Comments on: Replication tutorial notes - part 2 http://ebergen.net/wordpress/2008/04/14/replication-tutorial-notes-part-2/ You will probably want some waders, a pick axe, and one of those hats with a light on it before you go in here. Tue, 06 Jan 2009 09:40:59 +0000 http://wordpress.org/?v=2.6.2 By: Eric Bergen http://ebergen.net/wordpress/2008/04/14/replication-tutorial-notes-part-2/#comment-156329 Eric Bergen Tue, 22 Jul 2008 03:12:11 +0000 http://ebergen.net/wordpress/2008/04/14/replication-tutorial-notes-part-2/#comment-156329 The problem is that there is no longer a single source of truth for the data. If you restore a master from a backup then one master has records that aren't on another. It can get very messy very quickly. It's better to partition masters and have a single source of truth for each record at all times. The problem is that there is no longer a single source of truth for the data. If you restore a master from a backup then one master has records that aren’t on another. It can get very messy very quickly. It’s better to partition masters and have a single source of truth for each record at all times.

]]> By: Todd Lyons http://ebergen.net/wordpress/2008/04/14/replication-tutorial-notes-part-2/#comment-156239 Todd Lyons Mon, 21 Jul 2008 20:52:12 +0000 http://ebergen.net/wordpress/2008/04/14/replication-tutorial-notes-part-2/#comment-156239 "I don’t recommend using a hardware load balancer to manage write load between masters. There is a risk of sending writes to both masters at the same time. In a properly configured dual master setup half the writes will be rejected on a read only error. The worst case is that writes go to both masters causing replication error and inconsistent data." Why does configuring the offset setting for master-master configurations not solve this? Is there some inherent problem with master-master that will only be reached when you scale up in size and/or throughput? “I don’t recommend using a hardware load balancer to manage write load between masters. There is a risk of sending writes to both masters at the same time. In a properly configured dual master setup half the writes will be rejected on a read only error. The worst case is that writes go to both masters causing replication error and inconsistent data.”

Why does configuring the offset setting for master-master configurations not solve this? Is there some inherent problem with master-master that will only be reached when you scale up in size and/or throughput?

]]> By: Eric Bergen http://ebergen.net/wordpress/2008/04/14/replication-tutorial-notes-part-2/#comment-118012 Eric Bergen Wed, 16 Apr 2008 21:16:31 +0000 http://ebergen.net/wordpress/2008/04/14/replication-tutorial-notes-part-2/#comment-118012 What I mean is that if you choose to stick to row based replication there are a few types of queries that are going to be replicated as statements such as alter,create,drop table. This is unavoidable. What I mean is that if you choose to stick to row based replication there are a few types of queries that are going to be replicated as statements such as alter,create,drop table. This is unavoidable.

]]> By: Olly http://ebergen.net/wordpress/2008/04/14/replication-tutorial-notes-part-2/#comment-117502 Olly Tue, 15 Apr 2008 02:53:57 +0000 http://ebergen.net/wordpress/2008/04/14/replication-tutorial-notes-part-2/#comment-117502 Nice notes, I doubt I could write such good notes while sitting in a presentation. Found this here quite confusing: "The exception is things that can’t use row based replication like DML statements." Nice notes, I doubt I could write such good notes while sitting in a presentation.

Found this here quite confusing: “The exception is things that can’t use row based replication like DML statements.”

]]>