On Thu, Mar 25, 2010 at 4:04 PM, Merlin Moncure <mmoncure@gmail.com> wrote:
> There is very little reason to do this. both postgres and the
> operating system cache frequently used pages in memory already and
> they are pretty smart about it -- this leaves more memory for
> temporary demands like sorts, indexes, large result sets, etc. It's
> usually just as good to simply set fsync=off on the database in
> scenarios where you are ok with data loss following a crash and the
> system is performance critical.
(I work with the OP)
We've found that writes to the ram based DB are about 3x faster than
disk based (with fsync turned ON), but we were expecting them to be a
LOT faster than that and are wondering what we might be doing wrong.
--
“Don't eat anything you've ever seen advertised on TV”
- Michael Pollan, author of "In Defense of Food"