core 5 on my system. When I > use the FC5 boot disk with the kickstart directive > > linux ks=http://somedomain/ks.cfg > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From clumens at redhat.com Thu Feb 22 16:39:54 2007 From: clumens at redhat.com (Chris Lumens) Date: Thu, 22 Feb 2007 11:39:54 -0500 Subject: kickstart installation hardware support In-Reply-To: <7b5a52eb0702220749o4d014bc7oa8124ed5887f4ebe@mail.gmail.com> References: <7b5a52eb0702220749o4d014bc7oa8124ed5887f4ebe@mail.gmail.com> Message-ID: <20070222163954.GO4545@exeter.boston.redhat.com> > I am having some problems installing Fedora core 5 on my system. When I use > the FC5 boot disk with the kickstart directive > > linux ks=http://somedomain/ks.cfg > > The installation ignores the ks configuration completely and continues on to > the graphical installation. I do not know if it is relevent but I do not > have dhcp on my network where the machine so I need to setup the ip address > statically. Previous version of anaconda had a bug where if it was unable to get the kickstart file for any reason (bad URL, couldn't get network access, phase of the moon, etc.) it would just fall over to a regular interactive install with no notice at all. This should be fixed in F7 test 1. It will tell you there was trouble getting the kickstart file and prompt you to fix up the ks= parameter you provided to make another attempt. - Chris From hawke at hawkesnest.net Thu Feb 22 19:57:18 2007 From: hawke at hawkesnest.net (Alex Mauer) Date: Thu, 22 Feb 2007 13:57:18 -0600 Subject: kickstart install on "low" memory machine In-Reply-To: References: Message-ID: Shabazian, Chip wrote: > Check out rule-project.org That doesn't really help, as rule-project's installer doesn't support network booting or kickstart as far as I can see. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 252 bytes Desc: OpenPGP digital signature URL: From Eric.KENNEDYvanDAM at rvponp.fgov.be Fri Feb 23 07:49:05 2007 From: Eric.KENNEDYvanDAM at rvponp.fgov.be (KENNEDY VAN DAM Eric) Date: Fri, 23 Feb 2007 08:49:05 +0100 Subject: kickstart installation hardware support References: <7b5a52eb0702220749o4d014bc7oa8124ed5887f4ebe@mail.gmail.com> Message-ID: <2F4F96E90349F241809293630A893B3D0245AA70@MAIL03.rvponp.fgov.be> > De : kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] De la part de Steven Dalton > Envoy? : jeudi 22 f?vrier 2007 16:50 > ? : kickstart-list at redhat.com > Objet : kickstart installation hardware support > > > I am having some problems installing Fedora core 5 on my system. When I use the FC5 boot disk with the kickstart directive > > linux ks=http://somedomain/ks.cfg > > The installation ignores the ks configuration completely and continues on to the graphical installation. I do not know if it > is relevent but I do not > have dhcp on my network where the machine so I need to setup the ip address statically. Hi. As you don't have any DHCP server, you can try to configure ip on lilo/grub. Have a look to the ip kernel parameter: ip=:::::: In your case, I should try something like this: linux ks=http:///ks.cfg ip=:::::: Hope this can help Eric From William.Ramthun at Level3.com Fri Feb 23 16:01:18 2007 From: William.Ramthun at Level3.com (Ramthun, William) Date: Fri, 23 Feb 2007 09:01:18 -0700 Subject: "pump" times out -- extend a few seconds more? In-Reply-To: <40937.72.165.61.226.1168623154.squirrel@webmail.linuxpowered.net> Message-ID: > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of nate > Sent: Friday, January 12, 2007 10:33 AM > To: kickstart-list at redhat.com > Subject: Re: "pump" times out -- extend a few seconds more? > > Bill-Schoolcraft said: > > Hello Family, > > > > I have a series of switches, in racks, and one of them seems to be a bit > > slow so when kickstart goes for it's dhcp request, pump times out. The > > initial PXE takes much more time to get an "ack" at the initial contact > > than the same boxes in other racks with other switches but then finally > > starts chugging away -- it's the next dhcp request that fails, times out. > > > > Today I took the same box, draped a cat6 cable to another, less > > populated switch, and it kicked fine. > > > > (question) > > So my question is, how or where can I place an argument to tell > > kickstart's "pump" request to try longer before timing out. > > just curious, what type of switch? many older cisco switches have > STP on by default, if you set the port(s) to 'portfast' the links come > up immediately rather than waiting ~45 seconds. > > I'm not a cisco guy but if your switches are cisco the command > is something like > > int (interface) > spanning-tree portfast > ^Z > wr mem > > > e.g. on my old 3500s (IOS 12.0 is the latest they'll run) > int fa0/5 > spanning-tree portfast > > I haven't encountered any other switch vendors that have this > 'feature' enabled by default. You can tell if it's on without > even logging in(to the switch) by checking to see if the link > light for the port spends a long time in the orange color > before turning green. I've been told more modern cisco switches > ship with this feature off by default. > > as for extending the pump times, I'm not sure, so I can't > answer your specific question :) maybe the above will help a bit.. > > I did a quick search on pump timeout, and what I did see was > that it may be possible to extend the timeout by adding a > /etc/pump_device.conf and using the option 'timeout '. > > though you'll need to add that to the initrd image..not the > easiest thing in the world. > > nate > I am experiencing a similar problem; however, my situation is slightly different. I have HP DL385's connected to a 48 port 10/100 module in a Cisco 6509 (with Sup 2 routing module). I'm using the virtual CDROM via the HP ILO (Integrated Lights Out) interface to perform the initial boot. I mount a customized version of RHEL3U5 CD1. My kickstart file is specified as an url. The system boots and Anaconda starts, but before the link negotiation can complete Anaconda times-out and I am dumped into an interactive install. A couple of data points: 1) The HP DL385 on-board NICs are manufactured by Broadcom. I've read about a known bug with link negotiation between the Broadcom NICs and Cisco 48-port line cards. The link negotiation works, it just takes a "longer" time than with any other kind of NIC. 2) Setting the switch port NIC to enable "spanning-tree portfast" does shorten the link negotiation time, but not enough to get around my problem. Work Arounds: 1) I will prebuild the kickstart configs and place them on my customized RHEL3U5 CD. I've been unsuccessful at reading the kickstart config from the CDROM. The driver supporting the ILO virtual cdrom does not name the cdrom device as "cdrom". Still reading up on this. That's it so far, -Bill From Chip.Shabazian at bankofamerica.com Fri Feb 23 17:34:20 2007 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Fri, 23 Feb 2007 09:34:20 -0800 Subject: "pump" times out -- extend a few seconds more? In-Reply-To: Message-ID: We build HP DL 380/385/580/585 servers every day while connected to Cisco switches that we don't have any access to. Are you using RHEL 3 or 4? If 4, do you have spare NIC's in the box? RHEL 4 enumerates the bus backwards from RHEL 3, so if you have a NIC, it's going to be eth0 instead of the onboard... -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William Sent: Friday, February 23, 2007 8:01 AM To: Discussion list about Kickstart Subject: RE: "pump" times out -- extend a few seconds more? > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of nate > Sent: Friday, January 12, 2007 10:33 AM > To: kickstart-list at redhat.com > Subject: Re: "pump" times out -- extend a few seconds more? > > Bill-Schoolcraft said: > > Hello Family, > > > > I have a series of switches, in racks, and one of them seems to be a bit > > slow so when kickstart goes for it's dhcp request, pump times out. The > > initial PXE takes much more time to get an "ack" at the initial contact > > than the same boxes in other racks with other switches but then finally > > starts chugging away -- it's the next dhcp request that fails, times out. > > > > Today I took the same box, draped a cat6 cable to another, less > > populated switch, and it kicked