I am trying using a kickstart file to install CentOS 6.2 into a new virtual machine (the MBR sector of the harddisk is all zero), however I found that the installer cannot go through the harddisk partition. It failed after I chose "Use All Space" at harddisk partition option. The error message is: http://anony.ws/i/bMcTJ.png
"You have not defined a root partition (/), which is required for installation of CentOS to continue. You have not created a/boot/efi partition (note: I am using BIOS, not (U)EFI) This can happen if there is not enough space on your harddrive(s) for installation."
However, if I create a blank MBR partition table before before CentOS installation, then there is no problem. The same kickstart file works for CentOS 6.1 with blank harddisk. (with url parameter changed of course)
Below is the kickstart file I used (between dash lines)
------- url --url="http://ftp.twaren.net/Linux/CentOS/6.2/os/i386/" interactive timezone Asia/Hong_Kong firstboot --enable -------
Steps to reproduce: 1. create a new virtual machine with blank harddisk image (or, a real blank harddisk) 2. boot the netinstall iso 3. at boot menu, press tab and append ks=<kickstart_file_location> and press enter to boot 4. go through the boot option as usual until harddisk partition options 5. Choose "Use All Space" at harddisk partition options 6. error occurs
On Wed, Jan 4, 2012 at 07:25, lee_yiu_chung@yahoo.com lee_yiu_chung@yahoo.com wrote:
Below is the kickstart file I used (between dash lines)
url --url="http://ftp.twaren.net/Linux/CentOS/6.2/os/i386/" interactive timezone Asia/Hong_Kong firstboot --enable
clearpart --all --initlabel
於 2012/1/4 下午 11:21, John Broome 提到:
clearpart --all --initlabel
In fact I already tried this before sending this email, and it doesn't work. BTW, I don't need this option to install CentOS 6.1. This simply suppress the "unknown partition table format" warning before the GUI installer starts.
----- Original Message ----- | I am trying using a kickstart file to install CentOS 6.2 into a new | virtual machine (the MBR sector | of the harddisk is all zero), however I found that the installer | cannot go through the harddisk | partition. It failed after I chose "Use All Space" at harddisk | partition option. | The error message is: | http://anony.ws/i/bMcTJ.png | | "You have not defined a root partition (/), which is required for | installation of CentOS to continue. | You have not created a/boot/efi partition (note: I am using BIOS, not | (U)EFI) | This can happen if there is not enough space on your harddrive(s) for | installation." | | However, if I create a blank MBR partition table before before CentOS | installation, then there is no | problem. The same kickstart file works for CentOS 6.1 with blank | harddisk. (with url parameter | changed of course) | | Below is the kickstart file I used (between dash lines) | | ------- | url --url="http://ftp.twaren.net/Linux/CentOS/6.2/os/i386/" | interactive | timezone Asia/Hong_Kong | firstboot --enable | ------- | | Steps to reproduce: | 1. create a new virtual machine with blank harddisk image (or, a real | blank harddisk) | 2. boot the netinstall iso | 3. at boot menu, press tab and append ks=<kickstart_file_location> and | press enter to boot | 4. go through the boot option as usual until harddisk partition | options | 5. Choose "Use All Space" at harddisk partition options | 6. error occurs | _______________________________________________ | CentOS mailing list | CentOS@centos.org | http://lists.centos.org/mailman/listinfo/centos
clearpart --all --initlabel
part /boot --fstype=ext4 --size=1024 part pv.0 --grow --size=1 volgroup ROOTDISK --pesize=4096 pv.0 logvol swap --name=swap --vgname=ROOTDISK --recommended logvol / --fstype=ext4 --name=root --vgname=ROOTDISK --size=1 --grow
They that. Should work for you. Feel free to adjust as necessary
clearpart --all --initlabel
part /boot --fstype=ext4 --size=1024 part pv.0 --grow --size=1 volgroup ROOTDISK --pesize=4096 pv.0 logvol swap --name=swap --vgname=ROOTDISK --recommended logvol / --fstype=ext4 --name=root --vgname=ROOTDISK --size=1 --grow
Thanks for your suggestion, but I just wished to keep the kickstart file bare minimum. It is intended to save my little trouble in typing installation repository path and choosing timezones only and do nothing else. All other options would be asked interactively, including disk partition. (timezone and installation repository are the only options that are absolutely the same in every machine in a local environment :) )
Since this problem doesn't occur in 6.1, is it possible to be a installer bug in 6.2?
於 2012/1/4 下午 08:25, lee_yiu_chung@yahoo.com 提到:
I am trying using a kickstart file to install CentOS 6.2 into a new virtual machine (the MBR sector of the harddisk is all zero), however I found that the installer cannot go through the harddisk partition. It failed after I chose "Use All Space" at harddisk partition option. The error message is: http://anony.ws/i/bMcTJ.png
"You have not defined a root partition (/), which is required for installation of CentOS to continue. You have not created a/boot/efi partition (note: I am using BIOS, not (U)EFI) This can happen if there is not enough space on your harddrive(s) for installation."
However, if I create a blank MBR partition table before before CentOS installation, then there is no problem. The same kickstart file works for CentOS 6.1 with blank harddisk. (with url parameter changed of course)
Below is the kickstart file I used (between dash lines)
url --url="http://ftp.twaren.net/Linux/CentOS/6.2/os/i386/" interactive timezone Asia/Hong_Kong firstboot --enable
Steps to reproduce:
- create a new virtual machine with blank harddisk image (or, a real blank harddisk)
- boot the netinstall iso
- at boot menu, press tab and append ks=<kickstart_file_location> and press enter to boot
- go through the boot option as usual until harddisk partition options
- Choose "Use All Space" at harddisk partition options
- error occurs
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
I just tested with equivalent upstream version, and confirmed the same bug occurred in upstream. I opened a bugzilla ticket for this. https://bugzilla.redhat.com/show_bug.cgi?id=771806