On Tue, 2008-10-07 at 18:19 +0300, Heikki Linnakangas wrote:
> The bottom line is that hot standby is a big feature, and probably a big
> patch. No amount of version control will work around that. Finishing all
> that in a few weeks is a very ambitious goal. I wish you luck, and I
> wish I could do more to help you with that; it's a feature I'd really
> like to see in 8.4. Given how many iterations just this part of the
> patch needed, and it's still not there, the little project manager in me
> says that we really need to be seeing the complete patch or patches very
> soon, and do a first round of review well before the commit fest. The
> risk that we have to drop it as unfinished in November is very high
> otherwise.
Agreed and understood. The coding isn't the problem, its the risk of
being blindsided by a showstopper. Coding is quick when you know exactly
what you are trying to achieve.
That's one reason to strip out the bgwriter stuff. It's the postmaster
state change stuff that's most needed. Anyway, watch this space.
-- Simon Riggs www.2ndQuadrant.comPostgreSQL Training, Services and Support