1. Proposed title of this feature request

OpenShift installer to reject the nameserver coming from the additional network when doing IPI installation on OpenStack with multiple network.

2. What is the nature and description of the request?

The openshift IPI installer fails, if the DHCP is enabled on additional networks at the OpenStack level.

As a workaround, when DHCP is disabled at OpenStack level it doesn't add additional nameservers to resolv.conf file on the node. 

Following is the syntax for declaring additional network in the install-config file:
 

platform:   
  openstack:     
    additionalNetworkIDs: ['NW-1','NW-2']

 
Here, two additional networks are being used along with the primary network of the node.
Both these are DHCP enabled networks and comes along with 3 nameserver each.

The root cause observed:
DHCP enabled networks tends to populate multiple nameserver on the node. As a result the actual name server which resolve the API & ingress URLS is not added to the resolv.conf file.
There is a known limitation with  the resolv.conf, that it only acknowledges 3 nameservers. 

$ cat /etc/resolv.conf
# Generated by KNI resolv prepender NM dispatcher script
search ocp4labs.local local
nameserver 10.0.xx.1          ---------------> primary network NS (node ip)
nameserver 192.168.xx.1       ---------------> additional NW-1 NS
nameserver 192.168.xx.2 
# nameserver 192.168.xx.3 
# NOTE: the libc resolver may not support more than 3 nameservers.
# The nameservers listed below may not be recognized.
# nameserver 10.xx.0.1         -------------> NS which resolves the api & ingress URL
# nameserver 192.168.xx.1      -------------> additional NW-2 NS 
# nameserver 192.168.xx.2
# nameserver 192.168.xx.3 

3. Why does the customer need this? (List the business requirements here)

Customer wants to use additional network for their application specific requirements.

4. List any affected packages or components.

Installer, NetworkManager, RHCOS