On Thu, Dec 4, 2008 at 5:11 PM, Greg Smith <gsmith@gregsmith.com> wrote:
> On Thu, 4 Dec 2008, Kevin Grittner wrote:
>
>> I think there needs to be some easy way to choose an option which
>> yields a configuration similar to what we've had in recent production
>> releases -- something that will start up and allow minimal testing on
>> even a small machine.
>
> But that's the goal of what comes out of initdb already; I'm missing how
> that is something this script would even get involved in. Is your
> suggestion to add support for a minimal target that takes a tuned-up
> configuration file and returns it to that state, or did you have something
> else in mind?
I humbly suggest that the memory-related settings output by the tool
don't need to match what initdb outputs. But the values of
checkpoint_segments, constraint_exclusion, and
default_statistics_target probably should, given a small mixed-mode
database.
You've probably all figured out by now that I, personally, in my own
opinion, think that default_statistics_target = 10 is just fine for
such a database, but if a decision is made to change that number, so
be it. Just let's please change it both places, rather than letting
contrib/pgtune be a backdoor to get around not liking what initdb
does. And similarly with the other parameters...
...Robert