CONFERENCE CONFERENCE AND TUTORIAL AND TUTORIAL SCHEDULE SCHEDULE Thanks to our sponsors:

This is the interactive guide to Percona Live Europe 2019. Links are clickable, including links back to the timetables at the bottom of the pages. You can register at www.percona.com/live-registration Sections Daily Schedules Talks by Technology Keynotes Monday Tutorials Tuesday Talks Wednesday Talks Speakers

Spot talks by technology in the timetable: a color key

MySQL MongoDB

PostgreSQL Other , Multiple Databases, and Other Topics TUTORIALS DAY MONDAY, SEPTEMBER 30

ROOM ROOM ROOM ROOM A B 7 26 second floor PostgreSQL For Oracle and Accelerating Application MySQL 8.0 InnoDB Cluster: 9:00 MySQL DBAs and MySQL 101 Tutorial Part 1 Development with 9:00 Easiest Tutorial! For Beginners Amazon Aurora

12:00 LUNCH 12:00

Innodb Architecture and Accelerating Application Introduction to 1:30 Performance Optimization MySQL 101 Tutorial Part 2 Development with 1:30 PL/pgSQL Development Tutorial for MySQL 8 Amazon Aurora

4:30 WELCOME RECEPTION – EXPO HALL OPEN UNTIL 6PM 4:30

ROOM ROOM ROOM 8 9 10 Open Source MariaDB Server 10.4: Percona XtraDB Cluster 9:00 Performance Optimization 9:00 The Complete Tutorial Tutorial and Monitoring with PMM

12:00 LUNCH 12:00

Test Like a Boss: Deploy and A Journey with MongoDB HA. Getting Started with 1:30 Test Complex Topologies From Standalone to Kubernetes and Percona 1:30 With a Single Command Kubernetes Operator XtraDB Cluster

4:30 WELCOME RECEPTION – EXPO HALL OPEN UNTIL 6PM 4:30 TUESDAY, OCTOBER 1

ROOM Building Large, Scalable & ROOM Hot Topics, ROOM Performance & A Secure Deployments B New Features & Trends 7 Scalability 9:00 KEYNOTES 9:00

10:30 COFFEE BREAK 10:30

Our journey to better MySQL More Than a Query Language: MySQL 8.0 Performance: 11:00 Availability Using Global Transaction 11:00 SQL in the 21st Century Scalability & Benchmarks IDs, ProxySQL and Consul

12:00 LUNCH 12:00

Tarantool: Percona XtraDB Cluster 8.0 1:30 A NoSQL Database What’s New in PMM 2.0 1:30 (PXC-8.0) with SQL How to protect PII data in MySQL MyRocks and RocksDB Advanced New Indexing and Aggregation Pipeline 2:30 While Allowing Customers to 2:30 Features and Performance Capabilities in MongoDB 4.2 Access the Database

3:30 COFFEE BREAK 3:30

Lessons From PCI/DSS Compliance with MySQL: What’s New in 4:00 Building Automation for a 4:00 2019 Edition Percona Server for MongoDB Large Distributed Database InnoDB Management and MySQL InnoDB Cluster: 5:00 Backing up Wikipedia Databases Scalability Improvements 5:00 Advanced Configuration & Operation in MySQL 8.0 From Scheduled Downtime 6:00 to Self-Healing MySQL 8.0.18: Latest Updates Tracing and Profiling MySQL 6:00 in Less Than a Year COMMUNITY EVENT BY BOOKING.COM AT 7:00 7:00 HQ, HERENGRACHT 597, 1017CE AMSTERDAM TUESDAY, OCTOBER 1

ROOM Public, Private & ROOM Monitor, Manage, Maintain ROOM Reduce Costs & Complexity 8 Hybrid Cloud 9 Databases @ Scale 10 with OSS Databases 9:00 9:00

10:30 COFFEE BREAK 10:30

Vitess: Running Sharded MySQL Amazing Sandboxes Multi-Document Transactions 11:00 11:00 on Kubernetes with dbdeployer Through MongoDB shell

12:00 LUNCH 12:00

MySQL on Google Cloud: MongoDB Data Security: Why PostgreSQL is Becoming A 1:30 The Good, The Bad, 1:30 Custom Rules and Views Migration Target in Large Enterprises and The Ugly Pg_catalog Unrevealed! Comparison of 2:30 That Part of PostgreSQL You Are Enhancing MySQL Security 2:30 Kubernetes Operators for MySQL Probably Underusing

3:30 COFFEE BREAK 3:30

Join Heterogeneous Databases ClickHouse for Time-Series Improving Enterprises HA and 4:00 Using PostgreSQL 4:00 Real-time Analytics Disaster Recovery Solutions Foreign Data Wrappers VictoriaMetrics: Why and How Deep Dive on High Availability and Automatic 5:00 We Built A Fast and Scalable 5:00 Amazon Aurora Failover in PostgreSQL Open Source Time Series Database

ClickHouse Features to OpenCorporates: Providing 6:00 6:00 Blow Your Mind Transparency for the Public Benefit

COMMUNITY EVENT BY BOOKING.COM AT 7:00 7:00 HQ, HERENGRACHT 597, 1017CE AMSTERDAM WEDNESDAY, OCTOBER 2

ROOM Building Large, Scalable & ROOM Hot Topics, ROOM Performance & A Secure Deployments B New Features & Trends 7 Scalability Parted Ways with Partitioning? It’s Percona Server for MySQL 8.0: Billion Goods in Few Categories: 9:00 9:00 Time to Reconsider What It Is and How It Is Done How Histograms Save a Life? ClickHouse Materialized Views: 10:00 Percona Backup for MongoDB Databases, Crypto & Decentralization A Secret Weapon for 10:00 High Performance Analytics

10:30 COFFEE BREAK 10:30

Data Protection and OSS in What’s New on Sharding in 11:00 My First 90 Days With Vitess 11:00 the Age of GDPR MongoDB 4.2

10 Common Mistakes (Java) Securing MySQL with Vault Benchmarking Should 11:30 Developers Make When 11:30 and Table Level Encryption Never be Optional Writing SQL

12:00 LUNCH 12:00

Strength in Optimize and MySQL 8.0: 1:30 Numbers: Slack’s Database Troubleshoot MySQL 1:30 The New Replication Features Architecture Using PMM 2.0

Managing MySQL at Scale Graph Databases: Introduction, How to Instrument Your Code 2:30 2:30 in Standardization, Opportunities in Performance Schema

3:30 gdb Basics for MySQL DBAs 3:30

4:00 KEYNOTES, PRIZE GIVING, AND SUMMING UP 4:00 WEDNESDAY, OCTOBER 2

ROOM Public, Private & ROOM Monitor, Manage, Maintain ROOM Reduce Costs & Complexity 8 Hybrid Cloud 9 Databases @ Scale 10 with OSS Databases MySQL Has Gone Away: How a Modern Database Maintenance for 9:00 An In-Depth Look at the MySQL Gets Your Data Fast: 9:00 MongoDB Replica Sets Networking Implementation MariaDB Query Optimizer

MySQL Shell : The Best DBA tool? Fortify Your MySQL Data Security in Percona Distribution 10:00 Extend the Shell with the 10:00 AWS Using ProxySQL Firewall for PostgreSQL New Extension Infrastructure

10:30 COFFEE BREAK 10:30

Running PMM in Production Handling Transaction ID 11:00 TBC 11:00 at Tessi Wraparound in PostgreSQL

Top 10 Mistakes When Migrating Large Scale Deployment of 11:30 Protecting your Secrets 11:30 From Oracle to PostgreSQL SSL/TLS For MySQL

12:00 LUNCH 12:00

MongoDB Analysis with How to Upgrade Like a Boss Running ElasticSearch 1:30 1:30 Prometheus and Grafana to MySQL 8.0? at Large Scale

Becoming Cloud Native: Automatic Upgrade and The Database is Broken. 2:30 How Percona Brings Databases to New Error Logging 2:30 Now What? Kubernetes Using Operators in MySQL 8.0

PostgreSQL Plan at Execution Time: 3:30 JSON Array Indexes in MySQL 3:30 A Quick Show

4:00 4:00 TECHNOLOGIES

MySQL & MariaDB MySQL & MariaDB MySQL & MariaDB MONDAY TUESDAY WEDNESDAY MySQL 101 Tutorial Parts 1 & 2 How to protect PII data in MySQL Percona Server for MySQL 8.0: MySQL 8.0 InnoDB Cluster: While Allowing Customers to What It Is and How It Is Done Easiest Tutorial! Access the Database Billion Goods in Few Categories: MariaDB Server 10.4: Comparison of Kubernetes Operators How Histograms Save a Life? The Complete Tutorial for MySQL MySQL Has Gone Away: An In-Depth Percona XtraDB Cluster Tutorial Enhancing MySQL Security Look at the MySQL Networking Implementation Innodb Architecture and Performance PCI/DSS Compliance with MySQL: 2019 Optimization Tutorial for MySQL 8 Edition How a Modern Database Gets Your Data Fast: MariaDB Query Optimizer Getting Started with Kubernetes and Lessons From Building Automation Percona XtraDB Cluster for a Large Distributed Database My First 90 Days With Vitess Test Like a Boss: Deploy and Test Improving Enterprise HA and Disaster Fortify Your MySQL Data Security in Complex Topologies With a Single Recovery Solutions AWS Using ProxySQL Firewall Command Backing up Wikipedia Databases MySQL Shell : The Best DBA tool? Extend the Shell with the New Extension TUESDAY MySQL InnoDB Cluster: Advanced Configuration & Operation Infrastructure Our journey to better MySQL Securing MySQL with Vault and Table Availability Using Global Transaction InnoDB Management and Scalability Level Encryption IDs, ProxySQL and Consul Improvements in MySQL 8.0 Large Scale Deployment of SSL/TLS For MySQL 8.0 Performance: Scalability & From Scheduled Downtime to Self- MySQL Benchmarks Healing in Less Than a Year MySQL 8.0: The New Replication Vitess: Running Sharded MySQL MySQL 8.0.18: Latest Updates Features on Kubernetes Tracing and Profiling MySQL Amazing Sandboxes with dbdeployer OpenCorporates: Providing Percona XtraDB Cluster 8.0 (PXC-8.0) Transparency for the Public Benefit MySQL on Google Cloud: The Good, The Bad, and The Ugly

SEE WEBSITE FOR LATE CHANGES TECHNOLOGIES

MySQL & MariaDB MongoDB PostgreSQL WEDNESDAY MONDAY MONDAY Optimize and Troubleshoot MySQL Using A Journey with MongoDB HA. From PostgreSQL For Oracle and MySQL PMM 2.0 Standalone to Kubernetes Operator DBAs and For Beginners How to Upgrade Like a Boss to MySQL TUESDAY Introduction to PL/pgSQL Development 8.0? Multi-Document Transactions TUESDAY Managing MySQL at Scale in Facebook Through MongoDB shell Why PostgreSQL is Becoming A How to Instrument Your Code in MongoDB Data Security: Custom Rules Migration Target in Large Enterprises Performance Schema and Views Pg_catalog Unrevealed! That Part The Database is Broken. Now What? New Indexing and Aggregation Pipeline of PostgreSQL You Are Probably Automatic Upgrade and New Error Capabilities in MongoDB 4.2 Underusing Logging in MySQL 8.0 What’s New in Percona Server for Join Heterogeneous Databases Using gdb Basics for MySQL DBAs MongoDB PostgreSQL Foreign Data Wrappers JSON Array Indexes in MySQL WEDNESDAY High Availability and Automatic Failover in PostgreSQL Maintenance for MongoDB Replica Sets Percona Backup for MongoDB What’s New on Sharding in MongoDB 4.2 MongoDB Analysis with Prometheus and Grafana

SEE WEBSITE FOR LATE CHANGES TECHNOLOGIES

PostgreSQL Other Databases, Multiple Other Databases, Multiple WEDNESDAY Databases & Other Databases & Other Parted Ways with Partitioning? It’s Time Technologies Technologies to Reconsider MONDAY WEDNESDAY Percona Distribution for PostgreSQL Open Source Database Performance Databases, Crypto & Decentralization Handling Transaction ID Wraparound in Optimization and Monitoring with PMM ClickHouse Materialized Views: A Secret PostgreSQL Accelerating Application Development Weapon for High Performance Analytics Top 10 Mistakes When Migrating with Amazon Aurora Data Protection and OSS in the Age of From Oracle to PostgreSQL TUESDAY GDPR PostgreSQL Plan at Execution Time: Running PMM in Production at Tessi A Quick Show More Than a Query Language: SQL in the 21st Century Benchmarking Should Never be Optional Tarantool: A NoSQL Database with SQL 10 Common Mistakes (Java)Developers What’s New in PMM 2.0 Make When Writing SQL MyRocks and RocksDB Advanced Protecting your Secrets Features and Performance Strength in Numbers: Slack’s Database ClickHouse for Time-Series Real-Time Architecture Analytics Graph Databases: Introduction, Deep Dive on Amazon Aurora Standardization, Opportunities VictoriaMetrics: Why and How We Built Running ElasticSearch at Large Scale A Fast and Scalable Open Source Time Graph Databases: Introduction, Series Database Standardization, Opportunities ClickHouse Features to Becoming Cloud Native: How Percona Blow Your Mind Brings Databases to Kubernetes Using Operators

SEE WEBSITE FOR LATE CHANGES KEYNOTES

KEYNOTE ADDRESSES TUESDAY, OCTOBER 1

9:00AM Matt Yonkovit, Welcome to Percona Live Europe, 2019 CXO, Percona 9:05AM Peter Zaitsev The State of Open Source Databases CEO, Percona Peter explores the current topics and trends impacting the database ecosystem, as well as the businesses that run open-source databases as part of their infrastructure. Will cloud-native surpass public cloud usage? How Is the changing licensing around open source going to impact you? What keeps most of your peers up at night? What will impact your business in 2020 and beyond?

9:30AM Geir Høydalsvik, The State of the Dolphin Software Where are we with MySQL 8.0 one year after GA? How was the journey and how the Development Community impacted the road map? During this session come to discover our new Director, Oracle release way, allowing us to be even closer to users looking for performance, stability and innovation.

9:55AM John Boero, Unlocking the Cloud Operating Model Channel To unlock the fastest path to the value of the cloud, enterprises must consider how to Solutions industrialize the application delivery process across each layer of the cloud: embracing Engineer, the cloud operating model, and tuning people, processes, and tools to it. The essential Hashicorp implication of the transition to the cloud is the shift from static infrastructure to dynamic infrastructure: from a focus on configuration, and management of a static fleet of IT resources, to a focus on provisioning, securing, connecting, and running dynamic resources on demand. In this session, we’ll discuss what the Cloud Operating Model means for your enterprise. 10:25AM Matt Yonkovit, Announcements and Release to Sessions CXO, Percona

SEE THE WEBSITE FOR ANY LATE CHANGES KEYNOTES

KEYNOTE ADDRESSES WEDNESDAY, OCTOBER 2

4:00PM Matt Yonkovit, Welcome Back to Percona Live Europe, 2019 CXO, Percona 4:05PM Matt Yonkovit, Why Have Efforts to Simplify Led to More Complexity? CXO, Percona How to Survive What has Become an Increasingly Complex and Confusing World. A great deal of time and effort goes into making things easier, from simplification through automation, DevOps processes, and the emergence of new technology. Despite the best efforts of our brightest minds, systems still crash, databases are breached (seemingly daily), and smart people still do stupid things. Despite our best efforts, advancements have led to more complexity, less unity, and more confusing options than ever before. Why did this happen and how can we address it? 4:20PM Craig Kerstiens, Postgres: From Respected by DBAs to Loved by Application Head of product Developers for Citus Cloud, Ten years ago, Postgres was a safe and stable database, although not at the same level of Microsoft user-friendliness or rich feature set as other databases. But, that all started to change as it learned from MySQL and Mongo. We’ll examine some of that journey: how it went from being respected by DBAs to beloved by application developers. We’ll close with a look at what the next five years may hold.

4:45PM Panel Discussion Business Challenges Facing Modern Database Deployments We will engage with business leaders from around the globe and ask them about trends, changes, and issues they face day-to-day. We will ask them for their advice on how to build scalable, secure, and world-class applications that meet user needs.

5:10PM Matt Yonkovit, Wrap Up: See You At Percona Live 2020! CXO, Percona SEE THE WEBSITE FOR ANY LATE CHANGES TUTORIAL DAY MONDAY, SEPTEMBER 30 MONDAY

PostgreSQL For Oracle Following topics are covered in this MySQL 101 Tutorial tutorial. and MySQL DBAs and For 1. Brief history of PostgreSQL and its Parts 1 & 2 Beginners features Tom de Cooman, Avinash Vallarapu, Percona 2. PostgreSQL Architecture for MySQL/ Dim0 Vanoverbeke, Michal Nosek, Oracle DBAs and beginners. This tutorial is focused on teaching Percona 3. Installation of PostgreSQL PostgreSQL to professionals who are This full-day tutorial will be mainly hands- 4. Creating/Start/Stop a PostgreSQL already experienced with MySQL or Oracle on. We’ll show some slides on each of the Cluster and the components of a Data or SQL Server. Relational databases share various topics presented, but the goal is Directory a lot of similar concepts. You may find it to have YOU the student work in a MySQL 5. PostgreSQL parameters/settings you easier when the concepts of PostgreSQL environment as much as possible. need to know are related with Oracle/MySQL. Thus, I 6. PostgreSQL User Management What will we cover? Almost every aspect of would start with PostgreSQL Architecture 7. MVCC in PostgreSQL being a MySQL DBA, on MySQL 5.7: for Oracle/MySQL DBAs. 8. VACUUM and Transaction ID • We will install MySQL and configure the As we discuss from basics, even a Wraparound in PostgreSQL right privileges for your system beginner could adapt the concepts 9. Types of Indexes • Once we have a database running, we easily. It starts with an overview of the 10. EXPLAIN in PostgreSQL will get a better understanding of what our PostgreSQL architecture and a hands- 11. Logical and Physical Backups database is doing and look at the different on lab, helping the audience apply the 12. Types of Replication in PostgreSQL possible diagnostics and monitoring concepts presented, starting with a simple 13. Streaming replication and PITR options in MySQL using PMM and other and quick installation of PostgreSQL and (Point-in-time-recovery) tools. going all the way to setup and maintain a • We will troubleshoot issues standby (replica). • We will discuss and try some of the By the end of this tutorial, attendees different backup options, since not losing will have a good understanding of how any data is an important goal for a MySQL PostgreSQL works and feel confident DBA enough about it. • We will clarify and troubleshoot replication

RETURN TO TUTORIALS DAY TIMETABLE MONDAY

MySQL 8.0 InnoDB Cluster: MariaDB Server 10.4: Attendees will field fast-paced practice sessions on: Easiest Tutorial! The Complete Tutorial • Upgrading their MySQL servers to Frédéric Descamps, Oracle Colin Charles, Consultant MariaDB successfully. During this tutorial we will show how easy Let’s take a simple Oracle-backed • Learning the features that MariaDB it’s now to create, manage and monitor application, and migrate it to MariaDB offers beyond MySQL. a MySQL InnoDB Cluster. HA for MySQL Server 10.4. While we are at it, let’s learn • Understanding the compatibility between has never been so easy. We will migrate about how to use system versioned tables, MariaDB and MySQL, including how an existing environment with minimal Galera 4, the MyRocks storage engine, applications will work and react to it. downtime and show different tuning Flashback to rollback tables/databases to • Have an overview of storage engines, settings related to Group Replication. The an older snapshot, and also use the built-in but focus on what the most popular ones audience will also get familiar with the threadpool. can be used for. MySQL Shell. This and more is possible in 3 hours, as • Understanding MariaDB Server MariaDB Server 10.4 has been out for replication options sometime now (June 2019) and it has many The structure of this tutorial will be new features, some of which MySQL does practical, so please bring a laptop, ensure not have. Feature wise, it is important to you have the ability to run either SSH to know what MariaDB Server 10.4 has (e.g. a remote machine, or the ability to run system tables in the storage engine, “dbdeployer” on . There will be virtual ability to reload SSL certificates without machines provided as an alternative. a restart and more!) and what it lacks compared to MySQL 8.0 (group replication, the X Protocol, etc.). All participants (developer or operations personnel) will leave more knowledgable on how to better manage, observe, and secure their MariaDB Servers. In addition, there will be a focus on High Availability as well as backups (covering Mariabackup) & disaster recovery.

RETURN TO TUTORIALS DAY TIMETABLE MONDAY

Open Source Database Please bring a laptop with an ssh client Percona XtraDB Cluster as you will be connecting to a cloud- Performance Optimization and provisioned instance on which you will Tutorial Monitoring with PMM deploy PMM 2. Tibor Köröcz, Percona Michael Coburn, Vinicius Grippa, Percona XtraDB Cluster (PXC) is the Sveta Smirnova, Carlos Tutté, leading HA solution for MySQL with its Percona write-anywhere, synchronous replication capabilities. This half-day tutorial will This hands-on tutorial covers how to set introduce you to PXC, explain core up monitoring for MySQL, MongoDB, and concepts, discuss advanced techniques, PostgreSQL database servers using the cover common “gotchas”, and utilize Percona Monitoring and Management ProxySQL to create the most highly- (PMM) platform. available MySQL cluster solution currently PMM is an open-source platform for available. managing and monitoring MySQL, MongoDB, and PosgreSQL performance. It provides thorough time-based analysis for database servers to ensure that they work as efficiently as possible. You will learn: • MySQL, MongoDB, and PostgreSQL monitoring best practices • Metrics and time series • Data collection, management and visualization tools • Monitoring deployment • How to use graphs to spot performance issues • Query analytics for MySQL and MongoDB, and now PostgreSQL! • Alerting using Grafana • Trending and capacity planning • How to monitor HA

RETURN TO TUTORIALS DAY TIMETABLE MONDAY

Innodb Architecture and Introduction to PL/pgSQL Performance Optimization Development Tutorial for MySQL 8 Jim Mlodgenski AWS Peter Zaitsev, Percona When developing against a PostgreSQL MySQL 8 is the most commonly used database, many times you will use SQL storage engine for MySQL and Percona from your applications, but sometimes Server. It is the focus for the majority of you need more. You might need to run storage engine development and Percona some of your logic on the database server Server teams. for performance or you may want to just create some server-side functions to In this tutorial, we will look at the MySQL 8 simplify your application development. architecture, including new developments in MySQL 8 as well as Percona Server. We There is an infinite number of reasons will explain how it should be used to get the why you would want to write some code best performance from your application. that runs on the PostgreSQL server and PostgreSQL has a rich set of options when We will provide specific advice on server you want to do that, but most times, you configuration, schema design, application will use PL/pgSQL. The primary objective architecture, and hardware choices. of this tutorial is to introduce you to the This tutorial is updated from its previous PostgreSQL PL/pgSQL programming version, to cover new MySQL 8 and language and to give you the practical Percona Server features in more detail. skills to be a productive developer working with PostgreSQL databases. In the course of this tutorial, you will be introduced to a variety of fundamental PL/ pgSQL functions like: • How to declare and use PL/SQL Variables • How to create functions and procedures • How to write triggers • Handle run time errors And much more...

RETURN TO TUTORIALS DAY TIMETABLE MONDAY

Test Like a Boss: Deploy and Finally, we’ll see how to import existing Getting Started with deployments into a sandbox and how to Test Complex Topologies With clone a sandbox into another. Kubernetes and XtraDB a Single Command Cluster Giuseppe Maxia, dbdeployer Matthew Boehm, Percona Using MySQL databases or its derivatives Have you ever wanted to properly learn nowadays is much more than running a how to run Percona XtraDB Cluster on top main server in production. of Kubernetes? It’s much more than maintaining a master/ Join us in this 3 hour tutorial where we slave or group replication system. cover the quick basics of containers, New versions and new features are and then dive right in to the Percona coming at great speed, and users need Kubernetes Operator for XtraDB Cluster. to try the novelties quickly, possibly in You’ll learn: their in-house hardware. dbdeployer • How to install the operator, create a allows users to deploy simple and complex highly available production-ready XtraDB topologies with a single command, using Cluster. a single host, without interfering with • How to manage backups and recovery. existing servers in the same machine. • How to scale your cluster up and down In this tutorial you will learn how to to meet demand. install a test cluster with asynchronous replication, with group replication (single primary and multi primary), with NDB cluster, or other multi-master topologies such as fan-in and all-masters. We will also see how to replicate from single server to group replication, from group replication to master/slave, or from two NDB clusters. Also we’ll see how to download and deploy in a few commands, without need to use a browser, but just with dbdeployer’s assistance.

RETURN TO TUTORIALS DAY TIMETABLE MONDAY

A Journey with MongoDB Attendees will undertake practice sessions Accelerating Application on: HA From Standalone to • Basic concepts of MongoDB Development with Amazon Kubernetes Operator – A »»Installation, best practices Aurora Hands-On Tutorial configuration, user and roles Kathy Gibbs, John Nguyen, • How the replicaset works Amazon Web Services Adamo Tonete, Vinicius Grippa, »»What is a heartbeat, how does an Percona election works, hidden instances, read In this a hands-on tutorial, you’ll learn MongoDB® is the most popular NoSQL preference how to leverage the unique features of database in the world and the 5th in • MongoDB on Kubernetes (how to deploy, Amazon Aurora to build faster, more general. Thanks to its flexibility storing self-healing, auto-escaling) scalable database applications optimized data, horizontal scaling and high »»Create a replicaset with Percona for the cloud. We discuss architectural availability features MongoDB has its Kubernetes Operator for Percona Server best practices and features designed presence in the hall of fame. for MongoDB in 3 minutes with no manual to help you develop applications faster and reach the widest possible audience, We will discuss how data is stored on intervention to setup the replication including Aurora Serverless, read replica MongoDB and the internal aspects of the auto scaling, cross-region replicas, high availability and how to scale it. Finally, backtracking, fast database cloning, and one look closer to the future and how to Performance Insights. You’ll understand deploy MongoDB on Kubernetes. how to best take advantage of the Aurora It is expected that all participants platform’s capabilities to effectively (developer or operations personnel) accelerate application development. will leave more knowledgeable on how Please note: this is a half day tutorial that MongoDB works, main use cases and how will be repeated in morning and afternoon to setup a standalone, replicaset and how sessions. to use Percona MongoDB Operator for Kubernetes.

