Alex Mayrhofer <[EMAIL PROTECTED]> writes:
> a recent REINDEX reduced that index from about 3 pages to 18000.
That's not bloat, that's normal overhead. The traditional rule of thumb
for a btree is that at steady state, pages will be about 2/3rds full.
REINDEX packs pages to 90% IIRC, but you
Tom Lane wrote:
> Alex Mayrhofer <[EMAIL PROTECTED]> writes:
>> i have a bunch of indices over rather frequently updated large tables. Those
>> indices grow in size with the updates, so i frequently re-index them.
>
> That usually shouldn't be necessary since PG 7.4 or so. Do you have
> some stra
Alex Mayrhofer <[EMAIL PROTECTED]> writes:
> i have a bunch of indices over rather frequently updated large tables. Those
> indices grow in size with the updates, so i frequently re-index them.
That usually shouldn't be necessary since PG 7.4 or so. Do you have
some strange pattern of index key u
Hi,
i have a bunch of indices over rather frequently updated large tables. Those
indices grow in size with the updates, so i frequently re-index them.
Are there any plans to add REINDEX estimation/jobs to the autovacuum process
- or, alternatively, any options on reducing the growth rate of those
