From mdehaan at redhat.com Mon Mar 2 22:57:14 2009 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 02 Mar 2009 17:57:14 -0500 Subject: RHEL 5.3 Kickstart In-Reply-To: <15183618.2291235488898321.JavaMail.root@postal.insourcedsecurity.com> References: <15183618.2291235488898321.JavaMail.root@postal.insourcedsecurity.com> Message-ID: <49AC644A.4000605@redhat.com> Ian Lists wrote: > I am having an issue using disk 1 of RHEL 5.3x86_64 to start a install. I boot to the disk, put in the url to my kickstart server, set the IP and the kickstart looks like it starts. It grabs the kickstart file off the webserver, but when it starts looking for the install tree it fails. Has anything changed with the files it searches for in the tree? It doesn't seem like any of the files it is looking for are there. My tree is built from the 5.3 DVD. > > I see this in my apache logs. > File does not exist: /var/www/cobbler/ks_mirror/RHEL-5.3-x86_64/disc1 > File does not exist: /var/www/cobbler/ks_mirror/RHEL-5.3-x86_64/RELEASE-NOTES-en_US.UTF-8.html > File does not exist: /var/www/cobbler/ks_mirror/RHEL-5.3-x86_64/RELEASE-NOTES.en_US.UTF-8 > File does not exist: /var/www/cobbler/ks_mirror/RHEL-5.3-x86_64/RELEASE-NOTES-en_US.html > File does not exist: /var/www/cobbler/ks_mirror/RHEL-5.3-x86_64/RELEASE-NOTES.en_US > > > > Thanks, > Ian > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > The file request stuff is just installer noise and can be ignored. Looks like you have a cobbler question. (Next time use https://fedorahosted.org/mailman/listinfo/cobbler and you'll get a faster response) As you had to enter the IP and such, it sounds like you have a static configuration. Ultimately with what you've provided I don't have enough to go on, but it could very well be that the install tree is not configured correctly in the kickstart. (url --url=). It ultimately depends what you are using. cobbler profile getks --name=foo will let you view the kickstart file I'd need to know more about your network setup to know for sure. --Michael From masaiah.p at gmail.com Tue Mar 3 09:06:28 2009 From: masaiah.p at gmail.com (Masaiah) Date: 3 Mar 2009 09:06:28 -0000 Subject: Masaiah invites you to join Zorpia Message-ID: <20090303090628.29499.qmail@zorpia.com> Hi Discussion list about! Your friend Masaiah from , just invited you to his online photo albums and journals at Zorpia.com. So what is Zorpia? It is an online community that allows you to upload unlimited amount of photos, write journals and make friends. We also have a variety of skins in store for you so that you can customize your homepage freely. Join now for free! Please click the following link to join Zorpia: http://in.signup.zorpia.com/signup?invitation_key=20081273560f58f7bdb25e3c409baad3&referral=masaiah This message was delivered with the Masaiah's initiation. If you wish to discontinue receiving invitations from us, please click the following link: http://in.signup.zorpia.com/email/optout/kickstart-list at redhat.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From kanarip at kanarip.com Thu Mar 5 00:16:24 2009 From: kanarip at kanarip.com (Jeroen van Meeuwen) Date: Thu, 05 Mar 2009 01:16:24 +0100 Subject: --hostname not applied to system Message-ID: <49AF19D8.9070401@kanarip.com> Hi there, with Fedora 10 GA i386, I've found that network --bootproto=dhcp --device=eth0 --onboot=yes --hostname=host.domain.tld isn't always applied to the installed system. The installed system would come up with localhost.localdomain as the hostname. Is somebody else seeing this? Kind regards, Jeroen van Meeuwen -kanarip From kanarip at kanarip.com Thu Mar 5 12:33:14 2009 From: kanarip at kanarip.com (Jeroen van Meeuwen) Date: Thu, 05 Mar 2009 13:33:14 +0100 Subject: --hostname not applied to system In-Reply-To: <49AF19D8.9070401@kanarip.com> References: <49AF19D8.9070401@kanarip.com> Message-ID: <49AFC68A.1080800@kanarip.com> Jeroen van Meeuwen wrote: > Hi there, > > with Fedora 10 GA i386, I've found that > > network --bootproto=dhcp --device=eth0 --onboot=yes > --hostname=host.domain.tld > > isn't always applied to the installed system. The installed system would > come up with localhost.localdomain as the hostname. > From the cobbler mailing list, it seems that this is an issue in EL5 as well. Kind regards, Jeroen van Meeuwen -kanarip From rpm at squoggle.com Thu Mar 5 20:05:35 2009 From: rpm at squoggle.com (rpm at squoggle.com) Date: Thu, 05 Mar 2009 13:05:35 -0700 Subject: user/groups directive Message-ID: <20090305130535.0ry7euomqsgs08cw@host219.hostmonster.com> Adding the following to kickstart: user --name=mac --groups=wheel --password=blahblah --iscrypted user mac gets added to system but does not get added to wheel group Mac From chip.shabazian at bankofamerica.com Fri Mar 6 03:53:50 2009 From: chip.shabazian at bankofamerica.com (Shabazian, Chip) Date: Thu, 05 Mar 2009 19:53:50 -0800 Subject: --hostname not applied to system In-Reply-To: <49AFC68A.1080800@kanarip.com> References: <49AF19D8.9070401@kanarip.com> <49AFC68A.1080800@kanarip.com> Message-ID: See the presentation I did two years ago for a good way to set the hostname in the %post: www.shabazian.com/lw2007.pdf -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Jeroen van Meeuwen Sent: Thursday, March 05, 2009 4:33 AM To: Discussion list about Kickstart Subject: Re: --hostname not applied to system Jeroen van Meeuwen wrote: > Hi there, > > with Fedora 10 GA i386, I've found that > > network --bootproto=dhcp --device=eth0 --onboot=yes > --hostname=host.domain.tld > > isn't always applied to the installed system. The installed system would > come up with localhost.localdomain as the hostname. > From the cobbler mailing list, it seems that this is an issue in EL5 as well. Kind regards, Jeroen van Meeuwen -kanarip _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From kanarip at kanarip.com Fri Mar 6 12:09:03 2009 From: kanarip at kanarip.com (Jeroen van Meeuwen) Date: Fri, 06 Mar 2009 13:09:03 +0100 Subject: --hostname not applied to system In-Reply-To: References: <49AF19D8.9070401@kanarip.com> <49AFC68A.1080800@kanarip.com> Message-ID: <49B1125F.2010801@kanarip.com> Shabazian, Chip wrote: > See the presentation I did two years ago for a good way to set the > hostname in the %post: > > www.shabazian.com/lw2007.pdf > Thank you, but I'd rather find and fix the actual cause then supress the symptoms. Kind regards, Jeroen van Meeuwen -kanarip > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Jeroen van > Meeuwen > Sent: Thursday, March 05, 2009 4:33 AM > To: Discussion list about Kickstart > Subject: Re: --hostname not applied to system > > Jeroen van Meeuwen wrote: >> Hi there, >> >> with Fedora 10 GA i386, I've found that >> >> network --bootproto=dhcp --device=eth0 --onboot=yes >> --hostname=host.domain.tld >> >> isn't always applied to the installed system. The installed system > would >> come up with localhost.localdomain as the hostname. >> > > From the cobbler mailing list, it seems that this is an issue in EL5 as > > well. > > Kind regards, > > Jeroen van Meeuwen > -kanarip > > _______________________________________________ > 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 bda20 at cam.ac.uk Fri Mar 13 16:52:18 2009 From: bda20 at cam.ac.uk (Ben) Date: Fri, 13 Mar 2009 16:52:18 +0000 (GMT) Subject: USB devices not being seen in anaconda during RHEL4 kickstart Message-ID: Wotcha, I've got a pile of Sun x4140 and x4240 servers here which I'm required to install RHEL4 (update 7) on. I've already installed RHEL5.3 on them flawlessly, booting from a USB stick with a syslinux and .cfg kickstart setup on it for a HTTP-based install. The USB device appears as sdb. This works (for installing RHEL5.3 Server 32-bit): label wibbly kernel vmli5s32 append initrd=int5s32.img ks=hd:sdb1:/configs/wibbly.cfg console=ttyS0,9600 This does not (for installing RHEL4 u7 AS 32-bit): label gonk kernel vmli4s32 append initrd=int4s32.img ks=hd:sdb1:/configs/gonk.cfg console=ttyS0,9600 Now I'm trying it with RHEL4 and failing utterly. The machine itself will boot from the USB stick in the proper manner via syslinux and display the boot.msg. It'll accept keyboard input and boot the vmlinuz and initrd.img for AS4 (update 7) and start the process (see above). However at the point at which it should read from 'sdb1' to get the configuration (the IP, HTTP kickstart tree URL, etc.) it just presents me with the dialogue box for selecting the language to install in. Not only that but if I proceed manually and get to the partitioning stage there's no sign of anything but the internal (RAID) disk(s) (announced as sda). I've tried disabling the "Hand-off EHCI" setting and trying all the other USB-related options in the BIOS but nothing seems to make anaconda see USB devices. Does anyone have any suggestions, please? With thanks, Ben -- Unix Support, MISD, University of Cambridge, England Plugger of wire, typer of keyboard, imparter of Clue Life Is Short. It's All Good. From chip.shabazian at bankofamerica.com Fri Mar 13 16:59:51 2009 From: chip.shabazian at bankofamerica.com (Shabazian, Chip) Date: Fri, 13 Mar 2009 09:59:51 -0700 Subject: USB devices not being seen in anaconda during RHEL4 kickstart In-Reply-To: References: Message-ID: <200903131659.n2DGxsX5028181@sfdmzmailmx03.bankofamerica.com> If you don't need the USB drive for anything other than the initial boot, you could always embed the kickstart file in the initrd. