kamigerami / ecosystem

Different projects using technologies such as docker swarm, consul, kubernetes, rancher etc

EcoSystem

Vagrant - Docker Swarm (1.10) + Consul

In swarm_1.10_consul directory you can find the following "ecosystem architecture" :

A Consul, Dnsmasq, Docker Swarm, Docker compose, Artifactory ecosystem simulating a real world production setup

docker-compose - Rancher + External MYSQL db

In docker_compose_rancher_external_mysql directory you can find the following "ecosystem architechture":

A quick single-host Rancher-server with external mysql database ecosystem using docker-compose to bring everything up

Vagrant - Kubernetes + Controller + Worker

In vagrant_kubernetes_controller_worker directory you can find the following "ecosystem architechture":

A Vagrant ansible provisioned kubernetes cluster - 1 controller node with etcd, kube-controller, kube-apiserver, kube-scheduler and 2 worker nodes with kubectl

ecosystem's Issues

Add Swarm Manager to dockerswarm and agents to dockerhosts

Configure and add the swarm nodes + manager roles.

notify user of the different urls they can log in to after all plays are completed

Add a role to run at the end of everything which notifys all the urls that are available to visit. For example consul and shipyard web interface.

Add travis.CI build testing

Test travis CI build

Ansible fails downloading consul from the hashicorp path (wrong path)

Update using the correct path to download consul and update to latest (0.6.4)

Update to support ansible 2.x

Update the tasks and vagrantfile to add support for Ansible 2.0.
Changes needed are the "sudo: yes" which is being deprecated and need to be switched to "become: yes, become_method: sudo" instead.

Add registry or artifactory as a Docker registry

Swarm nodes don't get added to the cluster

The swarm nodes are not getting added to the cluster due to wrong ports in the configuration being exposed.

Update .travis-CI to run actual playbooks on CentOs

Since Travis don't support CentOs out of the box we will have to use docker containers to run our tests.
Create support for it.

make sure all hosts run Docker-engine and install swarm agents

Add Compose role to hosts running docker engine

Install Docker Compose on all hosts running the docker engine.

Vars used in template file are not using ansible_vars

Refactor the code to use ansible_vars in the template files for each host.

Run both Docker and Ansible provisioner dynamically

Currently a user can't provision using both ansible and docker dynamically.

Add shipyard role with rethinkdb and shipyard controller

Setup a role for shipyard which install rethinkdb as well as the shipyard controller and links to the db, swarm manager and discovery services.

Add a second swarm manager as a replica

Add second swarm manager as a replica for high availability

Update to support ansible 2.x

Update the tasks and vagrantfile to add support for Ansible 2.0.
Changes needed are the "sudo: yes" which is being deprecated and need to be switched to "become: yes, become_method: sudo" instead.

Add TLS for communication

Add support for container metrics / Data / monitoring / alerts

using for example prometheus, cadvisor, sensu or something similar

Update to use docker overlay network

Update from using --links to docker overlay network.

Create ansible_vars in vagrantfile and remove host_vars from ansible

I want the vagrant file to create the ansible_vars and not use the host_vars.