This MR will fix failing deployments via phabalicious which output only an error like Error 137. This could happen when k8s does a rollout, it keeps the old containers a bit around, while starting the new containers. Phabalicious might then pick up an old container, starting e.g. a drupal reset and k8s will kill the obsolete container in the meantime.
The proper fix is to get the latest revision from the deployment and then the pod-template hash from the replicaset with that revision. This will be added to the existing pod-selector and used to find the correct pod.
This MR will fix failing deployments via phabalicious which output only an error like Error 137. This could happen when k8s does a rollout, it keeps the old containers a bit around, while starting the new containers. Phabalicious might then pick up an old container, starting e.g. a drupal reset and k8s will kill the obsolete container in the meantime.
The proper fix is to get the latest revision from the deployment and then the pod-template hash from the replicaset with that revision. This will be added to the existing pod-selector and used to find the correct pod.
Please review!