[OpenAFS] cache performance

[Thomas Mueller]

On Thu, 24 Oct 2002, Nickolai Zeldovich wrote:

> > What do you think about using the mount options "noatime,data=3Dwriteba=
ck"=20
> > when using an ext3 filesystem for the AFS cache on linux systems.
> > I think this could increase the performance on a heavily used cache.
>=20
> The "noatime" option will not make any difference -- the kernel module
> already sets the NOATIME bit on all the cache file inodes.  I don't know
> for sure what the "data=3Dwriteback" option does, but guessing that it ma=
kes
> the buffer cache a write-back one, sure, it may increase your performance=
=2E
> Though I had the impression that Linux already did writeback caching..  O=
r
> maybe that was only for ext2, and ext3 enforces more traditional semantic=
s?

The default for ext3 is "data=3Dordered", which means=20
that all data is forced directly to the filesystem before associated=20
metadata updates are committed to the journal.
"data=3Dwriteback" means that filesystem data updates can be made lazily,=
=20
even after related metadata changes have been committed to the journal.

We will give it a trial  ...

Thanks,
Thomas.
=20
--=20
-------------------------------------------------
Thomas M=FCller, TU Chemnitz, URZ, D-09107 Chemnitz
Tel: +49 (0)371 5311755   Fax: +49 (0)371 5311629
-------------------------------------------------