sdhardy/klustr

Ansible project to create test clusters

README

Table of contents

• Introduction
• Creating your Lab
• Additional playbooks

Introduction

The playbooks are meant to create minimal clustered environments based on LXC for learning purposes.

Clustered technologies to be looked into initially include:

• Rabbit MQ
• Galera

Only the rabbit MQ, portion is implemented at this time.

Creating your Lab

1. Generate SSH key for the deployment host and add the key to known_hosts and authorized_keys

~# ssh-keygen -t rsa -b 4096 -f ~/.ssh/id_rsa -N ""
~# cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys
~# ssh-keyscan -H 192.168.1.112 >> ~/ssh/known_hosts

2. Clone the klustr repository to your deployment

~# git clone https://github.com/OchiengEd/klustr /opt/klustr

3. Create inventory file inside inventory directory, e.g.

~# cd /opt/klustr/playbooks
/opt/klustr/playbooks# vim inventory/hosts
[lab]
192.168.1.112

4. Choose which clusters you would like to have setup

/opt/klustr/playbooks# cat lab_overrides.yml 

install_rabbit_mq: true

install_galera: true

5. Run playbooks

/opt/klustr/playbooks# ansible-playbook main.yml 

6. Once the playbooks are done, ensure the desired clusters are up and running

/opt/klustr/playbooks# ansible galera -m shell -a "mysql -e \"show status like 'wsrep_cluster_%'\""
node1_galera_container | SUCCESS | rc=0 >>        
Variable_name   Value                               
wsrep_cluster_conf_id   2
wsrep_cluster_size      3
wsrep_cluster_state_uuid        85563195-9a43-11e8-9a09-7b7fb8465a9c
wsrep_cluster_status    Primary   

node2_galera_container | SUCCESS | rc=0 >>
Variable_name   Value
wsrep_cluster_conf_id   2
wsrep_cluster_size      3
wsrep_cluster_state_uuid        85563195-9a43-11e8-9a09-7b7fb8465a9c
wsrep_cluster_status    Primary

node3_galera_container | SUCCESS | rc=0 >>
Variable_name   Value
wsrep_cluster_conf_id   2
wsrep_cluster_size      3
wsrep_cluster_state_uuid        85563195-9a43-11e8-9a09-7b7fb8465a9c
wsrep_cluster_status    Primary

/opt/klustr/playbooks# ansible rabbit -m shell -a 'rabbitmqctl cluster_status'
node2_rabbit_mq_container | SUCCESS | rc=0 >>
Cluster status of node rabbit@node2_rabbit_mq_container ...
[{nodes,[{disc,[rabbit@node1_rabbit_mq_container,
                rabbit@node2_rabbit_mq_container,
                rabbit@node3_rabbit_mq_container]}]},
 {running_nodes,[rabbit@node3_rabbit_mq_container,
                 rabbit@node1_rabbit_mq_container,
                 rabbit@node2_rabbit_mq_container]},
 {partitions,[]}]
...done.

node1_rabbit_mq_container | SUCCESS | rc=0 >>
Cluster status of node rabbit@node1_rabbit_mq_container ...
[{nodes,[{disc,[rabbit@node1_rabbit_mq_container,
                rabbit@node2_rabbit_mq_container,
                rabbit@node3_rabbit_mq_container]}]},
 {running_nodes,[rabbit@node3_rabbit_mq_container,
                 rabbit@node2_rabbit_mq_container,
                 rabbit@node1_rabbit_mq_container]},
 {partitions,[]}]
...done.

node3_rabbit_mq_container | SUCCESS | rc=0 >>
Cluster status of node rabbit@node3_rabbit_mq_container ...
[{nodes,[{disc,[rabbit@node1_rabbit_mq_container,
                rabbit@node2_rabbit_mq_container,
                rabbit@node3_rabbit_mq_container]}]},
 {running_nodes,[rabbit@node1_rabbit_mq_container,
                 rabbit@node2_rabbit_mq_container,
                 rabbit@node3_rabbit_mq_container]},
 {partitions,[]}]
...done.

/opt/klustr/playbooks#

Additional Playbooks

The rabbitmq-breakfix.yml playbook has two tags - dummy-data and breakfix-1. The former will inject dummy data into message queues while the latter will break the cluster for you to fix.

/opt/klustr/playbooks# ansible-playbook rabbitmq-breakfix.yml --list-tags

playbook: rabbitmq-breakfix.yml

  play #1 (rabbit): Rabbit MQ breakfix  TAGS: []
      TASK TAGS: [breakfix-1, dummy-data]
/opt/klustr/playbooks#