[rt-users] RT related Postgres question

[Jesse Vincent]

> 
> ->  Seq Scan on tickets main  (cost=0.00..17539.97 rows=1485 width=168) 
> (actual time=0.040..846.740 rows=298633 loops=1)
>       Filter: ((((status)::text = 'open'::text) OR ((status)::text = 
> 'new'::text) OR ((status)::text = 'stalled'::text) OR ((status)::text = 
> 'resolved'::text)) AND (("type")::text = 'ticket'::text) AND 
> (effectiveid = id))
> 
> If however I remove the "main.EffectiveId=main.id" chunk of the query it 
> uses a far more efficient query.  Only a very small proportion of our 
> tickets are merged
> 
> Anybody here seen this and/or know where to start looking for how to 
> correct this?  I have run "VACUUM FULL ANALYZE" and tried playing with 
> the statistics on Tickets but to no avail.

That sequential scan looks funny. I suspect you really want an index on
it.

>