MySQL 8 High Availability with InnoDB Clusters

Copyright © 2017, Oracle and/or its affiliates. All rights reserved. |
Copyright © 2018, Oracle and/or its affiliates. All rights reserved.
MySQL 8
High Availability
with InnoDB Clusters
Miguel Araújo
Senior Software Developer
Alfredo Kojima
Senior Software Development Manager
MySQL Middleware and Clients
October 24, 2018

Safe Harbor Statement
The following is intended to outline our general product direction. It is intended for
information purposes only, and may not be incorporated into any contract. It is not a
commitment to deliver any material, code, or functionality, and should not be relied upon
in making purchasing decisions. The development, release, timing, and pricing of any
features or functionality described for Oracle’s products may change and remains at the
sole discretion of Oracle Corporation.

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 3
Who are we?

whoami
Miguel Araújo
• MySQL Senior Software Developer
• Lead developer for the AdminAPI at the
MySQL Shell team
• MySQL InnoDB Cluster evangelist
• Living in Portugal
• http://about.me/miguel-araujo
4

whoami
Alfredo Kojima
• MySQL Senior Software Development Manager
• Technical lead for MySQL Clients team: X Plugin,
Router, Shell, C Connectors
• MySQL InnoDB Cluster evangelist
• Living in the USA
5

Program Agenda
6
MySQL InnoDB Cluster
Get your workstation ready
Deploy a cluster
Test H/A
Advanced examples
1
2
3
4
5

Program Agenda
7
Deploy a cluster
Test H/A
Advanced examples
1
2
3
4
5

MySQL Group Replication
InnoDB cluster

InnoDB cluster
App Servers with
MySQL Router

InnoDB cluster
App Servers with
MySQL Router
MySQL Shell
Setup, Manage,
Orchestrate

InnoDB cluster
App Servers with
MySQL Router
MySQL Shell
Setup, Manage,
Orchestrate Clients

Architecture

Program Agenda
14
Deploy a cluster
Test H/A
Advanced examples
1
2
3
4
5

 Access the cookbook:
 Start Firefox
 http://insidemysql.com/idc
 Start the Shell and a Terminal

 Deploy 3 MySQL Server instances:
mysql1:3306
mysql2:3306
mysql3:3306
Username: root
Password: root
 Using MySQL official Docker images. Available at:
https://hub.docker.com/r/mysql/mysql-server

 Check the MySQL servers status
 Start a terminal
 Execute:
$ docker ps -a

Program Agenda
18
Deploy a cluster
Test H/A
1
2
3
4

Deploy a cluster
Configure instances for InnoDB Cluster usage
mysql-js> dba.checkInstanceConfiguration(“root@mysql1:3306”)
1 - Verify instances configuration

Deploy a cluster
mysql-js> dba.configureInstance(“root@mysql1:3306”)
2 - Configure instances
Note: When prompted about instance restart answer No

Deploy a cluster
$ docker restart docker_mysql1_1
3 - Restart instances manually
Note: Switch to the terminal to execute the following commands:

Deploy a cluster
Create MySQL InnoDB Cluster
mysql-js> c root@mysql1:3306
4 - Create the cluster
4.1 - Connect to the seed instance
mysql-js> var cluster = dba.createCluster(“oow18”)
4.2 - Create the cluster

Deploy a cluster
5 - Check the cluster status
mysql-js> cluster.status()
5.1 - Status
mysql-js> cluster.description()
5.2 - Description

Deploy a cluster
mysql-js> cluster.addInstance(“root@mysql2:3306”)
6 - Add instances to the cluster
6.1 - Add mysql2
mysql-js> cluster.addInstance(“root@mysql3:3306”)
6.2 - ... and mysql3

Deploy a cluster
7 - Re-check the cluster status

Deploy a cluster
Deploy MySQL Router
1 - Bootstrap the Router
$ mysqlrouter --bootstrap root@mysql1:3306 –d router_oow18
Note: Switch to the terminal to execute the following command:

Deploy a cluster
Deploy MySQL Router
2 - Start the Router
$ ./router_oow18/start.sh
Note: Execute the following on the same terminal window

