Environment info:
 * Fully disconnected OCP cluster version 4.14.22
 * NTP-servers configured via machine-config
 ** 10.171.8.5
 ** 10.171.8.4
 * External interface: infra-bond.94 (which comes br-ex after the ovs-configuration.sh is done) 

Description of problem:

 
 - Chronyd fails to synchonize time during the boot due networking is down at the time when chrony starts, marking both NTP-sources offline and chrony does not recover leaving node time unsynchronized.
 - External networking is not up before the ovs-configuration.service is ready, and only after that NTP-sources can be reached. But chrony has marked sources offline by then, and does not recover.
 - sos-report of one of the nodes:
 -- [https://access.redhat.com/support/cases/#/case/03776816/discussion?attachmentId=a096R00003HlPvjQAF]
 - Must-gather of the cluster is present here:
 -- [https://access.redhat.com/support/cases/#/case/03776816/discussion?attachmentId=a096R00003HlS2rQAF]

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

OCP 4.14.22

How reproducible:
Happens after every reboot

Steps to Reproduce:

1.Happens after every reboot

Actual results:
Initially the network is available, and chronyd-service starts. But soon after (before chronyd finishes the time-sync) the ovs-configure takes the bond-interface down and chrony fails and do not restore on its own once the ovs-configure.sh script is finished.

Expected results:
The chronyd should restore on its own once the ovs-configure.sh script is finished without .manual restart should not be required.

Additional info:

We have asked the customer to again share the latest sosreports and must-gather report from the time of issue to check more. These are yet to be provided.

Affected Platforms:

Is it an
customer issue