fsck errors on a running system that disappear in single-user mode

[Kris Kennaway]

On Fri, Mar 11, 2005 at 06:37:16PM +0100, Anthony Atkielski wrote:
> Jerry McAllister writes:
> 
> > At least most likely.    There could possibly a real inconsistency,
> > but it would be difficult on a running system to distinguish it 
> > from the artifacts you point out above.
> 
> But a real consistency would presumably persist with an fsck on a system
> in single-user mode, which is (I assume) quiescent, right?

Yes (single-user mode isn't the issue; the filesystem needs to be
unmounted or mounted read-only, or to be using UFS snapshots (in 5.x)
in order for fsck to have a consistent snapshot of the data that
doesn't change out from underneath it).

Kris