Custom PlaybookΒΆ

Your customized mantl.yml playbook should be used to deploy Mantl to your servers.

Below is an annotated playbook explaining the values:

# CHECK SECURITY - when customizing you should leave this in. If you
# take it out and forget to specify security.yml, security could be turned off
# on components in your cluster!
- include: "{{ playbook_dir }}/playbooks/check-requirements.yml"

# BASICS - we need every node in the cluster to have common software running to
# increase stability and enable service discovery. You can look at the
# documentation for each of these components in their README file in the
# `roles/` directory, or by checking the online documentation at
- hosts: all
    # consul_acl_datacenter should be set to the datacenter you want to control
    # Consul ACLs. If you only have one datacenter, set it to that or remove
    # this variable.
    # consul_acl_datacenter: your_primary_datacenter

    consul_servers_group: role=control
    - common
    - certificates
    - lvm
    - docker
    - logrotate
    - consul-template
    - nginx
    - consul
    - etcd
    - dnsmasq

# ROLES - after provisioning the software that's common to all hosts, we do
# specialized hosts. There are 4 built-in roles: control, workers (mesos),
# kubeworkers (kubernetes), and edge (for external traffic into the cluster).

# The control nodes are necessarily more complex than the worker nodes, and have
# ZooKeeper, Mesos, and Marathon leaders as well as Kubernetes master
# components. In addition, they control Vault to manage secrets in the cluster.
# These servers do not run applications by themselves, they only schedule work.
# That said, there should be at least 3 of them (and always an odd number) so
# that ZooKeeper can get and keep a quorum.
- hosts: role=control
  gather_facts: yes
    consul_servers_group: role=control
    mesos_leaders_group: role=control
    mesos_mode: leader
    zookeeper_server_group: role=control
    - vault
    - zookeeper
    - mesos
    - calico
    - marathon
    - mantlui
    - kubernetes
    - kubernetes-master

# The worker role itself has a minimal configuration, as it's designed mainly to
# run software that the Mesos leader shedules. It also forwards traffic to
# globally known ports configured through Marathon.
- hosts: role=worker
  # role=worker hosts are a subset of "all". Since we already gathered facts on
  # all servers, we can skip it here to speed up the deployment.
  gather_facts: no
    mesos_mode: follower
    zookeeper_server_group: role=control
    - mesos
    - calico

# The kubeworker role is similar to the worker role but are intended for
# Kubernetes workloads.
- hosts: role=kubeworker
  gather_facts: yes
    - calico
    - kubernetes
    - kubernetes-node

# The edge role exists solely for routing traffic into the cluster. Firewall
# settings should be such that web traffic (ports 80 and 443) is exposed to the
# world.
- hosts: role=edge
  gather_facts: yes
    # this is the domain that traefik will match on to do host-based HTTP
    # routing. Set it to a domain you control and add a star domain to route
    # traffic. (EG *.marathon.localhost)
    # For those migrating from haproxy, this variable serves the same purpose
    # and format as `haproxy_domain`.
    traefik_marathon_domain: marathon.localhost
    - traefik

Run this playbook with ansible-playbook -i plugins/inventory/ -e @security.yml /path/to/your/playbook.yml. It will take a while for everything to come up as fresh machines will have to download quite a few dependencies.