Hi, Michael
Thank you for your attention!
On Tue, 2021-11-30 at 17:29 +0900, Michael Paquier wrote:
> Hmm. Why should we care about invalid indexes at all, including
> pg_statio_all_indexes?
>
I think we should care about them at least because they are exists and
can consume resources. For example, invalid index is to be updated by
DML operations.
Of course we can exclude such indexes from a view using isvalid,
isready, islive fields. But in such case we should mention this in the
docs, and more important is that the new such states of indexes can
appear in the future causing change in a view definition. Counting all
indexes regardless of states seems more reasonable to me.
--
Andrei Zubkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company