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Ben Sent: Friday, March 13, 2009 9:52 AM To: RHEL4 mailing list Cc: kickstart-list at redhat.com Subject: USB devices not being seen in anaconda during RHEL4 kickstart Wotcha, I've got a pile of Sun x4140 and x4240 servers here which I'm required to install RHEL4 (update 7) on. I've already installed RHEL5.3 on them flawlessly, booting from a USB stick with a syslinux and .cfg kickstart setup on it for a HTTP-based install. The USB device appears as sdb. This works (for installing RHEL5.3 Server 32-bit): label wibbly kernel vmli5s32 append initrd=int5s32.img ks=hd:sdb1:/configs/wibbly.cfg console=ttyS0,9600 This does not (for installing RHEL4 u7 AS 32-bit): label gonk kernel vmli4s32 append initrd=int4s32.img ks=hd:sdb1:/configs/gonk.cfg console=ttyS0,9600 Now I'm trying it with RHEL4 and failing utterly. The machine itself will boot from the USB stick in the proper manner via syslinux and display the boot.msg. It'll accept keyboard input and boot the vmlinuz and initrd.img for AS4 (update 7) and start the process (see above). However at the point at which it should read from 'sdb1' to get the configuration (the IP, HTTP kickstart tree URL, etc.) it just presents me with the dialogue box for selecting the language to install in. Not only that but if I proceed manually and get to the partitioning stage there's no sign of anything but the internal (RAID) disk(s) (announced as sda). I've tried disabling the "Hand-off EHCI" setting and trying all the other USB-related options in the BIOS but nothing seems to make anaconda see USB devices. Does anyone have any suggestions, please? With thanks, Ben -- Unix Support, MISD, University of Cambridge, England Plugger of wire, typer of keyboard, imparter of Clue Life Is Short. It's All Good. _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From avalder at gmail.com Fri Mar 13 17:52:55 2009 From: avalder at gmail.com (Al Alder) Date: Fri, 13 Mar 2009 12:52:55 -0500 Subject: USB devices not being seen in anaconda during RHEL4 kickstart In-Reply-To: References: Message-ID: <1aa0c5a20903131052o1c2df14i8644ebf420043760@mail.gmail.com> Had a very similar problem with some HP servers and RHEL 4. RHEL 4 did not recognize the disk controller in the newer HP machines. Our solution was to go to RHEL 4.6 which had the driver for the newer disk controllers. So you are running into a similar problem with your USB stick. RHEL5.3 works because its initrd image has the driver for the USB. You can 1) rebuild your initrd image with the USB driver 2) just use RHEL 5.3 I investigated number 1, but couldn't find any documentation that explained how to that for the initial install boot initrd. Am sure I could have monkeyed with it and gotten it to work if I had had the time. We went with our option 2 and had to upgrade everything to RHEL 4.6. On Fri, Mar 13, 2009 at 11:52 AM, Ben wrote: > Wotcha, > > I've got a pile of Sun x4140 and x4240 servers here which I'm required to > install RHEL4 (update 7) on. I've already installed RHEL5.3 on them > flawlessly, booting from a USB stick with a syslinux and .cfg > kickstart setup on it for a HTTP-based install. The USB device appears as > sdb. > > This works (for installing RHEL5.3 Server 32-bit): > label wibbly > kernel vmli5s32 > append initrd=int5s32.img ks=hd:sdb1:/configs/wibbly.cfg > console=ttyS0,9600 > > This does not (for installing RHEL4 u7 AS 32-bit): > label gonk > kernel vmli4s32 > append initrd=int4s32.img ks=hd:sdb1:/configs/gonk.cfg console=ttyS0,9600 > > Now I'm trying it with RHEL4 and failing utterly. The machine itself will > boot from the USB stick in the proper manner via syslinux and display the > boot.msg. It'll accept keyboard input and boot the vmlinuz and initrd.img > for AS4 (update 7) and start the process (see above). However at the point > at which it should read from 'sdb1' to get the configuration (the IP, HTTP > kickstart tree URL, etc.) it just presents me with the dialogue box for > selecting the language to install in. Not only that but if I proceed > manually and get to the partitioning stage there's no sign of anything but > the internal (RAID) disk(s) (announced as sda). > > I've tried disabling the "Hand-off EHCI" setting and trying all the other > USB-related options in the BIOS but nothing seems to make anaconda see USB > devices. > > Does anyone have any suggestions, please? > > With thanks, > > Ben > -- > Unix Support, MISD, University of Cambridge, England > Plugger of wire, typer of keyboard, imparter of Clue > Life Is Short. It's All Good. > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- Later, Al Alder -------------- next part -------------- An HTML attachment was scrubbed... URL: From bda20 at cam.ac.uk Fri Mar 13 18:55:45 2009 From: bda20 at cam.ac.uk (Ben) Date: Fri, 13 Mar 2009 18:55:45 +0000 (GMT) Subject: USB devices not being seen in anaconda during RHEL4 kickstart In-Reply-To: <1aa0c5a20903131052o1c2df14i8644ebf420043760@mail.gmail.com> References: <1aa0c5a20903131052o1c2df14i8644ebf420043760@mail.gmail.com> Message-ID: On Fri, 13 Mar 2009, Al Alder wrote: > Had a very similar problem with some HP servers and RHEL 4. RHEL 4 did > not recognize the disk controller in the newer HP machines. Our solution > was to go to RHEL 4.6 which had the driver for the newer disk controllers. > So you are running into a similar problem with your USB stick. RHEL5.3 > works because its initrd image has the driver for the USB. You can > > 1) rebuild your initrd image with the USB driver > 2) just use RHEL 5.3 > > I investigated number 1, but couldn't find any documentation that > explained how to that for the initial install boot initrd. Am sure I > could have monkeyed with it and gotten it to work if I had had the time. I'm not sure I want to get into rolling my own special initrd, but needs must and all that. > We went with our option 2 and had to upgrade everything to RHEL 4.6. I can't use RHEL5.x I'm afraid. The software that goes on top of the OS is currently only certified for RHEL4. And as it is I'm using RHEL4 update 7, so theoretically I've got the most up-to-date support for devices that the RHEL4 stream offers. I wonder what's different about this USB controller Vs all USB controllers before it... My current plan on Monday is to try Chip Shabazian's suggestion of wrapping the ks.cfg up in the initrd.img and seeing if that works. I'm still listening out for suggestions and experience of other people with RHEL4 and Sun x4X40 servers. In case it matters x4100M2, x4200M2 and x4600M2 servers work perfectly well with USB kickstarted RHEL4 (-: Ben -- Unix Support, MISD, University of Cambridge, England Plugger of wire, typer of keyboard, imparter of Clue Life Is Short. It's All Good. From avalder at gmail.com Fri Mar 13 20:25:58 2009 From: avalder at gmail.com (Al Alder) Date: Fri, 13 Mar 2009 15:25:58 -0500 Subject: USB devices not being seen in anaconda during RHEL4 kickstart In-Reply-To: References: <1aa0c5a20903131052o1c2df14i8644ebf420043760@mail.gmail.com> Message-ID: <1aa0c5a20903131325h2642dec4i9bdf55a8d6f0e239@mail.gmail.com> Yep, I understand you have to stay with 4.7. In addition to Chip's solution, you might want to use the 5.3 initrd image on your USB stick and you would have to modify the 5.3 initrd image to specify your kickstart image instead of the 5.3 one. Just something you might want to consider if you have problems with Chip's solution. Whatever you do that works, please post what you did. On Fri, Mar 13, 2009 at 1:55 PM, Ben wrote: > On Fri, 13 Mar 2009, Al Alder wrote: > > Had a very similar problem with some HP servers and RHEL 4. RHEL 4 did >> not recognize the disk controller in the newer HP machines. Our solution >> was to go to RHEL 4.6 which had the driver for the newer disk controllers. >> So you are running into a similar problem with your USB stick. RHEL5.3 >> works because its initrd image has the driver for the USB. You can >> >> 1) rebuild your initrd image with the USB driver >> 2) just use RHEL 5.3 >> >> I investigated number 1, but couldn't find any documentation that >> explained how to that for the initial install boot initrd. Am sure I could >> have monkeyed with it and gotten it to work if I had had the time. >> > > I'm not sure I want to get into rolling my own special initrd, but needs > must and all that. > > > We went with our option 2 and had to upgrade everything to RHEL 4.6. >> > > I can't use RHEL5.x I'm afraid. The software that goes on top of the OS is > currently only certified for RHEL4. And as it is I'm using RHEL4 update 7, > so theoretically I've got the most up-to-date support for devices that the > RHEL4 stream offers. > > I wonder what's different about this USB controller Vs all USB controllers > before it... > > My current plan on Monday is to try Chip Shabazian's suggestion of wrapping > the ks.cfg up in the initrd.img and seeing if that works. I'm still > listening out for suggestions and experience of other people with RHEL4 and > Sun x4X40 servers. In case it matters x4100M2, x4200M2 and x4600M2 servers > work perfectly well with