fine. > > > > (question) > > So my question is, how or where can I place an argument to tell > > kickstart's "pump" request to try longer before timing out. > > just curious, what type of switch? many older cisco switches have STP > on by default, if you set the port(s) to 'portfast' the links come up > immediately rather than waiting ~45 seconds. > > I'm not a cisco guy but if your switches are cisco the command is > something like > > int (interface) > spanning-tree portfast > ^Z > wr mem > > > e.g. on my old 3500s (IOS 12.0 is the latest they'll run) int fa0/5 > spanning-tree portfast > > I haven't encountered any other switch vendors that have this > 'feature' enabled by default. You can tell if it's on without even > logging in(to the switch) by checking to see if the link light for the > port spends a long time in the orange color before turning green. I've > been told more modern cisco switches ship with this feature off by > default. > > as for extending the pump times, I'm not sure, so I can't answer your > specific question :) maybe the above will help a bit.. > > I did a quick search on pump timeout, and what I did see was that it > may be possible to extend the timeout by adding a > /etc/pump_device.conf and using the option 'timeout '. > > though you'll need to add that to the initrd image..not the easiest > thing in the world. > > nate > I am experiencing a similar problem; however, my situation is slightly different. I have HP DL385's connected to a 48 port 10/100 module in a Cisco 6509 (with Sup 2 routing module). I'm using the virtual CDROM via the HP ILO (Integrated Lights Out) interface to perform the initial boot. I mount a customized version of RHEL3U5 CD1. My kickstart file is specified as an url. The system boots and Anaconda starts, but before the link negotiation can complete Anaconda times-out and I am dumped into an interactive install. A couple of data points: 1) The HP DL385 on-board NICs are manufactured by Broadcom. I've read about a known bug with link negotiation between the Broadcom NICs and Cisco 48-port line cards. The link negotiation works, it just takes a "longer" time than with any other kind of NIC. 2) Setting the switch port NIC to enable "spanning-tree portfast" does shorten the link negotiation time, but not enough to get around my problem. Work Arounds: 1) I will prebuild the kickstart configs and place them on my customized RHEL3U5 CD. I've been unsuccessful at reading the kickstart config from the CDROM. The driver supporting the ILO virtual cdrom does not name the cdrom device as "cdrom". Still reading up on this. That's it so far, -Bill _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From William.Ramthun at Level3.com Fri Feb 23 17:49:47 2007 From: William.Ramthun at Level3.com (Ramthun, William) Date: Fri, 23 Feb 2007 10:49:47 -0700 Subject: "pump" times out -- extend a few seconds more? In-Reply-To: Message-ID: In our normal data centers, we build HP DL385/585 daily as well. The data center with issues is from an acquired company with older Cisco switches (5 years +). Installing RHEL 3 Update 5 The install client accesses the RHEL 3 media server via eth0. When Anaconda times-out, I can run thru the install setup manually and RHEL3U5 will install over the network. The problem is the Kickstart config is highly configured and doing a manual install is not really an option that I want to consider. What I want is to access the ks.cfg via the network. Which implies that the network needs to "come up" before Anaconda times out. OR Read the ks.cfg from the cdrom. But the cdrom is virually mounted over the ILO port and is not accessible like ks=cdrom:/ks.cfg from the boot prompt. I don't know what the device is. -Bill > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of Shabazian, Chip > Sent: Friday, February 23, 2007 10:34 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > We build HP DL 380/385/580/585 servers every day while connected to > Cisco switches that we don't have any access to. Are you using RHEL 3 > or 4? If 4, do you have spare NIC's in the box? RHEL 4 enumerates the > bus backwards from RHEL 3, so if you have a NIC, it's going to be eth0 > instead of the onboard... > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William > Sent: Friday, February 23, 2007 8:01 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > bounces at redhat.com] On Behalf Of nate > > Sent: Friday, January 12, 2007 10:33 AM > > To: kickstart-list at redhat.com > > Subject: Re: "pump" times out -- extend a few seconds more? > > > > Bill-Schoolcraft said: > > > Hello Family, > > > > > > I have a series of switches, in racks, and one of them seems to be a > bit > > > slow so when kickstart goes for it's dhcp request, pump times out. > The > > > initial PXE takes much more time to get an "ack" at the initial > contact > > > than the same boxes in other racks with other switches but then > finally > > > starts chugging away -- it's the next dhcp request that fails, times > out. > > > > > > Today I took the same box, draped a cat6 cable to another, less > > > populated switch, and it kicked fine. > > > > > > (question) > > > So my question is, how or where can I place an argument to tell > > > kickstart's "pump" request to try longer before timing out. > > > > just curious, what type of switch? many older cisco switches have STP > > on by default, if you set the port(s) to 'portfast' the links come up > > immediately rather than waiting ~45 seconds. > > > > I'm not a cisco guy but if your switches are cisco the command is > > something like > > > > int (interface) > > spanning-tree portfast > > ^Z > > wr mem > > > > > > e.g. on my old 3500s (IOS 12.0 is the latest they'll run) int fa0/5 > > spanning-tree portfast > > > > I haven't encountered any other switch vendors that have this > > 'feature' enabled by default. You can tell if it's on without even > > logging in(to the switch) by checking to see if the link light for the > > > port spends a long time in the orange color before turning green. I've > > > been told more modern cisco switches ship with this feature off by > > default. > > > > as for extending the pump times, I'm not sure, so I can't answer your > > specific question :) maybe the above will help a bit.. > > > > I did a quick search on pump timeout, and what I did see was that it > > may be possible to extend the timeout by adding a > > /etc/pump_device.conf and using the option 'timeout '. > > > > though you'll need to add that to the initrd image..not the easiest > > thing in the world. > > > > nate > > > > > I am experiencing a similar problem; however, my situation is slightly > different. > > I have HP DL385's connected to a 48 port 10/100 module in a Cisco 6509 > (with Sup 2 routing module). I'm using the virtual CDROM via the HP ILO > (Integrated Lights Out) interface to perform the initial boot. I mount > a customized version of RHEL3U5 CD1. My kickstart file is specified as > an url. The system boots and Anaconda starts, but before the link > negotiation can complete Anaconda times-out and I am dumped into an > interactive install. > > A couple of data points: > > 1) The HP DL385 on-board NICs are manufactured by Broadcom. I've read > about a known bug with link negotiation between the Broadcom NICs and > Cisco 48-port line cards. The link negotiation works, it just takes a > "longer" time than with any other kind of NIC. > > 2) Setting the switch port NIC to enable "spanning-tree portfast" does > shorten the link negotiation time, but not enough to get around my > problem. > > Work Arounds: > > 1) I will prebuild the kickstart configs and place them on my customized > RHEL3U5 CD. I've been unsuccessful at reading the kickstart config from > the CDROM. The driver supporting the ILO virtual cdrom does not name > the cdrom device as "cdrom". Still reading up on this. > > That's it so far, > -Bill > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From Chip.Shabazian at bankofamerica.com Fri Feb 23 17:56:09 2007 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Fri, 23 Feb 2007 09:56:09 -0800 Subject: "pump" times out -- extend a few seconds more? In-Reply-To: Message-ID: I think there was a "nicdelay" or something similar introduced around RHEL3U7, so you might want to try