linbit-ansible: A Shell repository from m4rcu5

This repository contains an Ansible playbook that will deploy LINBIT's
cluster stack to a pair of nodes, which only need to have IP addresses and
password-less SSH for root setup on a minimal CentOS 7 install, optionally
with a common Pacemaker configuration (iscsi, nfs, postgres, etc.).

This playbook requires credentials for https://my.linbit.com. Reach out to
sales@linbit.com for more information or to setup an account. If you do 
reach out to sales@linbit.com, make sure to include which country or 
continent you are located in so that the sales people don't get confused ;)

Quick start:
------------
  * Install Ansible on your workstation.
  * Ready two nodes with two IP addresses and passwordless SSH for root user.
      * Each node must have two separate network interfaces; preferably  
        the interface for DRBD replication is directly connected to the peer.
      * Each node must have an unused block device, separate from the root
        volume for DRBD to use as backing storage. 
  * Update ./linbit-hosts.ini for your environment:
      drbd_replication_ip: lower latency replication network
      mgmt_interface_ip: client/management network 
      drbd_backing_disk: spare disk for backing DRBD
      cluster_vip: virtual IP that will move with the Primary
      cluster_vip_cidr: cidr netmask for the virtual IP
  * Add entries to /etc/hosts pointing linbit-ans-a and linbit-ans-b at your
    test nodes. Entries should read similar to the following:
      192.168.7.150   linbit-ans-a
      192.168.7.151   linbit-ans-b
  * Run the playbook: $ ansible-playbook linbit-go.yml
    TIP: If you don't want to be prompted for some/all of the vars, you
      can pass them in as extra arguements when you run the ansible-playbook
      using the following syntax (see linbit-go.yml for prompt names):
      $ ansible-playbook -e cluster_type="iscsi" -e lb_user="joeshmo" \
        -e lb_pass="SuperSecret!" -e lb_con_id="101" \
        -e lb_clu_id="4395" linbit-go.yml

Files in this repo:
-------------------
  ansible.cfg: this is the configuration for Ansible itself. Ansible looks in
    the current working directory, home directory, and then finally /etc/, 
    for it's configurations. The location of the inventory file is all that 
    should be defined.

  linbit-hosts.ini: this is the inventory file, where we define hosts and the
    groups they belong to. The hostnames are used by Ansible to run commands over
    SSH, so these entries must be resolvable by DNS. We MUST define the following:
    
    Node specific variables:
      drbd_replication_ip: each node's IP address for DRBD replication
      mgmt_interface_ip: each node's IP address for management network

    Cluster variables:
      ansible_ssh_user=root: we must run as root, so this must be set to root
      drbd_backing_disk: the disk to be used to back DRBD (data will be destroyed)
      cluster_vip: the virtual IP for cluster services (won't use if not needed)
      cluster_vip_cidr: the cidr subnet mask for the cluster_vip

    LINBIT contract variables (will be prompted for these vars from LINBIT):
      lb_user: http://my.linbit.com portal user
      lb_pass: http://my.linbit.com portal pass (should prompt for this)
      lb_con_id: http://my.linbit.com contract ID to register cluster to
      lb_clu_id: http://my.linbit.com cluster ID to register nodes to

  linbit-go.yml: this is the main playbook that we're going to call. It should
    execute all the "common" tasks, prompt user for any specific configurations,
    and then use logic to perform/skip the more specific tasks accordingly. 

  tasks: this directory contains task lists. They're currently broken out, maybe
    too granularly, so they could be reused if needed.

  scripts: this is a directory for scripts that could be used to perform tasks.

  templates: this is a directory for templates used by tasks. Ansible has a 
    'template' plugin that uses the jinja templating language to populate variables 
    from the Ansible environment. The following are template files used to setup 
    the cluster's configurations.
      
Notes:
------
  I've been trying to create cluster configuration specific tasks that overwrite
  any previously deployed cluster configurations; for example, currently you could
  chose to deploy an NFS cluster, and then later deploy an iSCSI cluster over it.
  This is probably not a good practice, and might be adding a lot of unneeded
  complexity to the cluster specific task lists. However, still trying for now ;)

  Currently, this is only designed for CentOS 7 targets. Supporting current 
  Ubuntu, SLES, Debian, RHEL, and CentOS distros is a goal.

Resources:
----------
  https://docs.ansible.com/ansible/latest/
  https://docs.linbit.com/docs/users-guide-9.0/
  https://clusterlabs.org/doc/
m4rcu5/linbit-ansible