USB kickstarted RHEL4 (-: > > Ben > -- > Unix Support, MISD, University of Cambridge, England > Plugger of wire, typer of keyboard, imparter of Clue > Life Is Short. It's All Good. > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- Later, Al Alder -------------- next part -------------- An HTML attachment was scrubbed... URL: From chip.shabazian at bankofamerica.com Fri Mar 13 20:40:00 2009 From: chip.shabazian at bankofamerica.com (Shabazian, Chip) Date: Fri, 13 Mar 2009 13:40:00 -0700 Subject: USB devices not being seen in anaconda during RHEL4 kickstart In-Reply-To: <1aa0c5a20903131325h2642dec4i9bdf55a8d6f0e239@mail.gmail.com> References: <1aa0c5a20903131052o1c2df14i8644ebf420043760@mail.gmail.com> <1aa0c5a20903131325h2642dec4i9bdf55a8d6f0e239@mail.gmail.com> Message-ID: <200903132040.n2DKe45K025002@sfdmzmailmx03.bankofamerica.com> I have, in the past, replaced "loader" in the initrd with a newer version when we needed newer drivers but wanted to keep our existing kickstart tree. I would guess that you probably couldn't use the loader from RHEL 5 in a RHEL 4 initrd, but I guess you could try. ________________________________ From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Al Alder Sent: Friday, March 13, 2009 1:26 PM To: Discussion list about Kickstart Subject: Re: USB devices not being seen in anaconda during RHEL4 kickstart Yep, I understand you have to stay with 4.7. In addition to Chip's solution, you might want to use the 5.3 initrd image on your USB stick and you would have to modify the 5.3 initrd image to specify your kickstart image instead of the 5.3 one. Just something you might want to consider if you have problems with Chip's solution. Whatever you do that works, please post what you did. On Fri, Mar 13, 2009 at 1:55 PM, Ben wrote: On Fri, 13 Mar 2009, Al Alder wrote: Had a very similar problem with some HP servers and RHEL 4. RHEL 4 did not recognize the disk controller in the newer HP machines. Our solution was to go to RHEL 4.6 which had the driver for the newer disk controllers. So you are running into a similar problem with your USB stick. RHEL5.3 works because its initrd image has the driver for the USB. You can 1) rebuild your initrd image with the USB driver 2) just use RHEL 5.3 I investigated number 1, but couldn't find any documentation that explained how to that for the initial install boot initrd. Am sure I could have monkeyed with it and gotten it to work if I had had the time. I'm not sure I want to get into rolling my own special initrd, but needs must and all that. We went with our option 2 and had to upgrade everything to RHEL 4.6. I can't use RHEL5.x I'm afraid. The software that goes on top of the OS is currently only certified for RHEL4. And as it is I'm using RHEL4 update 7, so theoretically I've got the most up-to-date support for devices that the RHEL4 stream offers. I wonder what's different about this USB controller Vs all USB controllers before it... My current plan on Monday is to try Chip Shabazian's suggestion of wrapping the ks.cfg up in the initrd.img and seeing if that works. I'm still listening out for suggestions and experience of other people with RHEL4 and Sun x4X40 servers. In case it matters x4100M2, x4200M2 and x4600M2 servers work perfectly well with USB kickstarted RHEL4 (-: Ben -- Unix Support, MISD, University of Cambridge, England Plugger of wire, typer of keyboard, imparter of Clue Life Is Short. It's All Good. _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list -- Later, Al Alder -------------- next part -------------- An HTML attachment was scrubbed... URL: From avalder at gmail.com Fri Mar 13 21:19:20 2009 From: avalder at gmail.com (Al Alder) Date: Fri, 13 Mar 2009 16:19:20 -0500 Subject: USB devices not being seen in anaconda during RHEL4 kickstart In-Reply-To: <200903132040.n2DKe45K025002@sfdmzmailmx03.bankofamerica.com> References: <1aa0c5a20903131052o1c2df14i8644ebf420043760@mail.gmail.com> <1aa0c5a20903131325h2642dec4i9bdf55a8d6f0e239@mail.gmail.com> <200903132040.n2DKe45K025002@sfdmzmailmx03.bankofamerica.com> Message-ID: <1aa0c5a20903131419y2d2e876et690351e916cd6fca@mail.gmail.com> Using RHEL 5 in a RHEL 4 initrd will not work without modifying the RHEL 5 loader to point to RHEL 4. I tried using RHEL 5 loader in RHEL 4 which didn't work. Realized that the RHEL 5 loader would have to be modified, but I never took the time to pursue the modification. On Fri, Mar 13, 2009 at 3:40 PM, Shabazian, Chip < chip.shabazian at bankofamerica.com> wrote: > I have, in the past, replaced ?loader? in the initrd with a newer version > when we needed newer drivers but wanted to keep our existing kickstart > tree. I would guess that you probably couldn?t use the loader from RHEL 5 > in a RHEL 4 initrd, but I guess you could try. > > > ------------------------------ > > *From:* kickstart-list-bounces at redhat.com [mailto: > kickstart-list-bounces at redhat.com] *On Behalf Of *Al Alder > *Sent:* Friday, March 13, 2009 1:26 PM > *To:* Discussion list about Kickstart > *Subject:* Re: USB devices not being seen in anaconda during RHEL4 > kickstart > > > > Yep, I understand you have to stay with 4.7. > In addition to Chip's solution, you might want to > use the 5.3 initrd image on your USB stick and > you would have to modify the 5.3 initrd image to > specify your kickstart image instead of the 5.3 one. > Just something you might want to consider if you > have problems with Chip's solution. > Whatever you do that works, please post what you > did. > > On Fri, Mar 13, 2009 at 1:55 PM, Ben wrote: > > On Fri, 13 Mar 2009, Al Alder wrote: > > Had a very similar problem with some HP servers and RHEL 4. RHEL 4 did not > recognize the disk controller in the newer HP machines. Our solution was to > go to RHEL 4.6 which had the driver for the newer disk controllers. So you > are running into a similar problem with your USB stick. RHEL5.3 works > because its initrd image has the driver for the USB. You can > > 1) rebuild your initrd image with the USB driver > 2) just use RHEL 5.3 > > I investigated number 1, but couldn't find any documentation that explained > how to that for the initial install boot initrd. Am sure I could have > monkeyed with it and gotten it to work if I had had the time. > > > I'm not sure I want to get into rolling my own special initrd, but needs > must and all that. > > We went with our option 2 and had to upgrade everything to RHEL 4.6. > > > I can't use RHEL5.x I'm afraid. The software that goes on top of the OS is > currently only certified for RHEL4. And as it is I'm using RHEL4 update 7, > so theoretically I've got the most up-to-date support for devices that the > RHEL4 stream offers. > > I wonder what's different about this USB controller Vs all USB controllers > before it... > > My current plan on Monday is to try Chip Shabazian's suggestion of wrapping > the ks.cfg up in the initrd.img and seeing if that works. I'm still > listening out for suggestions and experience of other people with RHEL4 and > Sun x4X40 servers. In case it matters x4100M2, x4200M2 and x4600M2 servers > work perfectly well with USB kickstarted RHEL4 (-: > > Ben > -- > Unix Support, MISD, University of Cambridge, England > Plugger of wire, typer of keyboard, imparter of Clue > Life Is Short. It's All Good. > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > > -- > Later, > > Al Alder > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- Later, Al Alder -------------- next part -------------- An HTML attachment was scrubbed... URL: From debian at herakles.homelinux.org Sat Mar 14 00:12:30 2009 From: debian at herakles.homelinux.org (John Summerfield) Date: Sat, 14 Mar 2009 09:12:30 +0900 Subject: USB devices not being seen in anaconda during RHEL4 kickstart In-Reply-To: References: <1aa0c5a20903131052o1c2df14i8644ebf420043760@mail.gmail.com> Message-ID: <49BAF66E.4050105@herakles.homelinux.org> Ben wrote: > On Fri, 13 Mar 2009, Al Alder wrote: > >> Had a very similar problem with some HP servers and RHEL 4. RHEL 4 >> did not recognize the disk controller in the newer HP machines. Our >> solution was to go to RHEL 4.6 which had the driver for the newer disk >> controllers. So you are running into a similar problem with your USB >> stick. RHEL5.3 works because its initrd image has the driver for the >> USB. You can >> >> 1) rebuild your initrd image with the USB driver >> 2) just use RHEL 5.3 >> >> I investigated number 1, but couldn't find any documentation that >> explained how to that for the initial install boot initrd. Am sure I >> could have monkeyed with it and gotten it to work if I had had the time. > > I'm not sure I want to get into rolling my own special initrd, but needs > must and all that. Last I did it, it's not hard, it was a compressed CPIO archive. mkdir /tmp/initrd cd /tmp/initrd cat /boot/initrd | gunzip | cpio --extract -d { hack repack test } until done The hack is the hard bit:-) It's also a good way to get your ks file in there. