Bugzilla – Bug 619837
Boot fails after installation using autoyast on raid-1
Last modified: 2010-10-05 15:07:25 UTC
Created attachment 373734 [details] menu.lst and y2log after installation, rules and classes directories User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.9) Gecko/20100317 SUSE/3.5.9-0.1.1 Firefox/3.5.9 I use autoyast to create 3 raid-1 sets for /, swap and /data. The file /boot/grub/menu.lst contains bad lines like kernel (hd0,681)/boot/vmlinuz... where the 681 seems to be a random number. The system will boot after editing with grub to (hd0,0) for kernel and initrd (and maybe replacing resume=... with noresume, I am not yet sure with that). Reproducible: Always
Created attachment 373823 [details] Y2logs
Created attachment 373825 [details] menu.lst with incorrect root Has (hd0,5) as root device. Should be (hd0,0). Also, should the kernel root path read "root=/dev/md0" as in previous installs like 11.2?
I hit this one too. Seems like this issue re-appears every other release. In my case, the root device was set for (hd0,5) when it should have been (hd0,0). My files are attached. Please ask if you need additional info.
I performed another install and found that the boot setup on the installation summary screen is incorrect. There are two lines that read Boot from MBR enabled. 'disable' Boot from / is disabled. 'enable' If I enable "boot from /" the installation succeeds and grub is set up correctly. This should be enabled by default.
It seems that there is problem with perl-Bootloader and detection of "/" partition. Josef please check it.
The problem seems to go away when I define a <bootloader> section in my autoyast control files. menu.lst then contains a separate line root (hd0,0), and the kernel line has no hd() part. I need the bootloader section anyway, so for me, this part of the problem seems solved.
I have to correct my previous post, maybe it was just luck. Another test gave me a menu.lst with "kernel (hd 0,3)/boot ...", should have been (hd 0,0). The same file contains "root (hd0,0)" for the section defined in the xml file, which is correct.
Hi, it is related to combination of software raid, 11.3 and persistent device names. It is already fixed in update channel and should not appear in 11.4 (C) :) *** This bug has been marked as a duplicate of bug 626132 ***