Eric Lee Green <eric@badtux.org> writes:
> > You might try timing 1000 null-returning SELECTs against a populated
> > table and see how long they take, just to see if my hypothesis is
> > correct.
>
> Near instantaneously.
Hmmm. I think this is a situation where you're going to need to
profile the backend to see what's taking up all the time. It may be a
bad case of lock contention (though I can't see why), or something
really dumb may be going on. If you can provide a test case, I'm
willing to bet the PG gurus will be interested...
So just te be clear, what you're seeing is that batched INSERTs go
fast, and batched SELECTs go fast, but when you interleave them it's
really slow?
> I will create some more test programs tomorrow and post them if I can
> duplicate the problem in a way that won't cause Legal to have a heart attack
> :-}. Geez, I wish it was possible to make a living in the Open Source world
> nowdays, this business of NDA's and such sucks.
I hear that. ;)
-Doug
--
Doug McNaught Wireboard Industries http://www.wireboard.com/
Custom software development, systems and network consulting.
Java PostgreSQL Enhydra Python Zope Perl Apache Linux BSD...