#####The playbook install MySQL instances, set up replication and configure MHA Currently it supports the following:
- Linux user management
- Automated master-slave replication setup. You can add new slaves to a cluster online(xtrabackup stream)
- Semi-automated failover support with MHA
- my.cnf file creation based on available ram with templates
- custom my.cnf setup
- MySQL user management
###How to install
Requirements:
- ssh client
- ansible
- vagrant
- virtualbox
add your credentials to the encrypted system.yml (default password: 42)
ansible-vault edit group_vars/system.yml
you can change the password with
ansible-vault rekey group_vars/system.yml
###Spin up instances Spin up 4 instances, we will use them as:
- 1 master
- 2 slaves
- 1 MHA manager node
Set up ssh passwordless authentication from the host you are running ansible
Vagrant box: chef/centos-6.5
Can use:
vagrant up
from this directory to spawn 4 servers.
See setup/README.md for details.
Run the playbook from setup to install ssh keys and EPEL repo.
ansible-playbook -c paramiko -i setup/hosts setup/setup.yml --ask-pass --sudo
Define at least these tags:
- Name = anything
Edit the sample_cluster inventory file and change the ip addresses
~/github/ansible_mha$ cat clusters/sample_cluster
#ROLES
======
[system]
52.28.26.105 mycnf=plsc_s1 mysql_role=master nickname=dev-plsc-m
52.28.23.4 mycnf=plsc_s2 mysql_role=slave nickname=dev-plsc-s1
52.28.19.44 mysql_role=slave nickname=dev-plsc-s2
52.28.3.145 mysql_role=mha_manager
[system:vars]
cluster_id=dev-plscdemo
[databases:vars]
master_host=52.28.26.105
# Group to install and configure MySQL
[databases]
52.28.26.105 mysql_role=master server_id=1
52.28.23.4 mysql_role=slave server_id=2 backup_src=52.28.26.105
52.28.19.44 mysql_role=slave server_id=3 backup_src=52.28.26.105
run the playbook
ansible-playbook -i clusters/sample_cluster setup.yml -e bootstrap_enabled=true --ask-vault-pass#####All servers has read_only=1 set by default!!!
###MHA examples run these command on the ansible host: #####Connectivity test
ansible -i clusters/sample_cluster system --limit [ip_of_mha_manager_node] -a "masterha_check_ssh --conf=/etc/mha_dev-plscdemo.cnf"
#####Replication test
ansible -i clusters/sample_cluster system --limit [ip_of_mha_manager_node] -a "masterha_check_repl --conf=/etc/mha_dev-plscdemo.cnf"
#####Online failover You can see the current topology in the output of the replication test.
ansible -i clusters/sample_cluster system --limit [ip_of_mha_manager_node] -a "masterha_master_switch --master_state=alive --orig_master_is_new_slave --conf=/etc/mha_dev-plscdemo.cnf --interactive=0 --new_master_host=[ip address of the chosen slave ]"#####Failover with a dead master
ansible -i clusters/sample_cluster system --limit [ip_of_mha_manager_node] -a "masterha_master_switch --master_state=dead --conf=/etc/mha_dev-plscdemo.cnf --dead_master_host=[ip of the dead master] --interactive=0 --remove_dead_master_conf"
###TAGS #####'users' - User management If you want to add a new Linux user to the servers simply copy its public key to roles/common/files/users/ - filename will be the username. Once the file is deployed the users can connect to the servers with their own username and can sudo. If you want to delete a user, simply remove the file and run the playbook again. Run the following to apply the changes:
ansible-playbook -i clusters/sample_cluster setup.yml --tags users