[oclug]Postgresql tuning

[Milan Budimirovic]

Timothy Brier wrote:
> 
> As a follow up to my initial question, we ran some tests at our company and
> here is some of the details
> 
> I wrote a test to simulate our environment for which the question was
> submitted.  I had it reviewed by some skeptics who thought it was a "good
> test" and egged me on about another MS victory.
> 
> The test basically simulates - in this case 300 users  - except for the SQL
> server it topped out at 160 users - all doing inserts, updates, deletions,
> selects with back end calculations, functions and views.
> Each client request involved 1.2 million records in 7 tables.
> 
> I did turn off fsync(), because the other database SQL Server has a delayed
> writing feature as well.   After running the test against each server 4
> times - with the skeptics present and part of the test - the PostgreSQL
> system outshined the SQL server handily.   This is a summary of what
> occurred.
> 
> The client machines were all running Windows 2000.   Some were dual
> processors.   The lowest client processor used was 700 MHZ PIII.  All our
> connected to a 100mb network, no other network activity was present and the
> servers were not being used for another purpose.  Each client averaged a
> transfer of 800K per test.
> 
> The Web server is a 700 Mhz Celeron with 384K of Ram running Apache and
> ColdfusionMX.
> 

One question. What kind of hooks does ColdFusion have built into it for
connecting to and interfacing with a PostrgesSQL database? Does it allow
things like connection pooling, using place holders or binding columns?