As we know build in full text search is currently limited only to MyISAM search engine as well as has few other limits.

Today Sphinx Search plugin for MySQL was released which now provides fast and easy to use full text search solution for all storage engines. This version also adds a lot of other new features, including boolean search and distributed searching.

A while ago I already wrote about Sphinx Search Egine, comparing it to built in FullText search and Mnogosearch. I guess I should soon repeat tests, adding Lucene to the list for complete picture.

And if you do not feel like patching MySQL or use MySQL 5.1 beta to use sphinx as MySQL Storage Engine you can still use it old fashion way as separate server.

Many of you have experienced performance problems with JOINs in MySQL Cluster. However, in MySQL 5.1, it is possible to address this using MySQL replication and create a read slave where the JOINs are performed.

It is possible to set a MySQL Cluster looking like this (very schematic):
To read more on how to setup the replication link beetween the Master and the Slave, see http://dev.mysql.com/doc/refman/5.1/en/replication.html .

The trick here is to make sure that the tables created using the NDBCluster engine are created on the slave MySQL server as MyISAM or in another storage engine more performant resolving JOINs.

This way you can now issue complex JOINs on the slave instead of on the cluster and the replication …

The most powerful and efficient way to access data stored in the NDB storage engine is to use the NDBAPI, a C++ API which gives you direct access to data stored in MySQL Cluster.

Instead of having to go through the MySQL server to access data, you can access the data directly from your C++ application. Why would you like to do this? There are some reasons:

1. Data access with minimal overhead
   
2. Flexibility - posssible to hand-optimize queries. You control exactly when you want to send data to NDB Cluster.
   
3. Performance - Shorter response times, because of fewer network hops (no MySQL server as a proxy) and (1,2) above.

Typically, application developers that want soft realtime characteristics on data access use the NDBAPI, and use SQL (going through the MySQL server) for OAM type applications.

In this blog, I will try to show you how to start programming …

Today I uploaded version 1.01 of my e-book, No Lobbyists As Such - The War over Software Patents in the European Union. I just corrected a few minor errors and would like to express my gratitude for the corrections submitted by Alberto Barrionuevo and Péter Somogyi.

An hour ago I ended my 45 minutes MySQL Cluster Introduction talk at the Free and Open Source Software Conference (FrOSCon). I tried to really make this a compact, more or less complete overview. Apparently it worked!

MySQL Cluster is quite a beast to talk about. Doing this in 45 minutes, with 15min extra for questions, is not an easy task. You can check my presentation on the FrOSCon website of the presentation. I was troubled about the order of things, for example should I talk about components first, do the overview later? I feel it is OK as it is right now, though I have to cut a bit for making it shorter, and clearer.

A feedback after the talk was that I needed to touch the fact that Cluster is not really functional on Windows (yet). This made me think I indeed missed some nice information here and there.
Another comment from …

Right now I am in Skien, which is somewhere in the middle of the dark forests of Norway. Skien, a bustling town of a whopping 45.000 people is the seventh largest municipiality in Norway, and also the home of the PHP Vikinger Unconference.

My part on this Unconference was to hold a Nontalk, a session where I asked the audience to come up with things that they think are typical for Everyday PHP use and that currently do not have a name.

The following stuff are the slightly edited and commented notes I made in front of the audience while moderating our session.

Continue reading "phpvikinger.org: Things that have no name"

I’ve introduced a new design and Wordpress backend at www.vbmysql.com, feel free to take a look and provide feedback. Now that editing will be easier I think I’ll be pushing out some fresh content too.

MySQL doesn't allow referring to a table that's targeted for update in a FROM clause, which can be frustrating. There's a better way than creating endless temporary tables, though. This article explains how to update a table while selecting from it in a subquery.

I frequently get questions along the lines of "how many transactions per second MySQL can do" or "how many servers I need to handle 100.000 users" or "which hardware would be enough to handle my 40GB" database.

There are two things which are very interesting about these questions which make them funny and annoying at the same time

1) People asking these questions demand the answer. It could be internal Sales and Marketing people or even customers saying they typically get the answers from other vendors.

2) The best answer to this question will not be a number but something like "it depends". Seriously you need much more information than given in this example to give estimate which is correct at least to the order of magnitude.

Seriously in MySQL Consulting practice it would not be exception to have 1-2 orders of magnitude performance increase just by proper tuning - this means even application implementation …