that. Also, make sure that portfast is enabled on the switches. The other thing we do is set the speed/duplex via ethtool to avoid autonegotion delays: eth0_ethtool="autoneg=on speed=1000 duplex=full" When we have to build via iLO, we embed the kickstart file in the initrd and then use ks=file: -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William Sent: Friday, February 23, 2007 9:50 AM To: Discussion list about Kickstart Subject: RE: "pump" times out -- extend a few seconds more? In our normal data centers, we build HP DL385/585 daily as well. The data center with issues is from an acquired company with older Cisco switches (5 years +). Installing RHEL 3 Update 5 The install client accesses the RHEL 3 media server via eth0. When Anaconda times-out, I can run thru the install setup manually and RHEL3U5 will install over the network. The problem is the Kickstart config is highly configured and doing a manual install is not really an option that I want to consider. What I want is to access the ks.cfg via the network. Which implies that the network needs to "come up" before Anaconda times out. OR Read the ks.cfg from the cdrom. But the cdrom is virually mounted over the ILO port and is not accessible like ks=cdrom:/ks.cfg from the boot prompt. I don't know what the device is. -Bill > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of Shabazian, Chip > Sent: Friday, February 23, 2007 10:34 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > We build HP DL 380/385/580/585 servers every day while connected to > Cisco switches that we don't have any access to. Are you using RHEL 3 > or 4? If 4, do you have spare NIC's in the box? RHEL 4 enumerates the > bus backwards from RHEL 3, so if you have a NIC, it's going to be eth0 > instead of the onboard... > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William > Sent: Friday, February 23, 2007 8:01 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > bounces at redhat.com] On Behalf Of nate > > Sent: Friday, January 12, 2007 10:33 AM > > To: kickstart-list at redhat.com > > Subject: Re: "pump" times out -- extend a few seconds more? > > > > Bill-Schoolcraft said: > > > Hello Family, > > > > > > I have a series of switches, in racks, and one of them seems to be a > bit > > > slow so when kickstart goes for it's dhcp request, pump times out. > The > > > initial PXE takes much more time to get an "ack" at the initial > contact > > > than the same boxes in other racks with other switches but then > finally > > > starts chugging away -- it's the next dhcp request that fails, times > out. > > > > > > Today I took the same box, draped a cat6 cable to another, less > > > populated switch, and it kicked fine. > > > > > > (question) > > > So my question is, how or where can I place an argument to tell > > > kickstart's "pump" request to try longer before timing out. > > > > just curious, what type of switch? many older cisco switches have STP > > on by default, if you set the port(s) to 'portfast' the links come up > > immediately rather than waiting ~45 seconds. > > > > I'm not a cisco guy but if your switches are cisco the command is > > something like > > > > int (interface) > > spanning-tree portfast > > ^Z > > wr mem > > > > > > e.g. on my old 3500s (IOS 12.0 is the latest they'll run) int fa0/5 > > spanning-tree portfast > > > > I haven't encountered any other switch vendors that have this > > 'feature' enabled by default. You can tell if it's on without even > > logging in(to the switch) by checking to see if the link light for the > > > port spends a long time in the orange color before turning green. I've > > > been told more modern cisco switches ship with this feature off by > > default. > > > > as for extending the pump times, I'm not sure, so I can't answer your > > specific question :) maybe the above will help a bit.. > > > > I did a quick search on pump timeout, and what I did see was that it > > may be possible to extend the timeout by adding a > > /etc/pump_device.conf and using the option 'timeout '. > > > > though you'll need to add that to the initrd image..not the easiest > > thing in the world. > > > > nate > > > > > I am experiencing a similar problem; however, my situation is slightly > different. > > I have HP DL385's connected to a 48 port 10/100 module in a Cisco 6509 > (with Sup 2 routing module). I'm using the virtual CDROM via the HP ILO > (Integrated Lights Out) interface to perform the initial boot. I mount > a customized version of RHEL3U5 CD1. My kickstart file is specified as > an url. The system boots and Anaconda starts, but before the link > negotiation can complete Anaconda times-out and I am dumped into an > interactive install. > > A couple of data points: > > 1) The HP DL385 on-board NICs are manufactured by Broadcom. I've read > about a known bug with link negotiation between the Broadcom NICs and > Cisco 48-port line cards. The link negotiation works, it just takes a > "longer" time than with any other kind of NIC. > > 2) Setting the switch port NIC to enable "spanning-tree portfast" does > shorten the link negotiation time, but not enough to get around my > problem. > > Work Arounds: > > 1) I will prebuild the kickstart configs and place them on my customized > RHEL3U5 CD. I've been unsuccessful at reading the kickstart config from > the CDROM. The driver supporting the ILO virtual cdrom does not name > the cdrom device as "cdrom". Still reading up on this. > > That's it so far, > -Bill > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From Greg.Caetano at hp.com Fri Feb 23 18:56:02 2007 From: Greg.Caetano at hp.com (Caetano, Greg) Date: Fri, 23 Feb 2007 13:56:02 -0500 Subject: "pump" times out -- extend a few seconds more? In-Reply-To: References: Message-ID: Bill: An iso mounted via the Virtual Media appears to the OS as a USB Mass Storage device. /var/log/messages records: Kernel: usb 2-1: new full speed USB device using address 2 Kernel: Initializing USB Mass Storage driver Kernel: scsi2: SCSI emulation for USB Mass Storage devices kernel: Vendor: HP Model: Virtual CD-ROM Kernel: Type: CD-ROM Kernel: usbcore: registered new driver usb-storage Kernel: USB Mass Storage support registered Scsi.agent: cdrom at /devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1:1.0/host2/target2:0:0/2:0: 0:0 Kernel: sr0: scsi2-mmc drive 12x/12x cd/rw tray Kernel: Uniform CD-ROM driver revision: 3.20 Fstab-sync: added mount point /media/cdrom for /dev/scd0 regards Greg Caetano HP TSG Linux Solutions Alliances Engineering Chicago, IL greg.caetano at hp.com -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William Sent: Friday, February 23, 2007 11:50 AM To: Discussion list about Kickstart Subject: RE: "pump" times out -- extend a few seconds more? In our normal data centers, we build HP DL385/585 daily as well. The data center with issues is from an acquired company with older Cisco switches (5 years +). Installing RHEL 3 Update 5 The install client accesses the RHEL 3 media server via eth0. When Anaconda times-out, I can run thru the install setup manually and RHEL3U5 will install over the network. The problem is the Kickstart config is highly configured and doing a manual install is not really an option that I want to consider. What I want is to access the ks.cfg via the network. Which implies that the network needs to "come up" before Anaconda times out. OR Read the ks.cfg from the cdrom. But the cdrom is virually mounted over the ILO port and is not accessible like ks=cdrom:/ks.cfg from the boot prompt. I don't know what the device is. -Bill > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of Shabazian, Chip > Sent: Friday, February 23, 2007 10:34 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > We build HP DL 380/385/580/585 servers every day while connected to > Cisco switches that we don't have any access to. Are you using RHEL 3 > or 4? If 4, do you have spare NIC's in the box? RHEL 4 enumerates the > bus backwards from RHEL 3, so if you have a NIC, it's going to be eth0 > instead of the onboard... > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William > Sent: Friday, February 23, 2007 8:01 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > bounces at redhat.com] On Behalf Of nate > > Sent: Friday, January 12, 2007 10:33 AM > > To: kickstart-list at redhat.com > > Subject: Re: "pump" times out -- extend a few seconds more? > > > > Bill-Schoolcraft said: > > > Hello Family, > > > > > > I have a series of switches, in racks, and one of them seems to be a > bit > > > slow so when kickstart goes for it's dhcp request, pump times out. > The > > > initial PXE takes much more time to get an "ack" at the initial > contact > > > than the same boxes in other racks with other switches but then > finally > > > starts chugging away -- it's the next dhcp request that fails, times > out. > > > > > > Today I took the same box, draped a cat6 cable to another, less > > > populated switch, and it kicked fine. > > > > > > (question) > > > So my question is, how or where can I place an argument to tell > > > kickstart's "pump" request to try longer before timing out. > > > > just curious, what type of switch? many older cisco switches have STP > > on by default, if you set the port(s) to 'portfast' the links come up > > immediately rather than waiting ~45 seconds. > > > > I'm not a cisco guy but if your switches are cisco the command is > > something like > > > > int (interface) > > spanning-tree portfast > > ^Z > > wr mem > > > > > > e.g. on my old 3500s (IOS 12.0 is the latest they'll run) int fa0/5 > > spanning-tree portfast > > > > I haven't encountered any other switch vendors that have this > > 'feature' enabled by default. You can tell if it's on without even > > logging in(to the switch) by checking to see if the link light for the > > > port spends a long time in the orange color before turning green. I've > > > been told more modern cisco switches ship with this feature off by > > default. > > > > as for extending the pump times, I'm not sure, so I can't answer your > > specific question :) maybe the above will help a bit.. > > > > I did a quick search on pump timeout, and what I did see was that it > > may be possible to extend the timeout by adding a > > /etc/pump_device.conf and using the option 'timeout '. > > > > though you'll need to add that to the initrd image..not the easiest > > thing in the world. > > > > nate > > > > > I am experiencing a similar problem; however, my situation is slightly > different. > > I have HP DL385's connected to a 48 port 10/100 module in a Cisco 6509 > (with Sup 2 routing module). I'm using the virtual CDROM via the HP ILO > (Integrated Lights Out) interface to perform the initial boot. I mount > a customized version of RHEL3U5 CD1. My kickstart file is specified as > an url. The system boots and Anaconda starts, but before the link > negotiation can complete Anaconda times-out and I am dumped into an > interactive install. > > A couple of data points: > > 1) The HP DL385 on-board NICs are manufactured by Broadcom. I've read > about a known bug with link negotiation between the Broadcom NICs and > Cisco 48-port line cards. The link negotiation works, it just takes a > "longer" time than with any other kind of NIC. > > 2) Setting the switch port NIC to enable "spanning-tree portfast" does > shorten the link negotiation time, but not enough to get around my > problem. > > Work Arounds: > > 1) I will prebuild the kickstart configs and place them on my customized > RHEL3U5 CD. I've been unsuccessful at reading the kickstart config from > the CDROM. The driver supporting the ILO virtual cdrom does not name > the cdrom device as "cdrom". Still reading up on this. > > That's it so far, > -Bill > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From William.Ramthun at Level3.com Fri Feb 23 20:25:51 2007 From: William.Ramthun at Level3.com (Ramthun, William) Date: Fri, 23 Feb 2007 13:25:51 -0700 Subject: "pump" times out -- extend a few seconds more? In-Reply-To: Message-ID: I did a cursory search of nicedelay and I don't see how it is applicable for my situation... Maybe I just don't know how to use it. Portfast is enabled. I tried the eth0_ethtool option, but that didn't appear to help. I thought about placing the ks.cfg files in the initrd, but didn't really want to go there. Thanks > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of Shabazian, Chip > Sent: Friday, February 23, 2007 10:56 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > I think there was a "nicdelay" or something similar introduced around > RHEL3U7, so you might want to try that. Also, make sure that portfast > is enabled on the switches. The other thing we do is set the > speed/duplex via ethtool to avoid autonegotion delays: > > eth0_ethtool="autoneg=on speed=1000 duplex=full" > > When we have to build via iLO, we embed the kickstart file in the initrd > and then use ks=file: > > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William > Sent: Friday, February 23, 2007 9:50 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > In our normal data centers, we build HP DL385/585 daily as well. The > data center with issues is from an acquired company with older Cisco > switches (5 years +). > > Installing RHEL 3 Update 5 > > The install client accesses the RHEL 3 media server via eth0. When > Anaconda times-out, I can run thru the install setup manually and > RHEL3U5 will install over the network. The problem is the Kickstart > config is highly configured and doing a manual install is not really an > option that I want to consider. > > What I want is to access the ks.cfg via the network. Which implies that > the network needs to "come up" before Anaconda times out. > > OR > > Read the ks.cfg from the cdrom. But the cdrom is virually mounted over > the ILO port and is not accessible like ks=cdrom:/ks.cfg from the boot > prompt. I don't know what the device is. > -Bill > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > bounces at redhat.com] On Behalf Of Shabazian, Chip > > Sent: Friday, February 23, 2007 10:34 AM > > To: Discussion list about Kickstart > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > We build HP DL 380/385/580/585 servers every day while connected to > > Cisco switches that we don't have any access to. Are you using RHEL 3 > > > or 4? If 4, do you have spare NIC's in the box? RHEL 4 enumerates the > > > bus backwards from RHEL 3, so if you have a NIC, it's going to be eth0 > > > instead of the onboard... > > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com > > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, > William > > Sent: Friday, February 23, 2007 8:01 AM > > To: Discussion list about Kickstart > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > > -----Original Message----- > > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > > bounces at redhat.com] On Behalf Of nate > > > Sent: Friday, January 12, 2007 10:33 AM > > > To: kickstart-list at redhat.com > > > Subject: Re: "pump" times out -- extend a few seconds more? > > > > > > Bill-Schoolcraft said: > > > > Hello Family, > > > > > > > > I have a series of switches, in racks, and one of them seems to be > a > > bit > > > > slow so when kickstart goes for it's dhcp request, pump times out. > > The > > > > initial PXE takes much more time to get an "ack" at the initial > > contact > > > > than the same boxes in other racks with other switches but then > > finally > > > > starts chugging away -- it's the next dhcp request that fails, > times > > out. > > > > > > > > Today I took the same box, draped a cat6 cable to another, less > > > > populated switch, and it kicked fine. > > > > > > > > (question) > > > > So my question is, how or where can I place an argument to tell > > > > kickstart's "pump" request to try longer before timing out. > > > > > > just curious, what type of switch? many older cisco switches have > STP > > > on by default, if you set the port(s) to 'portfast' the links come > up > > > immediately rather than waiting ~45 seconds. > > > > > > I'm not a cisco guy but if your switches are cisco the command is > > > something like > > > > > > int (interface) > > > spanning-tree portfast > > > ^Z > > > wr mem > > > > > > > > > e.g. on my old 3500s (IOS 12.0 is the latest they'll run) int fa0/5 > > > spanning-tree portfast > > > > > > I haven't encountered any other switch vendors that have this > > > 'feature' enabled by default. You can tell if it's on without even > > > logging in(to the switch) by checking to see if the link light for > the > > > > > port spends a long time in the orange color before turning green. > I've > > > > > been told more modern cisco switches ship with this feature off by > > > default. > > > > > > as for extending the pump times, I'm not sure, so I can't answer > your > > > specific question :) maybe the above will help a bit.. > > > > > > I did a quick search on pump timeout, and what I did see was that it > > > > may be possible to extend the timeout by adding a > > > /etc/pump_device.conf and using the option 'timeout '. > > > > > > though you'll need to add that to the initrd image..not the easiest > > > thing in the world. > > > > > > nate > > > > > > > > > I am experiencing a similar problem; however, my situation is slightly > > > different. > > > > I have HP DL385's connected to a 48 port 10/100 module in a Cisco 6509 > > > (with Sup 2 routing module). I'm using the virtual CDROM via the HP > ILO > > (Integrated Lights Out) interface to perform the initial boot. I > mount > > a customized version of RHEL3U5 CD1. My kickstart file is specified > as > > an url. The system boots and Anaconda starts, but before the link > > negotiation can complete Anaconda times-out and I am dumped into an > > interactive install. > > > > A couple of data points: > > > > 1) The HP DL385 on-board NICs are manufactured by Broadcom. I've read > > > about a known bug with link negotiation between the Broadcom NICs and > > Cisco 48-port line cards. The link negotiation works, it just takes a > > > "longer" time than with any other kind of NIC. > > > > 2) Setting the switch port NIC to enable "spanning-tree portfast" does > > > shorten the link negotiation time, but not enough to get around my > > problem. > > > > Work Arounds: > > > > 1) I will prebuild the kickstart configs and place them on my > customized > > RHEL3U5 CD. I've been unsuccessful at reading the kickstart config > from > > the CDROM. The driver supporting the ILO virtual cdrom does not name > > the cdrom device as "cdrom". Still reading up on this. > > > > That's it so far, > > -Bill > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From Chip.Shabazian at bankofamerica.com Fri Feb 23 20:37:41 2007 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Fri, 23 Feb 2007 12:37:41 -0800 Subject: "pump" times out -- extend a few seconds more? In-Reply-To: Message-ID: It's applicable because you said your problem was: > the network needs to "come up" before Anaconda times out. That's exactly what nicdelay is supposed to help you with. The problem is, it takes ~ 30 seconds for the nic to negotiate with the switch. By delaying bringing the nic up until enough time has passed that the negotiation has taken place, you should then be able to configure the nic, get your kickstart file, and be off and running. Now, I say "supposed" to, because I've never had to use it, and heard mixed results from those who have. Remember, the nic gets completely cycled and needs to renegotiate 3 times during a build: at power on when getting the kickstart file when starting the build Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William Sent: Friday, February 23, 2007 12:26 PM To: Discussion list about Kickstart Subject: RE: "pump" times out -- extend a few seconds more? I did a cursory search of nicedelay and I don't see how it is applicable for my situation... Maybe I just don't know how to use it. Portfast is enabled. I tried the eth0_ethtool option, but that didn't appear to help. I thought about placing the ks.cfg files in the initrd, but didn't really want to go there. Thanks > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of Shabazian, Chip > Sent: Friday, February 23, 2007 10:56 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > I think there was a "nicdelay" or something similar introduced around > RHEL3U7, so you might want to try that. Also, make sure that portfast > is enabled on the switches. The other thing we do is set the > speed/duplex via ethtool to avoid autonegotion delays: > > eth0_ethtool="autoneg=on speed=1000 duplex=full" > > When we have to build via iLO, we embed the kickstart file in the initrd > and then use ks=file: > > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William > Sent: Friday, February 23, 2007 9:50 AM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > In our normal data centers, we build HP DL385/585 daily as well. The > data center with issues is from an acquired company with older Cisco > switches (5 years +). > > Installing RHEL 3 Update 5 > > The install client accesses the RHEL 3 media server via eth0. When > Anaconda times-out, I can run thru the install setup manually and > RHEL3U5 will install over the network. The problem is the Kickstart > config is highly configured and doing a manual install is not really an > option that I want to consider. > > What I want is to access the ks.cfg via the network. Which implies that > the network needs to "come up" before Anaconda times out. > > OR > > Read the ks.cfg from the cdrom. But the cdrom is virually mounted over > the ILO port and is not accessible like ks=cdrom:/ks.cfg from the boot > prompt. I don't know what the device is. > -Bill > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > bounces at redhat.com] On Behalf Of Shabazian, Chip > > Sent: Friday, February 23, 2007 10:34 AM > > To: Discussion list about Kickstart > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > We build HP DL 380/385/580/585 servers every day while connected to > > Cisco switches that we don't have any access to. Are you using RHEL 3 > > > or 4? If 4, do you have spare NIC's in the box? RHEL 4 enumerates the > > > bus backwards from RHEL 3, so if you have a NIC, it's going to be eth0 > > > instead of the onboard... > > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com > > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, > William > > Sent: Friday, February 23, 2007 8:01 AM > > To: Discussion list about Kickstart > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > > -----Original Message----- > > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > > bounces at redhat.com] On Behalf Of nate > > > Sent: Friday, January 12, 2007 10:33 AM > > > To: kickstart-list at redhat.com > > > Subject: Re: "pump" times out -- extend a few seconds more? > > > > > > Bill-Schoolcraft said: > > > > Hello Family, > > > > > > > > I have a series of switches, in racks, and one of them seems to be > a > > bit > > > > slow so when kickstart goes for it's dhcp request, pump times out. > > The > > > > initial PXE takes much more time to get an "ack" at the initial > > contact > > > > than the same boxes in other racks with other switches but then > > finally > > > > starts chugging away -- it's the next dhcp request that fails, > times > > out. > > > > > > > > Today I took the same box, draped a cat6 cable to another, less > > > > populated switch, and it kicked fine. > > > > > > > > (question) > > > > So my question is, how or where can I place an argument to tell > > > > kickstart's "pump" request to try longer before timing out. > > > > > > just curious, what type of switch? many older cisco switches have > STP > > > on by default, if you set the port(s) to 'portfast' the links come > up > > > immediately rather than waiting ~45 seconds. > > > > > > I'm not a cisco guy but if your switches are cisco the command is > > > something like > > > > > > int (interface) > > > spanning-tree portfast > > > ^Z > > > wr mem > > > > > > > > > e.g. on my old 3500s (IOS 12.0 is the latest they'll run) int fa0/5 > > > spanning-tree portfast > > > > > > I haven't encountered any other switch vendors that have this > > > 'feature' enabled by default. You can tell if it's on without even > > > logging in(to the switch) by checking to see if the link light for > the > > > > > port spends a long time in the orange color before turning green. > I've > > > > > been told more modern cisco switches ship with this feature off by > > > default. > > > > > > as for extending the pump times, I'm not sure, so I can't answer > your > > > specific question :) maybe the above will help a bit.. > > > > > > I did a quick search on pump timeout, and what I did see was that it > > > > may be possible to extend the timeout by adding a > > > /etc/pump_device.conf