Re: Initial prefetch performance testing

On Mon, 22 Sep 2008, Gregory Stark wrote:

> Hm, I'm disappointed with the 48-drive array here. I wonder why it maxed out
> at only 10x the bandwidth of one drive. I would expect more like 24x or more.

The ZFS RAID-Z implementation doesn't really scale that linearly. It's
rather hard to get the full bandwidth out of a X4500 with any single
process, and I haven't done any filesystem tuning to improve
things--everything is at the defaults.

> I'm quite surprised Solaris doesn't support posix_fadvise -- perhaps
> it's in some other version of Solaris?

Both the systems I used were standard Solaris 10 boxes and I'm not aware
of any changes in this area in the later OpenSolaris releases (which is
where I'd expect something like this to change first). The test program I
tried failed to find #ifdef POSIX_FADV_WILLNEED, and the message I saw
from you at
http://archives.postgresql.org/message-id/877imua265.fsf@oxford.xeocode.com
suggested you didn't find any fadvise either so I didn't look much
further.

The above is a cue for someone from Sun to chime in on this subject.

> I have an updated patch I'll be sending along shortly. You might want to test
> with that?

Obviously I've got everything setup to test right now, am currently
analyzing your earlier patch and the sequential scan fork that derived
from it. If you've got a later version of the bitmap heap scan one as
well, I'll replace the one I had been planning to test (your
bitmap-preread-v9) with that one when it's available.

--
* Greg Smith gsmith(at)gregsmith(dot)com http://www.gregsmith.com Baltimore, MD