Ahoi,Attending DOAG 2013 next week, Im going to give two talks (Applikationsvirtualisierung mit LXC and MySQL Replikation).Also Im invited to attend a MySQL Expert Panel, where you can let a bunch of MySQL Experts discuss your questions ;)DOAG is afaik the biggest Oracle Event (and even not driven by Oracle) in Europe.A good place to have nice discussions and learn a lot of new stuff :)
Hope to see youErkan Yanar

One of the best DBA questions I was ever asked in an interview, was to try and estimate required disk space for an application using a back-of-the-envelope calculation.

The idea behind a question like this, is that it is a thought exercise. Its aim is to better test competency than some of the simple questions that can be Googled for an answer (i.e. whether or not you know the answer is not important. You will in 2 minutes :)) Unfortunately I have also seen it negatively affect candidates that are not native English speakers, so be aware that your mileage may vary.

Today I wanted to try and recreate the back-of-the-envelope exercise that I was asked to complete. For simplicity the question will be restricted to database size and not include RAM or IOPS estimations.

Scope of Problem

“We are designing a new system to store application users for a very demanding system. There is only one table, but it is …

Economist says that Physics suggest that storms will get worse as the planet warms. Typhoon Haiyan in Philippines, bush-fires in Australia, floods in China, and extreme unpredictable weather across the planet is a sober reminder. Good news is that technology and awareness is rising, and so is the data. Database technologies are playing their part to intelligently store that data and enabling the stakeholders to analyze and get meaningful results to predict and counter the extreme conditions. This Log Buffer Edition appreciates these efforts.

Big Data:

Big Data Tools that You Need to Know About – Hadoop & NoSQL.

Dave Stokes is …

One of the most common errors in development is where a loop or a retrieval by index falls short or long by one unit, usually because of an oversight or a logic in coding.

Of the following snippets, which one will run 10 times?

/* #1 */    for (N = 0 ; N < 10; N++) printf("%d\n", N);

/* #2 */    for (N = 0 ; N <= 10; N++) printf("%d\n", N); 

/* #3 */    for (N = 1 ; N <= 10; N++) printf("%d\n", N);

/* #4 */    for (N = 1 ; N < 10; N++) printf("%d\n", N);

The question is deceptive, as there are two snippets that will run 10 times (1 and 3). But they will print different numbers. If you ware aiming for numbers from 1 to 10, only #3 is good.

After many years of programming, off-by-one errors are rare in my code, and I have been able to spot them or prevent them at first sight. That’s why I feel uneasy when I look at the way parallel replication is enabled in …

The problem

About two weeks ago I noticed my iPhone 5 had a faint, but ugly pink hue in the center of the screen. At first I suspected some software issue, because it looked as if the hue was exactly in the area an iOS7 table view uses for its cells to be displayed (i. e. full screen, but with a few points of inset on the left and right edges). Apparently someone must have left some kind of view with a red-ish color behind the table view that was now shining through, because of all the translucency effects iOS7 comes with.

Turns out though, that this was going on in all kinds of apps, including Instapaper, Pocket, Chrome, Safari Calendar etc. All those certainly would not put any views behind the actual content. Getting more suspicious I showed the problem to several colleagues who all did not see it at first, but once I had pointed it out to them (most visible on a white background) none of them could "unsee" it, …

Thanks to everyone who participated in Percona Live London 2013. We had a great event with nearly 20% higher attendance versus 2012. This year’s conference featured participation by representatives from Oracle including a compelling keynote by Tomas Ulin. The conference featured 10 tutorial sessions on Monday and 4 keynotes and 30 breakout sessions on Tuesday.

I personally found the keynotes highly interesting. Peter Zaitsev of Percona discussed the changing face …

Next Wednesday (November 20th, at 10 am PST), Ryan Lowe (Percona: Principle Consultant) and Andrew Aksyonoff (Sphinx: CEO and CTO) will be delivering a webinar on how to configure Sphinx for MySQL. “How to Optimally Configure Sphinx for MySQL” The discussion will center around getting started with, and seamlessly integrating, Sphinx into your MySQL-based applications. [...]

I have spent the better part of the last month at Big Data conferences trying to see behind the $2.5 million in marketing smoke to see what is really going to be showing up on the to-do list of DBAs. The first bit of news is that half the vendors at shows like Strata or Big Data Techon will probably be gone by this time next year. So picking a vendor right now is a little iffy. Hadoop’s ecosystem is flourishing and will surely be around for some time but the vendors are playing musical chairs.

But we are Open Source and we do not need vendors! Well, yes and no. The good folks at Cloudera and Horton Works have done you a big favor by providing wonderful tutorials that are worth your time to see. Recently two former MySQL-ers, Sarah Sproehnle and Ian Wrigley, have put together Udacity that concisely teaches Hadoop technology and Cloudera deserves a round of applause for this …

I recently had an issue with a client where they needed developers to have enforced lag on master/slave dev setup. There were issues where slave lag wasn't accounted for and with their previous development setup they never had lag. We needed to essentially always have some amount of slave lag so these bugs were exposed. My first thought was to use pt-slave-delay, which is part of the Percona toolkit. As an initial stab at solving the problem I configured pt-slave-delay for 30 seconds of replication lag and have it check every 15 seconds for the amount of current delay. I did something like this:

pt-slave-delay --delay 30s --interval 15s --daemonize 127.0.0.1 --log /var/log/pt-slave-status/pt-slave-delay.log

The problem was that, like many development environments, this development environement didn't have a very large amount of traffic. This caused the repliction lag to be uneven. Not really idea. …

Before there was information_schema and performance_schema, there were InnoDB Monitor tables.

What the monitor tables are, is a way of communicating directly to the InnoDB storage engine by creating a table of a special name. Additional debugging information is then written to MySQL’s error log. For example:

CREATE TABLE innodb_monitor (a INT) ENGINE=INNODB;

/* view to the error log */

DROP TABLE innodb_monitor;

There’s a number of problems with this approach:

1. By using a magic table name, there is no immediate feedback for typos or spelling errors in monitor table names. The user has to wait to see if the error log is written to, and then debug if not.
2. The monitors will continually write to the error log as long as the magic table still exists. This can reduce the …