On Tue, 23 Sep 2008, Gregory Stark wrote:
> I have *not* been able to observe any significant effect from
> POSIX_FADV_SEQUENTIAL but I'm not sure what circumstances it was a problem. It
> sounds like it's a peculiar situation which is not easy to reliably reproduce.
Zoltan, Hans-Juergen: would it be possible for you to try the latest
bitmap-preread-v18.diff.gz patch Greg Stark just sent over to the list?
It's at
http://archives.postgresql.org/message-id/87ljxjudu7.fsf@oxford.xeocode.com
as well. That's a refinement of the original strategy you used, and I'd
be curious to hear whether it still works usefully on the troublesome
workload you submitted your original patch against. Since none of the
rest of us have been successful so far replicating the large speed-up on
multiple concurrent sequential scans you reported, I think you're the best
candidate to see if there was any regression because of how the patch was
refactored.
I'm excited to see index scans in the new patch as well, since I've got
1TB of test data that gets navigated that way I can test with.
--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD