Ansible roles to install and manage pool of Indy Node

Quickstart

• Make sure you have AWS CLI installed and configured.
• Install requirements
• Run ansible-playbook pool_create.yml - this will create 4 EC2 instances and test_nodes directory with inventory and SSH keys.
• Run ansible-playbook -i test_nodes/hosts pool.yml - this will install and configure Indy Node pool on previously created EC2 instances.
• Run ssh -F test_nodes/ssh_config test_node_1 to login to first node and take a look around.
• Run ansible-playbook destroy.nodes.yml - this will terminate previously created AWS EC2 instances and clear config_pool and inventory_pool files.

Installation

• python 2.7 or 3.6
• required packages: pip install -r requirements.txt, options:
  • (recommended) inside python virtual environment
  • with --user inside user Python install directory
  • (not recommended) globally using sudo

Roles

aws_manage

Used to manage number of AWS instances.

Parameters:

• instance_count: number of instances in pool (provide 0 to destroy)
• aws_ec2_type (t2.micro): type of instances
• aws_region (eu-central-1): region of instances
• aws_tag_project (PoolAutomation): project name for created instances
• aws_tag_namespace (test): namespace of created instances
• aws_tag_group (default): ansible inventory group of created instances

Todos:

• extract key generation and inventory export to separate role
• make inventory span separate roles in namespace
• more tests

ansible_bootstrap

Installs python and sudo.

indy_node

Adds sovrin repository and installs and configures Indy Node.

Parameters:

• indy_node_channel (master): which release channel to use (master/rc/stable)
• indy_node_ver
• indy_plenum_ver

Todos:

• allow providing only indy_node_ver
• remove unused repositories when switching channels

Scripts

The directory scripts includes helper scripts. Please refer to scripts/README.md for more details.

Development

Install system packages:

• vagrant
• python 2.7 or 3.6, including dev package

Install python packages:

pip install -r requirements-dev.txt

Default development workflow would be:

• molecule lint
• molecule converge
• molecule verify
• molecule destroy

When you are ready you can run aggregative command molecule test.

Scenarios

By default scenarios based on docker are used. Also vagrant scenarios are available and might be run like molecule <command> -s vagrant.

In order to test all scenarios for some role go to a directory with the role (for example roles/indy_node) and run molecule test --all.

Vagrant scenarios specific

To speed up development and testing on vagrant VMs following workflow is recommended:

• After each change run molecule lint. This will quickly show some style recommendations and probably highlight some stupid mistakes.
• When lint is satisfied run molecule converge. This will spin up virtual machines (if neccessary) and run default playbook. This could be slow operation when running for the first time, but subsequent runs are much faster.
• When converge finishes without errors run molecule verify. This will start tests.
• Do more changes, running this lint-converge-verify sequence.
• When done (or there is suspicion that VM state gone very bad) run molecule destroy, this will destroy virtual machines.
• When virtual machines are running it's possible to login to them using molecule login --host name