Description of problem:

OpenShift 4.7.x fails during a VMware UPI installation.


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

Tested 4.7.0-4.7.3.


How reproducible:

Always.


Steps to Reproduce:
1. Install OpenShift
2. Wait for master nodes to come up.


Actual results:

Master nodes are in a permanent NotReady state.


Expected results:

OCP cluster installs successfully. 


Additional info:

The ovs-configuration.service fails on all master nodes. Output of journalctl -u ovs-configuration.service and logs from ovnkube-node container from one of the masters is attached.

$ oc logs -n openshift-ovn-kubernetes ovnkube-node-frwtm -c ovnkube-node | tail -n 10
I0323 11:59:22.921177  600528 gateway_iptables.go:46] Chain: "nat" in table: "POSTROUTING" already exists, skipping creation
I0323 11:59:22.930810  600528 ovs.go:166] exec(8): /usr/bin/ovs-vsctl --timeout=15 -- port-to-br br-ex
I0323 11:59:22.946460  600528 ovs.go:169] exec(8): stdout: ""
I0323 11:59:22.946506  600528 ovs.go:170] exec(8): stderr: "ovs-vsctl: no port named br-ex\n"
I0323 11:59:22.946526  600528 ovs.go:172] exec(8): err: exit status 1
I0323 11:59:22.946586  600528 ovs.go:166] exec(9): /usr/bin/ovs-vsctl --timeout=15 -- br-exists br-ex
I0323 11:59:22.960816  600528 ovs.go:169] exec(9): stdout: ""
I0323 11:59:22.960857  600528 ovs.go:170] exec(9): stderr: ""
I0323 11:59:22.960873  600528 ovs.go:172] exec(9): err: exit status 2
F0323 11:59:22.961023  600528 ovnkube.go:130] failed to convert br-ex to OVS bridge: Link not found