From: Tom Lane <tgl@sss.pgh.pa.us>
To: Keith Ouellette <Keith.Ouellette@Airgas.com>
cc: "pgsql-admin@postgresql.org" <pgsql-admin@postgresql.org>
Subject: Re: Wall replay
In-reply-to: <417C5AF7C228B94490192951394BEFE7AF396A@AIPHLEXDAG01B.airgas.com>
References: <417C5AF7C228B94490192951394BEFE7AF396A@AIPHLEXDAG01B.airgas.com>
Comments: In-reply-to Keith Ouellette <Keith.Ouellette@Airgas.com>
	message dated "Thu, 11 Apr 2013 21:01:31 -0000"
Date: Thu, 11 Apr 2013 17:55:34 -0400
Message-ID: <12450.1365717334@sss.pgh.pa.us>
Precedence: bulk
Sender: pgsql-admin-owner@postgresql.org

Keith Ouellette <Keith.Ouellette@Airgas.com> writes:
> I am realatively new to PostgreSQL. I have two servers in a Master/Slave relationship using WAL as the replication method. I have what I hope to be an easy question. If the Master goes down and before the slave is promoted, there may be a case where a log file may not have been shipped to the slave yet. If the new Master starts to process new transactions, is there a way to have it process the unshipped WAL files if they are copied to it? I do not see that case in the documentation.

You'd have to ship the remaining WAL over *before* the slave has come
up live, else it's too late.

If you're concerned about gaps of that sort, though, you should not be
using log file shipping at all.  Set up streaming replication instead.
(If the PG version you're using hasn't got that, high time to move
to something newer.)

			regards, tom lane


-- 
Sent via pgsql-admin mailing list (pgsql-admin@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin