[OpenAFS] Extremely poor write performance.

[Martin Schulz]

Derrick J Brashear <shadow@dementia.org> writes:

> On 16 Jan 2003, Martin Schulz wrote:
> 
> > Paul Blackburn <mpb@est.ibm.com> writes:
> > 
> > > It seems that you looked at the performance of 1 client with 1 server?
> > > Are your figures for a single access to a server?
> > 
> > That is the natural measuring a user would do. That is what is
> > important to *him*. That is what *he* will judge AFS on. 
> 
> I don't give my users a server to themselves. I wouldn't if it were NFS,
> either. 

Sorry if I was too imprecise: The volume a user can write to is likely
to be only his "personal" volume. (At least that is the case in our
setup, and I suppose others handle it in similar ways.) And the
rw-instance of that volume can only be on a single server. 

So if a user has to decide if he is satisfied with the present AFS
performance, he will naturally sit down in front of his client station
and start hammering on his rw-mounted volume. This is effectively a
one-client-one-server benchmark. 

This is a different focus from the overall view of an AFS
admin. That's all I wanted to state.

Was this clear now? 

> > How does this affect the "delete 10000 small files" performance? 
> 
> If you use a Solaris fileserver instead of Linux, file deletion
> performance probably improves, for instance. There are lots of variables.

Could you share with us the reason of your expectation? 

Yours,
-- 
Martin Schulz                             schulz@iwrmm.math.uni-karlsruhe.de
Uni Karlsruhe, Institut f. wissenschaftliches Rechnen u. math. Modellbildung
Engesser Str. 6, D-76128 Karlsruhe