Re: oddly high cost but low actual_time (but not in slony replicants)

On Wed, Mar 26, 2008 at 01:23:22PM -0500, Jeff Amiel wrote:
> explain analyze select * from table_a;
> "Seq Scan on table_a  (cost=100000000.00..100000001.02 rows=19
> width=103) (actual time=0.007..0.012 rows=19 loops=1)"
> "Total runtime: 0.040 ms"
>
> any ideas?

Looks like you've "disabled" seqscans.  Because seqscan is a fallback
method all disabling does is to make it very expensive.  To check:

  SHOW enable_seqscan;

If it's set then have a look in the config file, or maybe reset the
per-role setting using ALTER ROLE[1].


  Sam

 [1] http://www.postgresql.org/docs/current/static/sql-alterrole.html