Stefan Monnier wrote:
>
> > Stefan Monnier wrote:
> > >
> > > >>>>> "H" == H Peter Anvin <[EMAIL PROTECTED]> writes:
> > > > .. but has the same problems as direct mounts (effectively, a
> > > > sub-automount *is* a direct mount.)
> > >
> > > Really ?
> > > What if you restrict yourself to cases where the exported areas can
> > > not be nested (i.e. as is the case with Linux's knsfd but not unfsd) ?
> > > Then you can have:
> > >
> > > /net indirect automount mount point
> > > /net/hostA indirect automount mount point
> > > /net/hostA/mount indirect automount mount point
> > > /net/hostA/mount/point2 NFS mount
> > >
> > > Look ma: no direct mount.
> > >
> > > Stefan
> >
> > Doesn't work that way. You'd lose any other entries in
> > /new/hostA/mount.
>
> I'm not sure I understand. What entries ?
> Remember I said "no nested exported areas": if /mount/point2
> is exported, then /mount is not, hence /net/hostA/mount can only
> contain other mount points.
> What am I missing apart from the explicit restriction I added upfront ?
>
Because you get no guarantee that /mount/point2 isn't on a separate
filesystem. You probably *CAN* do this if NONE of the mount points are
hierarchially related, but that makes it pretty useless.
--
<[EMAIL PROTECTED]> at work, <[EMAIL PROTECTED]> in private!
"Unix gives you enough rope to shoot yourself in the foot."