and using the option 'timeout '. > > > > > > though you'll need to add that to the initrd image..not the easiest > > > thing in the world. > > > > > > nate > > > > > > > > > I am experiencing a similar problem; however, my situation is slightly > > > different. > > > > I have HP DL385's connected to a 48 port 10/100 module in a Cisco 6509 > > > (with Sup 2 routing module). I'm using the virtual CDROM via the HP > ILO > > (Integrated Lights Out) interface to perform the initial boot. I > mount > > a customized version of RHEL3U5 CD1. My kickstart file is specified > as > > an url. The system boots and Anaconda starts, but before the link > > negotiation can complete Anaconda times-out and I am dumped into an > > interactive install. > > > > A couple of data points: > > > > 1) The HP DL385 on-board NICs are manufactured by Broadcom. I've read > > > about a known bug with link negotiation between the Broadcom NICs and > > Cisco 48-port line cards. The link negotiation works, it just takes a > > > "longer" time than with any other kind of NIC. > > > > 2) Setting the switch port NIC to enable "spanning-tree portfast" does > > > shorten the link negotiation time, but not enough to get around my > > problem. > > > > Work Arounds: > > > > 1) I will prebuild the kickstart configs and place them on my > customized > > RHEL3U5 CD. I've been unsuccessful at reading the kickstart config > from > > the CDROM. The driver supporting the ILO virtual cdrom does not name > > the cdrom device as "cdrom". Still reading up on this. > > > > That's it so far, > > -Bill > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From debian at herakles.homelinux.org Fri Feb 23 21:12:30 2007 From: debian at herakles.homelinux.org (John Summerfield) Date: Sat, 24 Feb 2007 06:12:30 +0900 Subject: kickstart installation hardware support In-Reply-To: <2F4F96E90349F241809293630A893B3D0245AA70@MAIL03.rvponp.fgov.be> References: <7b5a52eb0702220749o4d014bc7oa8124ed5887f4ebe@mail.gmail.com> <2F4F96E90349F241809293630A893B3D0245AA70@MAIL03.rvponp.fgov.be> Message-ID: <45DF58BE.2030108@herakles.homelinux.org> KENNEDY VAN DAM Eric wrote: >> De : kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] De la part de Steven Dalton >> Envoy? : jeudi 22 f?vrier 2007 16:50 >> ? : kickstart-list at redhat.com >> Objet : kickstart installation hardware support >> >> >> I am having some problems installing Fedora core 5 on my system. When I use the FC5 boot disk with the kickstart directive >> >> linux ks=http://somedomain/ks.cfg >> >> The installation ignores the ks configuration completely and continues on to the graphical installation. I do not know if it > is relevent but I do not >> have dhcp on my network where the machine so I need to setup the ip address statically. > > > Hi. > > As you don't have any DHCP server, you can try to configure ip on lilo/grub. Have a look to the ip kernel parameter: > ip=:::::: I don't think the kernel's going to interpret this (when last I read up on this, it required the NIC driver to be built into the kernel, and for kernel support to be explicitly configured in). However, it is possible for the linuxrc script (and its successors) to interpret it. Certainly, IP= works in Ubuntu, but I've not explored why, I just assumed it's done in the scripts & didn't need to know one way or the other. > > In your case, I should try something like this: > > linux ks=http:///ks.cfg ip=:::::: _That_ would _never_ work on _my_ server. My server like names, not numbers - they are delicate souls:-) -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu Please do not reply off-list From Chip.Shabazian at bankofamerica.com Fri Feb 23 21:22:19 2007 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Fri, 23 Feb 2007 13:22:19 -0800 Subject: kickstart installation hardware support In-Reply-To: <45DF58BE.2030108@herakles.homelinux.org> Message-ID: You can use: boot: ks=http://xxxxxx ip=xxxxx netmask=xxxxx gateway=xxxxx dns=xxxxx One caveat, do not put TOO much on the boot: line, there is a 255 character limitation. Since you are supplying the dns server, you can use name resolution. Once the kickstart file is retrieved, the NIC will be cycled again, and it will use whatever is inside the kickstart file itself. Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of John Summerfield Sent: Friday, February 23, 2007 1:13 PM To: Discussion list about Kickstart Subject: Re: kickstart installation hardware support KENNEDY VAN DAM Eric wrote: >> De : kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] De la part de Steven Dalton >> Envoy? : jeudi 22 f?vrier 2007 16:50 >> ? : kickstart-list at redhat.com >> Objet : kickstart installation hardware support >> >> >> I am having some problems installing Fedora core 5 on my system. When >>I use the FC5 boot disk with the kickstart directive >> >> linux ks=http://somedomain/ks.cfg >> >> The installation ignores the ks configuration completely and continues on to the graphical installation. I do not know if it > is relevent but I do not >> have dhcp on my network where the machine so I need to setup the ip address statically. > > > Hi. > > As you don't have any DHCP server, you can try to configure ip on lilo/grub. Have a look to the ip kernel parameter: > ip=:: ip>:::: I don't think the kernel's going to interpret this (when last I read up on this, it required the NIC driver to be built into the kernel, and for kernel support to be explicitly configured in). However, it is possible for the linuxrc script (and its successors) to interpret it. Certainly, IP= works in Ubuntu, but I've not explored why, I just assumed it's done in the scripts & didn't need to know one way or the other. > > In your case, I should try something like this: > > linux ks=http:///ks.cfg ip=:::::: _That_ would _never_ work on _my_ server. My server like names, not numbers - they are delicate souls:-) -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu Please do not reply off-list _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From William.Ramthun at Level3.com Tue Feb 27 15:33:49 2007 From: William.Ramthun at Level3.com (Ramthun, William) Date: Tue, 27 Feb 2007 08:33:49 -0700 Subject: "pump" times out -- extend a few seconds more? In-Reply-To: Message-ID: Chip - thanks for the help. What I meant to say is that I am not sure how to use "nicedelay". Is this a boot: line option? I couldn't find any reference on how to use it. -Bill > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of Shabazian, Chip > Sent: Friday, February 23, 2007 1:38 PM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > It's applicable because you said your problem was: > > > the network needs to "come up" before Anaconda times out. > > That's exactly what nicdelay is supposed to help you with. The problem > is, it takes ~ 30 seconds for the nic to negotiate with the switch. By > delaying bringing the nic up until enough time has passed that the > negotiation has taken place, you should then be able to configure the > nic, get your kickstart file, and be off and running. > > Now, I say "supposed" to, because I've never had to use it, and heard > mixed results from those who have. > > Remember, the nic gets completely cycled and needs to renegotiate 3 > times during a build: > > at power on > when getting the kickstart file > when starting the build > > Chip > > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William > Sent: Friday, February 23, 2007 12:26 PM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > I did a cursory search of nicedelay and I don't see how it is applicable > for my situation... Maybe I just don't know how to use it. > > Portfast is enabled. > > I tried the eth0_ethtool option, but that didn't appear to help. > > I thought about placing the ks.cfg files in the initrd, but didn't > really want to go there. > > Thanks > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > bounces at redhat.com] On Behalf Of Shabazian, Chip > > Sent: Friday, February 23, 2007 10:56 AM > > To: Discussion list about Kickstart > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > I think there was a "nicdelay" or something similar introduced around > > RHEL3U7, so you might want to try that. Also, make sure that portfast > > > is enabled on the switches. The