Re: planner with index scan cost way off actual cost, - Mailing list pgsql-performance

From Guillaume Cottenceau
Subject Re: planner with index scan cost way off actual cost,
Date
Msg-id 87irq6e99n.fsf@meuh.mnc.lan
Whole thread Raw
In response to Re: planner with index scan cost way off actual cost,  (Scott Marlowe <smarlowe@g2switchworks.com>)
List pgsql-performance
Hi Scott,

Scott Marlowe <smarlowe 'at' g2switchworks.com> writes:

> On Wed, 2006-03-22 at 02:04, Guillaume Cottenceau wrote:

[...]

> > Yes, we use 7.4.5 actually, because "it just works", so production
> > wants to first deal with all the things that don't work before
> > upgrading. I have recently discovered about the background writer
> > of 8.x which could be a supplementary reason to push for an
> > ugprade though.
>
> Imagine you get a call from the manufacturer of your car.  There's a
> problem with the fuel pump, and, in a small percentage of accidents,
> your car could catch fire and kill everyone inside.
>
> Do you go in for the recall, or ignore it because you just want your car
> to "just work?"

Ah, this holy computer/OS/whatever-to-cars comparison.. How many
million electrons would the world save if computer people would
abandon it? :)

> In the case of the third number in postgresql releases, that's what
> you're talking about.  the updates that have come after the 7.4.5
> version, just talking 7.4 series here, have included a few crash and
> data loss fixes.  Rare, but possible.

I guess we didn't know that. I for myself have (a bit more)
excuses because I'm on the development side :) But I've passed
the information to the operation team, thank you.

--
Guillaume Cottenceau

pgsql-performance by date:

Previous
From: Csaba Nagy
Date:
Subject: Re: WAL logging of SELECT ... INTO command
Next
From: "Luke Lonergan"
Date:
Subject: Re: Intel C/C++ Compiler Tests