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From debian at herakles.homelinux.org Sat Mar 14 00:14:27 2009 From: debian at herakles.homelinux.org (John Summerfield) Date: Sat, 14 Mar 2009 09:14:27 +0900 Subject: USB devices not being seen in anaconda during RHEL4 kickstart In-Reply-To: <1aa0c5a20903131419y2d2e876et690351e916cd6fca@mail.gmail.com> References: <1aa0c5a20903131052o1c2df14i8644ebf420043760@mail.gmail.com> <1aa0c5a20903131325h2642dec4i9bdf55a8d6f0e239@mail.gmail.com> <200903132040.n2DKe45K025002@sfdmzmailmx03.bankofamerica.com> <1aa0c5a20903131419y2d2e876et690351e916cd6fca@mail.gmail.com> Message-ID: <49BAF6E3.3@herakles.homelinux.org> Al Alder wrote: > Using RHEL 5 in a RHEL 4 initrd will not work without > modifying the RHEL 5 loader to point to RHEL 4. > I tried using RHEL 5 loader in RHEL 4 which didn't > work. Realized that the RHEL 5 loader would have > to be modified, but I never took the time to pursue > the modification. ?? I don't understand that. _How_ did it fail? -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From bjs at redhat.com Sat Mar 14 00:23:29 2009 From: bjs at redhat.com (Bryan J Smith) Date: Fri, 13 Mar 2009 20:23:29 -0400 Subject: USB devices not being seen in anaconda during RHEL4 kickstart In-Reply-To: <49BAF6E3.3@herakles.homelinux.org> References: <1aa0c5a20903131052o1c2df14i8644ebf420043760@mail.gmail.com> <1aa0c5a20903131325h2642dec4i9bdf55a8d6f0e239@mail.gmail.com> <200903132040.n2DKe45K025002@sfdmzmailmx03.bankofamerica.com> <1aa0c5a20903131419y2d2e876et690351e916cd6fca@mail.gmail.com> <49BAF6E3.3@herakles.homelinux.org> Message-ID: <1236990209.4024.87.camel@localhost.localdomain> Al Alder wrote: > Using RHEL 5 in a RHEL 4 initrd will not work without > modifying the RHEL 5 loader to point to RHEL 4. > I tried using RHEL 5 loader in RHEL 4 which didn't > work. Realized that the RHEL 5 loader would have > to be modified, but I never took the time to pursue > the modification. On Sat, 2009-03-14 at 09:14 +0900, John Summerfield wrote: > ?? > I don't understand that. _How_ did it fail? Is he saying he was using GRUB from RHEL 5 that loads from a RHEL 5 root/usr to boot a RHEL 4 kernel/initrd in another partition and it failed? That sounds more like an issue with GRUB configuration than a failure of RHEL 5 v. 4 GRUB. I.e., did he change the booting reference for RHEL 4 to the correct boot/root slice (partition)? And not leave it the same for RHEL 5? -- Bryan J Smith Senior Consultant Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) -------------------------------------------------------- You already know Red Hat as the entity dedicated to 100% no-IP-strings-attached, community software development. But do you know where CIOs rate Red Hat versus other software and services firms for their own, direct needs? It's no comparison: http://www.redhat.com/promo/vendor/ From raqnerd at gmail.com Sun Mar 22 11:23:58 2009 From: raqnerd at gmail.com (Frank Kimbell) Date: Sun, 22 Mar 2009 12:23:58 +0100 Subject: NetworkManager? Message-ID: <846a2d7c0903220423g65c0b2e5ob3c5483b77c4e6b1@mail.gmail.com> Hi, I've built a custom build for Fedora9, which I've recently updated to Fedora10. Since then, the "NetworkManager" kicks in when initializing, just before formatting and installing RPM. The annoying thing is that it hangs there for a long time, then fails and asks to retry. When I hit retry, it seems to work. This is not what I want. I want it to skip a search for link (that's what it seems to be?), and just go on. I've configured the network settings manually, so I don't care for an automated procedure. How do I skip this? Thank you. Frank From keith at karsites.net Sun Mar 22 13:41:58 2009 From: keith at karsites.net (Keith Roberts) Date: Sun, 22 Mar 2009 13:41:58 +0000 (GMT) Subject: NetworkManager? In-Reply-To: <846a2d7c0903220423g65c0b2e5ob3c5483b77c4e6b1@mail.gmail.com> References: <846a2d7c0903220423g65c0b2e5ob3c5483b77c4e6b1@mail.gmail.com> Message-ID: On Sun, 22 Mar 2009, Frank Kimbell wrote: > To: kickstart-list at redhat.com > From: Frank Kimbell > Subject: NetworkManager? > > Hi, > > I've built a custom build for Fedora9, which I've recently updated to Fedora10. > Since then, the "NetworkManager" kicks in when initializing, just > before formatting and installing RPM. > > The annoying thing is that it hangs there for a long time, then fails > and asks to retry. When I hit retry, it seems to work. > This is not what I want. I want it to skip a search for link (that's > what it seems to be?), and just go on. I've configured the network > settings manually, so I don't care for an automated procedure. > > How do I skip this? > > Thank you. > Frank > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list Hi Frank. What does the network settings part of your kickstart file look like? Keith ----------------------------------------------------------------- Websites: http://www.php-debuggers.net http://www.karsites.net http://www.raised-from-the-dead.org.uk All email addresses are challenge-response protected with TMDA [http://tmda.net] ----------------------------------------------------------------- From raqnerd at gmail.com Sun Mar 22 15:29:05 2009 From: raqnerd at gmail.com (Frank Kimbell) Date: Sun, 22 Mar 2009 16:29:05 +0100 Subject: NetworkManager? Message-ID: <846a2d7c0903220829g37759denb07419a66b451cee@mail.gmail.com> > What does the network settings part of your kickstart file look like? network --device eth0 --mtu=1500 --bootproto static --ip 0.0.0.0 --netmask 255.255.255.0 --gateway 0.0.0.0 --nameserver 127.0.0.1 -- hostname svr.tester.com This has always worked before the recent update. (it sets IP to "0.0.0.0" and I change that later. But 0.0.0.0 is a correct value, and it did work before.) From virginian at blueyonder.co.uk Sun Mar 22 17:28:26 2009 From: virginian at blueyonder.co.uk (Virginian) Date: Sun, 22 Mar 2009 17:28:26 -0000 Subject: NetworkManager? References: <846a2d7c0903220423g65c0b2e5ob3c5483b77c4e6b1@mail.gmail.com> Message-ID: <792C0CF127444860AB97DBFB4FE53973@Desktop> chkconfig networkmanager off and service network manager stop. Just makes sure that network is chkconfig on at the appropriate levels (3 or 5 for xwindows) and service network start. ----- Original Message ----- From: "Frank Kimbell" To: Sent: Sunday, March 22, 2009 11:23 AM Subject: NetworkManager? > Hi, > > I've built a custom build for Fedora9, which I've recently updated to > Fedora10. > Since then, the "NetworkManager" kicks in when initializing, just > before formatting and installing RPM. > > The annoying thing is that it hangs there for a long time, then fails > and asks to retry. When I hit retry, it seems to work. > This is not what I want. I want it to skip a search for link (that's > what it seems to be?), and just go on. I've configured the network > settings manually, so I don't care for an automated procedure. > > How do I skip this? > > Thank you. > Frank > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From raqnerd at gmail.com Sun Mar 22 17:51:38 2009 From: raqnerd at gmail.com (Frank Kimbell) Date: Sun, 22 Mar 2009 18:51:38 +0100 Subject: NetworkManager? Message-ID: <846a2d7c0903221051p3ff07852p7fdaf2069a6cbb05@mail.gmail.com> > chkconfig networkmanager off and service network manager stop Are you talking AFTER the installation has taken place? I mean DURING the installation, when I start my kickstart install. If this is what you mean too, how do I do this in the kickstart anaconda? From bjs at redhat.com Mon Mar 23 01:20:12 2009 From: bjs at redhat.com (Bryan J Smith) Date: Sun, 22 Mar 2009 21:20:12 -0400 Subject: NetworkManager? In-Reply-To: <846a2d7c0903220829g37759denb07419a66b451cee@mail.gmail.com> References: <846a2d7c0903220829g37759denb07419a66b451cee@mail.gmail.com> Message-ID: <1237771212.3996.6.camel@localhost.localdomain> On Sun, 2009-03-22 at 16:29 +0100, Frank Kimbell wrote: > network --device eth0 --mtu=1500 --bootproto static --ip 0.0.0.0 > --netmask 255.255.255.0 --gateway 0.0.0.0 --nameserver 127.0.0.1 -- > hostname svr.tester.com > This has always worked before the recent update. (it sets IP to > "0.0.0.0" and I change that later. But 0.0.0.0 is a correct value, and > it did work before.) Not trying to question and I don't know the history, but if you need to bring an interface up, it never hurts to assign an address from the reserved IPv4 LINKLOCAL subnet, per IETF -- 169.254/16. It's not merely reserved for DHCP. The LINKLOCAL concept has a far greater purpose, including being the non-routable address to another, routable address assigned to the same interface. BTW, the last octets for the node often match those of the last two octets of the MAC address, although there is a formal way to derive them. -- Bryan J Smith Senior Consultant Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) -------------------------------------------------------- You already know Red Hat as the entity dedicated to 100% no-IP-strings-attached, community software development. But do you know where CIOs rate Red Hat versus other software and services firms for their own, direct needs? It's no comparison: http://www.redhat.com/promo/vendor/ From tom at ng23.net Tue Mar 24 11:20:16 2009 From: tom at ng23.net (Tom Brown) Date: Tue, 24 Mar 2009 11:20:16 +0000 Subject: specific package version in %packages Message-ID: <49C8C1F0.5080900@ng23.net> If i want foo to be at version 1 but in the repo at install time there is foo-1 and foo-2 in the %packages can i do %packages foo = 1 and i get foo-1 rather than the newer foo-2 ?? like i would in the .spec of an rpm? thanks From debian at herakles.homelinux.org Tue Mar 24 11:48:09 2009 From: debian at herakles.homelinux.org (John Summerfield) Date: Tue, 24 Mar 2009 20:48:09 +0900 Subject: specific package version in %packages In-Reply-To: <49C8C1F0.5080900@ng23.net> References: <49C8C1F0.5080900@ng23.net> Message-ID: <49C8C879.9090503@herakles.homelinux.org> Tom Brown wrote: > If i want foo to be at version 1 but in the repo at install time there > is foo-1 and foo-2 in the %packages can i do > > %packages > foo = 1 > > and i get foo-1 rather than the newer foo-2 ?? > > like i would in the .spec of an rpm? Depending on what you're installing, and without trying it, I would expect that anything that is accepted by rpm or yum will work. Why not just try it, it only takes a few minutes; I've just done two ks installs since dinner, and I can still tast desert. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From vandan-joshi at in.com Tue Mar 24 12:41:09 2009 From: vandan-joshi at in.com (Vandan Joshi) Date: Tue, 24 Mar 2009 18:11:09 +0530 Subject: how to ignore repositories Message-ID: <1237898469.281715cafa675bf359ebaa42cb44fa17@mail.in.com> Dear Friends,I am making a kickstart installation for fedora10. What I want is, It should not go for internet repositories, It should simply install the core packages resides in cdrom:/Packages directory.So How can i do tht ? means how can i disable repo in ks.cfg ?? Dear kickstartlist! Get Yourself a cool, short @in.com Email ID now! -------------- next part -------------- An HTML attachment was scrubbed... URL: From debian at herakles.homelinux.org Tue Mar 24 23:21:23 2009 From: debian at herakles.homelinux.org (John Summerfield) Date: Wed, 25 Mar 2009 08:21:23 +0900 Subject: how to ignore repositories In-Reply-To: <1237898469.281715cafa675bf359ebaa42cb44fa17@mail.in.com> References: <1237898469.281715cafa675bf359ebaa42cb44fa17@mail.in.com> Message-ID: <49C96AF3.2010707@herakles.homelinux.org> Vandan Joshi wrote: > Dear Friends,I am making a kickstart installation for fedora10. What I want is, It should not go for internet repositories, It should simply install the core packages resides in cdrom:/Packages directory.So How can i do tht ? means how can i disable repo in ks.cfg ?? Dear kickstartlist! Get Yourself a cool, short @in.com Email ID now! > I'd feel happier without your ad. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From vandan-joshi at in.com Wed Mar 25 06:53:43 2009 From: vandan-joshi at in.com (Vandan Joshi) Date: Wed, 25 Mar 2009 12:23:43 +0530 Subject: how to ignore repositories In-Reply-To: <49C96AF3.2010707@herakles.homelinux.org> Message-ID: <1237964023.9cac2ca53c5fe723c249d012d6091c50@mail.in.com> dear johnthnx for the replay.sorry for the add but its from in.com it self only, not mine.my only question is below..===========================================================================================Dear Friends,I am making a kickstart installation for fedora10. I have added a kickstart file ks.cfg in cdrom's root directory. At the time of kickstart installation its going to internet repositories. What I want is, It should not go for internet repositories, It should simply install the core packages resides in cdrom. So How can i do tht ? means how can i disable repo ?? Is there any option in kickstart file to disable this repos ??below is my ks.cfg file##############################################################################FILE: ks.cfg (fedora autoinstaller)##VERSION: DEVEL##PLATFORM: x86, AMD64, or Intel EM64T##AUTHOR: Vandan D. Joshi############################################################################installcdrom# Debugging (uncomment next line to deb ug in the interactive mode)#interactive# Language and input supportlang enUS.UTF8#langsupport default=enUS.UTF8 enUS.UTF8keyboard us#mouse generic3ps/2# XWindows (use "skipx" directive to skip XWindows configuration)skipx#services enabled=network disabled=NetworkManager# Network configurationnetwork device eth0 bootproto static ip 192.168.1.20 netmask 255.255.255.0 gateway 192.168.1.1 hostname vandan#network device eth0 onboot yes bootproto dhcp hostname googllies# Authentication and securityrootpw pass123firewall disabledselinux disabledauthconfig enableshadow enablemd5# timezonetimezone utc Asia/Kolkata# Boot loaderbootloader location=mbr# clear mbrzerombrclearpart all initlabel#part /boot fstype ext3 size=100#part /home fstype ext3 size=3000#part / fstype ext3 size=2000#part swap size=2000# Select packages%packages#@core# preinstallation script%pre# postinstallation script%post# installation over%end========================================================================= ================== Original message From:John Summerfield< debian at herakles.homelinux.org >Date: 25 Mar 09 04:51:23Subject:Re: how to ignore repositoriesTo: Discussion list about Kickstart Vandan Joshi wrote: > Dear Friends,I am making a kickstart installation for fedora10. What I want is, It should not go for internet repositories, It should simply install the core packages resides in cdrom:/Packages directory.So How can i do tht ? means how can i disable repo in ks.cfg ?? Dear kickstartlist! Get Yourself a cool, short @in.com Email ID now! > I'd feel happier without your ad. Cheers John spambait 1aaaaaaa at coco.merseine.nuZ1aaaaaaa@coco.merseine.nu Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smartquestions.html http://support.microsoft.com/kb/555375You cannot reply offlist:) Kickstartlist mailing list Kickstartlist at redhat.com https://www.redhat.com/mailman/listinfo/kickstartlist -------------- next part -------------- An HTML attachment was scrubbed... URL: From mrose at n-able.com Wed Mar 25 13:46:25 2009 From: mrose at n-able.com (Matt Rose) Date: Wed, 25 Mar 2009 09:46:25 -0400 Subject: [Fwd: network step always skipped in kickstart.py] Message-ID: <49CA35B1.9090509@n-able.com> I posted this to the anaconda list with echoing silence, so I was wondering if someone here could help me figure out exactly what was going on. Why does anaconda *always* skip the network configuration step if kickstart is used? -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- An embedded message was scrubbed... From: Matt Rose Subject: network step always skipped in kickstart.py Date: Tue, 24 Mar 2009 10:43:22 -0400 Size: 7043 URL: From kpowell at redhat.com Wed Mar 25 18:24:54 2009 From: kpowell at redhat.com (Kyle Powell) Date: Wed, 25 Mar 2009 14:24:54 -0400 Subject: [Fwd: network step always skipped in kickstart.py] In-Reply-To: <49CA35B1.9090509@n-able.com> References: <49CA35B1.9090509@n-able.com> Message-ID: <49CA76F6.5060407@redhat.com> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Matt Rose wrote: > I posted this to the anaconda list with echoing silence, so I was > wondering if someone here could help me figure out exactly what was > going on. Why does anaconda *always* skip the network configuration > step if kickstart is used? You patched anaconda and now it's not working. Your patch is only two lines. My guess would be self.ksdata.network is always evaluating as true, which is causing network to always be skipped. Add some debug code to confirm this. - -- Kyle Powell | Red Hat | Senior Consultant, RHCE | Cell: 571.215.4340 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iD8DBQFJynb27pTtanQdBU4RAg+cAJ9ROB1o/Oi8Cg5FyE5uT5TuIbTJ8gCdHhZM xlB6qKnvbH/gAGZZZZrjlTI= =o6zk -----END PGP SIGNATURE----- From mrose at n-able.com Wed Mar 25 19:37:38 2009 From: mrose at n-able.com (Matt Rose) Date: Wed, 25 Mar 2009 15:37:38 -0400 Subject: [Fwd: network step always skipped in kickstart.py] In-Reply-To: <49CA76F6.5060407@redhat.com> References: <49CA35B1.9090509@n-able.com> <49CA76F6.5060407@redhat.com> Message-ID: <49CA8802.2010806@n-able.com> Kyle Powell wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Matt Rose wrote: > >> I posted this to the anaconda list with echoing silence, so I was >> wondering if someone here could help me figure out exactly what was >> going on. Why does anaconda *always* skip the network configuration >> step if kickstart is used? >> > > You patched anaconda and now it's not working. Your patch is only two lines. My > guess would be self.ksdata.network is always evaluating as true, which is > causing network to always be skipped. Add some debug code to confirm this. > The debug code is what led me to write the patch. Well, that and some digging around in kickstart.py. If you read kickstart.py, it always calls dispatch.skipStep("network"), no matter whether or not the network configuration is specified. the patch I posted actually gets it to work as expected, and not call dispatch.skipStep("network") if there's no "network" command in the ks.cfg. without the patch, the network configuration step is *always* skipped, no matter whether I specify a network entry in ks.cfg or not. To put it in more "bug" terms, without the patch underneath, the following ks.cfg does *not* bring up the network configuration screen, even though I expect it to. install cdrom lang en_US.UTF-8 keyboard us rootpw --iscrypted "$1$Q8UIYpca$uHhNzcR8qKtzFAa4Z.zmq." firewall --disabled authconfig --enableshadow --enablemd5 text skipx selinux --disabled timezone --utc America/New_York reboot bootloader --location=mbr autopart %pre %packages @base %post Mind you, I'm not sure whether my expectations are correct. The patch I attached "fixes" this behaviour so that the network configuration dialog is brought up, if there is no "network" option in the ks.cfg Matt > - -- > Kyle Powell | Red Hat | Senior Consultant, RHCE | Cell: 571.215.4340 > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.7 (GNU/Linux) > Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org > > iD8DBQFJynb27pTtanQdBU4RAg+cAJ9ROB1o/Oi8Cg5FyE5uT5TuIbTJ8gCdHhZM > xlB6qKnvbH/gAGZZZZrjlTI= > =o6zk > -----END PGP SIGNATURE----- > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -------------- next part -------------- An HTML attachment was scrubbed... URL: From angus.clarke at gmail.com Thu Mar 26 11:05:25 2009 From: angus.clarke at gmail.com (Angus Clarke) Date: Thu, 26 Mar 2009 11:05:25 +0000 (GMT) Subject: Anaconda default behaviour on error Message-ID: <402529803.2986111238065525460.JavaMail.root@unx-s-manc3> Hi I use CGI scripts to auto-generate kickstart files depending on the target's IP address - $ENV{REMOTE_ADDR} I do this to avoid maintaining hundreds of KS files (normally one for each target server) My situation is slightly complex as I have dozens of target networks. We derive network information from predefined perl hash tables, where we lookup: - default route - dns server - target hostname (derived from DNS of the client's IP) I use this to generate a line such as the following in the default KS file: <-- snip network --device eth0 --bootproto static --ip 192.168.1.50 --netmask 255.255.255.0 --gateway 192.168.1.1 --nameserver 192.168.1.2 --hostname angus <-- snip Everything works great, until one of my guys rebuilds a server on some unknown network (ie not in the lookup table) When this occurs, anaconda is fed an incomplete kickstart file - it's default behaviour then is to prompt for input. The trouble I have with this is that we are usually rebuilding servers in remote locations without local hands! So, we affectively lose this server until our next site visit (lights out is not configured across all our sites) Is there some option or means of persuading anaconda to reboot on failures, prior to any changes being committed to the system? (ie. prior to fdisk activity of Kickstart file) Thanks in advance Angus From mrose at n-able.com Thu Mar 26 12:50:42 2009 From: mrose at n-able.com (Matt Rose) Date: Thu, 26 Mar 2009 08:50:42 -0400 Subject: Anaconda default behaviour on error In-Reply-To: <402529803.2986111238065525460.JavaMail.root@unx-s-manc3> References: <402529803.2986111238065525460.JavaMail.root@unx-s-manc3> Message-ID: <49CB7A22.7070907@n-able.com> what version of anaconda are you using for this? I personally don't see a way of altering the kickstart behaviour, except by getting down and dirty in the kickstart.py source file of anaconda, but maybe someone else has some ideas. Matt Angus Clarke wrote: > Hi > > I use CGI scripts to auto-generate kickstart files depending on the target's IP address - $ENV{REMOTE_ADDR} > I do this to avoid maintaining hundreds of KS files (normally one for each target server) > My situation is slightly complex as I have dozens of target networks. > > We derive network information from predefined perl hash tables, where we lookup: > - default route > - dns server > - target hostname (derived from DNS of the client's IP) > > I use this to generate a line such as the following in the default KS file: > > <-- snip > network --device eth0 --bootproto static --ip 192.168.1.50 --netmask 255.255.255.0 --gateway 192.168.1.1 --nameserver 192.168.1.2 --hostname angus > <-- snip > > Everything works great, until one of my guys rebuilds a server on some unknown network (ie not in the lookup table) > > When this occurs, anaconda is fed an incomplete kickstart file - it's default behaviour then is to prompt for input. > > The trouble I have with this is that we are usually rebuilding servers in remote locations without local hands! > So, we affectively lose this server until our next site visit (lights out is not configured across all our sites) > > Is there some option or means of persuading anaconda to reboot on failures, prior to any changes being committed to the system? (ie. prior to fdisk activity of Kickstart file) > > Thanks in advance > Angus > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -------------- next part -------------- An HTML attachment was scrubbed... URL: From tom at ng23.net Thu Mar 26 12:55:17 2009 From: tom at ng23.net (Tom Brown) Date: Thu, 26 Mar 2009 12:55:17 +0000 Subject: Anaconda default behaviour on error In-Reply-To: <402529803.2986111238065525460.JavaMail.root@unx-s-manc3> References: <402529803.2986111238065525460.JavaMail.root@unx-s-manc3> Message-ID: <49CB7B35.7080904@ng23.net> > > Is there some option or means of persuading anaconda to reboot on failures, prior to any changes being committed to the system? (ie. prior to fdisk activity of Kickstart file) > > use cobbler with koan and the -D option so that you can see what the kernel line will look like to the kickstar before you reboot the box so you can fix it and then reboot so avoid the machine not coming back up. From raqnerd at gmail.com Thu Mar 26 17:16:56 2009 From: raqnerd at gmail.com (Frank Kimbell) Date: Thu, 26 Mar 2009 18:16:56 +0100 Subject: NetworkManager? Message-ID: <846a2d7c0903261016h65f508f8o4988d667240f6629@mail.gmail.com> Does anyone know how to disable the NetworkManager in anaconda kickstarts? I've been reading through the python sources, but don't really find anything. Like Virginian said, I also turn off NetworkManager in a normal system, and let the network be handled by the init.d/network. I don't like NetworkManager. I turn that off using chkconfig off or del. Surely, I can turn this off in a kickstart somehow? Please advise. Thank you! From keith at karsites.net Thu Mar 26 19:08:18 2009 From: keith at karsites.net (Keith Roberts) Date: Thu, 26 Mar 2009 19:08:18 +0000 (GMT) Subject: NetworkManager? In-Reply-To: <846a2d7c0903261016h65f508f8o4988d667240f6629@mail.gmail.com> References: <846a2d7c0903261016h65f508f8o4988d667240f6629@mail.gmail.com> Message-ID: On Thu, 26 Mar 2009, Frank Kimbell wrote: > To: kickstart-list at redhat.com > From: Frank Kimbell > Subject: Re: NetworkManager? > > Does anyone know how to disable the NetworkManager in > anaconda kickstarts? I've been reading through the python > sources, but don't really find anything. > > Like Virginian said, I also turn off NetworkManager in a > normal system, and let the network be handled by the > init.d/network. I don't like NetworkManager. I turn that > off using chkconfig off or del. > > Surely, I can turn this off in a kickstart somehow? > > Please advise. > > Thank you! Well I don't think you want to turn off NM at the start of the installation. It's used to configure eth0 or whatever so you can do things like linux ks=http://server/ks.cfg But on F10 you can stop NM from being installed with: %packages ... ... ... # Don't install NetworkManager -NetworkManager %end Doesn't seem to break anything either :) Kind Regards, Keith Roberts ----------------------------------------------------------------- Websites: http://www.php-debuggers.net http://www.karsites.net http://www.raised-from-the-dead.org.uk All email addresses are challenge-response protected with TMDA [http://tmda.net] ----------------------------------------------------------------- From angus.clarke at gmail.com Fri Mar 27 09:58:40 2009 From: angus.clarke at gmail.com (Angus Clarke) Date: Fri, 27 Mar 2009 09:58:40 +0000 (GMT) Subject: Anaconda default behaviour on error In-Reply-To: <1423519052.3127901238147567476.JavaMail.root@unx-s-manc3> Message-ID: <2091057633.3128931238147920499.JavaMail.root@unx-s-manc3> Thanks Matt, We use stock packages that come with RHEL5.2. I suspect you are right with regard to modifying the behaviour - I guess I was hoping I had missed something and that someone would have the silver bullet fix to my issue :) Cheers Angus ----- Original Message ----- From: "Matt Rose" To: "Discussion list about Kickstart" Sent: Thursday, 26 March, 2009 13:50:42 GMT +01:00 Amsterdam / Berlin / Bern / Rome / Stockholm / Vienna Subject: Re: Anaconda default behaviour on error what version of anaconda are you using for this? I personally don't see a way of altering the kickstart behaviour, except by getting down and dirty in the kickstart.py source file of anaconda , but maybe someone else has some ideas. Matt Angus Clarke wrote: Hi I use CGI scripts to auto-generate kickstart files depending on the target's IP address - $ENV{REMOTE_ADDR} I do this to avoid maintaining hundreds of KS files (normally one for each target server) My situation is slightly complex as I have dozens of target networks. We derive network information from predefined perl hash tables, where we lookup: - default route - dns server - target hostname (derived from DNS of the client's IP) I use this to generate a line such as the following in the default KS file: <-- snip network --device eth0 --bootproto static --ip 192.168.1.50 --netmask 255.255.255.0 --gateway 192.168.1.1 --nameserver 192.168.1.2 --hostname angus <-- snip Everything works great, until one of my guys rebuilds a server on some unknown network (ie not in the lookup