Boy, the conference planners here must have some pull in high places, because they got a 5.9 earthquake to help them with their “Shake IT Up” slogan for Innovation Nation and GOSCON. I was in the upstairs conference hall, and it was pretty dramatic. I’ve never seen/felt/experienced anything quite like that before. All the walls were twisting and shimmying in different directions, and fixtures fell out of the ceiling. Just little ones — but I decided not to rush to the exit, as I was in a huge room and others were already rushing. I figured the odds of getting hurt in the rush were more than the odds of the building collapsing. But I did look upwards to see if I was going to get bonked on the head by a falling light fixture. High ceiling; time enough to dodge if I was lucky.

I met Keith Larson, Craig Sylvester, and several others whose names I’ll get wrong or misspell in the expo hall at the Oracle MySQL booth. I’m staying …

It seems to be a popular misconception that mk-slave-prefetch is designed to keep a MySQL replica server “warm” and ready to serve production traffic in case the master is demoted or fails. This is not what mk-slave-prefetch does. It’s related, and easy to confuse, but its purpose is different.

The mk-slave-prefetch tool is designed to try to execute a read-only approximation of the write workload that the replica is about to have to perform. It is meant to do this just a little bit before the replication thread (which can only be true if replication is lagging), so that when the replica replays writes to execute replication, it doesn’t have to wait for disk I/O.

Keeping caches warmed up for production traffic requires that the read workload, which does not flow through relay logs, is executed on the server. If you point mk-slave-prefetch at a server, you’re just double-executing the write …

I am writing this on my blog because I want to take the Percona hat off (partially) and wear my community hat a bit. My loyalty was to the MySQL community long before I was involved with Percona.

There are now three Percona Live events for which we have signed venue contracts and begun plans: London in October, Washington DC in January, and Santa Clara in April. I have insider knowledge of what’s going on with planning all three events, and I’m proud and happy that I have a community-member seat at the table.

The London event, which is October 24-25, will be a natural interpolation between the medium size of our very successful event earlier this year in New York, and the much larger Santa Clara event next year in April. We’re adding a day of tutorials and more speaking …

I’ll be presenting at Oracle Open World on the causes of downtime in MySQL, and how to prevent it. This is a research-based session that presents an easy-to-digest post-mortem of hundreds of emergency issues filed by Percona customers. The real causes and types of downtime surprised me quite a bit, and the preventions run counter to a lot of conventional wisdom. I’ll just give a preview by saying that you should consider it a top priority to monitor how full your disks are! On the other hand, despite the fact that every monitoring tool in existence shows the binary log cache hit rate, not a single emergency in Percona history has ever been attributed to that.

The agenda at OOW is mind-bogglingly huge (see Dave Stokes’s blog post, so here are the full …

Keith Larson (Oracle’s MySQL Community Manager) nominated me to become an Oracle ACE. Thank you Oracle — it feels very good to receive this award.

Related posts:

1. Oracle is not screwing MySQL
2. MySQL Enterprise/Community split could be renewed under Oracle
3. Migrating US Government applications from Oracle to MySQL …

Backing up databases has never been fun, not as fun as having a cool English Ale on the balcony on a hot summar day anyway, but MongoDB takes this one step further when it comes to annoyances.

In general, I often feel that many Open Source projects start with good intentions for what the project should do and how, but then more stuff is added as you require it, and suddenly what started out as a simple and fast application for a narrow usecase, suddenly turns into a bit of a mess. And the issue might well be that building fast, compact software for a specialized usecase, as they start out, is not the same as writing generic software, with a wide range of use cases, code that can easily be maintained and enhanced as we go along. And why should it not be like that? In many cases, this is just fine and the limited usecase is just what the project sets out to do, and it does it well. But sometimes this turns into something really …

Congratulations and thanks to Monty Program on the release of MariaDB 5.3! This looks like a great release. (I haven’t had a chance to try it yet.) I really welcome the query optimizer improvements; the list of changes is huge. There are a lot of fixes for problems I’ve seen for years. And there’s group commit, and high-resolution date/time types, and lots more.

And congratulations and thanks to Oracle on the next series of previews of what MySQL 5.6 will look like, including full-text search inside InnoDB, performance and scalability improvements, and usability and operational improvements. I love the batch of informational blog posts from the InnoDB team and others that usually comes out around such announcements. I …

I’ve been accepted to present at the brand-new and very exciting Postgres Open 2011 about system scaling, TCP traffic, and mathematical modeling. I’m really looking forward to it — it will be my first PostgreSQL conference in a couple of years! See you there.

Related posts:

1. Postgres folks, consider the 2011 MySQL conference
2. O’Reilly MySQL 2011 conference CfP is open
3. My sessions at the O’Reilly MySQL Conference 2011
4. …

If you have two or more database servers containing the same data, how do you know if the objects are identical. Furthermore, how can you be sure the data is the same on all of the servers? What is needed is a way to determine if the databases are in synch - all objects are present, the object definitions are the same, and the tables contain the same data. Synchronizing data can become a nightmare without the proper tools to quickly identify differences among objects and data in two databases. Perhaps a worst case (and more daunting) is trying find data that you suspect may be different but you don’t have any way of finding out.

This is where the new 'mysqldbcompare' utility comes in handy. The mysqldbcompare utility uses the mysqldiff functionality (mysqldiff allows you to find the differences in object definitions for two objects or a list of objects in two databases) and permits you to compare the object definitions and the data …

I’ve just sent an email to the Maatkit discussion list to announce a planned change to how Maatkit (and Aspersa) are developed. In short, Percona plans to create a Percona Toolkit of MySQL-related utilities, as a fork of Maatkit and Aspersa. I’m very happy about this change, and I welcome your responses to that thread on the discussion list.

Related posts:

1. Aspersa, a new opensource toolkit
2. Four companies to sponsor Maatkit development
3. How …