Deployment can gets stuck after recreating

After running this little script the output will be following:

--> running test #1 ...
deploymentconfig "test" deleted
deploymentconfig "test" created
--> status of #1 is Pending ...
--> status of #1 is Running ...
--> running test #2 ...
deploymentconfig "test" deleted
pod "test-1-deploy" deleted
pod "test-1-s7vq4" deleted
deploymentconfig "test" created
--> status of #2 is Running ...
--> running test #3 ...
deploymentconfig "test" deleted
pod "test-1-deploy" deleted
deploymentconfig "test" created
--> status of #3 is Failed ...

What I think happen is that the "test-1-deploy" pod is in Terminating status (or it is being terminated by GC), however it is not fast enough and that means the next DC will fail to create deployer pod (because it already exists) and it never retries.

Before we giving up on creating a new deployer pod for new DC, we should perhaps check the ownerRef and verify that the test-1-deploy which already exists is meant to be for the current DC. We should reconcile and recreate that Pod for the new DC?

Reproduce the deployment recreation sequence with the linked script and observe the deployer pod lifecycle. Trace the deployment controller's deployer-pod creation and reconciliation path, then verify owner references and retry behavior when the previous pod is still terminating or already exists.