Re: Syncer failure after fsck on all partitions

> From: Kent Stewart <kstewart_at_owt.com>
> Date: Mon, 1 Sep 2003 17:52:13 -0700
> 
> On Monday 01 September 2003 05:38 pm, Kevin Oberman wrote:
> > This is really sort of two problems, but I'll focus on the primary
> > concern here and maybe the more basic (but less significant) problem
> > will get addressed at some point. (Or I'll PR it.)
> >
> > I have noticed that my stable system never seems to leave the
> > filesystems clean after a crash on CURRENT. I get a message that
> > background FSCK is starting in 60 seconds, but I see no indication
> > that it is ever run. All partitions are UFS and have soft-update.
> >
> > After I notice that re-boots are reporting all ufs partitions as
> > unclean, I shut down and boot standalone. I then do "fsck -y" on all
> > partitions. Many simply mark the partition clean with no fixes while
> > others do find the typical problems and fix them.
> >
> > Finally I do a reboot. No partition has ever been mounted except root
> > which is read-only. Syncer reports 1 or more buffers remain and
> > reports this until it gives up.
> >
> > After the reboot, all partitions mount cleanly and all seems normal.
> >
> > I really feel uncomfortable about this and can't understand why the
> > syncer should have any buffers to handle when no partition has been
> > mounted RW. Do I need to worry?
> 
> Did you fsck from single user mode? There were problems a while back 
> where you had to check the file system in the foreground mode.

I only fsck from standalone. I have never thought that a full fsck on a
mounted filesystem was safe. As I said, all disks were unmounted
except root which was still mounted read-only.

I see that Andy Farkas reported the same issue. (Sorry, but I missed
it.) 
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman_at_es.net			Phone: +1 510 486-8634