Re: initdb and fsync

On mån, 2012-06-18 at 20:57 +0200, Andres Freund wrote:
> I don't think the difference in initdb cost is relevant when running
> the regression tests. Should it prove to be we can re-add -N after a
> week or two in the buildfarm machines.

Keep in mind that the regression tests are not only run on the
buildfarm, and that pg_regress is not only used for the main regression
tests.  When you're dealing with things in contrib or src/pl or external
modules, then adding 3 seconds to a 5 second test is not so nice.

>  I just remember that there were several OS specific regression when
> adding the pre-syncing for createdb.

Sure, it would be useful to test this, but then we should also turn on
fsync in the backend.  Make it a separate slow mode (or a separate fast
mode).