RETURN TO TUTORIALS DAY TIMETABLE CONFERENCE DAY ONE TUESDAY, OCTOBER 1 TUESDAY

Our Journey to Better MySQL More Than a Query Language: MySQL 8.0 Performance: Availability Using Global SQL in the 21st Century Scalability & Benchmarks Transaction IDs, ProxySQL Markus Winand, winand.at Dimitri Kravtchuk, Oracle and Consul Did you know the purely relational dogma This session will be interesting to everyone Stephane Combaudon, Yuriy of SQL was already abandoned in 1999? looking for the latest news about MySQL 8.0 Performance: Olshanetsky, Rakuten Intelligence The last SQL standard that was limited to the relational idea was SQL-92. From Since MySQL 8.0 we moved to “continuous Data is core to our business at Rakuten SQL:1999 onwards, the SQL language was release” model so with every update Intelligence. And while databases not being extended with non-relational operations many new improvements are delivered available can hurt us badly, our system and non-relational data structures. As but how does it also improve MySQL 8.0 was not initially designed with MySQL high much as this move was discussed at Performance ? ;-) availability in mind. Each MySQL master that time, it took decades until database The latest benchmark results obtained switch involved stopping and restarting vendors caught up with this idiomatic with MySQL 8.0 will be in center of the applications, which at some point was no change. Many SQL users haven’t heard of talk because every benchmark workload longer a valid option. it until today. for MySQL is a headache and “problem to No existing HA solution worked well for This talk provides the big picture on resolve” and each resolved problem is a us, therefore we focused on implementing the evolution of the SQL standard and potential gain in your production ! the glue between existing bricks, the introduces some selected modern SQL Many important internal design changes main ones being: ProxySQL (connection features by example. You will see that SQL are coming with MySQL 8.0 management between the applications has changed as much as our requirements • how to bring them all in action most and MySQL), Consul (service discovery) have changed over the past decades. efficiently ? and MySQL GTIDs (easy tracking of • what kind of trade-offs to expect, what is transactions across servers). already good, and what is not yet” ? We will share our failures and how we • how well MySQL 8.0 is able to use incrementally improved our system to the latest new HW and how can it run reach our availability goals. efficiently on small boxes ? • could you really speed-up your IO by deploying your data on the latest flash storage ? These and many other questions are answered during this talk + proven by benchmark results... and as usual, some surprises to expect ;-)) RETURN TO TUESDAY’S TIMETABLE TUESDAY

Vitess: Running Sharded Amazing Sandboxes with Multi-Document Transactions MySQL dbdeployer Through MongoDB Shell on Kubernetes Giuseppe Maxia, dbdeployer Aayushi Mangal, Percona Sugu Sougoumarane, PlanetScale You probably know that you can deploy a Transactions are a well known and widely Vitess has continued to evolve into a MySQL database or a group of databases used feature of RDBMS systems, but massively scalable sharded solution for in replication in your laptop using MongoDB is a document-oriented non- the cloud. It’s is now used for storing core dbdeployer. RDBMS database. Every single document business data for companies like Slack, But unless you have been paying close is atomic in MongoDB,. It is widely known Square, JD.com, and many others. attention, there are a few brilliancies for its simplicity of importing data, and the sharding of data across many servers. This session will cover the high-level that may have escaped you. This session This talk includes transactions in detail features of Vitess with a focus on what will show the basic usage of dbdeployer, and internals of transactions oplogs. makes it cloud-native. before delving into the advanced operations, such as: importing running Key points covered: We’ll conclude with a demo of the powerful databases into a sandbox, replicating materialized views feature that most • Transactions on-the-fly between single sandboxes, or sharded systems have yet to solve. • Prerequisites between groups, or NDB clusters. You • writeConflicts can also clone sandboxes with a single • Considerations and best practices command (sometimes with just half a • Hands-on through MongoDB shell command) and run commands in several servers at once. Want to run a query on all slaves? You will see that! How about on all masters? You’ve got it! Want to change a sandbox server configuration temporarily or permanently? You will see that also! Do you want to make your own commands with fine tuning options depending on database flavor and version? dbdeployer sandboxes are built for this purpose: you just need to see how to use the force that has been waiting for you.

RETURN TO TUESDAY’S TIMETABLE TUESDAY

Tarantool: Percona XtraDB Cluster 8.0 What’s New in PMM 2.0 A NoSQL Database with SQL (PXC 8.0) Michael Coburn, Rachel Pescador, Pavel Lapaev, Tarantool Krunal Bauskar, Percona Percona We will present a short history and Percona XtraDB Cluster 8.0 (PXC-8.0) is PMM 2 represents a significant advance overview of Tarantool, an open source the latest addition to PXC family. in terms of monitoring for Open Source in- DBMS. We will explain why it is Starting with MySQL/PS-8.0, upstream Databases. Come to this session in order cool to have a DBMS in the same address have made a lot of significant changes to learn about the following: space as your application server, why including atomic DDL, replication channel, • Query Analytics improvements Tarantool is in fact single threaded and locking algorithm changes, etc.... • Architectural changes other main features. Then, we’ll explain, During this session, we will explore: • API and GUI configuration why a no-SQL DBMS supports SQL. • how these changes affect PXC-8.0 • what has changed in PXC-8.0 • new/deprecated features • important bugs and more.

If you are already a PXC user or planning to consider it, attend this session to find out more about PXC-8.0.

RETURN TO TUESDAY’S TIMETABLE TUESDAY

MySQL on Google Cloud: MongoDB Data Security: Why PostgreSQL is Becoming The Good, The Bad, and The Custom Rules and Views a Migration Target in Large Ugly Adamo Tonete, Percona Enterprises Jeremy Cole, Shopify, Inc. In this presentation, we will discuss how Jobin Augustine, Percona Running MySQL in the cloud isn’t magic - to create custom rules when the default 1. Integration with Other RDBMS systems, it can be brilliant, but it can also be a real rules are not enough for the application. Oracle, MSSQL, and MySQL challenge. We’ll learn about some of the Have you needed to give a more permissive 2. Integration with Hadoop and NoSQL big wins that can be had in the cloud (such rule to a user just because of this user databases as elasticity and easy provisioning). But, wanted to run a specific command? 3. Compatibility layer: Example of OraFCE we’ll talk about the dark underbelly as well Also, we will discuss how to use view for 4. Migration tools: Example of Ora2Pg - and face some of the challenges to run hiding fields from users when we don’t 5. How PostgreSQL is suitable for Agile a realistic MySQL installation in the cloud want them to read all the collection. If you development and Knowledge Portability. (errrr... performance?). have concerns about security come to this 6. PostgreSQL derived projects and talk. freedom of choice.

RETURN TO TUESDAY’S TIMETABLE TUESDAY

How to Protect PII Data MyRocks and RocksDB New Indexing and Aggregation in MySQL While Allowing Advanced Features and Pipeline Capabilities in Customers to Access the Performance MongoDB 4.2 Database Yoshinori Matsunobu, Facebook Antonios Giannopoulos, Alexander Rubin, VirtualHealth We have been running MyRocks in ObjectRocket by Rackspace Protecting PII data is one of the biggest Facebook’s largest database (UDB) and MongoDB 4.2 comes GA soon delivering challenges of security minded DBA. But backends for a while. some amazing new features on multiple when you add a requirement to allow We have made significant improvements areas. In this talk, we will focus on the new customers to have a direct access to a in both MyRocks and RocksDB, especially capabilities of the aggregation framework. database replica it will make a life of a DBA from performance, efficiency, reliability We are going to cover the new operators nightmare. At VirtualHealth we deals with and operations point of view. and expressions. At the same time, we will medical data. In this talk I will share an In this session, the speaker will go explore how updates commands can now experience of how to implement customer through several important features and use the aggregation framework operators. authentication, authorization and data performance improvements. We are also going to present aggregation isolation in MySQL. I will cover such topics framework improvements focusing on the as: on-demand materialized views. Finally, we • Authentication of external users: how to are going to explore the wildcard indexes create a secure connection to MySQL and introduced in MongoDB 4.2 and how they manage users (LDAP, external auth, etc) change the way we design documents and • Authorization and data isolation build queries/aggregations. We will also • Auditing the logins and login attempts make a reference to the new index build • Encryption at rest, field level encryption system. and key rotation • Data masking and data de-identification (for dev systems)

RETURN TO TUESDAY’S TIMETABLE TUESDAY

Comparison of Kubernetes Pg_catalog Unrevealed! Enhancing MySQL Security Operators for MySQL That Part of PostgreSQL You Vinicius Grippa, Percona Mykola Marzhan, Percona Are Probably Underusing Security is always a challenge when it Running databases in Kubernetes attracts Boriss Mejias, 2ndQuadrant comes to data protection but regulations a lot of attention today. Orchestration like GDPR and HIPAA brings a new layer PostgreSQL users want to know how on top of it with severe fines for breaches of MySQL on Kubernetes is no way a the system is performing. This is valid straightforward process. There are and rules more restrictive to access for DBAs, system administrators, and and manipulate personal data. Join us in several good MySQL based solutions in developers. They want to know if the the open-source world, made by Oracle, this presentation to check security best system needs some improvements, but practices, traditional and new features Presslabs, and Percona. Having a common also to generate the occasional report base, they differ in self-healing capabilities, available for MySQL 8. requested by managers. Standard • On this talk, DBAs and sysadmins will multi-master and backup/restore support, monitoring tools will monitor the CPU, RAM etc. So let’s make a fair comparison to walkthrough since the beginning in the and I/O consumption, but they won’t be application where data is born until the figure out the pros and cons of their able to tell you if indexes are being used, current state. end it is archived on the backup. The topics if tables are bloated, what is the lag of are presented below: replication, what are the status of sessions • SO/Cloud security connected, and many other interesting and • SSL useful things from PostgreSQL. • Password Management This talk is intended to review the • TDE (Transparent Data Encryption) possibilities offered by the PostgreSQL • Audit Plugin catalog. Plenty of information is available • MySQL 8 features there but it is unknown to many users, • New caching_sha2_password and therefore unused. We will see how • Roles to exploit the catalog, how to send the • Password Management information to other monitoring tools, • FIPS mode and how the tables in pg_catalog are • Using Xtrabackup to encrypt the fundamentally interconnected to other backups topics such as performance, replication, We will share our experience of working MVCC, security, the universe and with 2000 support customers and help the everything. audience to become familiar with all the security concepts and methods, and give you the necessary knowledge to apply to your environment. RETURN TO TUESDAY’S TIMETABLE TUESDAY

PCI/DSS Compliance with What’s New in Percona Server MySQL: 2019 Edition for MongoDB Carlos Tutte, Percona Akira Kugonane, Adamo Tonete, In the past decade there was a boom in Percona online transactions, and because of this News on the free-and-open enterprise criminals moved online to exploit billions of features in PSMDB (v4.0+) and the recent insecure consumer information records. v4.2 release. Payment Card Industry Data Security Data-at-rest Encryption requires key Standard (PCI/DSS) dictates norms about management (no point storing the storing, processing and transmitting keys with the data!). We’ll demonstrate cardholder data and originated in an effort how v4.0.10-5 of Percona’s fork uses to reduce and prevent credit card fraud. integration with Hashicorp Vault to provide Discover which tools/plugins/settings you this. need to use to comply with PCI/DSS when Percona Server for MongoDB 4.2.0 has using MySQL, including: been released merging all of MongoDB’s • Protect cardholder data from upstream community edition code. unauthorised use The headline features of distributed • Enforce strong controls around transactions and update-capable privileged users and data access aggregation pipelines are in (see other • Implement centralized, automated role- talks!). In this presentation we’ll share based access control, authorization, and some other important changes that all authentication MongoDB DBAs should be aware of. • Provide system and database auditing, • A new throttle mechanism that works on and database activity monitoring primary nodes; • Encryption in MySQL: TDR, LUKS, AES, • A merging of foreground and DES, asymmetric key background index build methods Attendees will gain a solid grasp of the • The eight-year wait to be able to list all possibilities MySQL offers to someone client ops (active or idle) is finally over. looking to become PCI/DSS certified, as • Wildcard Indexes well as the limitations it comes with. Don’t miss this talk!

RETURN TO TUESDAY’S TIMETABLE TUESDAY

Lessons From Building In this talk, we will outline the natural ClickHouse for Time-Series evolution of automation software. We Automation for a Large believe, at scale automation software, has Real-time Analytics Distributed Database the same sets of rigorous requirements as Alexander Zaitsev, Altinity Ameet Kotian, Leigh Johnson, any other production software but there ClickHouse is a very fast general purpose Slack Inc are easily identifiable patterns common open source DBMS. Starting from web to automation software - things like - a analytics use case several years ago, it Automation is a critical investment when scheduler/state-machine to execute a quickly develops to other areas, including scaling infrastructure operations. If you series of steps, resiliency features such time series applications where ClickHouse delay automating, your company’s growth as timeout and retries, method of failure manages to outperform dedicated time will end up supported by a mountain of handling and escalation, audit features series solutions. In this talk I will explain human toil. such as logging and actions history. Finally different approaches of using ClickHouse Everyday operations like provisioning, some sort of interface clients like web-ui for time series, highlight ClickHouse host replacements start of as steps in or command line. features that help to design an efficient a runbook. Later, you might code-ify Once such an automation framework schema, present some benchmarks, these processes with scripts. At some is available, we can build self-healing and walk through real life optimisation point, there is enough complexity and infrastructure on top it by feeding it techniques. requirements for your automation that reliable and robust failure signals. it evolves into a piece of full-fledged production software. The automation of stateful services, like databases, in particular comes with a special set of problems, including: • Risk of data loss • Strict SLOs for availability • Longer provisioning times due to data restore process and migrations The automation needs additional level of safety built-in to minimize data loss risk. This is a common problem without a common, universal solution. We needed to figure out an answer to this problem while scaling out a stateful MySQL database fleet to thousands of hosts.

RETURN TO TUESDAY’S TIMETABLE TUESDAY

Improving Enterprise HA and We will discuss the right way to analyze Join Heterogeneous the platform, and how to finally setup a Disaster Recovery Solutions MySQL/Maria/Percona cluster. Databases Using PostgreSQL Marco Tusa, Percona The presentation will also answer the Foreign Data Wrappers Who nowadays does not know of HA question on how setup multiple DR sites. Ibrar Ahmed, Percona solutions based on MySQL/Maria/ PostgreSQL provides a way to Percona cluster (Galera)? communicate with external data sources. Even so, the rollout of this very powerful This could be another PostgreSQL solution is often not correctly designed, instance or any other database. The other especially when trying to solve not database might be a relational database only local HA, but also to meet Disaster such as Clickhouse, MySQL or Oracle; or Recovery scenarios. any NoSQL database such as MongoDB During the many experiences since or Hadoop. To achieve this, PostgreSQL my first MySQL+Galera installation, I implements ISO Standard call SQL-MED in have faced many misconceptions when the form of Foreign Data Wrappers (FDW). customers have needed to design solid/ This presentation will explain in detail how trustable/efficient DR solutions. PostgreSQL FDWs work. It will include a Foremost is the lack of consideration detailed explanation of simple features and of what can be really achieved without will introduce more advanced features breaking the laws of physics. that were added in recent versions of This presentation wants to clarify once for PostgreSQL. Examples of these would be all WHY some of the current misconception to show how aggregate pushdown and join are ...just that, “misconceptions”. pushdown work in PostgreSQL. At the same time, I will illustrate this talk The talk will include working examples using an architecture solution, that we of these advanced features and have identified as the real answer for DR demonstrating their use with different scenarios and that we have successfully databases. These examples show how data implemented many times. from different database flavors can be I will start from customer requirements, used by PostgreSQL, including those from and show how we have guided our heterogeneous relational databases, and customers through the avaialable options showing NoSQL joins. to finally identify the correct solution.