other thing we do is set the > > speed/duplex via ethtool to avoid autonegotion delays: > > > > eth0_ethtool="autoneg=on speed=1000 duplex=full" > > > > When we have to build via iLO, we embed the kickstart file in the > initrd > > and then use ks=file: > > > > > > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com > > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, > William > > Sent: Friday, February 23, 2007 9:50 AM > > To: Discussion list about Kickstart > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > In our normal data centers, we build HP DL385/585 daily as well. The > > data center with issues is from an acquired company with older Cisco > > switches (5 years +). > > > > Installing RHEL 3 Update 5 > > > > The install client accesses the RHEL 3 media server via eth0. When > > Anaconda times-out, I can run thru the install setup manually and > > RHEL3U5 will install over the network. The problem is the Kickstart > > config is highly configured and doing a manual install is not really > an > > option that I want to consider. > > > > What I want is to access the ks.cfg via the network. Which implies > that > > the network needs to "come up" before Anaconda times out. > > > > OR > > > > Read the ks.cfg from the cdrom. But the cdrom is virually mounted > over > > the ILO port and is not accessible like ks=cdrom:/ks.cfg from the boot > > > prompt. I don't know what the device is. > > -Bill > > > > > -----Original Message----- > > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > > bounces at redhat.com] On Behalf Of Shabazian, Chip > > > Sent: Friday, February 23, 2007 10:34 AM > > > To: Discussion list about Kickstart > > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > > > We build HP DL 380/385/580/585 servers every day while connected to > > > Cisco switches that we don't have any access to. Are you using RHEL > 3 > > > > > or 4? If 4, do you have spare NIC's in the box? RHEL 4 enumerates > the > > > > > bus backwards from RHEL 3, so if you have a NIC, it's going to be > eth0 > > > > > instead of the onboard... > > > > > > -----Original Message----- > > > From: kickstart-list-bounces at redhat.com > > > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, > > William > > > Sent: Friday, February 23, 2007 8:01 AM > > > To: Discussion list about Kickstart > > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > > > > -----Original Message----- > > > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > > > bounces at redhat.com] On Behalf Of nate > > > > Sent: Friday, January 12, 2007 10:33 AM > > > > To: kickstart-list at redhat.com > > > > Subject: Re: "pump" times out -- extend a few seconds more? > > > > > > > > Bill-Schoolcraft said: > > > > > Hello Family, > > > > > > > > > > I have a series of switches, in racks, and one of them seems to > be > > a > > > bit > > > > > slow so when kickstart goes for it's dhcp request, pump times > out. > > > The > > > > > initial PXE takes much more time to get an "ack" at the initial > > > contact > > > > > than the same boxes in other racks with other switches but then > > > finally > > > > > starts chugging away -- it's the next dhcp request that fails, > > times > > > out. > > > > > > > > > > Today I took the same box, draped a cat6 cable to another, less > > > > > populated switch, and it kicked fine. > > > > > > > > > > (question) > > > > > So my question is, how or where can I place an argument to tell > > > > > kickstart's "pump" request to try longer before timing out. > > > > > > > > just curious, what type of switch? many older cisco switches have > > STP > > > > on by default, if you set the port(s) to 'portfast' the links come > > up > > > > immediately rather than waiting ~45 seconds. > > > > > > > > I'm not a cisco guy but if your switches are cisco the command is > > > > something like > > > > > > > > int (interface) > > > > spanning-tree portfast > > > > ^Z > > > > wr mem > > > > > > > > > > > > e.g. on my old 3500s (IOS 12.0 is the latest they'll run) int > fa0/5 > > > > spanning-tree portfast > > > > > > > > I haven't encountered any other switch vendors that have this > > > > 'feature' enabled by default. You can tell if it's on without even > > > > > logging in(to the switch) by checking to see if the link light for > > the > > > > > > > port spends a long time in the orange color before turning green. > > I've > > > > > > > been told more modern cisco switches ship with this feature off by > > > > > default. > > > > > > > > as for extending the pump times, I'm not sure, so I can't answer > > your > > > > specific question :) maybe the above will help a bit.. > > > > > > > > I did a quick search on pump timeout, and what I did see was that > it > > > > > > may be possible to extend the timeout by adding a > > > > /etc/pump_device.conf and using the option 'timeout number>'. > > > > > > > > though you'll need to add that to the initrd image..not the > easiest > > > > thing in the world. > > > > > > > > nate > > > > > > > > > > > > > I am experiencing a similar problem; however, my situation is > slightly > > > > > different. > > > > > > I have HP DL385's connected to a 48 port 10/100 module in a Cisco > 6509 > > > > > (with Sup 2 routing module). I'm using the virtual CDROM via the HP > > ILO > > > (Integrated Lights Out) interface to perform the initial boot. I > > mount > > > a customized version of RHEL3U5 CD1. My kickstart file is specified > > as > > > an url. The system boots and Anaconda starts, but before the link > > > negotiation can complete Anaconda times-out and I am dumped into an > > > interactive install. > > > > > > A couple of data points: > > > > > > 1) The HP DL385 on-board NICs are manufactured by Broadcom. I've > read > > > > > about a known bug with link negotiation between the Broadcom NICs > and > > > Cisco 48-port line cards. The link negotiation works, it just takes > a > > > > > "longer" time than with any other kind of NIC. > > > > > > 2) Setting the switch port NIC to enable "spanning-tree portfast" > does > > > > > shorten the link negotiation time, but not enough to get around my > > > problem. > > > > > > Work Arounds: > > > > > > 1) I will prebuild the kickstart configs and place them on my > > customized > > > RHEL3U5 CD. I've been unsuccessful at reading the kickstart config > > from > > > the CDROM. The driver supporting the ILO virtual cdrom does not > name > > > the cdrom device as "cdrom". Still reading up on this. > > > > > > That's it so far, > > > -Bill > > > > > > _______________________________________________ > > > Kickstart-list mailing list > > > Kickstart-list at redhat.com > > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > > > _______________________________________________ > > > Kickstart-list mailing list > > > Kickstart-list at redhat.com > > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From Chip.Shabazian at bankofamerica.com Tue Feb 27 19:32:20 2007 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Tue, 27 Feb 2007 11:32:20 -0800 Subject: "pump" times out -- extend a few seconds more? In-Reply-To: Message-ID: Like I said, I've never used it, but it is SUPPOSED to exist: https://rhn.redhat.com/errata/RHEA-2006-0443.html nicdelay= If you want to download the srpm and take a look at loader, I think that's where you will find the code. Chip -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William Sent: Tuesday, February 27, 2007 7:34 AM To: Discussion list about Kickstart Subject: RE: "pump" times out -- extend a few seconds more? Chip - thanks for the help. What I meant to say is that I am not sure how to use "nicedelay". Is this a boot: line option? I couldn't find any reference on how to use it. -Bill > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of Shabazian, Chip > Sent: Friday, February 23, 2007 1:38 PM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > It's applicable because you said your problem was: > > > the network needs to "come up" before Anaconda times out. > > That's exactly what nicdelay is supposed to help you with. The problem > is, it takes ~ 30 seconds for the nic to negotiate with the switch. By > delaying bringing the nic up until enough time has passed that the > negotiation has taken place, you should then be able to configure the > nic, get