[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#351768: Installation report: lilo fails with "mapped boot device cannot be on multiple real devices" (LVM2)



Package: debian-installer
Version: 20051111

Hi,

I can't find any reference to this anywhere in the BTS. I am currently downloading the latest daily image in case it has been fixed there. Will follow this up tomorrow (or tonight, if successful). 

My system is as follows: 

hda: IDE HD (3.6Gb)
hdb: CDROM
hdc: IDE HD (3.0Gb)

hda1 and hdc1 form LVM2 volume group VG1 (6.6Gb)

VG1 contains 4 logical volumes: 

VG1-Root (2.7Gb)  (this is a partition I'm trying to rescue with a new install, but that's another story)
VG1-Swap (128Mb)
VG1-Swap2 (128Mb)
VG1-NewRoot (1.5Gb)

I booted from the netinst CD, chose expert mode. All stages went ok - found the hardware and the LVs. I chose to install on VG1-NewRoot after formatting it. Base system installed fine, kernel 2.6.12 installed fine, initrd built fine. The option to install grub was removed after installing the base system (on the Alt-F4 log there was a note to that effect, something like "root is on LVM, cannot install grub"). 

So I chose lilo, and it failed. Checking Alt-F4 showed the error "mapped boot device cannot be on multiple real devices". I can only find two references to this on google (none in the Debian mailing list archives or BTS), both in old patches from times when lilo didn't yet work with libdevmapper: 

https://www.redhat.com/archives/linux-lvm/2002-December/msg00140.html

and

http://evms.sourceforge.net/patches/2.4.0/lilo-22.6-devmapper.patch

I checked the system using the Alt-F2 console. The /target/etc/lilo.conf looked fine (boot=/dev/hda, root=/dev/mapper/VG1-NewRoot). /target/dev/mapper looks like this

VG1-Root	254 0
VG1-Swap	254 1
VG1-Swap2	254 2
VG1-NewRoot	254 3

... and /proc/devices shows block device 254 as device-mapper.

So, am I doing something wrong? Kernel is 2.6.12, the one on the 20051111 installer image. 

Regards,

CC



Reply to: