Re: Mini improvement: statement_cost_limit - Mailing list pgsql-hackers

From Josh Berkus
Subject Re: Mini improvement: statement_cost_limit
Date
Msg-id 200808041354.01115.josh@agliodbs.com
Whole thread Raw
In response to Re: Mini improvement: statement_cost_limit  (Gregory Stark <stark@enterprisedb.com>)
List pgsql-hackers
Greg,

> For such an application this would be a major foot-gun which would give
> a false sense of security simultaneously causing random outages and not
> providing even the protection you're counting on.

Hmmm.  That sounds like a call for some testing.  While our cost estimation 
has some issues, I don't think it's unreliable as all that.  And it's easy 
enough to document the limitations.  If it's 80% accurate, then it's 
fixing more problems than it causes.  If it's 30% accurate, then obviously 
it's a bad idea.

Personally, I don't see much use for this interactively, because an 
experienced DBA can easily enough run an EXPLAIN before running the query.  
I usually do, on production systems.

-- 
--Josh

Josh Berkus
PostgreSQL
San Francisco


pgsql-hackers by date:

Previous
From: Simon Riggs
Date:
Subject: Re: Mini improvement: statement_cost_limit
Next
From: Simon Riggs
Date:
Subject: Automatic Client Failover