Use PostgreSQL as Ansible Inventory CMDB

Pigsty v2.7 normally reads its Ansible inventory from pigsty.yml. The release also ships three scripts that can load that inventory into the pigsty schema of the meta database and switch Ansible to a generated dynamic inventory.

The implementation in this page is tied to the v2.7.0 objects bin/inventory_load, bin/inventory_cmdb, and bin/inventory_conf.

Prerequisites

The CMDB schema is created from files/cmdb.sql when the default pg-meta cluster is provisioned. Complete infra.yml and the pg-meta database initialization first, then verify that the meta service is reachable.

By default, ansible.cfg contains:

[defaults]
inventory = pigsty.yml

Load the YAML inventory

bin/inventory_load reads a YAML inventory and replaces the rows in pigsty.group, pigsty.host, pigsty.group_var, pigsty.host_var, and pigsty.global_var. Back up or review the target CMDB before loading a new inventory.

bin/inventory_load                                      # use $PIGSTY_HOME/pigsty.yml and $METADB_URL
bin/inventory_load -p files/pigsty/demo.yml             # load another v2.7 inventory
bin/inventory_load -p pigsty.yml \
  -d postgresql://dbuser_meta:[email protected]:5432/meta

The script accepts -p|--path for the YAML file and -d|--data for the PostgreSQL connection string. Its defaults come from PIGSTY_HOME and METADB_URL.

Switch inventory source

Switch Ansible to the CMDB-backed inventory.sh:

bin/inventory_cmdb
ansible-inventory --list

bin/inventory_cmdb writes inventory.sh, makes it executable, and changes the inventory entry in ansible.cfg. The generated script queries pigsty.inventory through the meta PostgreSQL service.

Switch back to the static pigsty.yml inventory with:

bin/inventory_conf
ansible-inventory --list

These scripts only change the inventory source. Continue to edit and validate the source configuration deliberately; loading a new YAML file is not an automatic two-way synchronization mechanism.