Description of problem:

When changing the service IP of the registry, all of the masters services need to be restarted.  I have heard that this is because the registry's IP is cached.

Having to restart the master's services should not be needed in this case.  The service IP should be looked up dynamically and not cached.


Version-Release number of selected component (if applicable):

3.1.1.6

How reproducible:

Very

Steps to Reproduce:
1. Create registry according to docs (https://docs.openshift.com/enterprise/3.1/install_config/install/docker_registry.html#advanced-overriding-the-registry-configuration)
2.  Delete registry service
3. Re-create registry service

Actual results:

The cluster will not use the new registry service IP until the master's services are restarted.

HEre are some results:

Error: build error: Failed to push image. Response from registry is: unable to ping registry endpoint https://172.30.22.223:5000/v0/
v2 ping attempt failed with error: Get https://172.30.22.223:5000/v2/: dial tcp 172.30.22.223:5000: no route to host
 v1 ping attempt failed with error: Get https://172.30.22.223:5000/v1/_ping: dial tcp 172.30.22.223:5000: no route to host



Expected results:

Registry could be looked up and pushed to the right registry ip