As you might have seen MySQL QA Team has published their benchmarks for MySQL 5.0.72 and 5.1.30.

It is interesting to compare with results I posted previously

The quote from the results mentioned above:

"Maybe you’ve seen some claims by others in the MySQL community that MySQL 5.1 runs slower than MySQL 5.0. Maybe you’ve also seen some claims by others in the MySQL community that MySQL 5.1 runs faster than MySQL 5.0.

Guess what? They’re both right. "

But is it really what results are telling us ?

I do not think so. When you're doing benchmarks you should be comparing best performance settings for given application and conditions. For example it is unfair to …

I have had the privilege to go to the O'Reilly Open Source Convention 4 of the past 5 years.  During that time, it has been held in beautiful Portland, Oregon.  Well, the OSCON 2009 web site is up and they are moving it to San Jose, CA.  I have never been to the San Jose Convention Center.  The pictures look nice.  I have only been to San Jose at all to get off a plane and go to the Santa Clara Convention Center.  I hope San Jose has more to offer than Santa …

Why a new tool?

The MySQL test framework needs a tool that will increase the production of test cases. The proposed tool will be a command line client that records queries and responses to then create the .test and .result files. There were a couple of proposed options that would also achieve a similar objective. The first was using the MySQL command line client and logging the client-server interaction, then creating the appropriate test files. Using this method, writing test cases for concurrent connections would have to be serialized and some concurrent test cases cannot be written in this fashion. Another option was patching the MySQL command line client to record queries and responses like my proposed tool. This would work well with concurrent connections, but a tool would have to be built to convert the log file to .test and .result files regardless. So instead of implementing these alternatives that would take more time …

After weeks of revising UML diagrams I think it is actually time to write some code and prove myself to be at least somewhat productive. Here is the final UML diagram for the tool that I will be creating.

As I go forth and begin coding I have decided to use the extreme programming practice of prototyping and testing. This will allow me to modify the design if need be during the project depending on what incompatibilities are found, but I will hopefully find them faster than if I implemented this in a modular fashion.

As always, any comments are more than welcome.

A few days ago the mysql proxy tool was released (link) and it does exactly what the name would suggest. But it does boast a few features that have changed the development of the test creation tool.

• It has a Lua interpreter that allows user written scripts or modules to be run by the proxy. These scripts can intercept queries and responses between the client and the MySQL server, and also manipulate them in whatever way necessary. There are examples on the wiki of how to use the scripts.
• Instead of writing an entirely new command line client, we can use the standard mysql client and focus on the scripts that will intercept the client-server communication and translate that into test cases.

For these reasons the test creation tool will not be written in C/C++ but instead as Lua modules that the proxy will run. This will hopefully …

Via Ditesh:

Interesting aside: the bus stopped in a ${RANDOM} town in Johor for food, and at the cashier, I spotted the cashier using KC POS which had a prominent “Powered by MySQL” text and the Sakila logo. Very cool!

This was a small town coffee shop, using a cash register, powered by the mighty Sakila. Similar to the chain of restaurants, Old Town White Coffee.

The whole blog post from Ditesh itself is interesting, but knowing you can find MySQL just about anywhere, showing the ubiquity of the database, just makes you proud to know, you work at/on/with MySQL.

This week Zmanda was invited to participate in the launch of OpenSolaris 2008.11. In his interview Chander Kant, CEO of Zmanda, explained why OpenSolaris is a big step forward from backup and recovery perspective both as a data source for backup and as a repository for backup images.

The advanced snapshot capabilities of ZFS enables fast and scalable backups for today’s most demanding workloads for servers and easy to use data protection for workstations. For example, the new Time Slider feature in OpenSolaris provides an automatic way to backup your data on the local disk with intuitive browsing and recovering of files from snapshots using the GNOME file manager. That is great, but what if you loose your local disk? Considering that Time Slider does not support network attached disk yet, you might be in danger …

We’re closing in on features left to be implemented in 5.1 and keep improving on stability. All object editors are in place now (some of them still in a basic version though) and a bunch of memleaks has been fixed already. Additionally we now statically link the ctemplate-library and therefore you won’t have to care about installing/maintaining that package on your own anymore.

We provide bin and src packages for Ubuntu and Fedora9 (both 32bit) at the moment. Please grab the latest packages from our official download page and Test-drive our latest offspring. Let us know about any problems on using or building (also for other distributions) this alpha verison.

Wasn’t there a release just some days ago, you might ask? Yes, indeed - but we had to create a quick follow-up because a nasty bug found its way through our test and into the recently released 5.0.28. A bit of debug code from our latest canvas changes was left behind and causes Workbench to crash under certain circumstances. Once discovered we found and fixed the root of the problem quite quickly. Additionally we introduced some new icons last time which didn’t find their way into the installer script as well … so frankly its a quickly delieverd follow up to fix 2 things we missed on last release.
We strongly recommend to update to this release asap.

Download version 5.0.29 from our Download Page or - for our SE users - use the integrated update wizard to get you up2date.

A couple of people have asked for a means to set the list of INSERT statements of a table from data already in an external file. While we haven’t implemented that feature in Workbench as of now, that’s something that can be easily added as a Lua plugin.

The plugin is really simple, but I’ll take the opportunity to use it as a quick tutorial on how to write a plugin from scratch, using the mini-IDE in Workbench 5.0  Note that this tutorial is meant for MySQL Workbench 5.0; Workbench 5.1 has some differences in how plugins are handled and that is not covered here.

1.  Start Workbench and Open the “GRT Shell IDE”

2. Click on “New GRT File” button

3. Select the “Lua GRT Module File” option, “Table_Plugin” template and fill in a name for the file, eg: table_utils

4. …