Re: Integrity on large sites - Mailing list pgsql-general

From Ron Johnson
Subject Re: Integrity on large sites
Date
Msg-id 4653A811.4060007@cox.net
Whole thread Raw
In response to Integrity on large sites  (Naz Gassiep <naz@mira.net>)
List pgsql-general
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 05/22/07 21:12, Naz Gassiep wrote:
> I'm working in a project at the moment that is using MySQL, and people
> keep making assertions like this one:
>
> "*Really* big sites don't ever have referential integrity. Or if the few
> spots they do (like with financial transactions) it's implemented on the
> application level (via, say, optimistic locking), never the database level."
>
> This sounds to me like MySQLish. A large DB working with no RI would
> give me nightmares. Is it really true that large sites turn RI off to
> improve performance, or is that just a MySQL thing where it gets turned
> off just because MySQL allows you to turn it off and improve
> performance? Can you even turn RI off in PostgreSQL? Does Oracle, DB2 or
> MSSQL allow you to turn it off? Am I just being naive in thinking that
> everyone runs their DBs with RI in production?

Allow you to turn it off???

RI as in foreign keys or RI as in primary keys?

FKs are not implemented on our big transactional systems that use
Rdb/VMS.  Originally this was because the extra load would slow down
a system that needed every ounce of speed back on late 1990s technology.

Now we have (some) faster hardware, but even higher posting volumes.

- --
Ron Johnson, Jr.
Jefferson LA  USA

Give a man a fish, and he eats for a day.
Hit him with a fish, and he goes away for good!

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFGU6gRS9HxQb37XmcRAjSnAJwN8XhCxsHyeJHqxzi/k0Dj6O8fVACdGxrd
R1hfrTh9ifDivr51AGt1NNQ=
=CSLd
-----END PGP SIGNATURE-----

pgsql-general by date:

Previous
From: "Richard P. Welty"
Date:
Subject: Re: Integrity on large sites
Next
From: Michael Glaesemann
Date:
Subject: Re: Using a trigger with an object-relational manager