Re: Why should such a simple query over indexed columns be so slow?

[Alessandro Gagliardi <alessandro@xxxxxxxx>]

Hm. Well, it looks like setting enable_seqscan=false is session specific, so it seems like I can use it with this query alone; but it sounds like even if that works, it's a bad practice. (Is that true?)

My effective_cache_size is 1530000kB

On Mon, Jan 30, 2012 at 12:50 PM, Claudio Freire <klaussfreire@xxxxxxxxx> wrote:

On Mon, Jan 30, 2012 at 5:35 PM, Alessandro Gagliardi
<alessandro@xxxxxxxx> wrote:
> To answer your (non-)question about Heroku, it's a cloud service, so I don't
> host PostgreSQL myself. I'm not sure how much I can mess with things like
> GUC since I don't even have access to the "postgres" database on the server.
> I am a long time SQL user but new to Postgres so I welcome suggestions on
> where to start with that sort of thing. Setting enable_seqscan=false made a
> huge difference, so I think I'll start there.

It's not a good idea to abuse of the enable_stuff settings, they're
for debugging, not for general use. In particular, disable sequential
scans everywhere can have a disastrous effect on performance.

It sounds as if PG had a misconfigured effective_cache_size. What does
"show effective_cache_size" tell you?