RETURN TO TUESDAY’S TIMETABLE TUESDAY

Backing up Wikipedia MySQL InnoDB Cluster: InnoDB Management and Databases Advanced Configuration & Scalability Improvements in Jaime Crespo, Manuel Arostegui, Operation MySQL 8.00 Wikipedia Frédéric Descamps, Pedro Gomes, Sunny Bains, Oracle Imagine a world where all Wikipedia Oracle At the beginning there was one general articles disappear due to a human error or MySQL InnoDB Cluster is a very easy HA tablespace... Then we added a tablespace software bug. Sounds unreal? According solution to deploy. per table... What have we designed and to some estimations, it would take an However it’s also a very customizable delivered in MySQL 8.0 to manage even excess of hundreds of million person- solution able to respond to most better your tablespaces? hours to be written again. To prevent that needs. During this session I will give an Come to this session to learn everything scenario from ever happening, our SRE overview of settings that you may tune you need about InnoDB tablespaces. team at Wikimedia recently refactored the like those related to quorum lost, level of relational database recovery system. consistency, but also some you may not In this session, we will discuss how we know like how to change recovery system, backup 550TB of MariaDB data without effect of increasing the event horizon. impacting the 15 billion page views We will also discus about maintenance per month we get. We will cover what operations like how to stream large were our initial plans to replace the transactions, how to deal with DDL in old infrastructure, how we achieved multi-primary environments... recovering 2TB databases in less than 30 minutes while maintaining per-table granularity, as well as the different types of backups we implemented. Lastly, we will talk about lessons learned, what went well, how our original plans changed and future work.

RETURN TO TUESDAY’S TIMETABLE TUESDAY

Deep Dive on Amazon Aurora High Availability and Automatic 1. How the implementation of HA differs Yoav Eilat, Amazon Web Services for each type of replication available in Failover in PostgreSQL PostgreSQL Amazon Aurora is a MySQL- and Avinash Vallarapu, Percona 2. How to combine HAproxy with etcd PostgreSQL-compatible relational PostgreSQL is an advanced open source plus a detailed explanation of the RAFT database designed from the ground up database that is completely community Algorithm for HA using Patroni. to benefit from the cloud. Aurora offers driven. Continuous development and 3. What is stolon, pgpool-II and repmgr and a novel architecture that moves the performance improvements while making how can they be helpful. monolithic database stack to a service- it feature rich gave PostgreSQL a great 4. Some other tools admins may wish to oriented architecture, pushing the lowest market penetration. When a database explore. layers of the database into a distributed, server is deployed in production, it 5. What are the tools and solutions that multi-tenant, log-structured storage is important to achieve several 9’s of help you achieve automatic failover in AWS service. We’ll cover key Aurora capabilities availability. It starts with a replication and other cloud-based environments. such as low latency replicas, instant crash setup that is stable and some tool that 6. How to avoid huge data loss during recovery, in-place rewind, copy-on-write could manage a failover transparent to failovers. cloning, zero-downtime patching, multi- application. This talk starts with a list of master, and serverless, and look under the replication types in PostgreSQL followed hood to explain how they work. by a detailed understanding of some of the widely used open source tools for high availability and automatic failover in PostgreSQL. The following topics will be discussed during this talk:

RETURN TO TUESDAY’S TIMETABLE TUESDAY

VictoriaMetrics: Why and How From Scheduled Downtime to MySQL 8.0.18: Latest Updates We Built Fast and Scalable Self-Healing in Less Than a Norvald Ryeng, Oracle Open Source Time Series Year MySQL 8.0.18 is almost here! In this talk Károly Nagy, Salesforce we’ll focus on two of the new Database features: hash joins and EXPLAIN Roman Khavronenko, Aliaksandr Infrastructure automation is not easy, ANALYZE. This talk will cover how especially for stateful services like MySQL Valialkin, VictoriaMetrics these features work, both on the surface (or any other database for that matter). and on the inside, and how you How did we get the idea of creating new It goes way beyond the capabilities of time series database? Ansible, Chef, SaltStack or other similar can use them to improve your queries and The talk describes typical evolution of tools. In this session I’ going to show you make them go faster. application monitoring during the last few how we went from fully manual operations Both features are the result of major years: it starts with zero monitoring, then to a self-healing system in less than a year refactoring of how the MySQL logging is added, then the observability at Salesforce. Having done this at several executor works. In addition to explaining is improved with app-specific metrics, companies already I’ve seen the common and demonstrating the features altering and graphing. The evolution meets mistakes that can break your system themselves, we’ll look at how the various stumbling blocks - scalability, and make your well intended scheduler/ investment in a new iterator based performance, storage space usage. They scripts/orchestrator a ticking bomb. I will executor prepares MySQL for a future may result in creating new TSDB such as share how to avoid these problems and with faster queries, greater plan VictoriaMetrics. build a robust and scalable automation flexibility and even more SQL features. framework that’s been battle tested at companies such as Booking.com and Dropbox. We will cover: • Tool comparison • Centralised vs decentralised system • Concurrency handling • Best practices and anti-patterns

RETURN TO TUESDAY’S TIMETABLE TUESDAY

Tracing and Profiling MySQL ClickHouse Features to Blow OpenCorporates: Providing Valerii Kravchuk, MariaDB Your Mind Transparency for the Public Corporation Aleksei Milovidov, Yandex Benefit While troubleshooting MySQL performance I will present a subset of the most notable Ben Tullis, OpenCorporates problems it is important to find out where ClickHouse features over the last half of OpenCorporates is one of the largest open CPU time is spent inside mysqld process. year: The process of investigation should have databases made available for the public • data skipping indices, including full text benefit in the world. It hosts structured as small influence as possible on the indices (with performance evaluation and server we try to troubleshoot. data on over 100 million companies insights on implementation) and is used regularly by journalists, Performance_schema introduced in MySQL • custom compression codecs for time investigators, governments, even banks. 5.5 is supposed to provide detailed enough series data; instrumentation for most cases. But it MySQL continues to be a crucial part • HDFS and Parquet integration; of our technology stack and the size of comes with a cost, require careful sizing of • fuzzy string search (it is really fast performance counters, and the process of the primary database has increased by fuzzy string search); multiple substring around 800% in the past four years. instrumenting the code is not yet complete matching; This talk describes our journey and the even for MySQL itself, to say nothing about • sampling profiler on the query level; challenges, pitfalls and crucial business 3rd party storage engines, plugins and • z-curve indexing; decisions along the way which led us to libraries like Galera. • table and columns TTL; This is when profilers come handy. Poor our current implementation using Percona Man’s Profiler (PMP, aka pt-pmp) and perf XtraDB Cluster and ProxySQL on top of can be easily used while studying MySQL ZFS. performance problems. Basic usage steps are presented and several typical use cases are discussed. Tracing system calls (that is needed for MySQL troubleshooting sometimes) is usually performed with strace. But recent Linux kernels with eBPF support allows to trace these and more with less performance impact and collect more details. Brief introduction to bpftrace is presented.

RETURN TO TUESDAY’S TIMETABLE TUESDAY

RETURN TO TUESDAY’S TIMETABLE CONFERENCE DAY TWO WEDNESDAY, OCTOBER 2 WEDNESDAY

Parted Ways with Partitioning? The way we will answer these questions Percona Server for MySQL: is by using some of the most useful It’s Time to Reconsider... performance measurement tools available What It Is and How It Is Done Feike Steenbergen, Timescale in PostgreSQL: EXPLAIN and pg_stat_ Satya Bodapati, Yura Sorokin, How PostgreSQL Revived statements. Percona Partitioning to Improve Database We’ll use a publicly available data set This talk will explain what exactly Percona and see how both the performance and Server is and how it is similar to / different Performance the query plans have changed in the past from Oracle MySQL Server. It will also give Partitioning has improved drastically in releases of PostgreSQL. you a brief description of some unique PostgreSQL in the last 3 releases. features it already has and may have in the Whereas a few years ago partitioning near future. would only be considered to solve specific In addition, it will reveal a few secrets pain points, it is now in such a state that about what Percona Dev Team does on you can use the power of partitioning in regular basis to keep up-to-date with almost any database design to improve Oracle MySQL, to constantly improve code performance, to aid in data retention quality and to fight with bugs. policies, and to have fewer operational headaches. In this talk, we’ll briefly go through the history of partitioning, and then we will take a deep dive into what this means for your next schema (re)design. More specifically, we will discuss: • Exploring design patterns • Validating your database and schema designs • What to look for in your explain plans • How to ensure that partitioning pruning is done at the correct time

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Billion Goods in Few MySQL Has Gone Away: An How a Modern Database Gets Categories: How Histograms In-Depth Look at the MySQL Your Data Fast: MariaDB Save a Life? Networking Implementation Query Optimizer Sveta Smirnova, Percona Iwo Panowicz, Percona Vicențiu Ciorbaru, MariaDB We store data with an intention to use We’ll take a look at the implementation of Foundation it: search, retrieve, group, sort... To do the network layer in MySQL 8. A modern database query optimizer needs it effectively the MySQL Optimizer uses • What is the purpose of Virtual I/O to be able to handle a lot of different index statistics when compiles the query Library? workloads, especially when it comes to execution plan. This approach works • How do packets move from a socket to general-purpose databases. Most of the excellently unless your data distribution is Storage Engines? time, it does a good job; however, certain not even. • How network related algorithms work in patterns are hard to predict. With so many Last year I worked on several tickets MySQL? moving parts, it’s hard for the average DBA where data follow the same pattern: • poll(2), select(2), ppoll(2), epoll(2) - what or database developer to come up with millions of popular products fit into a you can find in MySQL source code? a good explanation for why the optimizer couple of categories and rest used the • What’s the real role of MySQL network chooses certain query plans. rest. We had a hard time to find a solution variables (like net_read_timeout, net_ Vicențiu Ciorbaru dives deep into the for retrieving goods fast. We offered write_timeout, back_log etc.)? Why usually optimizer, covering strategies for dynamic workarounds for version 5.7. However are defaults the best choice? query rewriting, including derived-table new MariaDB and MySQL 8.0 feature: • How not closing connections by merging and condition push down and histograms, - would work better, cleaner applications affects MySQL? moving HAVING clauses into WHERE; and faster. The idea of the talk was born. • When to tune MySQL network layer and temporary table reuse; join optimizations Of course, histograms are not a panacea how it affects MySQL? via IN/EXISTS subqueries; condition and do not help in all situations. • MySQL has gone away from a low-level selectivity estimations via pre-collected I will discuss: perspective. histograms; and improved analytics via • SSL/TLS, is it slower? And why? • how index statistics physically stored by Window functions and CTEs. Armed with • When to tune network the storage engine this extra information, you’ll be able to stack? • which data exchanged with the Optimizer better tune problematic queries and better • why it is not enough to make correct model a database schema, especially for index choice large-scale datasets. • when histograms can help and when The talk will focus on MariaDB’s optimizer they cannot point of view; however, he examines other • differences between MySQL and database systems where it’s relevant. MariaDB histograms RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Maintenance for MongoDB Percona Backup for MongoDB Databases, Crypto & Replica Sets Akira Kurogane, Percona Decentralization Igor Donchovski, Pythian Percona’s open-source solution saves Caleb James DeLisle, CJDNS Redundancy and high availability are the consistent sharded cluster backups to S3 The first major revolution in computing basis for all production deployments. (or S3-compatible) object stores or locally- was lead by the database, and from the With MongoDB high availability is mounted filesystem. It also provides a mainframe era to the rise of web2.0, the achieved with replica sets which provides single-command action to restores those venerable database has been at the heart automatic failover in case the Primary in-place. It is compatible with MongoDB® of every information system. However, goes down. In this session we will review Community edition as well as Percona new technologies such as federated social multiple maintenance scenarios that Server for MongoDB, networking and blockchain threaten will include the proper steps for keeping Please join to hear about use cases and to sideline the database in favor of the the high availability while we perform the simple topology used for PBM agents. protocol, even calling into question it’s the maintenance steps without causing future relevance. downtime. I will argue that in the era of This session will cover Database upgrades, decentralization, the database is more OS server patching, Hardware upgrades, relevant than ever, and that the next Network maintenance and more. revolution in computing will be the • How MongoDB HA works general purpose decentralized database, • Replica sets components/deployment where all manor of decentralized apps typologies from federated social networking to • Database upgrades cryptocurrency can coexist as schemas in • System patching/upgrade a DBMS. • Network maintenance • Add/Remove members to the replica set • Reconfiguring replica set members • Building indexes • Backups and restores

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

ClickHouse Materialized Fortify Your MySQL Data We will discuss: • Access using not standard port Views: A Secret Weapon for Security in AWS Using • Implement selective query access High Performance Analytics ProxySQL Firewall • Define accessibility by location/ip/id Robert Hodges, Altinity Marco Tusa, Percona • Reduce to minimum cost of filtering • Automate the query discovery Materialized views are a powerful addition In Percona we care about you, and we are to data warehouses that automatically constantly looking for the right solution transform data into new forms and who can answer your needs. improve query performance by orders of Securing our data is a complex topic. We magnitude. ClickHouse has a particularly can build a very strong protection around good implementation. This talk introduces our data, but nothing will prevent the ClickHouse materialized views using one WHO could potentially access it to standard examples that should be in the compromise the data integrity or to expose toolkit of any data warehouse designer. We it. start with simple aggregation and extend This because we either under estimate to time series use cases like the ‘last point the control we can or should impose, or problem,’ reclustering data for better because we think to do not have the tools performance, and even pulling data from to perform such control. Kafka queues. Through these concrete Nowadays to be able to control and examples you’ll understand the benefits of manage what can access our data is a materialized views as well as the plumbing must, while how to do with standard tools that makes them work in large, distributed it is a nightmare. data warehouses. The presentation will guide you in a journey, there you will discover how implementing a quite robust protection, more than what you thought was possible. Even more, it is possible and your performances will even improve. Cool right?

