Jobs remain Running in Ansible Automation Platform after execution node reboot

Issue

• After an execution node reboot, Automation Controller jobs remain in running status with no finish time. Lab example (job id 12):

  "status": "running",
  "finished": null,
  "execution_node": "10.0.x.x",
  "work_unit_id": "100108119KoOuCIUb",
  "job_explanation": ""
  

  Note: /api/controller/v2/jobs/<job-id>/

• The execution node returns to ready and continues to report consumed capacity for the stuck job. Lab example:

  "node_type": "execution",
  "node_state": "ready",
  "jobs_running": 1,
  "consumed_capacity": 2,
  "errors": ""
  

  Note: /api/controller/v2/instances/?hostname=<execution-node>

• On the execution node, no ansible-runner / command-runner worker process remains for the job (ps shows no matches), but Receptor still stores the work unit as Running with a stale PID. Lab example from the Receptor status file:

  {"State":1,"Detail":"Running: PID 23","StdoutSize":4733,"WorkType":"ansible-runner","ExtraData":{"Pid":12,"Params":"worker --private-data-dir=/home/cloud-user/aap/controller/data/job_execution/awx_12_dg33gsth --delete"}}
  

  Note: /var/lib/receptor/<execution-node>/<work-unit-id>/status

• Jobs do not automatically fail, time out, or get reaped after the reboot. Worker capacity stays consumed until an administrator cancels the job and/or cleans up job execution data on the node.