<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: Towards a New Modular Replication Architecture</title>
	<atom:link href="http://www.joinfu.com/2009/04/towards-a-new-modular-replication-architecture/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.joinfu.com/2009/04/towards-a-new-modular-replication-architecture/</link>
	<description>the art of sql</description>
	<lastBuildDate>Mon, 23 Jan 2012 21:05:00 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.3.1</generator>
	<item>
		<title>By: Drizzle Replication &#8211; The Command Message &#171; join-fu!</title>
		<link>http://www.joinfu.com/2009/04/towards-a-new-modular-replication-architecture/comment-page-1/#comment-215308</link>
		<dc:creator>Drizzle Replication &#8211; The Command Message &#171; join-fu!</dc:creator>
		<pubDate>Sat, 13 Mar 2010 04:06:07 +0000</pubDate>
		<guid isPermaLink="false">http://joinfu.com/2009/04/towards-a-new-modular-replication-architecture#comment-215308</guid>
		<description>[...] in development, and a lot more. Before jumping in, you may want to read a quick overview about the concepts of Drizzle replication [...]</description>
		<content:encoded><![CDATA[<p>[...] in development, and a lot more. Before jumping in, you may want to read a quick overview about the concepts of Drizzle replication [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Eric Day</title>
		<link>http://www.joinfu.com/2009/04/towards-a-new-modular-replication-architecture/comment-page-1/#comment-182559</link>
		<dc:creator>Eric Day</dc:creator>
		<pubDate>Fri, 03 Apr 2009 17:15:01 +0000</pubDate>
		<guid isPermaLink="false">http://joinfu.com/2009/04/towards-a-new-modular-replication-architecture#comment-182559</guid>
		<description>Looks great Jay! This is not only a great post for the replication design, but how Drizzle modules will work in general.</description>
		<content:encoded><![CDATA[<p>Looks great Jay! This is not only a great post for the replication design, but how Drizzle modules will work in general.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jay Pipes</title>
		<link>http://www.joinfu.com/2009/04/towards-a-new-modular-replication-architecture/comment-page-1/#comment-182554</link>
		<dc:creator>Jay Pipes</dc:creator>
		<pubDate>Fri, 03 Apr 2009 15:16:53 +0000</pubDate>
		<guid isPermaLink="false">http://joinfu.com/2009/04/towards-a-new-modular-replication-architecture#comment-182554</guid>
		<description>Excellent question, Sergey!

The end goal is to simply have create table map to InsertRecordEvent, alter table to UpdateRecordEvent, and drop table map to DeleteRecordEvent.  With a real data-dictionary (materialized information_schema) then all a DDL statement would be is an event which changes the records in the information_schema...

For administrative commands, we have a similar dilemma though, and we&#039;ve been discussing that issue on the mailing list:

https://lists.launchpad.net/drizzle-discuss/msg03523.html

Would be good to get your input on that!

Cheers!

jay</description>
		<content:encoded><![CDATA[<p>Excellent question, Sergey!</p>
<p>The end goal is to simply have create table map to InsertRecordEvent, alter table to UpdateRecordEvent, and drop table map to DeleteRecordEvent.  With a real data-dictionary (materialized information_schema) then all a DDL statement would be is an event which changes the records in the information_schema&#8230;</p>
<p>For administrative commands, we have a similar dilemma though, and we&#8217;ve been discussing that issue on the mailing list:</p>
<p><a href="https://lists.launchpad.net/drizzle-discuss/msg03523.html" rel="nofollow">https://lists.launchpad.net/drizzle-discuss/msg03523.html</a></p>
<p>Would be good to get your input on that!</p>
<p>Cheers!</p>
<p>jay</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mark Callaghan</title>
		<link>http://www.joinfu.com/2009/04/towards-a-new-modular-replication-architecture/comment-page-1/#comment-182553</link>
		<dc:creator>Mark Callaghan</dc:creator>
		<pubDate>Fri, 03 Apr 2009 15:14:26 +0000</pubDate>
		<guid isPermaLink="false">http://joinfu.com/2009/04/towards-a-new-modular-replication-architecture#comment-182553</guid>
		<description>Jay,
I couldn&#039;t agree with you more about the editorial on design. Today I am trying to debug some of our extensions to replication and I really, really do not enjoy dealing with that code. And in 5.1, more complexity has been added to it (RBR).

MySQL already supports sync replication, but only between a storage engine and the binlog. You need to provide more details on the new interfaces if we are to understand when sync and semi-sync can be done. For sync to be possible, methods are needed for Prepare and Commit as the storage and the implementation of the new interface are both resource managers.</description>
		<content:encoded><![CDATA[<p>Jay,<br />
I couldn&#8217;t agree with you more about the editorial on design. Today I am trying to debug some of our extensions to replication and I really, really do not enjoy dealing with that code. And in 5.1, more complexity has been added to it (RBR).</p>
<p>MySQL already supports sync replication, but only between a storage engine and the binlog. You need to provide more details on the new interfaces if we are to understand when sync and semi-sync can be done. For sync to be possible, methods are needed for Prepare and Commit as the storage and the implementation of the new interface are both resource managers.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Sergey Petrunya</title>
		<link>http://www.joinfu.com/2009/04/towards-a-new-modular-replication-architecture/comment-page-1/#comment-182546</link>
		<dc:creator>Sergey Petrunya</dc:creator>
		<pubDate>Fri, 03 Apr 2009 13:31:49 +0000</pubDate>
		<guid isPermaLink="false">http://joinfu.com/2009/04/towards-a-new-modular-replication-architecture#comment-182546</guid>
		<description>InsertRecordEvent, DeleteRecordEvent, UpdateRecordEvent. What about table creation/alter/drop events?</description>
		<content:encoded><![CDATA[<p>InsertRecordEvent, DeleteRecordEvent, UpdateRecordEvent. What about table creation/alter/drop events?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Roland Bouman</title>
		<link>http://www.joinfu.com/2009/04/towards-a-new-modular-replication-architecture/comment-page-1/#comment-182541</link>
		<dc:creator>Roland Bouman</dc:creator>
		<pubDate>Fri, 03 Apr 2009 10:08:28 +0000</pubDate>
		<guid isPermaLink="false">http://joinfu.com/2009/04/towards-a-new-modular-replication-architecture#comment-182541</guid>
		<description>Hi!

great post, thanks!</description>
		<content:encoded><![CDATA[<p>Hi!</p>
<p>great post, thanks!</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Stewart Smith</title>
		<link>http://www.joinfu.com/2009/04/towards-a-new-modular-replication-architecture/comment-page-1/#comment-182527</link>
		<dc:creator>Stewart Smith</dc:creator>
		<pubDate>Fri, 03 Apr 2009 01:40:49 +0000</pubDate>
		<guid isPermaLink="false">http://joinfu.com/2009/04/towards-a-new-modular-replication-architecture#comment-182527</guid>
		<description>I think the most important thing in API design is not to be easy to use but to be hard to misuse.

e.g. the field-&gt;store methods can be said to be &quot;easy to use&quot; (at least for some things) but they are certainly very *easy* to misuse and end up with garbage or totally not what you expected.

Rusty&#039;s API levels is a good read.</description>
		<content:encoded><![CDATA[<p>I think the most important thing in API design is not to be easy to use but to be hard to misuse.</p>
<p>e.g. the field->store methods can be said to be &#8220;easy to use&#8221; (at least for some things) but they are certainly very *easy* to misuse and end up with garbage or totally not what you expected.</p>
<p>Rusty&#8217;s API levels is a good read.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

