Description of problem:

Customer run multiple master OSE with proxy-mode set to iptables on nodes.
In case one master become unavailable, pods start failing during name resolution.

It turns out to be a failure to update iptables rules after master node goes down. iptables rule is not removed and some of the DNS requests are still routed to a failed node.
This issue probably affects not only DNS but also other services provided by then master.

Version-Release number of selected component (if applicable):
3.1

How reproducible:


Steps to Reproduce:
1.
start multiple-master OSE
2.
halt one master
3.
check iptables on a node

Actual results:
iptables still have failed master listed

Expected results:
only alive services do have corresponding records in iptables

Additional info: