Re: Four issues why "old elephants" lack performance: Explanation sought Four issues why "old elephants" lack performance: Explanation sought - Mailing list pgsql-general

From Karsten Hilbert
Subject Re: Four issues why "old elephants" lack performance: Explanation sought Four issues why "old elephants" lack performance: Explanation sought
Date
Msg-id 20120226163852.GA2512@hermes.hilbert.loc
Whole thread Raw
In response to Re: Four issues why "old elephants" lack performance: Explanation sought Four issues why "old elephants" lack performance: Explanation sought  (Andy Colson <andy@squeakycode.net>)
Responses Re: Four issues why "old elephants" lack performance: Explanation sought Four issues why "old elephants" lack performance: Explanation sought
List pgsql-general
On Sun, Feb 26, 2012 at 08:37:54AM -0600, Andy Colson wrote:

>> 3. WAL logging
>
> PG writes a transaction twice.  Once to WAL and once to
> the DB.  WAL is a simple and quick write, and is only ever
> used if your computer crashes and PG has to re-play
> transactions to get the db into a good/known state.  Its a
> safety measure that doesn't really take much time, and I
> don't think I've heard of anyone being WAL bound.  Although
> it does increase IO ops, it's not the biggest usage of IO.
> This one falls under "lets be safe" which is something NoSQL
> did away with.  Its not something I want to give up,
> personally.  I like using a net.

And, one could still effectively disable WAL by using
unlogged tables.

Karsten
--
GPG key ID E4071346 @ gpg-keyserver.de
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346

pgsql-general by date:

Previous
From: Clodoaldo Neto
Date:
Subject: Constant value for a partitioned table query inside a plpgsql function
Next
From: Tom Lane
Date:
Subject: Re: How to debugging a an external C function(IMMUTABLE STRICT )