On Sep 17, 2008, at 9:30 AM, Heikki Linnakangas wrote:
> I think we'd still need to WAL log operations that decrease the
> amount of free space on page. Otherwise, after we have partial
> vacuum, we might never revisit a page, and update the FSM, even
> though there's usable space on the page, leaving the space
> forgotten forever.
ISTM that it would be better to deal with such corner cases via
periodic non-partial vacuums, done by something like autovac, and
probably done with an ever higher-than-normal vacuum_cost_delay
setting so as to minimize performance. That's likely a lot less
wasteful than further compounding lock contention for the WAL. Even
if it does result in more overall IO, you have to trade a *lot* of IO
to balance out the impact of lock contention.
--
Decibel!, aka Jim C. Nasby, Database Architect decibel@decibel.org
Give your computer some brain candy! www.distributed.net Team #1828