table) When this occurs, anaconda is fed an incomplete kickstart file - it's default behaviour then is to prompt for input. The trouble I have with this is that we are usually rebuilding servers in remote locations without local hands! So, we affectively lose this server until our next site visit (lights out is not configured across all our sites) Is there some option or means of persuading anaconda to reboot on failures, prior to any changes being committed to the system? (ie. prior to fdisk activity of Kickstart file) Thanks in advance Angus _______________________________________________ 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 angus.clarke at gmail.com Fri Mar 27 10:11:05 2009 From: angus.clarke at gmail.com (angus.clarke at gmail.com) Date: Fri, 27 Mar 2009 10:11:05 +0000 (GMT) Subject: Anaconda default behaviour on error In-Reply-To: <77547407.3130871238148521613.JavaMail.root@unx-s-manc3> Message-ID: <822605354.3131191238148665236.JavaMail.root@unx-s-manc3> Thanks Tom Cobbler looks interesting and is new to me! I really should peruse the Fedora website more often. Im not sure your suggestion direclty tackles my issue as these are blind remote installs - we only know there is a problem when we realise the machine is not available after the usual 20mins (or so) and then notice the lack of web logs that you would expect to be seen when serving RPM packages from the web server. Its always after the event that human error is noticed/acknowledged ;) I will certainly further investigate Cobbler in any case. On further digging, it seems the default behaviour of anaconda/kickstart method is not optionable (not a configuration parameter.) I can see that the stancepoint of prompting for user input on some error is probably safest for most situations - just not mine. Cheers Angus ----- Original Message ----- From: "Tom Brown" To: "Discussion list about Kickstart" Sent: Thursday, 26 March, 2009 13:55:17 GMT +01:00 Amsterdam / Berlin / Bern / Rome / Stockholm / Vienna Subject: Re: Anaconda default behaviour on error > > Is there some option or means of persuading anaconda to reboot on failures, prior to any changes being committed to the system? (ie. prior to fdisk activity of Kickstart file) > > use cobbler with koan and the -D option so that you can see what the kernel line will look like to the kickstar before you reboot the box so you can fix it and then reboot so avoid the machine not coming back up. _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From tom at ng23.net Fri Mar 27 10:29:40 2009 From: tom at ng23.net (Tom Brown) Date: Fri, 27 Mar 2009 10:29:40 +0000 Subject: Anaconda default behaviour on error In-Reply-To: <822605354.3131191238148665236.JavaMail.root@unx-s-manc3> References: <822605354.3131191238148665236.JavaMail.root@unx-s-manc3> Message-ID: <49CCAA94.7060900@ng23.net> > > Im not sure your suggestion direclty tackles my issue as these are blind remote installs - we only know there is a problem when we realise the machine is not available after the usual 20mins (or so) and then notice the lack of web logs that you would expect to be seen when serving RPM packages from the web server. Its always after the event that human error is noticed/acknowledged ;) > > I will certainly further investigate Cobbler in any case. > > it does in a roundabout manner as with koan you can see what the the kernel command line will look like BEFORE you reboot and therefore you'll be able to see if the network info that will be used during the kickstart is correct and therefore wether the machine will come back or not. As an example this is what a 'rebuild' kernel command line could look like from koan title kick1238146211 root (hd0,0) kernel /vmlinuz ro root=/dev/VolGroup00/LogVol00 acpi=False console=xvc0 ksdevice=eth0 lang= text ks=http://192.168.11.1/cblr/svc/op/ks/system/virt02 netmask=255.255.255.0 dns=192.168.11.1 ip=192.168.11.4 kssendmac gateway=192.168.11.1 as an example of course - cobbler is very very very good tom From kanarip at kanarip.com Mon Mar 30 11:59:42 2009 From: kanarip at kanarip.com (Jeroen van Meeuwen) Date: Mon, 30 Mar 2009 13:59:42 +0200 Subject: [Fwd: network step always skipped in kickstart.py] In-Reply-To: <49CA8802.2010806@n-able.com> References: <49CA35B1.9090509@n-able.com> <49CA76F6.5060407@redhat.com> <49CA8802.2010806@n-able.com> Message-ID: <49D0B42E.9080005@kanarip.com> On 03/25/2009 08:37 PM, Matt Rose wrote: > Mind you, I'm not sure whether my expectations are correct. > > The patch I attached "fixes" this behaviour so that the network > configuration dialog is brought up, if there is no "network" option in > the ks.cfg > Has your network been preconfigured in any way? I'm thinking ksdevice=, ip= (and friends) on the cmdline or something. In those cases, a kickstart that does not have a "network" line should not bring up the network configuration dialog. Kind regards, Jeroen van Meeuwen -kanarip From mrose at n-able.com Mon Mar 30 13:11:15 2009 From: mrose at n-able.com (Matt Rose) Date: Mon, 30 Mar 2009 09:11:15 -0400 Subject: [Fwd: network step always skipped in kickstart.py] In-Reply-To: <49D0B42E.9080005@kanarip.com> References: <49CA35B1.9090509@n-able.com> <49CA76F6.5060407@redhat.com> <49CA8802.2010806@n-able.com> <49D0B42E.9080005@kanarip.com> Message-ID: <49D0C4F3.9010706@n-able.com> First of all, thanks for responding, I was pretty sure that the issue had fallen into the void. It doesn't *really* matter to me, as I've fixed it in my source code, but I know that it's bitten other people, as it was originally brought to my attention in the centos IRC channel. That's what I thought at first, because my test platform uses kickstart and updates over http, so I figured that it was just as you suspected, but I tried it with both the kickstart burned onto CD, and the ks=floppy boot method, and both times, it failed to configure the network. There's one scenario left that I haven't tested, which is to boot up on a network without a DHCP server (I haven't tested it for a couple of reasons[1]), and see if it asks for a network configuration. However based on my reading of the anaconda code; the fact that kickstart.py *always* calls dispatch.skipStep("network"), and the commit message from a commit that was subsequently retracted in the anaconda git repo[2], I'm sure that, no matter what I do, if kickstart is invoked, it will not ask me to configure the network. I guess I have to open up an enhancement request on Fedora to get this fixed to behave as I think it should... 1. I'd have to set up a virtual network on a private LAN, which would take too much time for a test that I know would fail, and it's kinda beside the point, just because the use case I'm envisioning separates install-time network configuration from permanent network config 2. http://preview.tinyurl.com/dyv7y9 and the commit message states: This patch adds support for a new network bootproto. The point of this is to work around our basic assumption that no network line in the kickstart file means you get dhcp. Some environments may want the assumption that no network line means you get prompted for network configuration. That's what this patch adds. and the commit is reverted two weeks later. After reading the bug referenced the commit msg, I kinda understand what happened. (It's a confusing bug). Jeroen van Meeuwen wrote: > On 03/25/2009 08:37 PM, Matt Rose wrote: > >> Mind you, I'm not sure whether my expectations are correct. >> >> The patch I attached "fixes" this behaviour so that the network >> configuration dialog is brought up, if there is no "network" option in >> the ks.cfg >> >> > > Has your network been preconfigured in any way? I'm thinking ksdevice=, > ip= (and friends) on the cmdline or something. > > In those cases, a kickstart that does not have a "network" line should > not bring up the network configuration dialog. > > Kind regards, > > Jeroen van Meeuwen > -kanarip > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -------------- next part -------------- An HTML attachment was scrubbed... URL: From steven.hajducko at digitalinsight.com Mon Mar 30 17:08:22 2009 From: steven.hajducko at digitalinsight.com (Hajducko, Steven) Date: Mon, 30 Mar 2009 10:08:22 -0700 Subject: Scripting the url --url install method? Message-ID: <21F8C5D33731F44BB9DCC396A12F9F2110A32D15@WLVEXM01.corp.ad.diginsite.com> Is it possible to use a %pre script in any way to customize the install method? I'm attempting to create a kickstart file that we can move around our network, however, some of our networks require different URL's to grab the ks file and to find the install media. What I have currently isn't working as anaconda doesn't seem to like that it can't find the install method before it runs the %pre section. Here is what I have currently - install %include /tmp/url ... %pre # Append network hostname as some tiers don't have DNS - if we don't have DNS # then the hostname becomes 'localhost' when using the network information from the # boot line HOSTNAME=$( cat /proc/cmdline | sed -e s'/ /\n/g' | grep '^hostname=' ) echo "network --${HOSTNAME}" > /tmp/hostname # Get our architecture ARCH=$(uname -i) # Figure out the kickstart URL from the boot line so we can find the install media KSURL=$(cat /proc/cmdline | sed -e 