From: Marcus Bointon
Date: October 27 2011 9:55am
Subject: Re: Master-Master -> duplicate entry
List-Archive: http://lists.mysql.com/replication/2244
Message-Id: <4F60E0D4-716F-4095-97B6-D0F4093549C5@synchromedia.co.uk>
MIME-Version: 1.0 (Apple Message framework v1251.1)
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: quoted-printable

On 27 Oct 2011, at 09:51, Johan De Meersman wrote:

> Say what? If that's the case, you haven't set up your replication =
correctly. There are quite a few reasons to not use both masters =
concurrently, but this is not one of them.

Not at all. The point of redundancy (via replication or DRBD) is to =
maintain consistent and available data in the face of a server or =
replication failure. If you only write to one master at a time you have =
that. If you write to both you have neither - i.e. it's actually worse =
than having no redundancy at all, hence my statement. Because there is =
no performance gain and there is an increased risk of data loss it's not =
just pointless, it's actively bad. It's a bit like using RAID-0, but =
without the speedup! With writes to both there's no way of avoiding the =
split-brain scenario you describe and it's really hard to recover from =
(I've been there!). No advantages plus lots of disadvantages sounds like =
a bad combo to me.

There are some attempts to introduce a semi-synchronous replication =
system for MySQL that waits until transactions are replicated (possibly =
to a quorum of slaves, much like Cassandra) before committing (or at =
least before returning a result to the client).

Marcus
--=20
Marcus Bointon
Synchromedia Limited: Creators of http://www.smartmessages.net/
UK info@hand CRM solutions
marcus@stripped | http://www.synchromedia.co.uk/