Neil,
> [If my comments below have already been raised, sorry -- I've been away,
> and missed the earlier drafts of the release.]
Howdy! How was the internship?
> I don't think the main intent of the new wire protocol is to "increase the
> speed of data transfers" -- does it actually make a significant
> improvement? (I'm not sure, having been away for the summer).
Somebody else inserted this; I'll let them defend it.
> > HIGH AVAILABILITY: Expansion of PostgreSQL's Free Space Map disk
> > management feature to support continuous index maintenence and the Auto
> > Vacuum Daemon are the last "puzzle pieces" in providing 99.999% uptime
> > for PostgreSQL databases.
>
> I certainly wouldn't call the autovacuum stuff the "last piece" in HA for
> PostgreSQL, and ISTM you just picked 5 nines right out of thin air.
Hmmm .... I see what you mean. I was trying to cut length by not having pgavd
as a seperate sentence. And yes, the 99.999 stuff is made up, although
completely possible now ...
If you have a suggestion for re-wording that paragraph around more solid
information, I'd like to see it. We want to get the point across that
vacuuming indexes + pgavd = no shutting the database down or lokcing tables
except for version upgrades, ever. And it has to be translated into
marketing-hypespeak.
--
Josh Berkus
Aglio Database Solutions
San Francisco