Global Transaction IDs (GTIDs) are one of my favorite features of MySQL 5.6. The main limitation is that you must stop all the servers at the same time to allow GTID-replication. Not everyone can afford to take a downtime so this requirement has been a showstopper for many people. Starting with Percona Server 5.6.22-72.0 enabling GTID replication can be done without almost no downtime. Let’s see how to do it.

Implementation of the Facebook patch

Finding a solution to migrate to GTIDs with no downtime is not a new idea, and several companies have already developed their own patch. The 2 best known implementations are the one from Facebook and the one from …

We often get questions about how we measure queries inside MySQL (and now PostgreSQL too). Many of these questions reveal assumptions about the metrics we provide and our methods for obtaining them. I thought it would be both interesting and helpful to clarify how we really do this. This article may get a bit dense, but by the end you will have a better sense of our database management platform’s unique approach.

VividCortex is a complex system that captures and analyzes several different types of data. It’s not solely one thing or another. The magic comes from the intersection of the various useful kinds of data:

• Unlike database monitoring tools that simply draw graphs of status counters, VividCortex measures queries the server executes.
• Unlike log analysis tools that just aggregate queries along …

The performance_schema is a powerful tool for analyzing MySQL performance and behavior. One aspect of the performance_schema is that the view of the data is "right now", and very little historical information is present. You can see that there are 10 threads connected right now, but what about five minutes ago?

ps_history
ps_history is a set of stored routines and events for periodically collecting the data in the performance_schema into another schema called ps_history. The ps_history schema contains a copy of each performance_schema view as a real table, and timestamp and server_id columns have been added to each table. Periodically (by default every 30 seconds) the performance_schema data is written into the history tables.

ps_history comes as one script (setup.sql) which will create the ps_history schema, the tables within it, and …

The performance_schema is a powerful tool for analyzing MySQL performance and behavior. One aspect of the performance_schema is that the view of the data is "right now", and very little historical information is present. You can see that there are 10 threads connected right now, but what about five minutes ago?

ps_history
ps_history is a set of stored routines and events for periodically collecting the data in the performance_schema into another schema called ps_history. The ps_history schema contains a copy of each performance_schema view as a real table, and timestamp and server_id columns have been added to each table. Periodically (by default every 30 seconds) the performance_schema data is written into the history tables.

ps_history comes as one script (setup.sql) which will create the ps_history schema, the tables within it, and …

Few times I heard from MySQL users, customers and Sales consultants about their wish to have SQL command, allowing to backup MySQL server. I asked why calling mysqbackup or mysqldump does not work for them and these people had reasons. One of them is security: this way you don't need to give shell access to MySQL server for the person who is supposed to do backup.

MySQL does not have such a command, except for MySQL Cluster. There are instead multiple programs which can make backup if run externally, such as mysqlbackup of mysqldump.

That time I wrote a UDF which runs any external command (https://github.com/svetasmirnova/run_external). So customers could run something similar to:

mysql> SELECT run_external(concat('$HOME/build/mysql-trunk/bin/mysqldump  --socket=', @@socket, ' …

In my previous post, I talked about implementing multi-threaded replication (MTR) using Percona Server 5.6. The server pairs that are utilizing MTR are also exclusively using the TokuDB storage engine.

I find TokuDB to be a fascinating engine. I can tell I will need to re-watch our Dbhangops session where Tim Callaghan talked about the differences between B-Tree and Fractal Tree indexes. There’s also a session on how compression works in …

…

Monitoring tools offer alerts based on thresholds and time-series charts of status counters, but they do not tell us what we should monitor. On Tuesday, February 10th at 2 PM EST, Baron Schwartz will discuss the guiding principles needed to answer these questions.

The following questions will be answered:

• What status counters from MySQL are central and core, and which are peripheral?
• What is the meaning of MySQL status metrics?
• Which subsystems inside MySQL are the most common causes of problems in production?
• What is the unit of work-getting-done in MySQL, and how can you measure it?
• Which open-source tools do a good job at monitoring in the way we recommend at VividCortex?
• Which new and/or popular open-source tools should you evaluate when choosing a solution?

To find out more information and register, click …

With JSON functionality, alternate protocols (HTTP, memcache), a move towards saner defaults and crash safety, pluggable logging etc it really looks like MySQL is following what we did in Drizzle years ago, which is great!

I recently updated the High Performance MySQL website to modernize it. I am impressed at how easy it is these days to get a great little brochure site hosted. It used to be a lot more work. I used a variety of tools and services to do this and decided to share this for people who are interested. Hopefully you’ll add comments and point me towards more tools and tips to make these things even easier for me in the future!

Before

The website used to be a WordPress blog. This used to be my go-to solution for everything. It used to be the easiest way to whip together something quickly and put it online.

But WordPress sites have a bunch of problems.

• They need a database. Database down, website down.
• They need care and feeding. WordPress is a major hack target and you have to update it or you’re going to end up serving malware and ads without …