Nomad-TuringpPi

This repository is related to the talk I made for the HashiTalks DACH day in 2020 https://events.hashicorp.com/hashitalksdach The video in German can be found here https://www.youtube.com/watch?v=ppDfZLaFoSE at 8:37:20

The slides can be found here

Architecture

Hardware

The infrastructure is made up of

1 [Turing PI v1] https://turingpi.com/v1/ and 7 [RapberryPi 3+ compute modules]https://www.raspberrypi.org/products/compute-module-3/?resellerType=home
1 Vagrant Cluster running on an Ubuntu Laptop.

Software

Nomad Enterprise 1.0-beta3+ent
Consul 1.9
Ansible
Vagrant
Terraform

Preparing the Infrastructure

mkdir tpi-demo
cd tpi-demo
git clone https://github.com/lhaig/nomad-turingpi.git

Deploy the Vagrant cluster using the Vagrantfile provided.

cd vagrant
vagrant up

Install and configure your TuringPi cluster with Ubuntu 18.04 I added the node addresses to my local DNS environment.

Deploy Nomad and Consul to the Servers and Clients

Open the hosts.ini file in your editor and make changes to the file with the details of your Vagrant and TuringPi servers and Hosts.

Once updated make sure you can SSH into each of the servers without using usernames and passwords. Then run the playbook below to install Nomad and Consul on your servers.

cd ansible-turing-pi
ansible-playbook -i hosts.ini turingpi.yml

Run the jobs in this order

Traefik Proxy

Run the terraform to deploy the traefik proxy.

Copy the terraform.tfvars.example file and rename it to terraform.tfvars Edit the terraform.tfvars file and update the nomad_addr variable with the URL or IP address of your nomad cluster. Deploy traefik with the commad below

cd ../jobs/terraform/traefik_proxy
terraform plan
terraform apply -auto-approve

You should now see the traefik proxy in your server console.

Prometheus

cd ../nomad
nomad job run prometheus.nomad

Grafana

cd ../nomad
nomad job run grafana.nomad

Deploy Grafana Dashboards.

Run the terraform to deploy the Grafana dashboards.

Copy the terraform.tfvars.example file and rename it to terraform.tfvars Edit the terraform.tfvars file and update the 2 variables with the URL or IP address of your deployed servers. Deploy the dashboards with the commad below

cd ../terraform/grafana_dashboards
terraform plan
terraform apply -auto-approve

Login to Grafana

You should now be able to login to Grafana using

http://{SERVERADDERSS}:3000

using the username admin and the password admin

Multi cluster federation.

Federate the 2 clusters with the following command

nomad server join {NAME-OR-IP-OF-VAGRANT-SERVER}:4648

You should get a federated message Then you can yse the redis-mr.nomad job to deploy across both clusters.