Re: Config for fast huge cascaded updates

From: Andrew Sullivan <ajs(at)crankycanuck(dot)ca>
To: pgsql-general(at)postgresql(dot)org
Subject: Re: Config for fast huge cascaded updates
Date: 2017-06-27 13:28:43
Message-ID: 20170627132843.ranvsfdk5zpclsbt@crankycanuck.ca
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Mon, Jun 26, 2017 at 07:26:08PM -0700, Joshua D. Drake wrote:

> Alternatively, and ONLY do this if you take a backup right before hand, you
> can set the table unlogged, make the changes and assuming success, make the
> table logged again. That will great increase the write speed and reduce wal
> segment churn.

Note that this is not for just that table, but for all of the
implicated ones because of the CASCADE statements. It sounds like the
OP is basically rewriting a significant chunk of the entire database,
so nothing is going to be super fast: all those CASCADEs have to fire
and all those other tables need to be updated too.

> However, if that fails, the table is dead. You will have to reload it from
> backup.

Right, and that goes for all the affected tables.

Best regards,

A

--
Andrew Sullivan
ajs(at)crankycanuck(dot)ca

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Albe Laurenz 2017-06-27 15:36:20 Re: Re: Unable to understand index only scan as it is not happening for one table while it happens for other
Previous Message Adrian Klaver 2017-06-27 13:08:45 Re: Accessing DB2 tables from postgresql