On 6 January 2013 16:29, Jeff Janes <jeff.janes@gmail.com> wrote:
> Worse, this over-punishment of bloat is more likely to penalize partial
> indexes. Since they are vacuumed on the table's schedule, not their own
> schedule, they likely get vacuumed less often relative to the amount of
> turn-over they experience and so have higher steady-state bloat. (I'm
> assuming the partial index is on the particularly hot rows, which I would
> expect is how partial indexes would generally be used)
That's an interesting thought. Thanks for noticing that.
-- Simon Riggs http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services