cdelgehier.XtraDB-Cluster

Ansible Role: `ansible-role-XtraDB-Cluster`

This Ansible role helps you set up a Percona XtraDB Cluster. This role specifically handles:

Installing the necessary packages
Securing connections
Bootstrapping the cluster

Requirements

No special requirements are needed.

Role Variables

Basics

Variable	Default	Comments
`xtradb_bind_address`	-	The IP address to listen on
`xtradb_bind_interface`	`eth0`	The network interface for the cluster
`xtradb_cluster_name`	`foo`	Name of the cluster
`xtradb_configured`	`xtradb_datadir/configured`	Used for idempotency
`xtradb_datadir`	`/var/lib/mysql`	Data directory
`xtradb_master_node`	`groups[xtradb_nodes_group][0]`	The node designated as master
`xtradb_mysql_user`	`mysql`	User to run Galera
`xtradb_nodes_group`	`xtradb-cluster-nodes`	Group of nodes in the cluster
`xtradb_root_password`	`root`	Password for the root user
`xtradb_root_user`	`root`	The root user
`xtradb_secured`	`xtradb_datadir/secured`	Used for idempotency
`xtradb_service`	`mysql`	Linux service name
`xtradb_sst_password`	`sstpassword`	Password for the `xtradb_sst_user`
`xtradb_sst_user`	`sstuser`	User for state snapshot transfer
`xtradb_swappiness`	`0`	The swappiness value
`xtradb_databases`	[]	List of database names to add
`xtradb_users`	[]	List of users to add
`xtradb_version`	`57`	Version of XtraDB

MySQL Parameters

For additional details, refer to the MariaDB Server System Variables.

Variable	Default	Comments
`xtradb_binlog_format`	`ROW`	Format for binary logging
`xtradb_character_set_server`	`utf`	Character set used
`xtradb_collation_server`	`utf8_general_ci`	Collation
`xtradb_default_storage_engine`	`InnoDB`	Default storage engine setting
`xtradb_innodb_autoinc_lock_mode`	`2`	Lock modes for auto-increment
`xtradb_innodb_buffer_pool_instances`		Number of buffer pool instances
`xtradb_innodb_buffer_pool_size`		Size of the InnoDB buffer pool
`xtradb_innodb_file_per_table`	`on`	Enables separate tablespace for each table
`xtradb_innodb_flush_log_at_trx_commit`		Determines logging behavior during commits
`xtradb_innodb_log_buffer_size`		Size of the log buffer
`xtradb_max_connections`	`4096`	Maximum number of connections
`xtradb_slow_query_log`	`0`	Set to 1 to enable slow query logging
`xtradb_skip_name_resolve`	`1`	Only use IP addresses, set to 0 for hostnames

Adding Databases

Databases are defined with a dictionary containing name: (required) and init_script: (optional). The init script initializes the database with tables and values.

xtradb_databases:
  - name: keystone
  - name: mydb
    init_script: files/init_mydb.sql

Adding Users

Users are defined with a dictionary containing name:, password:, priv:, and optionally host:.

xtradb_users:
  - name: keystone
    password: KEYSTONE_DBPASS
    priv: 'keystone.*:SUPER'
  - name: cdelgehier
    password: yolo
    priv: 'mydb.*:ALL'
    host: '192.168.1.%'

Dependencies

There are no dependencies required.

Example Playbook

- hosts: db
  gather_facts: true
  become: true
  roles:
    - role: ansible-role-XtraDB-Cluster
      xtradb_cluster_name: "prod-customer"
      xtradb_sst_user: sstuser
      xtradb_sst_password: s3cr3t
      xtradb_root_password: yolo
      xtradb_bind_interface: eth0

- hosts: db
  gather_facts: true
  become: true
  roles:
    - role: ansible-role-XtraDB-Cluster
      xtradb_cluster_name: "prod-customer"
      xtradb_bind_address: "{{ ansible_default_ipv4.address }}"
      xtradb_master_node: "{{ hostvars[ groups['db'][0] ].ansible_default_ipv4.address }}"

[db]
node1 ansible_host=192.168.1.173
node2 ansible_host=192.168.1.156
node3 ansible_host=192.168.1.154

Testing

You can test using two environments: one with Vagrant and the other with Docker, suitable for automated tests. Check the README files in the corresponding branches for setup instructions.

Docker: docker-tests

Remove

To start fresh and remove all traces:

ansible db -m shell -a 'rm -rf /var/lib/mysql /var/log/mysqld.log /etc/percona-xtradb-cluster.conf.d ; yum remove Percona* -y'

Contributing

Feel free to report issues or suggest features in the Issues section. Pull requests are welcomed. To submit a PR, fork the project and create a topic branch for your change, then push to your fork and create a PR on GitHub.