After this post I've got a question how one can tell if his outer join was converted to inner. You can find it out by looking at the warning generated by EXPLAIN EXTENDED. If the outer join wasn't converted, you'll see it in the rewritten query in the warning:

mysql> explain extended select * from t1 left join (t2, t3) on t2.a= t1.a;
...
3 rows in set, 1 warning (0.00 sec)

mysql> show warnings\G
*************************** 1. row ***************************
  Level: Note
   Code: 1003
Message: select `test`.`t1`.`a` AS `a`,`test`.`t2`.`a` AS `a`,`test`.`t3`.`a`
AS `a` from `test`.`t1` left join (`test`.`t2` join `test`.`t3`) on ((`test`.`t2`.
`a` = `test`.`t1`.`a`)) where 1

In this query LEFT JOIN is not converted to inner.

Now let's try a query where outer join will be converted:

mysql> explain extended select * from t1 left join (t2, t3) on …

I’ll be speaking at the upcoming 2007 MySQL Conference & Expo (Why they dropped the word User, who knows), this time with Guy Harrison (Author of many books including MySQL Stored Procedures). We will be talking on MySQL for Oracle DBAs and Developers.

Anyway, good friend Paul McCullagh, creator of PBXT will be also speaking on PrimeBase XT: Design and Implementation of a Transactional Storage Engine. He coined to me in an email “CU at the UC”. I’ve done a further level of refactoring, and added marketing. You …

Changing hats


Note: I have updated this article as of April 24, 2008 to indicate that you do not have to set auto_increment_increment or auto_increment_offset for the slaves. If you see this in the diagrams, please ignore.

The master sets the auto_increment value in it's binary log which the slave reads via replication. I may write an update to this article at some point soon.


I have put on the hat of a MySQL user, a customer, as opposed to a developer (Federated Storage Engine, DBD::mysql) these last couple months with my new employer, Grazr. One of the things we need is a whole replication set-up, with read-only slaves, and write masters. Thats master(s), with an 's', plural. We need a set-up where we have two data centers. In each of the data centers, there is a master and its slaves, all on the same network. That master needs …

Welcome to the 33rd edition of Log Buffer, the weekly review of database blogs. We’ll look at some of the standout blog items from the week gone by. To start, you might be surprised by what is in your DBMS. An item on the Oracle Contractors Blog asks whether artificial intelligence is in Oracle, [...]

So far, I've never blogged about the work I've been doing for the MySQL Certification Team. Time to change that: I'd like to tell you a little bit more about the things I've been doing since I joined MySQL AB.

I started in July 2006 and from the on, I've been almost exclusively occupied with activities to develop certification for MySQL 5.1 Cluster.
A summary of the things I have done so far:

• 
  
• Working with the training department to design an outline for the cluster exam
  
• Interviewing MySQL Cluster developers to check all kinds of facts and details of the behaviour of MySQL 5.1 Cluster
  
• Creating, reviewing …

Alfresco, the leader in open source document management, has now made a subtle but significant change in their licensing.  Effective immediately, Alfresco is being released under the GPL rather than their previous Mozilla-derived license. 

While there's nothing wrong with the Mozilla license they used, my take is, it had two things that were slowing things down.  First of all, it's yet-another-license, even if it is based on a reasonably well-known and accepted license.  So inevitably that means that lawyers start asking lots of questions, at least compared to the GPL which is fairly well understood.  (The GPL is not perfect, but in my view, it's the best going.  And partly because it's widely used by Linux, MySQL and others, it's understood.)  The second issue is that …

Changing hats

I have put on the hat of a MySQL user, a customer, as opposed to a developer (Federated Storage Engine, DBD::mysql) these last couple months with my new employer, Grazr. One of the things we need is a whole replication set-up, with read-only slaves, and write masters. Thats master(s), with an 's', plural. We need a set-up where we have two data centers. In each of the data centers, there is a master and it's slaves, all on the same network. That master needs to replicate to another slave, which is also a master, in its own data center, where it has its own slaves. The idea is that the web servers running mod_perl applications in each data center send all their read-only queries to slaves running on the same server, and all write-only queries to the master in that data center.

How does one go about setting up a multi-master replication scheme, also having slaves in the mix? There are several articles …

Solid has released 5.0.27-0059. It is primarily a bug fixing release. You can get it from the usual place (http://dev.soliddb.com/download/).

Today Alfresco announced that it has gone 100% GPL. As you might imagine, I'm ecstatic about the move (and have been working toward this since I joined Alfresco in 2005). Actually, I think it would be fair to say that everyone in the company is ecstatic about the move. It makes things easier for marketing, for engineering, and for sales. (Gasp! Did he really say "GPL" and "sales" in the same paragraph? :-)

As CNET has covered, this move is in part designed to help expand our already healthy community. But it's also about driving a stake into the heart of our outdated, expensive, proprietary competitors, as John Powell, Alfresco's CEO …

I recently found a great article on mysql performance blog giving you a short and precise description of what to tune in the first place.

This is a must!