Etcdadm does not reconcile main node's ip change during bootstrap phase

During bootstrapping etcdadm cluster, if the main etcd node's ip changes, etcdadm does not reconcile the change and update the secret with the new etcd's main node ip. This causes subsequent etcd node's bootstrap process to use a broken etcd node's ip which causes etcdadmcluster bootstrap to fail.

Etcdadm controller logs does print that etcd failed to come up, but says it might be transient since etcd is just trying to come up during bootstrap.

Start by tracing the etcdadm controller's bootstrap flow and the logs that report the main etcd node failed to come up. Verify how the main node IP is stored in the secret and how subsequent node bootstrap reads it. Done means an IP change during bootstrap is reconciled and later nodes use the updated address.

Written by the indexing model from the issue text.