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Percona Distribution for MySQL Shell: The Best DBA Data Protection and OSS in the PostgreSQL Tool? Extend the Shell With the Age of GDPR Ibrar Ahmed, Percona New Extension Infrastructure Cristina DeLisle, DPO XWiki SAS Percona has announced the release of Frédéric Descamps, Oracle Privacy is becoming central in shaping Percona Distribution for PostgreSQL. MySQL Shell 8.0.17 has been released the future of tech and the data protection Alongside the PostgreSQL distribution, our with a new extension framework. Join legislation has contributed significantly package incorporates other tools that we this session to discover how to write your to making this happen. Privacy by default feel native PostgreSQL lacks: pgBackrest, extension. I will also give you an overview and design are core principles of how a better alternative to PostgreSQL pg_ of several useful extensions, how to find software should be envisioned. The GDPR basebackup, and Patroni to support an HA them and how to install them. that came into the spotlight last year has solution. a strong case to become a standard even Finally I will show you how to create your outside European borders, influencing the Percona provides the rpm and own. packages for PostgreSQL and these tools. way we protect personal data. However I will discuss how to enable a repository its impact might be, its implementation to use the Percona distribution, and how is still in its infancy. OSS has found itself some of the extensions work. facing the situation and one aspect which Currently, we are at the beta stage, stay is particularly interesting on the tech tuned for the GA! side is how to incorporate the principles of privacy by default and design into the software that we build. This talk is an overview of how the GDPR has impacted OSS, how to apply the model of data controller and data processor on the contexts of various communities and infrastructure providers. Also, it will tackle aspects regarding what do we mean by privacy by default and by design, how could we envision them applied in our OSS. It will bring examples from which we might find something interesting to learn from, regardless if we are looking at them as mistakes, best practices or just ways of doing things.

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

What’s New On Sharding in My First 90 Days with Vitess TO BE CONFIRMED MongoDB 4.2 Morgan Tocker, PlanetScale Antonios Gianopoulos, I have worked with MySQL for several ObjectRocket by Rackspace years, but my adventures with Vitess have only just started. With this experience MongoDB 4.2 comes GA soon delivering still fresh in my mind, I want to share some amazing new features on multiple with other MySQL users looking to walk areas. In this talk, we will focus on changes down the same path what I found most related to sharded clusters. We are going interesting to discover :-) to cover distributed transactions & mutable This talk will start with the basic shard keys providing examples that will architecture of Vitess. From there, I will reveal the internals of those new features. go over some of the basic concepts and We will provide best practices around the terminology that I found useful to learn new sharding features and we will cover such as Topos, Keyspaces, VSchemas, and other minor changes related to it. Tablets. Finally, we will talk about MySQL compatibility. This includes both what schemas and SQL you can expect to be supported.

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Running PMM in Production at Handling Transaction ID This talks covers the following topics: • Transaction IDs in PostgreSQL Tessi Wraparound in PostgreSQL • How does a wraparound happen? Valentin Traën, Jonathan Gourdan, Avinash Vallarapu, Percona • How is it avoided? Tessi & A transaction ID in PostgreSQL is a 32 bit • What happens before the age of 10 Michael Coburn, Percona unsigned integer. 2.1 billion transactions million and 1 million of 2.1 billion? • Should template databases be Have you struggled with problem in the past are visible and 2.1 billion in vacuumed? identification using MySQL? the future are not visible. What it means is, a tuple inserted by a transaction ID : • How to monitor and how to tune? The sooner problems are identified means n is not visible to a transaction id : (2.1 • How to repair damage? the faster the environment is stabilized. billion + n + 1). This is called a transaction Come listen to how Tessi Document ID wraparound. However, there is a Services uses PMM + MySQL InnoDB mechanism in PostgreSQL that avoids Cluster to visualize cluster performance this from happening. But, there can be and to identify & analyze slow queries. a situation when it could go so near to We’ll also show some of the dashboard and that age where downtime is needed to custom scripting improvements made to perform a maintenance task. In this talk, extend PMM. the audience would get introduced to the internals of transaction ID’s with a detailed understanding of transaction ID wraparound. And then understand how this can be avoided through parameter tuning and routine monitoring and maintenance. And finally move to the procedure involved in repairing the damage when a database is no more available for writes.

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Securing MySQL with Vault Benchmarking Should Never Finally you can also see the process of optimizing your Mean Time Till Repair as a and Table Level Encryption Be Optional benchmark exercise, which can be part of Kevin Markwardt, Pythian Art van Scheppingen, Messagebird your disaster recovery plan. With all of the leaks of personal Benchmarking is evaluating your database information it is very important to system or environment by comparing it make sure that your data is secure and with a set baseline. In this session you will protected. In this talk I will go into and learn why benchmarking any (database) demonstrate how to secure your MySQL system is not optional and should never user accounts with Hashicorp Vault, to be skipped. Benchmarking takes time do MySQL user account rotation and and often, as projects are often rushed expiration. You can then integrate it with into production, skipped due to time LDAP. There are a lot of users that do not constraints and this means you will be use command line interface that may want flying blind as you don’t know what your to access these environment. This can all limits are. Also a common misconception be done with Vaults built in GUI interface. is that benchmarking is only to be used for Finally I will demonstrate how to encrypt finding the optimal performance of your your tables and store the key in Vault by database, while it can give you a whole base encoding your key and placing it into additional array of insights on your entire a KV store, or by just using the Percona environment: resource planning, predict keyring_vault plugin. behaviour and even give insight in the Mean Time Till Repair (MTTR). Leveraging benchmarks for resource planning will give you insight into how much capacity you need or when you are expected to scale or upgrade. Benchmarks can also be utilized to put an application or architecture under a lot of stress. This can predict how the application or entire architecture will behave when it’s resources near exhaustion and this can help you to identify issues early on when this would otherwise happen when it hits production.

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

10 Common Mistakes (Java) Top 10 Mistakes When Large Scale Deployment of Developers Make when Migrating From Oracle to SSL/TLS For MySQL Writing SQL PostgreSQL Daniël van Eeden Charly Batista, Percona Jim Mlodgenski, Amazon Web Deploying SSL/TLS with MySQL at Booking. It’s easy for Java developers (and users Services com on thousands of servers is not without issues. of other OO languages) to mix object- As more and more people are moving In this session I’ll tell you what steps we oriented thinking and imperative thinking. to PostgreSQL from Oracle, a pattern of took, what problems we hit, and how we But when it comes to writing SQL the mistakes is emerging. They can be caused improved various parts of the MySQL nightmare begins! Firstly because SQL is by the tools being used or just by not ecosystem while doing so. a declarative language and it has nothing understanding how PostgreSQL is different to do with either OO or imperative thinking than Oracle. In this talk, we will discuss the To start we go over the basics: Which TLS and as for one point it makes it relatively top mistakes people generally make when settings are there in MySQL and MariaDB easy to express a condition in SQL it is not moving to PostgreSQL from Oracle and and how does this differ from HTTPS as so easy to express it optimally and even what the correct course of action. used in browsers. And why do we want TLS worse to translate it to the OO paradigm. in the first place? Is TLS and SSL the same For another point, they need to think in thing? terms of set and relational algebra, even if The first set of problems is inside MySQL: unconsciously! YaSSL vs. OpenSSL, verification issues and In this talk, we’ll see the most common reloading of certificates. mistakes that OO, and in particular Java, The second set of problems is inside developers make when writing SQL code Connectors: I’ll touch on DBD:: (), and how we can avoid them. Go-MySQL-Driver, libmysqlclient () Not all connectors have the same options and defaults. I’ll go into TLSv1.2 support. The third set of problems is tools: Using the require_secure_transport option caused issues with Percona Toolkit and Orchestrator. I’ll also cover: RSA v.s EC, security issues I found and how I wrote a Proxy for MySQL

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Protecting Your Secrets Strength in Numbers: MySQL 8.0: The New John Boero, Hashicorp Slack’s Database Architecture Replication Features When we talk about secret management Josh Varner, Guido Iaquinti, Slack Luis Soares, Oracle what we’re really talking about is Traditionally, database reliability has MySQL 8.0 replication reduces operation managing a set of different credentials. focused heavily on the stability of a single costs and minimizes administration What we mean when we talk about these server or a small number of servers. At overhead. MySQL DBAs and DevOps can credentials is anything that might grant Slack, however, we’ve built a database then spend more time expanding their you authentication to a system. Some architecture that instead focuses on infrastructure, rather than catering for it. examples of this might be usernames and an approach of strength in numbers. Many different areas have been enhanced. passwords, it might be things like database We’ve designed this architecture to allow For example security, operations, fail- credentials. This falls into the realm of failures to happen and to limit the effects over, observability, failure detection, secrets, and these are things we want to of those failures instead of attempting to consistency, split-brain protection and manage carefully. We want to understand eliminate them entirely. We’ve also built primary election and flexible replication who has access to them, we want to this architecture from the ground up to workflows. understand who’s been using these be cloud-first, and to use the nuances things, and in the case of most of these, This session highlights the new replication of running databases in the cloud to our features in MySQL 8.0. we want some story around how we can advantage. periodically rotate these. In this session, Those that were released pre and post- In this talk, we will review the architecture John will present his recommendations GA. Come and learn, directly from the we’ve built, specificlaly focusing on our on how to protect your secrets and avoid engineers, how the new features help you Vitess infrastructure. Vitess currently secret sprawl. operate, sustain and extend your MySQL handles nearly half of Slack’s database Replication infrastructure. traffic. We’ll talk about how we’ve built this system to withstand common failures and continue to perform at our scale.

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Optimize and Troubleshoot MongoDB Analysis with How To Upgrade Like a Boss to MySQL using PMM 2 Prometheus and Grafana MySQL 8.0? Peter Zaitsev, Percona Akira Kurogane, Percona Alkin Tezuysal, Percona Optimizing MySQL performance and Prometheus and Grafana are spreading This talk will explain best practices for troubleshooting MySQL problems are through DevOps organizations around upgrade techniques in MySQL. In deep two of the most critical and challenging the world. We’ll show how they been put dive, we will go over how to upgrade tasks for MySQL DBAs. The databases together in PMM2 to create MongoDB successfully to MySQL 8.0. Explain MySQL powering your applications need to be monitoring timeseries graphs and alert 8.0 upgrade specific challenges. Go over able to handle changing traffic workloads triggers equal to any of the commercial gotchas and best practices. Review the while remaining responsive and stable solutions that currently exist. latest version of MySQL 8.0 and bug so that you can deliver an excellent • Quick demo of the classic DBA tasks reports. user experience. Further, DBAs are also • Some new MongoDB DBA tricks expected to find cost-efficient means of • The ‘math’ in MongoDB metrics: solving these issues. »» the base counters in the server process In this presentation, we will demonstrate »» the prometheus timeseries metrics how using recently released PMM version expressions 2 enables you to solve these challenges, • Demonstration of fine-detail which is built on free and open-source performance bottleneck analysis software. We will look at specific, common MySQL problems and review.

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Running ElasticSearch at and what to alert on to make sure that Managing MySQL at Scale in your cluster is stable and … more. Large Scale All of this will be followed by some Q/A Facebook Alex Cercel, Palantir Technologies where I might be able to offer additional Pradeep Nayak, Junyi Lu, Facebook ElasticSearch is a high performance information that could help with other Section1: Terminology. Quick intro to search engine based on Apache Lucene. topics. shard/replicaset/instance/host and the It is growing in popularity which means I look forward to welcoming as many scale of FB that more and more users are running people as possible to the wonderful world Section2: Lifecycle of an instance part of their search patterns against the of search. Section3: Tools that are used for moving information stored on this datastore. instance between states: MPS Copy, What is quite difficult to find in one place is rebuild_db etc a comprehensive set of best practices that Section4: Shard movements(OLM) and will assist in efficiently running large scale touch base on balancing clusters. Section5: Test Infrastructure: Shadow and From my experience I’ve seen Merlindb many bottlenecks created by some misconfigured parameter or rogue approach and I would like to offer some tips and tricks so others won’t feel the same unnecessary pain. I’ve also seen how a tweek to a parameter nobody would have thought about, ended up to be a game changer. In this 50 minutes talk I will cover tuning parameters and approaches that will assist the audience in running fast, efficient and well written queries on stable and performant clusters. I will cover topics like schema design, talk about some different Garbage Collection mechanisms but also offer a good start on how to pick the right number of shards and replicas, role separation, what to look for when it comes to monitoring the cluster RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Graph Databases: without having to worry about competing How to Instrument Your Code languages and standards. I’ll try to sort Introduction, Standardization, through this and show what is currently in Performance Schema Opportunities going on in graph database languages. Georgi Kodinov, Oracle Peter Eisentraut, 2ndQuadrant Having all of the server code instrumented In this talk I want to present current in performance schema helps users developments in the world of graph diagnose problems without having to databases and discuss how SQL databases use a debugger. But you don’t have to be could evolve to work well with graph data. a server developer to instrument your code. MySQL 8.0 performance schema Graph databases are a major database APIs allow good instrumentation of all paradigm next to relational databases, code, including the code living in plugins object databases, hierarchical databases, and components. You can even be adding and so on. While there are popular your own performance schema tables standalone graph database products, and expose your own volatile data and some vendors from other fields including statistics into these. This session aims traditional SQL as well as NoSQL have to walk you through the process of begun to extend their products with graph instrumenting your code that is to run data capabilities and interfaces. Relational inside the MySQL server process. And give databases have been successfully you tips on how to do this most efficiently extended to support nonrelational uses and where to go for further details. such as document stores, so it is worth discussing whether and how graph database capabilities could also be integrated into SQL databases. At the moment, the graph database world suffers from an overabundance of different query languages. There are, however, efforts underway to unify some of these and standardize them under the umbrella of standardization organizations such as W3C and ISO. This would make it easier for SQL database products, whose primary focus is not graph data, to add support for graph data query languages RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

