torpor-playbooks

Ansible playbooks that complement torpor (a multi-distro test-VM harness). Designed to run from Ascender / AWX against host inventories that torpor populates.

Content here stays focused on the Ansible side; the torpor CLI and matrix tooling live in the main torpor repo.

Playbooks

gather_facts_ephemeral.yml

Same behaviour as the standard gather_facts.yml from ctrliq/ascender-playbooks, but with ignore_unreachable: true at the play level.

Intended for host groups that are expected to be off most of the time — test VMs, game servers, and similar ephemeral fleets. Hosts that happen to be awake still contribute their package and service facts; hosts that are asleep don't mark the job as failed. This keeps the run status honest: "failed" means "something went wrong," not "half the fleet is doing exactly what it's supposed to be doing."

linux-weekly-update.yml

Package-manager-abstracted weekly update cycle for a multi-distro fleet. Targets the linux_test_vms group (override with --limit) and:

1. Starts every VM in the group via qm start on the PVE host.
2. Runs the distro's native package manager (dnf / apt / zypper) to update all packages.
3. Reboots if a new kernel landed (RedHat: needs-restarting -r, Debian: /run/reboot-required, Suse: zypper ps -s).
4. Emits a per-host JSON report.
5. Shuts everything back down.

Safe to run repeatedly. Schedule on Ascender / AWX / a cron.

inventory/example.yml

Starter Ansible inventory showing the group structure torpor expects (linux_test_vms parent, per-vendor child groups). Copy and populate with your real hosts.

Usage

Point an Ascender / AWX Project at this repo (public HTTPS clone, no credentials needed), then create a Job Template that references the desired playbook.

Or run directly with ansible-playbook:

ansible-playbook ~/torpor-playbooks/linux-weekly-update.yml \
  -i /path/to/your/inventory.yml --limit rocky9 --check