--nextPart2422088.co9N6hHnLq
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

On Tuesday 29 March 2005 11:13, Karsten Kruse wrote:
> I guess that your writeback cache is active. In that case softdep can not
> safe you from powerfailure because the disk writes when it's in the mood,
> not when it got the command to write.

The disk in question (at the time, I no longer own this configuration) is a=
n=20
external RAID controller, with external disks.  The controller nor the disk=
s=20
were power cycled.

With softdep, I lost the filesystem 8/10 times.  Without softdep, I lost it=
=20
0/10 times.  This was done by hitting the reset button mid-compile=20
in /usr/src, at random points.  Usually, this was due to a MP-related locku=
p,=20
but not always.

> But even in this case a fs with softdep is safer than one without- It
> safes you when you accidently press reset or a kernel-panic occurs (the
> dirty stuff in the writeback cache can be written to disk as long as you
> have power).
>
> Performance is a bonus, i think the safety alone is worth it.

I don't understand how you can conclude that delaying writes to disks is sa=
fer=20
in any way.  softdep is supposed to order those writes and delay them in=20
order to do that ordering, and to make it so reliable systems are faster. =
=20
However, it seems to be failing in the error case -- and that error case is=
=20
not uncommon when tracking -current.

=2D-Michael

--nextPart2422088.co9N6hHnLq
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (NetBSD)

iD8DBQBCSYhDl6Nz7kJWYWYRAmgUAJ0XZXEPeBZ2nStxXEe8BKY4bQ+pYwCfUGyK
6iZaDTA6yB7fgNeEygGprfM=
=QATA
-----END PGP SIGNATURE-----

--nextPart2422088.co9N6hHnLq--