Taking another shot at the PostgreSQL/filesystem benchmarks

Alright.

To anyone who didn't get the news the first time:
The first set of benchmarks were terribly skewed because FreeBSD
didn't properly work with the hardware I was using. Thanks to
those who pointed the problem out to me.

I have scrounged new hardware, and insured that FreeBSD is working
properly with it.  And I have rerun many of the tests (not all).

Thanks again to all who responded with helpful tips and pointers.

I'm hoping to work on these tests a little bit at a time, adding
additional tests and their results, so feel free to stop back and
check the page every so often.  I don't intend to announce future
updates on these lists unless there is something that seems
considerably important found.

If you've made a suggestion on additional tests to run, please
rest assured that I got your email and have added the test to my
list of things to try.  I simply got more emails than I could
respond to.  I apologize to everyone who didn't get a personal
response.

http://www.potentialtech.com/wmoran/postgresql.php

Again, feedback is welcome, but if it's of any magnitude similar
to what I just got, I doubt I'll be able to respond to everyone.
Please don't be offended.  I'm rather surprised at how popular
this information was.

--
Bill Moran
Potential Technologies
http://www.potentialtech.com