Bugzilla – Bug 530833
lvm support in mkinitrd broken
Last modified: 2009-11-13 07:55:08 UTC
A system using lvm currently can't boot. This is because /sbin/vgscan is missing in the initrd because it's not mentioned in the #%programs line of /lib/mkinitrd/scripts/boot-lvm2.sh. In fact the latest sources for mkinitrd don't contain even a trace of vgscan/vgchange. How is lvm support supposed to function?
Would you care to clarify which version you are referring to ? It might be a duplicate of bnc#505670 ?
This is factory M5 and it definitely won't boot so something is missing. If you're in the suse network you can check yourself, the machine is paradies.suse.de.
hello, I've got a problem that sounds like this one When I upgrade from ms5 to ms6 and a kernel update is installed after the reboot it fails to boot. This happens after each kernel update, I fix this by booting a suse 11.1 dvd *I run x86_64 and the ms5 dvd image did not fit on the disk therefore I used the 11.1 version *because I did not want to wait for the long dvd download for ms-6 I used the 11.1 version again for the latest update. my partitions /dev/sda1 boot (ext3) /dev/sda2 lvm (system) /dev/system/root (reiserfs) /dev/system/home (reiserfs) after booting to rescue mode I fix the initrd image through the following steps mount /dev/system/root /mnt mount /dev/sda1 /mnt/boot mount --bind /dev /mnt/dev chroot /mnt mount /proc mount /sys mkinitrd reboot my system and the system works again I am not sure if this because of the lack of lvm or the lack of reiserfs support in the mkinitrd image
oops ignore my entry...... wrong bug
I copied the vmlinuz and initrd file from paradise.suse.de, and it boots fine on my laptop. Is there any other steps needed to reproduce this issue ?
Not that I know. See if anything is missing from paradies.suse.de, but every time I have to (re)boot I have to go into the shell for vgscan and vgchange in order to continue booting.
May be this is related to #544361 or #525237 (which are most likely duplicates) pth@ please check
Created attachment 322955 [details] bad initrd-2.6.31.3-1-default
Created attachment 322956 [details] y2logs created on first boot post-fresh install from Factory http I think my problem is this bug. Yesterday I did zypper dup on a mixed regular & LVM VIA/Socket A system last dup'd from Factory about the time the final 11.1RC was announced ready for GM.I had no problem booting the new kernel/initrd 2.6.31.3-pae. However, the system was still using legacy IDE instead of libata. The dup filled the / volume to 100%, so I backed up selected files and directories to the /home volume, booted Fedora, reformatted the Factory / volume with b1024 J size=400 I128 to improve space efficiency, copied the backups back, which included the new pae kernel/initrd & /lib/modules, and did a minimalist install. Install used the default kernel instead of pae. First boot failed, with a message scrolling the screen too fast too read, something about not finding any VolGroup. It's up now on the yesterday's pae initrd, but using IDE instead of LIBATA. I have no idea what to do to try to fix this. Last time I tried to do a manual initrd conversion from IDE to LIBATA I got nowhere, with no help from the Factory mailing list or Factory IRC.
Created attachment 322957 [details] approximate partition structure
Close it as a duplicate of #505670 *** This bug has been marked as a duplicate of bug 505670 ***