Program Agenda
28
Deploy a cluster
Test H/A
Advanced examples
1
2
3
4
5

Test H/A
Check query routing
Router listening on:
R/W connections: localhost:6446
R/O connections: localhost:6447

Test H/A
Check query routing
1 - Connect to your cluster via the Router
mysql-js> c root@localhost:6446
1.1 - Connect to the R/W Router port
mysql-js> var cluster = dba.getCluster()
1.2 - Get the cluster handle and check it status

Test H/A
Check query routing
2 - Test query routing
$ mysqlsh root@localhost:6446 –-sql -e “select @@hostname”
2.1 - Where are the R/W queries being redirected too?
Note: Check the hostnames mapping of docker
$ docker inspect <@@hostname> | grep compose.service

Test H/A
Check query routing
2 - Test query routing
2.2 - Where are the R/O being redirected too?
---> Repeat the previous commands, using now the R/O port: localhost:6447

Test H/A
Test failover
1 - Shutdown/kill the primary instance
mysql-js> sql
mysql-sql> SHUTDOWN;

Test H/A
Test failover
1 - Shutdown/kill the primary instance
Automatic failover takes place:
New primary automatic elected
Router automatically redirects R/W queries to new primary

Test H/A
Test failover
2 - Re-check the cluster status
mysql-sql> js
mysql-js> c root@localhost:6446

Test H/A
Test failover
3 - Restart killed instance
$ docker restart <@@hostname>
3.1 - Restart MySQL instance
3.2 – Re-check the cluster status

Program Agenda
37
Deploy a cluster
Test H/A
Advanced examples
1
2
3
4
5

Advanced examples
Quorum loss
Loss of quorum happens when:
 A majority of cluster members is lost
 Remaining members cannot achieve
consensus to take a decision

Test H/A
Quorum loss
1 - Simulate quorum loss scenario
$ docker kill <@@hostname_mysql1>
$ docker kill <@@hostname_mysql2>
1.1 – Kill MySQL instances

Test H/A
Quorum loss
2 – Restore from quorum loss scenario
$ docker restart <@@hostname_mysql1>
$ docker restart <@@hostname_mysql2>
2.1 – Restart MySQL instances
mysql-js> cluster.forceQuorumUsingPartitionOf(“root@mysql2:3306”)
2.2 – Restore quorum

Test H/A
Quorum loss
2 – Restore from quorum loss scenario
mysql-js> cluster.rejoinInstance(“root@mysql1:3306”)
mysql-js> cluster.rejoinInstance(“root@mysql3:3306”)
2.3 – Rejoin MySQL instances back to the cluster

Advanced examples
Complete outage
A complete outage happens when:
 All cluster members are lost
 Even if the members are restarted the cluster
ceased to exist

Test H/A
Complete outage
1 - Simulate complete outage scenario
$ docker kill $(docker ps -a –q)
1.1 – Kill all MySQL instances
$ docker restart $(docker ps -a –q)
1.2 – Restart all MySQL instances

Test H/A
Complete outage
2 – Restore from complete outage scenario
2.1 – Try to get the cluster handle
mysql-js> var cluster = dba.rebootClusterFromCompleteOutage()
2.2 – Restore cluster from complete outage

Test H/A
Complete outage
2 – Restore from complete outage scenario

Resources
• MySQL InnoDB Cluster Userguide:
– http://dev.mysql.com/doc/refman/en/mysql-innodb-cluster-userguide.html
• MySQL Shell Userguide:
– https://dev.mysql.com/doc/refman/en/mysql-shell.html
• APIs Reference manuals
– JavaScript: https://dev.mysql.com/doc/dev/mysqlsh-api-javascript/
– Python: https://dev.mysql.com/doc/dev/mysqlsh-api-python/
• Blogging
– http://mysqlserverteam.com/category/high-availability
46

MySQL 8 High Availability with InnoDB Clusters

MySQL 8 High Availability with InnoDB Clusters

More Related Content

What's hot (20)

Similar to MySQL 8 High Availability with InnoDB Clusters (20)

More from Miguel Araújo (13)

Recently uploaded (20)

MySQL 8 High Availability with InnoDB Clusters