Re: reboot/halt fails to unmount /stop filesystem translators?

[David Walter]

Grant Bowman <grantbow@grantbow.com> writes:

> * Marcus Brinkmann <Marcus.Brinkmann@ruhr-uni-bochum.de> [020512 11:34]:
> > On Sun, May 12, 2002 at 01:54:01AM -0400, David Walter wrote:
> > > When rebooting the HURD I haven't found a dependable way to ensure
> > > that it doesn't come back up with unclean filesystems.
> > [...]
> > 
> > Note, the filesystem must be clean to begin with.  So do one e2fsck (for
> > example from a parallel system, or in read only single user mode), and then
> > reboot into single user mode, login, reboot.  If it was clean and then
> > becomes untidy, there is a bug.
> 
> This was happening very consistently for me when running from the libio
> tarball.  I upgraded the glibc and gcc package and it seems fine now.  I
> am happy to be unable to reproduce this problem now.

To followup on the help that I got and to clarify.

I actually had an fstab file in /etc.

the problem in my case turned out to be that the pass option was zero.

was:

# <file system> <mount point>   <type>  <options>               <dump>  <pass>
/dev/hd0s12     /               ext2    rw                      1       0

is:
# <file system> <mount point>   <type>  <options>               <dump>  <pass>
/dev/hd0s12     /               ext2    rw                      1       1

I suppose the  problem revolves around the dependency  on an mtab, but
mtab  isn't  quite there yet   for the  HURD,  AFAIK?  Is this a  fair
assumption?

Thanks so much for the help though. Much better behaviour.

-- 
Hope springs eternal!

/^\
\ /     ASCII RIBBON CAMPAIGN
 X        AGAINST HTML MAIL
/ \