This story focuses on thoroughly testing Podman's behavior after a host reboot, specifically regarding its database refresh mechanism. When a host running Podman is rebooted, Podman is expected to clear any stale state from the previous boot and reinitialize its database, reflecting the current state of containers, images, and volumes.

The goal is to design and execute test cases that simulate various reboot scenarios (e.g., graceful shutdown, unexpected power loss) and then verify that Podman correctly identifies and cleans up residual data, ensuring a clean and consistent state upon restart. This includes checking for:

• Removal of containers that were running before the reboot but are not configured for auto-start.

• Correct status of containers configured for auto-start.

• Integrity of Podman's internal database.

• Absence of orphaned resources (e.g., network interfaces, volumes not associated with active containers).

Acceptance Criteria

• AC1: The test environment is configured to allow for simulated host reboots (e.g., using virtual machines, reboot command, or similar tools).

• AC2: A test case is executed where Podman containers are running (some configured for auto-start, some not) before a simulated graceful host reboot.

• AC3: After the graceful reboot, Podman's database is verified to be in a consistent state, with non-auto-starting containers correctly removed and auto-starting containers in their expected state.

• AC4: A test case is executed where Podman containers are running before a simulated unexpected host power loss (e.g., abrupt VM shutdown).

• AC5: After the unexpected reboot, Podman's database is verified to be in a consistent state, with non-auto-starting containers correctly removed and auto-starting containers in their expected state, demonstrating resilience to crashes.

• AC6: Verification steps include checking Podman logs for any errors or warnings related to database refresh or state cleanup during and after reboot.

• AC7: A method is established to confirm that no orphaned Podman-related resources (e.g., podman ps -a, podman images, podman volume ls, ip a, brctl show) persist from the previous boot that should have been wiped.