There’s one thing that has always bugged me when I review configuration files. People always seem to want to set certain buffer settings to ridiculously high values without really understanding what they’re doing.

Case in point: sort_buffer_size.

The misconception seems to be that setting this value to a high number will always improve a servers performance because everything works better with more memory, right? For some variables (key_buffer, innodb_buffer_pool) maybe, but the entire length of a sort buffer is allocated when ORDER BY is used. In practice, I find it best to leave it at default and watch sort_merge_passes in your status counter to determine if you need to adjust this value. Also, adjust in small amounts.

…

I will be presenting at Percona Live 2014 and I’m excited to share and discuss the latest and greatest features and improvements that we have made to MySQL/InnoDB in 5.7. Great performance improvements, there are some new exciting compression features that we are working on,  GIS support,  temporary table performance etc.. There is a long list. Also, we are always interested to hear about user issues and priorities so that we can address them and/or work them into our plan. Your feedback is very important for us, if you want to influence the direction of InnoDB development then you need to talk to me .

Percona Toolkit has a great tool, pt-query-digest, that can use tcpdump data. Capturing raw tcp data can be taxing on a server, however, when you see the following message:

64000 packets received by filter
12000 packets dropped by kernel

When there is a significant amount of user cpu% being used, the kernel will drop packets you are trying to capture, leading to a partial picture and missing data. I’ve found that if you write it using the native tcpdump format, it’s more efficient and you drop less. There are also recommendations on Stack Overflow on how to help prevent this.

To perform a capture for a specific length of time, here’s the trick I …

Some people think that performance tuning is all about reducing CPU or memory utilization. However CPU and memory are designed to be used. You should concentrate your efforts on improving the user experience by reducing response times. You can do this by optimizing your queries and by enabling your system to serve more requests.

To learn more about improving the performance of the MySQL database, consider taking the MySQL Performance Tuning course.

You can take this 4-day instructor-led course through the following delivery methods:

• Training-on-Demand: Start training within 24-hours of registering, following lectures at your own pace via streaming video and booking time on a lab environment at a time that suits you.
…

You may think that you have to buy, install, and get up to speed on a new database if you want to work with large amounts of data, but you can do more than you think with the MySQL you already have.
Register Now!

SPEAKER: Jon Tobin, Tokutek
DATE: Thursday, March 20th
TIME: 1pm ET

Without having to change your application or do special tuning you can increase performance and save significant time and money when you need to scale.

Join Tokutek’s Jon Tobin as he demonstrates how to use MySQL or MariaDB in Big Data applications by simply upgrading the storage engine with TokuDB, and how to effectively evaluate TokuDB for increased performance, compression and agility.

During this webinar you will learn:

• How to dramatically increase …

Parallel replication is in MariaDB 10.0. I did some benchmarks on the code in 10.0.9. The results are quite good! Here is a graph that shows a 10-times improvement when enabling parallel replication:

The graph shows the transaction per second as a function of number of slave worker threads, when the slave is executing events from the master at full speed. The master binlog was generated with sysbench oltp.lua. When the binlog is enabled on the slave and made crash-safe (--sync-binlog=1 --innodb-flush-log-at-trx-commit=1), the slave is about ten times faster at 12 worker threads and above compared to the old single-threaded replication.

These results are for in-order parallel replication. With in-order, transactions are committed on the …

If you have a busy schedule or too much going on to attend a live training session, Training-on-Demand is for you. With Training-on-Demand, you can watch recorded classroom sessions taught by top Oracle instructors and gain hands-on experience with a dedicated lab environment.

In viewing the classroom sessions, you access streaming lectures on your desktop or tablet, with the ability to fast-forward, pause, rewind and search.

MySQL for Database Administrators is available as training-on-demand. This course is designed for DBAs and other database professionals. Learn to configure the MySQL Server, set up replication and security, perform database backups and performance tuning and protect MySQL …

(This is a cross-post from the Engine Yard blog. The original article appears here.)

Here is a question I've actually been asked a few times:

"I am writing a batch processing script which modifies data as part of an ongoing process that is scheduled by cron. I have the ability to group a number of modifications together into a transaction, but I'm not sure what the correct number is?"

First off, I think that this question is interesting not just in the context of batch processing, but it equally applies to all parts of the application. If you are designing a high throughput system for MySQL, there are actually some potential pain points that you can design your way around.

Potential Pain Points

Here are the situations where the size of the transaction could impact performance:

Very Small …

Few weeks ago I asked my friends who speak both English and Russian if it is worth translating slides about Performance Schema which I prepared for a seminar at Devconf 2013. They said it is. Today I finished translation and uploaded slides to SlideShare.

Strictly speaking simple translation of slides is not enough, because they were created for the seminar where I was going to explain what they mean. I think I need to repeat same seminar, this time in English language. But if you have rough imagination about what Performance Schema is and need hints for practical use you will find such suggestions in the slides. You will also find ready-to-use queries which you can use to troubleshoot most frequent performance issues.

Enjoy!

MySQL is very popular and growing on the Windows platform. As a consequence we want to further staff our Windows efforts with a full time Windows performance architect. The Windows performance architect will be an integral part of the MySQL Server development team and also work closely with our Windows Experience team. Work will be centered around profiling and tuning large multi-threaded service applications on the Windows platform, debug and remove limitations and develop new code.

Key qualifications are:

• 5+ years C/C++ software development experience on Windows platform
• Ability to understand large pieces of code fast
• Strong understanding of long running programs (for example Operating System Kernels, Filesystems, high end networking equipment)
• Strong understanding of multi-threaded service applications
• Proven skills in designing and coding complex environments
…