Skip to content

tower/job-poll-resume

Operator opt-in to resume tower-invoke-curl polling on an existing job ID

When set to true, instructs the tower-invoke-curl task to skip the initial launch POST and re-enter the polling loop using the existing tower/job-id already recorded on the machine.

Use this after a max-time-expired termination if the Tower job is expected to still be running on the AWX / Tower server and you want to continue waiting for it rather than launching a new one.

Preconditions checked at task start:

  • tower/job-poll-state must equal max-time-expired
  • tower/job-id must be set and non-empty
  • inventory/tower/id must be set on the machine (captured from the launch response by tower/output-to-inventory)

The inventory/tower/id check ensures the launch-response audit trail is still available before re-entering the polling loop. If an operator has removed the launch-response inventory params, resume fails with an error directing them to either repopulate inventory/tower/id or start fresh by clearing state.

If any precondition fails, the task exits 1 with an error message describing the required state.

This parameter is always cleared on any terminal exit (success, failure, or timeout). Each resume is therefore a single explicit operator action.

To start a fresh Tower job instead of resuming, clear tower/job-poll-state and tower/job-id before re-triggering the stage. See tower/job-poll-state for the full workflow.

Objects that Reference tower/job-poll-resume

Objects used by tower/job-poll-resume

Not used in current content packs