[OpenAFS] vice partition sizes, and large disks

Jeffrey Hutzelman jhutz@cmu.edu
Wed, 20 Jun 2001 15:13:36 -0400 (EDT)


On 16 Jun 2001, Sam Hartman wrote:

> >>>>> "Derrick" == Derrick J Brashear <shadow@dementia.org> writes:
> 
>     Derrick> On Sat, 16 Jun 2001, chas williams wrote:
>     >> i am not convinced that rx is the problem.  especially for the
>     >> vos dump.
> 
>     Derrick> nor am i An argument here would be interesting.  The
> 
> profiling data I posted as well as the discussions on openafs-devel
> very strongly pointed at RX sucking, or at least the RX implementation
> sucking.  I'd be happy to run more tests if you can propose some.

Some people claimed rx sucks.  Some even posted data.  However, some
experiments we ran here suggested that the problem was _not_ rx.  We
managed to move some significant quantities of data around at reasonable
speeds using rx.  I'm pretty sure we were also able to see performance
problems on vol-dump directly to local disk.

The vos dump performance issue does not appear to have a simple answer.
I'm not at all convinced that rx has serious performance problems.

-- Jeff