Re: Optimise PostgreSQL for fast testing

On Thu, Feb 23, 2012 at 11:15 AM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
> On Thu, Feb 23, 2012 at 5:20 PM, Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com> wrote:
>
>> He's probably doing automated continuous integration testing.  Two
>> jobs ago we had a setup to do that and had 40k tests.  The whole test
>> suite took about 30 minutes to runm and kicked off automatically when
>> the last one finished and anyone touched any code.
>
> Having lots of tests is a good thing. Bring 'em on.
>
> If you use SQLite for that, then it all runs in a single thread and it
> could easily take 30 minutes or longer.
>
> Now all you have to do is parallelise the tests and everything can
> work 10 times quicker and it would be much faster than the time SQLite
> produced.

It's funny how once you start thinking of how to optimize to run 8 or
16 or more concurrent tests, you sometimes forget that doing that same
thing to some simpler tools might result in very poor performance til
you have to run the tests on the old system and start wondering why
you ever thought it was fast.