clone 428399 -1 tags 428399 -patch stop On Sat, Jul 28, 2007 at 06:21:55PM +0200, Václav J??za wrote: > tags 428399 +patch > tags 428399 -moreinfo > thanks > Hi, > I had the problem, which ended up with the same message, but wasn't > caused by dirty first sector of ext3 (I use xfs):
so please file a new bug and don't post to unrelated bug reports. keeping cc for this message, but on next reply please take care to post to the cloned bug report, thanks! > The problem was in two things: > 1) init/local, in function get_fstype, contains: this function is cleaned up in latest git see http://git.debian.org/?p=kernel/initramfs-tools.git > 2) I had old version of klibc-utils (1.4.xx). When I tried "fstype /dev/sda8" > in the busybox shell, it ended up > with "Invalid seek" error. Then ROOTFSTYPE was set up to unknown, and that > caused mount command to fail. > > In my case then, even if I had working vol_id, this broken utility fstype was > called instead. > > To fix the problem, I suggest (patch attached): > 1) switch if and else branch above no fstype supports *more* fs than vol_id, we had a bunch of failures due to it, it is *not* ready. and actually far from ready as mkfs.ext3 will create a partition vol_id will happily identify as vfat. > 2) bump dependency on klibc-utils your klibc-utils was pre-etch release, but ok i'll pump the version to the one in etch. not as much as your suggestion. regards -- maks