's/ /\n/g' | grep '^ks=' | cut -d= -f2 | cut -d/ -f1,2,3) echo "url --url=${KSURL}/yum/rhel5-server-${ARCH}/disc1" > /tmp/url ... We boot from the DVD and enter the network information on the linux: boot command, so ultimately our boot line looks something like: linux ks=http://yum.example.com:9300/ks/hp-rhel5.cfg hostname=test1.example.com ip=1.1.1.2 netmask=255.255.255.0 gateway=1.1.1.1 dns=1.1.1.3 I don't see any errors when viewing the logs - it just prompts me for the install method. I've also tried something like the following in hopes that when it re-read the kickstart file, the URL line existed and it would use that instead - unfortunately that didn't work either. install %include /tmp/url cdrom Any help would be greatly appreciated. ..................................................................... Steven Hajducko Systems Engineer, Systems Integration w: 818.597.6443 m: 805.377.9074 This email may contain confidential and privileged material for the sole use of the intended recipient. Any review or distribution by others is strictly prohibited. If you are not the intended recipient please contact the sender and delete all copies. From kpowell at redhat.com Mon Mar 30 19:16:20 2009 From: kpowell at redhat.com (Kyle Powell) Date: Mon, 30 Mar 2009 15:16:20 -0400 Subject: Scripting the url --url install method? In-Reply-To: <21F8C5D33731F44BB9DCC396A12F9F2110A32D15@WLVEXM01.corp.ad.diginsite.com> References: <21F8C5D33731F44BB9DCC396A12F9F2110A32D15@WLVEXM01.corp.ad.diginsite.com> Message-ID: <49D11A84.3050403@redhat.com> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hajducko, Steven wrote: > Is it possible to use a %pre script in any way to customize the install > method? No, anaconda wants the url to be in the ks.cfg when it's read. > We boot from the DVD and enter the network information on the linux: boot > command, so ultimately our boot line looks something like: > > linux ks=http://yum.example.com:9300/ks/hp-rhel5.cfg hostname=test1.example.com ip=1.1.1.2 netmask=255.255.255.0 gateway=1.1.1.1 dns=1.1.1.3 > > Any help would be greatly appreciated. You could use a cgi wrapper to insert the proper hostname in the kickstart file as it's sent to the client. linux ks=http://yum.example.com:9300/cgi-bin/ks?ks=ks/hp-rhel5.cfg hostname=... /var/www/cgi-bin/ks (Off the top of my head, I haven't tested this): #!/usr/bin/perl use strict; use warnings; use CGI qw/:standard/; use LWP::Simple; print "Content-type: text/plain\n\n"; if (url_param('ks')) { my $ks_file = url_param('ks'); $ks_file =~ s/^\/*//; my $url = "http://127.0.0.1:9300/$ks_file"; my $content = get($url); if ($content) { my $hostname = `hostname -f`; chomp $hostname; # Here is where you would mangle the url as you wish # I'm simply changing the hostname $content =~ s,^\s*url --url http://[^/]*,url --url http://$hostname:9300,m; print $content; } } The kickstart file could have anything for the hostname in the --url line: url --url http://THIS_WILL_BE_CHANGED/path/to/files and the ks cgi would change to line to: url --url http://yum.example.com:9300/path/to/files as it's sent to the client. You could change the ks cgi to mangle the url further in order to change the path or port or whatever. I hardcoded the port above, but if you use different ports you could make the cgi script detect on which port the request came in and set that port in the output url as well. HTH. - -- Kyle Powell | Red Hat | Senior Consultant, RHCE | Cell: 571.215.4340 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iD8DBQFJ0RqE7pTtanQdBU4RAl0lAJ9t9TVGNrlaQhfj+APgvSM1zD3MQQCfeYar rOozIZ0DjRqWxc6d/zE7pSY= =HQq4 -----END PGP SIGNATURE----- From hellfrozeover at gmail.com Tue Mar 31 21:47:15 2009 From: hellfrozeover at gmail.com (Tom H) Date: Tue, 31 Mar 2009 22:47:15 +0100 Subject: Installing a system to a directory only Message-ID: <970066760903311447qc18caedw6890c2bb2a40898@mail.gmail.com> Hi all, I'm trying to use anaconda to build system images for Amazon's EC2 service. For this, you need to build a standard Linux system, then run a tool to package up the system into several files. The whole image building and packaging process is a bit adhoc, so I'm trying to automate it. My idea is to start with a kickstart file, then have anaconda build the system into a directory (rather than build a system from scratch), then run the packager on that directory only. To summarise: kickstart file -> [anaconda install] -> full system image in a directory -> [EC2 packager] -> final system image Looking at various websites, the --rootpath option seems to be the answer. On a CentOS 5.2 i386 machine, with the 11.1.2.113-1.el5.centos.2 version of anaconda installed, I tried this: mkdir -p /mnt/runtime/usr/sbin/ LANG=en /usr/sbin/anaconda -T -m http://mirrors.dedipower.com/centos/5.2/os/i386/ --rootpath=/root/buildroot The installer displays the text menu, asks for the early options, but then dies just after the "Checking dependencies in packages selected for installation..." step: Traceback (most recent call first): File "/usr/lib/anaconda/urlinstall.py", line 95, in systemMounted self.loopbackFile = "%s%s%s" % (chroot, File "/usr/lib/anaconda/yuminstall.py", line 1328, in doPreInstall if self.method.systemMounted (anaconda.id.fsset, anaconda.rootPath): File "/usr/lib/anaconda/backend.py", line 180, in doPreInstall anaconda.backend.doPreInstall(anaconda) ++ File "/usr/lib/anaconda/dispatch.py", line 201, in moveStep rc = stepFunc(self.anaconda) File "/usr/lib/anaconda/dispatch.py", line 124, in gotoNext self.moveStep() File "/usr/lib/anaconda/text.py", line 588, in run anaconda.dispatch.gotoNext() File "/usr/sbin/anaconda", line 982, in ? anaconda.intf.run(anaconda) IndexError: list index out of range The last statement run is: self.loopbackFile = "%s%s%s" % (chroot, fsset.filesystemSpace(chroot)[0][0], "/rhinstall-stage2.img") With this being the problem: (Pdb) print fsset.filesystemSpace(chroot) [] The same problem has been noted before: http://www.linux-archive.org/centos/101098-anaconda-test-problem.html http://www.linuxsir.org/bbs/thread330760.html Does anyone have any ideas? Tom From jason at rampaginggeek.com Tue Mar 31 22:44:10 2009 From: jason at rampaginggeek.com (Jason Edgecombe) Date: Tue, 31 Mar 2009 18:44:10 -0400 Subject: Installing a system to a directory only In-Reply-To: <970066760903311447qc18caedw6890c2bb2a40898@mail.gmail.com> References: <970066760903311447qc18caedw6890c2bb2a40898@mail.gmail.com> Message-ID: <49D29CBA.9020709@rampaginggeek.com> Tom H wrote: > Hi all, > > I'm trying to use anaconda to build system images for Amazon's EC2 > service. For this, you need to build a standard Linux system, then run > a tool to package up the system into several files. > > The whole image building and packaging process is a bit adhoc, so I'm > trying to automate it. My idea is to start with a kickstart file, then > have anaconda build the system into a directory (rather than build a > system from scratch), then run the packager on that directory only. To > summarise: > > kickstart file -> [anaconda install] -> full system image in a > directory -> [EC2 packager] -> final system image > > Looking at various websites, the --rootpath option seems to be the answer. > > On a CentOS 5.2 i386 machine, with the 11.1.2.113-1.el5.centos.2 > version of anaconda installed, I tried this: > > mkdir -p /mnt/runtime/usr/sbin/ > LANG=en /usr/sbin/anaconda -T -m > http://mirrors.dedipower.com/centos/5.2/os/i386/ > --rootpath=/root/buildroot > > The installer displays the text menu, asks for the early options, but > then dies just after the "Checking dependencies in packages selected > for installation..." step: > > Traceback (most recent call first): > File "/usr/lib/anaconda/urlinstall.py", line 95, in systemMounted > self.loopbackFile = "%s%s%s" % (chroot, > File "/usr/lib/anaconda/yuminstall.py", line 1328, in doPreInstall > if self.method.systemMounted (anaconda.id.fsset, anaconda.rootPath): > File "/usr/lib/anaconda/backend.py", line 180, in doPreInstall > anaconda.backend.doPreInstall(anaconda) > ++ File "/usr/lib/anaconda/dispatch.py", line 201, in moveStep > rc = stepFunc(self.anaconda) > File "/usr/lib/anaconda/dispatch.py", line 124, in gotoNext > self.moveStep() > File "/usr/lib/anaconda/text.py", line 588, in run > anaconda.dispatch.gotoNext() > File "/usr/sbin/anaconda", line 982, in ? > anaconda.intf.run(anaconda) > IndexError: list index out of range > > The last statement run is: > self.loopbackFile = "%s%s%s" % (chroot, > fsset.filesystemSpace(chroot)[0][0], > "/rhinstall-stage2.img") > > With this being the problem: > (Pdb) print fsset.filesystemSpace(chroot) > [] > > The same problem has been noted before: > > http://www.linux-archive.org/centos/101098-anaconda-test-problem.html > http://www.linuxsir.org/bbs/thread330760.html > > Does anyone have any ideas? > Tom, I think that anaconda is the wrong tool for your task. I think you should look at using the "yum" command with the "--installroot" option. What you want sounds similar to how a livecd is built. Take a look at this for an example: http://www.centos.org/modules/newbb/viewtopic.php?topic_id=18293 You also might want to look at revisor. It's a tool to automate the creation of livecd's. http://revisor.fedoraunity.org/ Jason