(based on a conversation with a colleague, and a bit of Twitter)

A Conundrum

A developer colleague paged me with this:

mysql> select
UNIX_TIMESTAMP("2021-03-26 03:07:00" + INTERVAL 2 YEAR) -
UNIX_TIMESTAMP("2021-03-26 02:07:00" + INTERVAL 2 YEAR) as delta\G
delta: 420

It is obviously wrong, and weirdly so. It only works for “2 year”, not with other values:

mysql> select
UNIX_TIMESTAMP("2021-03-26 03:07:00" + INTERVAL 1-11 year_month) -
UNIX_TIMESTAMP("2021-03-26 02:07:00" + INTERVAL 1-11 year_month) as delta\G
delta: 3600

mysql> select
UNIX_TIMESTAMP("2021-03-26 03:07:00" + INTERVAL 1-12 year_month) -
UNIX_TIMESTAMP("2021-03-26 02:07:00" + INTERVAL 1-12 year_month) as delta\G
delta: 3600

mysql> select
UNIX_TIMESTAMP("2021-03-26 03:07:00" + INTERVAL 1-13 year_month) -
UNIX_TIMESTAMP("2021-03-26 02:07:00" + INTERVAL 1-13 …

The tables in PERFORMANCE_SCHEMA (P_S) are not actually tables. You should not think of them as tables, even if your SQL works on them. You should not JOIN them, and you should not GROUP or ORDER BY them.

Unlocked memory buffers without indexes

The stuff in P_S has been created with “keep the impact on production small” in mind. That is, from a users point of view, you can think of them as unlocked memory buffers – the values in there change as you look at them, and there are precisely zero stability guarantees.

There are also no indexes.

Unstable comparisons

When sorting a table for a GROUP BY or ORDER BY, it may be necessary to compare the value of one row to other rows multiple times in order to determine where the row goes. The value compared to other rows can change while this happens, and will change more often the more load the server has. The end …

At work, I am in an ongoing discussion with a number of people on the Observability of Outliers. It started with the age-old question “How do I find slow queries in my application?” aka “What would I want from tooling to get that data and where should that tooling sit?”

As a developer, I just want to automatically identify and isolate slow queries!

Where I work, we do have SolarWinds Database Performance Monitor aka Vividcortex to find slow queries, so that helps. But that collects data at the database, which means you get to see slow queries, but maybe not application context.

There is also work done by a few developers which instead collects query strings, query execution times and query counts at the application. This has access to the call stack, so it can tell you which code generated the query that was slow.

It …

We are proud to announce that Percona was selected as a participating organization for the Google Summer of Code (GSoC) 2020 program, this is our second year as a participating org with the GSoC program.

GSoC is a great program to involve young student developers in open source projects. We participated in the program in 2019 for the first time and we were really happy and satisfied with the results.
Percona Platform Engineering team decided to participate again for the 2020 program and we are glad and really happy to inform you that we were selected and welcome the student to work with our team during the summer of 2020 on their GSoC Project.

Preparations

We started planning for GSoC around November-December 2019, with the help from our Product Management team, we were able to shortlist a few ideas which we thought were really the right fit for …

This session at Percona Live ONLINE was presented by Akshay Suryawanshi, Senior Production Engineer at Shopify, and Jeremy Cole, Senior Staff Production Engineer – Datastores at Shopify.

Shopify is an online and on-premise commerce platform, founded in 2006. Shopify is used by more than a million merchants, and hundreds of billions of dollars of sales have happened on the platform since its inception. The company is a large user of MySQL, and the Black Friday and Cyber Monday weekends are their peak dates during the year, handling hundreds of billions of queries with MySQL. This year’s presentation was an opportunity to talk about the company’s challenges and progress over the last twelve months.

Key Google Cloud concepts from the presentation

As part of the presentation, it’s important to understand the naming conventions that exist around …

MySQL, arguably the most popular relational database, is used pretty extensively at the popular professional social network LinkedIn. At Percona Live ONLINE 2020, the company’s flagship event held online for the first time due to the Covid-19 pandemic, Karthik Appigatla from LinkedIN’s database SRE team discussed the company’s approach to securing their database deployment without introducing operational hiccups or adversely affecting performance.

Instead of just performing admin duties, Karthik’s team builds automated tools to scale their infrastructure, and he talked about some of these tailored tools in his presentation. The database SREs on his team also work with the developers at LinkedIn and help them streamline their applications to make best use of the database.

Talking about LinkedIn’s reliance on MySQL, Karthik said that not only do all their infrastructural tools rely on MySQL, many of the internal …

Long months of the pandemic lockdown have brought to life many great online events enabling the MySQL community to get together and stay informed about the very recent developments and innovations available to MySQL users. It isn’t over yet! Next Thursday, June 4th, Percona & ProxySQL are co-hosting the ProxySQL Tech Talks with Percona virtual meetup covering ProxySQL, MySQL and Percona XtraDB Cluster.

The attendees are invited to participate in the two-hour deep-dive event with plenty of time for questions and answers (we will have two 40-minute sessions + 20 minutes allocated for Q&A). Get prepared, come with your burning questions and true war stories – we’ll have our speakers answer and comment on them! And here come the speakers:

• René Cannaò, ProxySQL author and CEO of ProxySQL …

In my last blog post I showed you how to get the entire Percona “Stack” up and running on a Raspberry Pi. This time around, I would like to show the impact on performance between using an SSD hard disk and a standard hard disk.

Disk performance is a key factor in Percona Server for MySQL (or any RDB platform) performance on a Raspberry Pi 4.

Test set up

Each test below was run three times per Hard Disk and I took the best of the three for comparison.

Hardware

• Raspberry Pi 4+ with 4GB ram.
• Disk 1: USB3 Western Digital My Passport Ultra, 1TB
• Disk 2: USB3 KEXIN 240GB Portable External SSD Drive

Hardware stayed consistent …

Starting at version MySQL5.6+ by default innodb_file_per_table is enabled and all data is stored in separate tablespaces.

It provides some advantages. I will highlight some of them:

• You can reclaim disk space when truncating or dropping a table stored in a file-per-table tablespace. Truncating or dropping tables stored in the shared system tablespace creates free space internally in the system tablespace data files (ibdata files) which can only be used for new InnoDB data.
• You can store specific tables on separate storage devices, for I/O optimization, space management, or backup purposes. …

This is the first in a series of blog articles in which I will discuss the changing landscape of open source and money. Or, more specifically, open source databases and money. And even more specifically MySQL and its all variants (AWS Aurora, MariaDB, Percona Server, RDS/MySQL) and money. But before going too deep into what is changing, let’s review all the traditional business models in and around the MySQL marketplace.

In general, these are the following types of companies in the MySQL commercial ecosystem, sorted by total annual revenue and addressable market size:

• Developers who do not aim to monetize the open source code, just provide value to others and hope to get development and other contributions in return. This is the purest form of open source. For example, all …