Richard Neill wrote:
> Now, I understand that increasing checkpoint_segments is generally a
> good thing (subject to some limit), but doesn't that just mean that
> instead of say a 1 second outage every minute, it's a 10 second outage
> every 10 minutes?
That was the case in versions before 8.3. Now, the I/O is spread out
over most of the next checkpoint's time period. So what actually
happens is that all the I/O that happens over 10 minutes will be spread
out over the next five minutes of time. With the defaults, there's so
little time between checkpoints under heavy writes that the spreading
doesn't have enough room to work, leading to higher write bursts.
--
Greg Smith 2ndQuadrant Baltimore, MD
PostgreSQL Training, Services and Support
greg@2ndQuadrant.com www.2ndQuadrant.com