--zkS4nBu7xsy2hAaK
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Aug 22, 2005 at 01:47:15AM +0200, Hubert Feyrer wrote:
> On Mon, 22 Aug 2005, Daniel Carosone wrote:
> >Looking at this made me realise one 'problem' to be avoided in future
> >work on wedges if they're named like this: presented with an fstab
> >listing filesystems on /dev/dk* like this, fsck will try to check
> >multiple partitions of the same physical volume in parallel
> >competition with eachother.
>=20
> Isn't that what the last column of fstab is good for sorting out?

Unless it was changed recently, that only pays attention to 0,1,n;
listing something at 5 won't make it run later than something at 2 or
3.  The above might be a motivation for 'fixing' that behaviour, but
something smarter would be nice.

There was some discussion about a general interface to query wedge
parentage, having fsck build up a dependency tree back to physical
devices would be a great use for this information, especially if it
was supported by other stacking logical block drivers (raid, cgd, etc)
so that fsck could see through them to physical spindles.

--
Dan.

--zkS4nBu7xsy2hAaK
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (NetBSD)

iD8DBQFDCRUiEAVxvV4N66cRApQzAJ9wynaHiMiba0eEtFIwLxDjwxqbsQCg8DBP
n5z1GNI5vXkhtnyCxitrHOE=
=pg/i
-----END PGP SIGNATURE-----

--zkS4nBu7xsy2hAaK--