your kickstart file, and be off and running. > > Now, I say "supposed" to, because I've never had to use it, and heard > mixed results from those who have. > > Remember, the nic gets completely cycled and needs to renegotiate 3 > times during a build: > > at power on > when getting the kickstart file > when starting the build > > Chip > > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, William > Sent: Friday, February 23, 2007 12:26 PM > To: Discussion list about Kickstart > Subject: RE: "pump" times out -- extend a few seconds more? > > I did a cursory search of nicedelay and I don't see how it is applicable > for my situation... Maybe I just don't know how to use it. > > Portfast is enabled. > > I tried the eth0_ethtool option, but that didn't appear to help. > > I thought about placing the ks.cfg files in the initrd, but didn't > really want to go there. > > Thanks > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > bounces at redhat.com] On Behalf Of Shabazian, Chip > > Sent: Friday, February 23, 2007 10:56 AM > > To: Discussion list about Kickstart > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > I think there was a "nicdelay" or something similar introduced around > > RHEL3U7, so you might want to try that. Also, make sure that portfast > > > is enabled on the switches. The other thing we do is set the > > speed/duplex via ethtool to avoid autonegotion delays: > > > > eth0_ethtool="autoneg=on speed=1000 duplex=full" > > > > When we have to build via iLO, we embed the kickstart file in the > initrd > > and then use ks=file: > > > > > > > > -----Original Message----- > > From: kickstart-list-bounces at redhat.com > > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, > William > > Sent: Friday, February 23, 2007 9:50 AM > > To: Discussion list about Kickstart > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > In our normal data centers, we build HP DL385/585 daily as well. The > > data center with issues is from an acquired company with older Cisco > > switches (5 years +). > > > > Installing RHEL 3 Update 5 > > > > The install client accesses the RHEL 3 media server via eth0. When > > Anaconda times-out, I can run thru the install setup manually and > > RHEL3U5 will install over the network. The problem is the Kickstart > > config is highly configured and doing a manual install is not really > an > > option that I want to consider. > > > > What I want is to access the ks.cfg via the network. Which implies > that > > the network needs to "come up" before Anaconda times out. > > > > OR > > > > Read the ks.cfg from the cdrom. But the cdrom is virually mounted > over > > the ILO port and is not accessible like ks=cdrom:/ks.cfg from the boot > > > prompt. I don't know what the device is. > > -Bill > > > > > -----Original Message----- > > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > > bounces at redhat.com] On Behalf Of Shabazian, Chip > > > Sent: Friday, February 23, 2007 10:34 AM > > > To: Discussion list about Kickstart > > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > > > We build HP DL 380/385/580/585 servers every day while connected to > > > Cisco switches that we don't have any access to. Are you using RHEL > 3 > > > > > or 4? If 4, do you have spare NIC's in the box? RHEL 4 enumerates > the > > > > > bus backwards from RHEL 3, so if you have a NIC, it's going to be > eth0 > > > > > instead of the onboard... > > > > > > -----Original Message----- > > > From: kickstart-list-bounces at redhat.com > > > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ramthun, > > William > > > Sent: Friday, February 23, 2007 8:01 AM > > > To: Discussion list about Kickstart > > > Subject: RE: "pump" times out -- extend a few seconds more? > > > > > > > -----Original Message----- > > > > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > > > > bounces at redhat.com] On Behalf Of nate > > > > Sent: Friday, January 12, 2007 10:33 AM > > > > To: kickstart-list at redhat.com > > > > Subject: Re: "pump" times out -- extend a few seconds more? > > > > > > > > Bill-Schoolcraft said: > > > > > Hello Family, > > > > > > > > > > I have a series of switches, in racks, and one of them seems to > be > > a > > > bit > > > > > slow so when kickstart goes for it's dhcp request, pump times > out. > > > The > > > > > initial PXE takes much more time to get an "ack" at the initial > > > contact > > > > > than the same boxes in other racks with other switches but then > > > finally > > > > > starts chugging away -- it's the next dhcp request that fails, > > times > > > out. > > > > > > > > > > Today I took the same box, draped a cat6 cable to another, less > > > > > populated switch, and it kicked fine. > > > > > > > > > > (question) > > > > > So my question is, how or where can I place an argument to tell > > > > > kickstart's "pump" request to try longer before timing out. > > > > > > > > just curious, what type of switch? many older cisco switches have > > STP > > > > on by default, if you set the port(s) to 'portfast' the links come > > up > > > > immediately rather than waiting ~45 seconds. > > > > > > > > I'm not a cisco guy but if your switches are cisco the command is > > > > something like > > > > > > > > int (interface) > > > > spanning-tree portfast > > > > ^Z > > > > wr mem > > > > > > > > > > > > e.g. on my old 3500s (IOS 12.0 is the latest they'll run) int > fa0/5 > > > > spanning-tree portfast > > > > > > > > I haven't encountered any other switch vendors that have this > > > > 'feature' enabled by default. You can tell if it's on without even > > > > > logging in(to the switch) by checking to see if the link light for > > the > > > > > > > port spends a long time in the orange color before turning green. > > I've > > > > > > > been told more modern cisco switches ship with this feature off by > > > > > default. > > > > > > > > as for extending the pump times, I'm not sure, so I can't answer > > your > > > > specific question :) maybe the above will help a bit.. > > > > > > > > I did a quick search on pump timeout, and what I did see was that > it > > > > > > may be possible to extend the timeout by adding a > > > > /etc/pump_device.conf and using the option 'timeout number>'. > > > > > > > > though you'll need to add that to the initrd image..not the > easiest > > > > thing in the world. > > > > > > > > nate > > > > > > > > > > > > > I am experiencing a similar problem; however, my situation is > slightly > > > > > different. > > > > > > I have HP DL385's connected to a 48 port 10/100 module in a Cisco > 6509 > > > > > (with Sup 2 routing module). I'm using the virtual CDROM via the HP > > ILO > > > (Integrated Lights Out) interface to perform the initial boot. I > > mount > > > a customized version of RHEL3U5 CD1. My kickstart file is specified > > as > > > an url. The system boots and Anaconda starts, but before the link > > > negotiation can complete Anaconda times-out and I am dumped into an > > > interactive install. > > > > > > A couple of data points: > > > > > > 1) The HP DL385 on-board NICs are manufactured by Broadcom. I've > read > > > > > about a known bug with link negotiation between the Broadcom NICs > and > > > Cisco 48-port line cards. The link negotiation works, it just takes > a > > > > > "longer" time than with any other kind of NIC. > > > > > > 2) Setting the switch port NIC to enable "spanning-tree portfast" > does > > > > > shorten the link negotiation time, but not enough to get around my > > > problem. > > > > > > Work Arounds: > > > > > > 1) I will prebuild the kickstart configs and place them on my > > customized > > > RHEL3U5 CD. I've been unsuccessful at reading the kickstart config > > from > > > the CDROM. The driver supporting the ILO virtual cdrom does not > name > > > the cdrom device as "cdrom". Still reading up on this. > > > > > > That's it so far, > > > -Bill > > > > > > _______________________________________________ > > > Kickstart-list mailing list > > > Kickstart-list at redhat.com > > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > > > _______________________________________________ > > > Kickstart-list mailing list > > > Kickstart-list at redhat.com > > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > 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