Re: test / live environment, major performance difference - Mailing list pgsql-performance

From Tom Lane
Subject Re: test / live environment, major performance difference
Date
Msg-id 8385.1181664586@sss.pgh.pa.us
Whole thread Raw
In response to Re: test / live environment, major performance difference  (Christo Du Preez <christo@mecola.com>)
Responses Re: test / live environment, major performance difference  (Christo Du Preez <christo@mecola.com>)
List pgsql-performance
Christo Du Preez <christo@mecola.com> writes:
> On my laptop the explain analyze looks like this:

> "Index Scan using fki_layertype_parentid on layertype  (cost=0.00..8.27
> rows=1 width=109)"
> "  Index Cond: (parentid = 300)"

OK ...

> and on the problem server:

> "Seq Scan on layertype  (cost=0.00..20.39 rows=655 width=110)"
> "  Filter: (parentid = 300)"

The server thinks that every row of the table matches the WHERE clause.
That being the case, it's making the right choice to use a seqscan.
The question is why is the rows estimate so far off?  Have you ANALYZEd
the table lately?

            regards, tom lane

pgsql-performance by date:

Previous
From: Alvaro Herrera
Date:
Subject: Re: VACUUM vs auto-vacuum daemon
Next
From: Bill Moran
Date:
Subject: Re: VACUUM vs auto-vacuum daemon