Just Launched: You can now import projects and releases from Google Code onto SourceForge
We are excited to release new functionality to enable a 1-click import from Google Code onto the Allura platform on SourceForge. You can import tickets, wikis, source, releases, and more with a few simple steps. Read More
From: Henry Nestler <Henry.Ne@Arcor.de> - 2006-02-17 13:14:35
Mark G. Woodruff wrote:
> Just upgraded from 0.6.2 (gentoo 2.6.10) to 0.6.3
> (gentoo 2.6.11) and I'm having a very odd problem.
> Every other time I start colinux, it refuses to boot,
> saying it can't read the root drive. If I run the
> colinux-daemon again, it boots and runs fine. If I
> change my colinux.xml file to use the old disk image,
> it boots and runs fine as well, although with the slow
> console bug.
> <initrd path="initrd.gz" />
the problem exist, mostly on second boot. I had it only in a time, if I
used my SuSE 9.0 partion. For me the 'umount ... sleep 1 ... sync ...
sleep 1' helps. I'm not know, what this is.
Perhaps is an old 'umount' version in busybox?
You sayed, the the 'chroot /mnt/linux mount -a' is needed for some
users, that have /lib/modles on separate partion.
I hate the 'mount -a' in initrd scripts! This give me many troubles,
because it mounts my 'nfs' and 'cofs'. To solve this I have added my
fstab options to 'user' for this mount points.
Problem I see here is, that the rootfs is double mounts. Ones for
/mnt/linux and the second in chroot as '/'. In the 'umount' it is a
next problem. The kernel daemon for jurnaling need some more time for
update, before you can umount the previous level.
We should better grep for '/lib' and '/lib/modules' and only mount this
stuff, and not alls, and please not the rootfs '/' again.