RE: NFS performance

[James Rich <james@xxxxxxxxxxx>]

On Mon, 17 Aug 2009, Evan Harris wrote:

> I've seen this kind of attitude many times and in most cases until people
> were prepared to deal with the facts and data they were unable to solve
> their problem.
>
> What you need to do is measure all the parts of the process on both machines
> and any equipment in between and identify which part is the hold up using
> benchmarks and data. Locate the problem with numbers not prejudice.
>
> It may well be the iSeries that is the problem, but with your approach if it
> isn't you'll *never* get it fixed.

I'm not sure what kind of attitude you're referring to.  The performance 
of NFS on my iSeries is significantly less than that of my other machines 
and I posted a question asking how to resolve that.  I tried the 
suggestions offered and posted my suspicions as to the cause.  The tests 
I've done back up my suspicions.  I don't know everything, thus I post 
here.  I believe I am following the scientific method:  make an 
observation, form a hypothesis, and do tests to confirm the hypothesis.

James Rich

if you want to understand why that is, there are many good books on
the design of operating systems. please pass them along to redmond
when you're done reading them :)
	- Paul Davis on ardour-dev