Becoming Cloud Native: How The DB is Broken. Now What? Automatic Upgrade and New Percona Brings Databases to Carmen Mason, Vitalsource Error Logging in MySQL 8.0 Kubernetes Using Operators Technologies, Allan Mason, Ivan Ståle Deraas, Oracle Tyler Duzan, Percona Groenewold, Pythian We will cover upgrade from MySQL 5.7 This talk covers some of the challenges Sometimes the only piece of information to MySQL 8.0 (8.0.15), going from legacy we sought to address by creating a that you are given is that the database is meta data storage to transactional Kubernetes Operator for Percona XtraDB not working correctly. What do you do data dictionary. We will cover the new Cluster and Percona Server for MongoDB, now? Panic? Call Mom? But, you’re the possibilities for automation of upgrade, as well as a look into the current state DBA. What you do next is crucial. and the major advances upgrade speed of the Operators, a brief demonstration In this talk, we’ll cover the tools that and reliability as well as new consistency of their capabilities, and a preview of the you’ll use to determine what other agents checks in the MySQL upgarde checker. roadmap for the remainder of the year. may be working against the database MySQL 8.0 has pluggable error log. We will This is a great way to find out more about performance. We’ll share our own talk about the traditional error logger and the viability of running databases within a experiences with crazy scenarios that the JSON error logger, which empowers containerized environment, how you can stumped us, and the tools that we used to user with advanced filtering. self-service databases in a cloud-vendor figure out the root cause. agnostic way. You also get the chance to Some examples include how systemd ask questions of the Product Manager and can ruin your day, NUMA imbalance, and provide feedback on your experiences in issues unique to AWS, and GCP. the Kubernetes ecosystem.

RETURN TO WEDNESDAY’S TIMETABLE WEDNESDAY

PostgreSQL Plan at execution gdb basics for MySQL DBAs time : A quick show Valerii Kravchuk, MariaDB Jobin Augustine, Percona Corporation A quick demo of decision makings Sometimes MySQL DBAs have to work with happening during the execution phase in stalled, hanged or unresponsive MySQL LIGHTNING PostgreSQL latest versions. instance, where their usual SQL-based tricks do not work any more. Sometimes they can not even connect to check what’s TALKS going on inside server. In other cases they know what to do and everything still works, but they have to implement changes to read-only server variables. Server restart is often not JSON Array Indexes in MySQL an option in production, as it means Norvald Ryeng, Oracle some downtime and may cause negative MySQL 8.0.17 has the ability to create performance impact. indexes over JSON arrays to speed up In these cases one could do something your queries. In this session we’ll show given read and write access to server examples and explain all you need to know memory/internals. Here comes gdb, that, about this new feature: How are array alone with careful reading of the source indexes created? How do they work? When code helps to often resolve the problems are they used? Are there any limitations? described above. During this session I’ll explain what can be done with gdb when server already is in troubles, and how to use gdb to “see” and understand MySQL internals better.

RETURN TO WEDNESDAY’S TIMETABLE SPEAKERS SPEAKERS

Ibrar Ahmed Manuel Arostegui Sunny Bains Senior Software Engineer & Database Administrator, Wikipedia Director of Software Development, PostgreSQL Consultant, Percona Manuel Aróstegui is a DBA working Oracle Ibrar Ahmed is a Senior Software Engineer for the Wikimedia Foundation, the non- Sunny Bains joined the Oracle/InnoDB in Percona. Prior to coming to open source profit organization maintaining Wikipedia team in 2006. He currently manages development, he had vast experience and other free knowledge collaborative the InnoDB team. He has been writing in software design and development. projects. software for longer than he can remember. The main focus was on the system level Backing up Wikipedia Databases Prior to joining InnoDB he wrote a embedded development. After joining database engine for a distributed database EnterpriseDB, an Enterprise PostgreSQL’s Jobin Augustine product at his previous employer. The company in 2006 he started his career Senior Support Engineer, Percona networked database was used in a mission in open source development specifically critical environment in the airline industry One of the people behind the PostgreSQL in PostgreSQL. He has contributed to the with very strict service level guarantees. Distribution: BigSQL. Jobin has always PostgreSQL community as well as other been an active participant in the InnoDB Management and Scalability open source communities. His contribution OpenSource communities, especially Improvements in MySQL 8.0 is from the main performance feature database communities and as an enhancements to various PostgreSQL’s OpenSource advocate. He is one of the top Charly Batista modules. In the database field, he has bloggers in the PostgreSQL community. Technical Lead, Percona experience in other well-known databases He is a contributor to various OpenSource MySQL, Oracle, and NoSQL databases such Charly worked as Java Architect for many Projects and he loves to code in C++ and as MongoDB and Hadoop. His experience is years and using many different database Python in his free time. He worked for not limited to core databases, but with the technologies. He helped to design some OpenSCG for 2 Years and Dell for 10 years tools related to databases like Hive, HBase of the features of the system used in the as a Database expert. and Spark. He also worked on integrating Brazilian Postal Service, the largest Java these tools with PostgreSQL. He joined Why PostgreSQL is Becoming A project in Latin America in that time. He Percona in the month of July 2018. He also Migration Target in Large Enterprises also helped to design the database of the authored multiple books on PostgreSQL. PostgreSQL Plan at Execution Time: Brazilian REDESIM project, the system that is responsible for the municipalities Join Heterogeneous Databases Using A Quick Show taxation in Brazil. He now lives in China and PostgreSQL Foreign Data Wrappers works as Senior Engineer at Percona. Percona Distribution for PostgreSQL 10 Common Mistakes (Java)Developers Make When Writing SQL

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Krunal Bauskar Satya Bodapati John Boero PXC Product Lead, Percona Senior Software Engineer, Percona Channel Solutions Engineer, Krunal has been working in MySQL space Satya works with Percona Server Hashicorp for over a decade now. He has actively Engineering team. He is responsible for John Boero is from Chicago living in contributed to MySQL community through Percona Server features, bug fixes, etc. He London as a Channel Solution Engineer at different projects he has worked on. has an overall experience of 12 years in Hashicorp. He helps educate Hashicorp’s He continues to contribute to Percona MySQL. He joined Percona in 2018. Before international partner network on XtraDB Cluster making it one of the joining Percona, Satya worked with the Hashicorp technologies and enables them first choice when it comes to MySQL HA InnoDB Development team at Oracle for 6 to sell to their customer base. solutions. In past he was also associated years. He has good knowledge of InnoDB Protecting your Secrets with InnoDB team @ MySQL/Oracle where internals and the new Data Dictionary. he authored some of the most important Loves to do C++ coding and write Keynote features of MySQL-5.7 like temporary algorithms. table revamp, atomic truncate, undo log Percona Server for MySQL 8.0: Alex Cercel truncate, etc... He has also worked with What It Is and How It Is Done Site Reliability Engineer, Palantir Yahoo! Labs researching on bigdata Technologies problems and MySQL based database Matthew Boehm Seasoned Linux, Windows, Network startup which is now part of Teradata. Senior Instructor, Percona administrator, DBA, recently turned Percona XtraDB Cluster 8.0 (PXC-8.0) Matthew has been with Percona for almost SRE with 15 years experience in the 7 years. During that time he has fulfilled industry. For over 4 years I’ve been the roles of Senior Consultant, Senior involved in running some of the largest Architect, and now as Senior Instructor, implementations of MongoDB and he delivers world-class training to clients ElasticSearch in the world. All round geek all across the globe. His areas of speciality with a genuine passion for anything that include Percona XtraDB Cluster, ProxySQL, runs on or involves computers. From the MySQL Backups and Recovery. OS to games. Getting Started with Kubernetes and Running ElasticSearch at Large Scale Percona XtraDB Cluster

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Colin Charles Michael Coburn Stephane Combaudon Consultant, Galera Cluster Product manager, Percona Senior DB engineer, Rakuten Colin Charles is a Consultant at Codership, Michael joined Percona as a Consultant Intelligence the makers of Galera Cluster. Previously, in 2012, and progressed through various Stephane has worked for almost 15 Colin was on the founding team of MariaDB roles including Managing Consultant, years in open source environments, Server, and has been around the MySQL Principal Architect, Technical Account and has spent most of the last 10 years ecosystem including being an early Manager, and Technical Support solving performance or high-availability employee at MySQL, and worked actively Engineer, and is now leading the Product problems with MySQL and more recently on the Fedora and OpenOffice.org projects. Management of Percona Monitoring and MongoDB. He is now working for Rakuten Colin has been a MySQL user since 2000. Management. Intelligence, a company that helps you MariaDB Server 10.4: Open Source Database Performance keep track of your online purchases. The Complete Tutorial Optimization and Monitoring with PMM His main role is to automate MySQL and What’s New in PMM 2.0 MongoDB operations, and from time to Vicențiu Ciorbaru time, he is also very excited to kill a nasty Software Developer Team Lead, Running PMM in Production at Tessi query or add a missing index on a 500GB MariaDB Foundation table. When not in front of a computer, Jeremy Cole Stephane likes spending time with his Vicențiu Ciorbaru works for the MariaDB Sr. Staff Production Engineer, family. Foundation as a Server Developer and Shopify, Inc. Our journey to better MySQL de facto team-lead of junior developers Availability Using Global Transaction and new contributors. Vicențiu’s past Jeremy Cole is a pioneer of MySQL projects include Roles in MariaDB, Window scalability having worked for on MySQL IDs, ProxySQL and Consul Functions, Custom Aggregate functions. for 19 years at MySQL AB, Yahoo!, Proven He also ensures MariaDB remains stable Scaling (providing advanced MySQL and up-to-date in Linux distributions consulting), Twitter, Google, and now by handling distribution specific bugs, working as a Sr. Staff Production Engineer especially in Debian. in the Datastores team at Shopify. How a Modern Database Gets Your Data MySQL on Google Cloud: The Good, The Fast: MariaDB Query Optimizer Bad, and The Ugly

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Jaime Crespo Caleb James DeLisle Cristina DeLisle Sr. Database Administrator, Protocol Designer, CJDNS Office and Legal Administrator, DPO Wikipedia Caleb James DeLisle founded the cjdns XWiki SAS Jaime Crespo is a DBA working for the Open Source project for cryptographically Cristina became interested in software Wikimedia Foundation, the non-profit assured IP packet routing, he then went when she began to work at XWiki and organization maintaining Wikipedia on to co-develop ncry.pt, the first Zero CryptPad, open source projects that made and other free knowledge collaborative Knowledge pastebin using a secure key her more tech aware, coming from a legal projects. Lately his focus has been on sharing technique which would later find background. She is a data protection recovery and database automation. it’s way into ZeroBin, Mega.co.nz and officer and in this role she’s following the Backing up Wikipedia Databases others. During his time as a researcher GDPR’s evolution, striving to ensure on a at XWiki SAS, he authored the first daily basis compliance with the applicable Tom De Cooman realtime collaborative editor to make data protection rules. use of Nakamoto Blockchains for state Data Protection and OSS in the Age of SDM Team Manager, Percona synchronization and used this to launch GDPR Tom has been active in Unix system’s the CryptPad project, a google docs like administration for 10+ years. Before editor where the all content is encrypted Ståle Deraas joining Percona, he has been working and the server has no access to the keys. for several consultancy companies and Software Development Director, His Open Source work has been featured system’s integrators, always working, Oracle in Wired Magazine (USA), NewScientist focusing and offering solutions based on (USA), Neue Zurcher Zeitung (Switzerland), Ståle Deraas has been working with OSS. Kommersant (Russia). MySQL Database team since 2008. MySQL 101 Tutorial Parts 1 & 2 He is currently employed by Oracle, Databases, Crypto & Decentralization based in Norway. He is Senior Software Development Manager for the runtime team of the MySQL Database. He has a background in the database industry, working for Sun Microsystems on Java DB, and prior to that on Software Development Tools. He has a B.Sc Hons degree in Computing Science from the University of Glasgow. Automatic Upgrade and New Error Logging in MySQL 8.0

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Frédéric Descamps Igor Donchovski Operators MySQL Community Manager, Lead Database Consultant, Pythian Yoav Eilat Oracle MongoDB Certified DBA supporting the Senior Product Manager, Amazon “@lefred” has been consulting OpenSource next-generation of database solutions in and MySQL for over 15 years. After both MySQL and MongoDB. With a masters Web Services graduating in Management Information degree in Software Engineering, Igor Deep Dive on Amazon Aurora Technology, Frédéric Descamps started his enjoys the variety of challenges he faces career as a developer for an ERP under while working at Pythian, rather than Peter Eisentraut HPUX. He will then opt for a career in the working on the same thing every day. Deputy CTO, 2ndQuadrant world of open-source by joining one of When he’s not working, he can be found Peter has been a contributor to the first Belgian start-up dedicated 100% biking PostgreSQL development since 1999. He to free projects around GNU/Linux. It is Maintenance for MongoDB Replica Sets is a committer, core team member, and is in 2011 that lefred joined Percona, one involved in many aspects of the community. of the leading MySQL-based specialists. Tyler Duzan He works at 2ndQuadrant, a global service He decided to join the MySQL Community Product Manager, Percona Team in 2016 as a MySQL Community and support provider for PostgreSQL. Manager for EMEA & APAC. Frédéric is Prior to joining Percona as a Product Graph Databases: Introduction, also a regular speaker of OpenSource Manager, Tyler spent almost 13 years as Standardization, Opportunities Conferences. His blog mostly dedicated to an operations and security engineer in a MySQL is lefred.be variety of different industries. Deciding to Daniël van Eeden take his analytical mindset and strategic Database Administrator, MySQL 8.0 InnoDB Cluster: focus into new territory, Tyler is applying Easiest Tutorial! his knowledge to solving business Booking.com MySQL InnoDB Cluster: Advanced problems for Percona customers with Open Source enthusiast. Contributor to Configuration & Operation inventive solutions combining technology various projects in the MySQL ecosystem. MySQL Shell : The Best DBA tool? and services. Large Scale Deployment of SSL/TLS For Extend the Shell with the New Extension Becoming Cloud Native: How Percona MySQL Infrastructure Brings Databases to Kubernetes Using

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Antonios Gianopoulos Kathy Gibbs Jonathan Gourdon Senior Database Administrator, Senior Database Specialist Tessi ObjectRocket by Rackspace Solutions Architect Jonathan started his career as a Software Antonios Giannopoulos is a Senior Accelerating Application Development engineer in Bank IT. He joined Tessi in Database Administrator at ObjectRocket with Amazon Aurora 2009. supporting MongoDB, Elastic and As a constant learner, he progressed CockroachDB. Antonios has 18 years Pedro Gomes through various roles within the company of experience in databases and system Software Engineer, Oracle and is currently working as an Automation engineering. He really enjoys challenges Leader. His main role is to secure and in sharding and schema design and loves Pedro is graduated and did his masters automate solutions deployments, including migrations from Relational to NoSQL world. on Minho University where he did his monitoring, such as MySQL environments dissertation on distributed systems. It was or NoSQL databases (Cassandra). New Indexing and Aggregation Pipeline a mainly about the use of NoSQL databases He works continuously towards the Capabilities in MongoDB 4.2 in the industry and it included a ORM proof objective of providing production ready What’s New on Sharding in MongoDB of concept. He joined Oracle on September workflows. 4.2 2012 after having worked for one year for the University Distributed System Lab In his free time, he loves climbing and doing some research. being in nature. MySQL InnoDB Cluster: Advanced Running PMM in Production at Tessi Configuration & Operation

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Vinicius Grippa Ivan Groenewold Kenny Gryp Senior Support Engineer, Percona Lead Database Consultant, Pythian Principal Product Manager, Oracle Vinicius Grippa is a Percona Senior Group Kenny is a MySQL expert with a lot Support Engineer. Vinicius has a Ivan has been supporting mission-critical of experience in performance and Bachelor’s degree in Computer Science environments for top of the line companies architecture. While he has focused and has been working with databases for over 15 years, while becoming a on MySQL and automation for over a for 10 years. He has experience in certified professional for Oracle, MySQL, decade, he has also developed expertise designing databases for mission-critical MongoDB, Google Cloud and Amazon Web in other areas. He’s a well-rounded, applications and in the last few years Services in the process. multidimensional nerd who is skilled in has become a specialist in MySQL and Also co-developed the Database-as- many things: making a coffee on the side MongoDB ecosystems. Working in the a-Service platform for the biggest of a mountain, electrical building codes Support team, he has helped Percona e-commerce site of Latin America for single family dwellings in the US, the customers with hundreds of different (MELI:NSQ) perfect ratio of cumin and asafoetida in cases featuring a very wide range of tadka dal, and providing occupational scenarios and complexities. Vinicius is also The Database is Broken. Now What? therapy to a shy rabbit named Winston. active in the OS community, participating He’s adept at talking, so don’t feel the need in virtual rooms like Slack, and speaking to limit in-session questions to MySQL. at conferences, and MeetUps in Europe, MySQL 8.0 InnoDB Cluster: North and South America. Easiest Tutorial! Open Source Database Performance Optimization and Monitoring with PMM A Journey with MongoDB HA. From Standalone to Kubernetes Operator Enhancing MySQL Security

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Robert Hodges Geir Høydalsvik Guido Iaquinti CEO, Altinity Software Development Director, Operations Engineer, Slack Robert Hodges is CEO of Altinity, which MySQL Server, Oracle I’m a system engineer with academic offers software and services for eir Høydalsvik has been working with background and experience in high ClickHouse, a high performance OSS MySQL Database team since 2008. volume/high availability Internet data warehouse. His experience in He is currently employed by Oracle, architectures. I’m a technology enthusiast data management includes relational based in Norway. He is Senior Software excited about open source software. My database work at Sybase, developing SaaS Development Director and responsible passion is to develop, scale and automate applications on Oracle, and over 10 years for the development and maintenance of complex systems. working on replication and clustering MySQL Database. He has a background Strength in Numbers: Slack’s Database products for MySQL. He was CEO of in the database industry, working for Architecture Continuent, Inc., when it was acquired by the database startup company Clustra VMware in 2014. After 4 years working Inc. on the Clustra database and for Leigh Johnson on multi-tenant cloud services, disaster Sun Microsystems on Java DB. He has Staff Database Reliability Engineer, recovery, and security he returned to data a Master degree in Computer Science Slack management at Altinity in 2019. and a PhD in Software Engineering from ClickHouse Materialized Views: A Secret the Norwegian University of Science and Leigh Johnson is a Staff Database Weapon for High Performance Analytics Technology. Reliability Engineer at Slack and a Google Developer Expert in Machine Learning. Her Keynote passion is automating herself out of a job, so she can spend all of her time with her corgi. Lessons From Building Automation for a Large Distributed Database

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Craig Kerstiens Ameet Kotian Valerii Kravchuk Head of Product for Citus Cloud, Staff Database Reliability Engineer, Principal Support Engineer, Microsoft Slack MariaDB Corporation Craig currently works on product for open Ameet Kotian is a Staff Database Reliability Valerii Kravchuk supported MySQL users source databases at Microsoft. He joined Engineer at Slack, where he is responsible since 2005, in MySQL AB, Sun, Oracle, Microsoft by way of Citus Data, where for operating the infrastructure used to Percona and (since 2016) MariaDB he ran product including their database store all of Slack’s data. Previously, he was Corporation. as a service. Prior to joining Citus he ran one of the first site reliability engineers at Tracing and Profiling MySQL Twitter, where he spent close to five years product for Heroku Postgres. He’s actively gdb Basics for MySQL DBAs involved within the Postgres community, working on the graph storage service and blogs about it regularly, and curates Twitter’s internal distributed database. Dimitri Kravtchuk Postgres Weekly. Lessons From Building Automation MySQL Performance Architect, for a Large Distributed Database Keynote Oracle Roman Khavronenko Tibor Köröcz Over 25 years in IT industry, mainly in COO, VictoriaMetrics Architect, Percona Databases and Systems Performance. Since 2011 working full time on MySQL Tibi joined Percona in 2015 as a VictoriaMetrics: Why and How We Built Performance at Oracle. Previous 15 Consultant. Before joining Percona, A Fast and Scalable Open Source Time years at Sun Benchmark Center in Paris. among many other things, he worked Series Database Previously at Informix Benchmark Team in at the world’s largest car hire booking France. Author of dim_STAT Performance service as a Senior Database Engineer. He Georgi Kodinov Monitoring & Analyzing tool, actively enjoys trying and working with the latest Team Lead, MySQL Server General publishing MySQL Performance related technologies and applications which can articles at http://dimitrik.free.fr/blog. Team, Oracle help or work with MySQL together. In his Georgi “Joro” Kodinov started working on spare time he likes to spend time with his MySQL 8.0 Performance: Scalability & MySQL in 2006 as a part of the optimizer friends, travel around the world and play Benchmarks team. He has moved to lead the Server ultimate frisbee. General team, working on security, Percona XtraDB Cluster Tutorial monitoring and the client-server protocol. Joro has background in banking IT. How to Instrument Your Code in Performance Schema

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Akira Kurogane Junyi Lu Kevin Markwardt MongoDB Technical Lead, Percona Production Engineer, Facebook Project Engineer, Pythian Akira Kurogane is an experienced Junyi Lu is currently working at Facebook Kevin Markwardt has twenty years database and search engine developer as a production engineer for the MySQL of system administration experience who advocates for long-term goals in the infrastructure. Responsible for developing ranging from MySQL, Linux, Windows, IT industry. automations which will manage tons of and VMware. Over the last six years he What’s New in Percona Server for MySQL instances. Prior to this, he worked has been dedicated to MySQL and Linux MongoDB as a MySQL DBA at dianping.com, which administration with a focus on scripting, is one of the biggest online-to-offline automation, HA, and cloud solutions. Kevin Percona Backup for MongoDB platforms in China. has lead and assisted with many projects MongoDB Analysis with Prometheus Managing MySQL at Scale in Facebook focusing on larger scale implementations and Grafana of technologies, including ProxySQL, Aayushi Mangal Orchestrator, Pacemaker, GCP, AWS RDS, Pavel Lapaev Support Engineer, Percona and MySQL. Kevin Markwardt is a certified GCP Product Manager, Tarantool Aayushi Mangal is a Support Engineer Professional Cloud Architect, and a Background of finance and audit with Percona. Previously, she worked as certified AWS Solutions Architect - turned into several years as a technical a NoSQL Database Administrator for the Associate. Currently he is a Project trainer and telecom business trainer retail based MNC. Her experience includes Engineer at Pythian specializing in MySQL in Netcracker. Last two years works in work as a database developer, database and large scale client projects. One of his Tarantool as a product manager. administrator, and as a web designer. new directives is Postgres and is currently Living in India, Aayushi is fond of nature, Tarantool: A NoSQL Database with SQL supporting multiple internal production sightseeing and peace. Postgres instances. Multi-Document Transactions Securing MySQL with Vault and Table Through MongoDB Shell Level Encryption

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Mykola Marzhan Carmen Mason Giuseppe Maxia Director of Server Engineering, Senior Database Administrator Software explorer, dbdeployer Percona VitalSource Technologie Software Explorer working with a virtual Mykola is Kubernetes and Clouds lover Carmen Mason works at VitalSource team and maker of exploration tools. and, currently, his goal to bring databases Technologies as their Senior Database Make difficult things easy. into Kubernetes world. Since 2004, most of Administrator, where she helps decide MySQL community member for long, long his career has focused on the development on and plays with the latest cloud toys, time. of monitoring, update and deployment and works with some pretty brilliant folks Four times receiver of MySQL community systems. on a daily basis, and enjoys ROOT on her awards, twice as contributor, and twice as Comparison of Kubernetes Operators database instances. Her favorite thing application of the year developer. about her job is the challenge and the for MySQL Test Like a Boss: Deploy and Test people she works with. Complex Topologies With a Single Allan Mason The Database is Broken. Now What? Command Database Consultant, Pythian Amazing Sandboxes with dbdeployer Group Yoshinori Matsunobu Production Engineer, Facebook Allan Mason enjoys constantly learning solutions to current and future issues; Yoshinori Matsunobu is a Production utilizing that information to help people and Engineer and now a tech lead of MySQL businesses. He works for Pythian doing team at Facebook. Yoshinori has been exactly what he loves in a challenging around MySQL community for over 10 environment. He is passionate about years, and was a senior consultant at teaching and automation. MySQL Inc since 2006 to 2010. Yoshinori created a couple of useful open source The Database is Broken. Now What? product/tools, including MHA (automated MySQL master failover tool), quickstack and MyRocks (RocksDB storage engine for MySQL). MyRocks and RocksDB Advanced Features and Performance

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Boriss Mejias Aleksei Milovidov Károly Nagy PostgreSQL Consultant, Team Lead, Yandex Lead Database Engineer, 2ndQuadrant Since 2008 I was working on development Salesforce I’m a holistic system software engineer, of data processing pipeline of Yandex Karoly has been working with Salesforce PostgreSQL consultant and trainer, Metrica - web analytics system. since 2018 to improve the MySQL free software activist, and headbanger. Since 2015 I am responsible for automation and help them scale the I have been working with PostgreSQL development of ClickHouse - open-source database infrastructure. Before that he since version 9.1. First, as part of my job column-oriented database management was working with Dropbox and Booking. related to other projects, and now with full system, used for realtime analytical com on similar projects. His enthusiam dedication since I joined 2ndQuadrant in reports (https://clickhouse.yandex/). towards large scale distributed and 2017. I have a background as teaching- I have 11 years experience with autonomous systems had helped him assistant and researcher on self- development of specialized data structures developing time-tested frameworks managing distributed systems, and I have using C++ programming language. to alleviate the day to day burden of presented in different conferences. ClickHouse Features to operations. In open source, I have contributed to the Blow Your Mind From Scheduled Downtime to Self- Mozart-Oz programming language and Healing in Less Than a Year Alfresco communities. In 2018 I have Jim Mlodgenski started the PostgreSQL User Group in Principal Database Engineer, Belgium. Amazon Web Services Pg_catalog Unrevealed! That Part of PostgreSQL You Are Probably Principal database engineer at Amazon Underusing Web Services, Co-Chair - PostrgesConf, Co-organizer of Philly PUG & NYC PUG, former CTO at OpenSCG Introduction to PL/pgSQL Development Top 10 Mistakes When Migrating From Oracle to PostgreSQL

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Pradeep Nayak Michal Nosek Yuriy Olshanetsky Production Engineer, Facebook Senior Solutions Engineer, Percona Senior DB Engineer, Rakuten Pradeep is a Production Engineer During ten years of his career, Michal took Intelligence at Facebook and works with MySQL different roles from a software engineer Yurii started his career as DBA in Infrastructure. He loves hacking code in and business analyst to a technical Rakuten Intelligence 3 years ago. Rakuten python and builds bots to do things for him. sales consultant, always staying close to Intelligence is a company that helps you While he is not working, he enjoys traveling technology. He has hands-on experience keep track of your online purchases and and clicking pictures. with a broad range of programming save money. His main role is to automate Managing MySQL at Scale in Facebook languages and database technologies in MySQL and MongoDB operations, isolate different verticals. DBs from the human factor, improve John Nguyen As a Solutions Engineer at Percona, security and of course help people to make Senior Partner Solutions Architect Michal’s objective is to bridge the gap good queries. In free time Yurii prefers to between business and engineering by be with his family and traveling. Accelerating Application Development providing customers with the best open- Our journey to better MySQL with Amazon Aurora source software solutions and Percona Availability Using Global Transaction services. IDs, ProxySQL and Consul Michal holds a BS degree in Computer and Information Science from Cracow University of Technology. He lives in Gdansk, Poland with his wife. In his spare time, Michal enjoys traveling and sports. MySQL 101 Tutorial Parts 1 & 2

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Iwo Panowicz Vítor Pombeiro Norvald Ryeng Senior Support Engineer, Percona Support Engineer, Microsoft Software Development Senior Iwo is a Percona Support Engineer MySQL, MariaDB & PostGreSQL Manager, Oracle (EMEA) since August 2017. Prior to joining as PaaS on Azure Norvald has been at Oracle since 2011 and Percona Iwo was a System Administrator is the MySQL optimizer for more than 10 years with a focus Alexander Rubin team lead, responsible for query on high availability, performance and Director, Data Architecture, optimization and execution. He holds a automation. He also was helping many VirtualHealth PhD in Computer and Information Science companies in designing infrastructure and from the Norwegian University of Science ha/failover solutions. Alexander has over 10 years industry and Technology. MySQL Has Gone Away: An In-Depth experience with the MySQL database and related technologies. His specialties MySQL 8.0.1: Latest Updates Look at the MySQL Networking are performance tuning, full text search, Implementation JSON Array Indexes in MySQL high availability, database infrastructure architecture and data warehouses. He has Rachel Pescador helped many MySQL customers design Product Marketing Manager, extremely high performance databases Percona with optimized schema and queries. Rachel Pescador is a Product Marketing How to protect PII data in MySQL Manager at Percona. She is responsible While Allowing Customers to for highlighting Percona’s open source Access the Database database expertise by promoting Percona’s support, consulting, managed services, and software solutions. Prior to joining Percona, Rachel spent 18 years in marketing and business development for some of the world’s largest and most well- known professional services firms. What’s New in PMM 2.0

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Art van Scheppingen Sveta Smirnova Luis Soares Database Engineer, Messagebird Principal Support Escalation MySQL Replication Director, Oracle Art van Scheppingen is Database Engineer Specialist, Percona Luis Soares is the MySQL Replication at Messagebird. He’s a pragmatic MySQL Sveta Smirnova is a MySQL Support Director at Oracle. His research interests and Database expert with over 18 years Engineer with over than 10 years include replication technologies, experience in web development. He experience. Her main professional dependable systems and high availability. previously worked as a Senior Support interests are problem-solving, working Before joining the MySQL team, he was Engineer at Severalnines. Prior to with tricky issues, bugs, finding patterns both a post-graduate student and a Severalnines he worked at Spil Games as which can solve typical issues quicker, researcher at the University of Minho, Head of Database Engineering. teaching others how to deal with MySQL Portugal, where he designed and He has experience with a broad spectrum issues, bugs and gotchas effectively. implemented group-based replication of (open source) databases: from MySQL She is the author of the book “MySQL protocols. Before that, he worked for a to MongoDB, Cassandra to Hadoop and Troubleshooting” and JSON UDF functions Portuguese multi-national company as a from Sphinx Search to SOLR. He regularly for MySQL. system analyst. presents his work and projects at various Open Source Database Performance MySQL 8.0: The New Replication conferences (Percona Live, Community Optimization and Monitoring with PMM Features Open House for MongoDB, FOSDEM) and related meetups. Billion Goods in Few Categories: Yura Sorokin How Histograms Save a Life? Benchmarking Should Never be Optional Senior Software Engineer, Percona Yura is a Senior Software Engineer at Percona. He is a primary developer who implemented “Compressed Columns with Dictionaries” for Percona Server. Before joining in July 2015 he was leading backend dev team for a cloud file service with a focus on client-side encryption. He has 15+ years of software development experience, primarily in C++. Yura holds Master degree in Computer Science from National Technical University of Ukraine. He lives in Kyiv, Ukraine. Percona Server for MySQL 8.0: What It Is and How It Is Done SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Sugu Sougoumarane Feike Steenbergen Alkin Tezuysal CTO, PlanetScale Software Engineer, Timescale Sr. Technical Manager, Percona Sugu is the co-creator of Vitess, and has Feike has been working with PostgreSQL Alkin Tezuysal is a Senior Technical been working on it since 2010. Prior to for over a decade. Performance related Manager in Percona’s Managed Services Vitess, Sugu worked on scalability at issues have been a recurring theme; team. He has extensive experience in YouTube and was also part of PayPal whether working in retail (Zalando, enterprise relational databases working in the early days. His recent interest is bol.com), finance (Adyen) or now with in various sectors for large corporations. in distributed systems and consensus TimescaleDB, whether with OLTP or with With more than 20 years of industry algorithms. He occasionally shares his OLAP databases. experience, he has acquired skills for thoughts on his blog ssougou.blogspot. He’s one of the organizing members of The managing large projects from the ground com. Netherlands PostgreSQL Meetup. up to production. For the past 10 years, Vitess: Running Sharded MySQL Feike was part of the team that built he’s been focusing on e-commerce, SaaS on Kubernetes Patroni, allowing fast failovers of and MySQL technologies. He managed and PostgreSQL without human intervention. architected database topologies for high He has given PostgreSQL related trainings volume site at eBay Intl. He has several almost every year in the past 7 years. years of experience in 24X7 support and operational tasks as well as improving Parted Ways with Partitioning? It’s Time database systems for major companies. He to Reconsider has led MySQL global operations team on Tier 1/2/3 support for MySQL customers. In 2016 he joined Percona’s expert technical management team. How to Upgrade Like a Boss to MySQL 8.0?

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Morgan Tocker Adamo Tonete Valentin Traën Community Development Manager, Senior Support Engineer, Percona Administrateur Bases de données, PlanetScale Adamo joined Percona in 2015, after Tessi Morgan Tocker is the Community working as a MongoDB/MySQL Database Valentin has been working for Tessi for Development Manager at PlanetScale. He Administrator for three years. As the more than 4 years, and as the DBA for 2 has worked in the MySQL space for the main database member of a startup, years. He is mainly responsible for the last 15 years, and was previously Product he was responsible for suggesting maintenance of Tessi’s databases but Manager for the MySQL Server at Oracle. the best architecture and data flows the person who gets the job when a new My First 90 Days With Vitess for a worldwide company in a 7/24 query or table is needed. He is studying an environment. Before that, he worked engineering course at the famous French as a Microsoft SQL Server DBA in a school, the CNAM and will graduate in large e-commerce company, mainly on 2020. performance tuning and automation. Valentin specializes in MySQL database Adamo has almost ten years of engineering but also works with experience working as a DBA and in the other database systems (Cassandra, past five years, he has moved to NoSQL PostgreSQL, Oracle, SQL Server ....). technologies without giving up relational I currently live in France, in Limoges. databases. Running PMM in Production at Tessi A Journey with MongoDB HA. From Standalone to Kubernetes Operator Ben Tullis MongoDB Data Security: Custom Rules Head of Infrastructure, and Views OpenCorporates Ltd What’s New in Percona Server for With over 20 years’ experience focused MongoDB on Linux systems management and Open Source infrastructure projects behind him, Ben has for the last six years played a key role in crafting the platforms underpinning OpenCorporates, the largest open database of companies in the world. OpenCorporates: Providing Transparency for the Public Benefit

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Marco Tusa defining the Organization’s guidelines for Carlos Tutté Consultant, Percona the dissemination of information from the Support Engineer, Percona technology and the management point of Marco Tusa had his own international Carlos joined Percona on February 2018, view.He has participated in field missions practice for the past twenty eight years. being part of AMER support team. in order to perform analysis, reviews and His experience and expertise are in a evaluation of the status of local projects, From Montevido, Uruguay graduated as wide variety of information technology providing local support and advice. He a Compute Engineer on 2012. Working on and information management fields, had collaborates with MIT Media Lab IT since 2010 taking different roles and cover research, development, analysis, (Massachusetts Institute of Technology technologies, first as a developer, and then quality control, project management and laboratory) and FAO as Sustainable moving to DBA and sysadmin. team management.Marco is currently Information Technology for developing Open Source Database Performance working at Percona as High Availability countries Specialist in relation with the Optimization and Monitoring with PMM Practice Manager, previously working FAO’s Special Program for Food Security at Percona as manager of the Consulting PCI/DSS Compliance with MySQL: 2019 for Senegal. Rapid Response Team on October 2013. Edition He has being working as employee Improving Enterprise HA and Disaster for the SUN Microsystems as MySQL Recovery Solutions Aliaksandr Valialkin Professional Service manager for South Fortify Your MySQL Data Security in CTO, VictoriaMetrics Europe., and previously in MySQL AB.He AWS Using ProxySQL Firewall Go developer, fasthttp author, has worked with the Food and Agriculture VictoriaMetrics founder and core Organization of the United Nation since developer. 1994, leading the development of the VictoriaMetrics: Why and How We Built Organization’s hyper textual environment. Team leader for the FAO corporate A Fast and Scalable Open Source Time database support.For several years he Series Database has led the development group in the WAICENT/Faoinfo team. He has assisted in

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Avinash Vallarapu Dimitri Vanoverbeke Markus Winand PostgreSQL Technical Lead, Senior Solutions Engineer, Percona SQL Renaissance Ambassador, Percona At the age of 7, Dimitri received his first winand.at Avinash Vallarapu joined Percona in computer, since then he has felt addicted Markus Winand is the SQL Renaissance the month of May 2018. Before joining to anything with a digital pulse. Dimitri Ambassador. He is on a mission to Percona, Avi worked as a Database has been active in IT professionally since introduce developers to the evolution of Architect at OpenSCG for 2 Years and as a 2003 in which he took various roles from SQL in the 21st century. His first book “SQL DBA Lead at Dell for 10 Years in Database internal system engineering to consulting. Performance Explained” is available in five technologies such as PostgreSQL, Oracle, Prior to joining Percona, Dimitri worked languages and can be read for free on his MySQL and MongoDB. He has given as a Open Source consultant for a leading website use-the-index-luke.com. Markus is several talks and trainings on PostgreSQL. Open Source software consulting firm currently working on his second book that He has good experience in performing in Belgium. During his career, Dimitri is likewise available free of charge online Architectural Health Checks, Performance became familiar with a broad range of at modern-.com . Markus can be hired Tuning and Migrations to PostgreSQL open source solutions and with the devops as trainer, speaker and consultant via Environments. philosophy. Whenever he’s not glued to winand.at. PostgreSQL For Oracle and MySQL his computer screen, he enjoys travelling, More resources: use-the-index-luke.com/ cultural activities, basketball and the great DBAs and For Beginners outdoors. Dimitri is living with his girlfriend More Than a Query Language: High Availability and Automatic Failover in the beautiful city of Ghent, Belgium. SQL in the 21st Century in PostgreSQL MySQL 101 Tutorial Parts 1 & 2 Handling Transaction ID Wraparound in PostgreSQL Josh Varner Database Reliability Engineering Manager, Slack Strength in Numbers: Slack’s Database Architecture

SEE THE WEBSITE FOR ANY LATE CHANGES SPEAKERS

Matt Yonkovit Alexander Zaitsev success with MySQL. Percona was named CXO, Percona CTO, Altinity to the Inc. 5000 in 2013. Peter was an early employee at MySQL AB, eventually Before joining Percona, Matt worked Alexander Zaitsev is a co-founder and leading the company’s High-Performance at MySQL and Sun Microsystems as CTO of Altinity. He has been involved in Group. A serial entrepreneur, Peter co- a Solution Architect, helping to build software development, alongside academic founded his first startup while attending architectures for Fortune 500 and top research, since 1997. Alexander’s Moscow State University where he Alexa-rated companies. Since the mid-’90s, interests include distributed architecture, majored in Computer Science. he has worked with relational databases databases and analytics. His focus is as a DBA, system administrator, and all on building analytics solutions using As CEO of Percona, Peter enjoys mixing around techie. During his career as a DBA, database management systems capable business leadership with hands-on Matt has designed and supported massive of processing petabytes of data, such technical expertise. Peter is the co-author database deployments that ran on MySQL, as Vertica and ClickHouse. Alexander of High-Performance MySQL published by Oracle, DB2, SQL Server, Sybase, Postgres, has a Master’s degree in mathematics O’Reilly, one of the most popular books on and the AS/400. and computer science from Lomonosov MySQL performance. Peter blogs regularly on MySQLPerformanceBlog.com and Matt has a passion for performance Moscow State University. speaks frequently at conferences. and spends his free time doing things ClickHouse for Time-Series Real-time like benchmarking, finding outside-the- Analytics Peter lives in North Carolina with his wife box solutions for today’s performance and two children. In his spare time, Peter challenges and trying to make things just Peter Zaitsev enjoys travel and spending time outdoors. go faster. CEO, Percona Keynote Keynotes & MC Peter co-founded Percona in 2006, Innodb Architecture and Performance assuming the role of CEO. Percona helps Optimization Tutorial for MySQL 8 companies of all sizes maximize their Optimize and Troubleshoot MySQL Using PMM 2.0

SEE THE WEBSITE FOR ANY LATE CHANGES CONFERENCE CONFERENCE AND TUTORIAL AND TUTORIAL SCHEDULE SCHEDULE