Ansible

Ansible is installed by default on all INFRA nodes and can be used to manage the entire deployment.

Pigsty implements automation based on Ansible, following the Infrastructure-as-Code philosophy.

Ansible knowledge is useful for managing databases and infrastructure, but not required. You only need to know how to execute Playbooks - YAML files that define a series of automated tasks.

Installation

Pigsty automatically installs ansible and its dependencies during the bootstrap process. For manual installation, use the following commands:

# Debian / Ubuntu
sudo apt install -y ansible python3-jmespath

# EL 10
sudo dnf install -y ansible python-jmespath

# EL 8/9
sudo dnf install -y ansible python3.12-jmespath

# EL 7
sudo yum install -y ansible python-jmespath

macOS

macOS users can install using Homebrew:

brew install ansible
pip3 install jmespath

Basic Usage

To run a playbook, simply execute ./path/to/playbook.yml. Here are the most commonly used Ansible command-line parameters:

Limiting Hosts

Use -l|--limit <pattern> to limit execution to specific groups, hosts, or patterns:

./node.yml                      # Execute on all nodes
./pgsql.yml -l pg-test          # Only execute on pg-test cluster
./pgsql.yml -l pg-*             # Execute on all clusters starting with pg-
./pgsql.yml -l 10.10.10.10      # Only execute on specific IP host

Running playbooks without host limits can be very dangerous! By default, most playbooks execute on all hosts. Use with caution!

Limiting Tasks

Use -t|--tags <tags> to only execute task subsets with specified tags:

./infra.yml -t repo           # Only execute tasks to create local repo
./infra.yml -t repo_upstream  # Only execute tasks to add upstream repos
./node.yml -t node_pkg        # Only execute tasks to install node packages
./pgsql.yml -t pg_hba         # Only execute tasks to render pg_hba.conf

Passing Variables

Use -e|--extra-vars <key=value> to override variables at runtime:

./pgsql.yml -e pg_clean=true         # Force clean existing PG instances
./pgsql-rm.yml -e pg_rm_pkg=false    # Keep packages when uninstalling
./node.yml -e '{"node_tune":"tiny"}' # Pass variables in JSON format
./pgsql.yml -e @/path/to/config.yml  # Load variables from YAML file

Specifying Inventory

By default, Ansible uses pigsty.yml in the current directory as the inventory. Use -i|--inventory <path> to specify a different config file:

./pgsql.yml -i files/pigsty/full.yml -l pg-test

[!NOTE]

To permanently change the default config file path, modify the inventory parameter in ansible.cfg.