On Fri, 23 Jan 2009, Merlin Moncure wrote:
> Note, while sequential write speeds are a good indication of general
> raid crappyness, they are not the main driver of your low pgbench
> results (buy they may be involved with poor insert performance) That is
> coming from your seek performance, which is also lousy at ~130 according
> to bonnie.
The bonnie seek test includes a write component to it: "In 10% of cases,
it is dirtied and written back with write(2)." Since ~130 isn't that much
worse than normal for a cheap 7200 RPM drive (near 200), it could easily
be the case that the awful performance on those writes is pulling down the
seek score. Sequential writes are low on the scale of things that matter
in a database context, but if they're screwed up bad enough it can ripple
into more important areas.
--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD