Re: Vacuum goes worse

[Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>]

Brian Herlihy <btherl(at)yahoo(dot)com(dot)au> writes:
> Would it make sense to show the FSM stats for individual table vaccums as w=
> ell?  I'm wondering if the reason they aren't shown is because it wouldn't =
> be useful or isn't practical, or just that it hasn't been done.

It was intentionally omitted in the original design, on the grounds that
after a single-table VACUUM there's no very good reason to think that
the global FSM stats are sufficiently complete to be accurate.  Of
course, in a multi-database installation the same charge could be
leveled against the situation after a single-database VACUUM, so maybe
there's not a lot of strength in the argument.

IIRC the code change would be trivial, it's just a matter of judgment
whether the extra output is useful/trustworthy.

			regards, tom lane