Christopher Browne wrote:
>swm@linuxworld.com.au (Gavin Sherry) wrote:
>
>
>>I guess the main point is, if something major like this ships in the
>>backend it says to users that the problem has gone away. pg_autovacuum is
>>a good contrib style solution: it addresses a problem users have and
>>attempts to solve it the way other users might try and solve it. When you
>>consider it in the backend, it looks like a workaround. I think users are
>>better served by solving the real problem.
>>
>>
>
>Hear, hear!
>
>It seems to me that the point in time at which it is *really*
>appropriate to put this into the backend is when the new GUC variable
>"dead_tuple_map_size" (akin to FSM) is introduced, and there is a new
>sort of 'VACUUM DEAD TUPLES' command which goes through the DTPM (Dead
>Tuple Page Map).
>
>In THAT case, there would be the ability to do a VACUUM on the "dead
>bits" of the table that consists of 50M rows without having to go
>through the 49M rows that haven't been touched in months.
>
>
This will make VACUUM less painful, but it doesn't eliminate the need /
desire for autovacuum. I agree this would be good, but I see it as a
separate issue.