This week I already had two serious performance regression cases when upgrading from MySQL 4.0 and 4.1 to MySQL 5.0. By serious I mean several times performance difference not just 5-10% you often see for simple queries due to generally fatter code.

The problem in both cases was MySQL 5.0 broken group commit bug.

First I should note I am extremely unhappy how MySQL handled this problem. While working for MySQL we spotted this problem early in MySQL 5.0 release cycle as it was introduced and reported it to everyone we could inside the company - this was over 2 years ago. Few months later I created a bug for this issue to get more public attention to the problem and giving extra motivation to MySQL to fix it. Few months later I blogged about this problem with more …

Just short message that patch enables microsecond resolution in slow-log (see more http://www.mysqlperformanceblog.com/2006/09/06/slow-query-log-analyzes-tools/) for 5.0.37 is available here:
www.mysqlperformanceblog.com/files/patches/patch.slow-micro.5.0.37.diff

The patch for 5.0.41 :
www.mysqlperformanceblog.com/files/patches/patch.slow-micro.5.0.41.diff

There is a popular myth about the SQL GROUP BY clause. The myth holds that 'standard SQL' requires columns referenced in the SELECT list of a query to also appear in the GROUP BY clause, unless these columns appear exclusively in an aggregated expression. MySQL is often accused of violating this standard.

In this article I will attempt to debunk this myth, and to provide a more balanced view regarding MySQL's treatment of GROUP BY at the same time.

MySQL Archiver is the implementation of the efficient forward-only archiving and purging strategies I wrote about more than a year ago. It nibbles rows from a table, then inserts them into another table and/or writes them to a file. The object is to do this without interfering with critical online transaction-processing (OLTP) queries. Several people have asked me to release this code, which I originally wrote for my employer. As it turns out, the delay has been fruitful.

I came across a list of possible questions for a security panel and thought it may be interesting to apply some of them to MySQL and see what happens .

* What is the biggest challenge in data security?

I would suggest it is always the same problem with security - human error. It seems that no matter how well you protect the data, people will be the weakest link in the chain. This could come from social engineering, or to the point where the admin does silly things like deleting/dropping data from the wrong area.

How can we address this problem with MySQL? There are some configuration options that let you force safer interaction like safe-updates etc, but it will never be enough. Backups are probably the best answer.

read more

I came across a list of possible questions for a security panel and thought it may be interesting to apply some of them to MySQL and see what happens .

* What is the biggest challenge in data security?

I would suggest it is always the same problem with security - human error. It seems that no matter how well you protect the data, people will be the weakest link in the chain. This could come from social engineering, or to the point where the admin does silly things like deleting/dropping data from the wrong area.

How can we address this problem with MySQL? There are some configuration options that let you force safer interaction like safe-updates etc, but it will never be enough. Backups are probably the best answer.

read more

OK, this is maybe no big news and maybe old, but I needed an excuse to blog something!

Since MySQL 5.1.15 the /etc/mysql/my.cnf has been added as default location after /etc/my.cnf. This is great and something I like keeping /etc a bit cleaner, but it might be tricky for some not knowing!

I found this the hard way and a bit with red cheeks (I'm a MySQL Support Engineer after all..). After installing postfix-mysql Ubuntu package, which also sets up common MySQL stuff, I couldn't connect to my fresh installed 5.1.19. Confused, I saw it was using a socket setting I didn't specify. So, /etc/mysql/my.cnf was read extra to /etc/my.cnf! I said 'doh!' and saw it was good.

Tip: leave /etc/my.cnf there for a while with a warning like "You're living in the past dude! Check /etc/mysql/my.cnf", or something.

Next would be to read /etc/mysql/config.ini for MySQL Cluster too, but work is on its way to …

Tonight, I went for the first time to the NY Tech Meetup. (Meetup's founder is also Fotolog's co-founder).

Ever wanted to create your own television station And be able to professionally produce live shows just like your favorite cable tv channel? What about being able to instantly add any video from YouTube, Revver or other video sharing sites? And what if all this was free? Close your mouth, I can see your mouth watering already.

Check out what I honestly think has the potential to be ground breaking: Mogulus. I had the chance to meet and see Max Haot, CEO of Mogulus, at the NY Tech Meetup. The entire interface is flash based and therefore no download is necessary.

I think Mogulus seems to be using S3 and …

I know this is a bit old, but I’ve been trying to catch up with all the new stories after a conference, vacation, broken laptop and loads of work.

RedHat Enterprise Linux 5 (RHEL 5) was launched some months ago. I’ve never been a big fan of RedHat in terms of technology. I guess is quite good now, but RPMs scared me years ago and I’m not over it yet.

One of the things I liked is that RedHat proved that the KISS principle doesn’t only apply to software development, but to marketing and sales too. This is their new SLA (service level agreement). Can’t be simpler.

That reminds me of MySQL and their all-you-can-eat support package (MySQL Enterprise Unlimited), easy to understand and with a …

The weather was great here in Seattle this last weekend. Nice and sunny,..

The Google Gears stuff is still on the brain :)

Pulled the source code to take a look at it. The database code is laid out in the "database" directory. It is split up for each browser, with a "common" directory (though little seems to be common). The code is written in C++, with an object type of "GearsDatabase" representing the database itself. The code is lightly commented, but the naming conventions make the code self readable. There is more database code then what is found in this directory though, so there is not a full encapsulation of the database code.

And the real question I wanted to answer for myself?

What about making MySQL work?

We could make MySQL work and reuse the Google code. Not cleanly though... there interfaces weren't …