On Thu, 2007-09-06 at 21:50 +0100, Gregory Stark wrote:
> > - Improve scalability of recovery for large I/O bound databases
>
> That seems too vague for the TODO. Did you have specific items in mind?
I think we should parallelise recovery. Heikki wanted to do this another
way, so I worded it vaguely (i.e. as a requirement) to cover multiple
approaches. Some of that was discussed on -hackers already (where?).
-- Simon Riggs 2ndQuadrant http://www.2ndQuadrant.com