OpenShift pods in crashloopbackoff due to dynatrace prestart hook sigabrt

Issue

• Upgrade failed from 4.5 to 4.6; cluster nodes unable to upgrade due to failed service ovs-vswitchd.service. Accessing the affected nodes via ssh shows:

  [systemd]
  Failed Units: 1
  ovs-vswitchd.service
  

• The journal logs of the failed ovs-vswitchd.service throws the following error:

  Feb 06 08:41:37 master2 sh[1882425]: /usr/bin/chown: changing group of '/usr/share/openvswitch': Read-only file system
  Feb 06 08:41:37 master2 ovs-ctl[1882427]: /usr/share/openvswitch/scripts/ovs-lib: line 519: /usr/share/openvswitch/scripts/ovs-kmod-ctl: Permission denied
  Feb 06 08:41:37 master2 systemd[1]: ovs-vswitchd.service: Control process exited, code=exited status=1
  Feb 06 08:41:37 master2 systemd[1]: ovs-vswitchd.service: Failed with result 'exit-code'.
  Feb 06 08:41:37 master2 systemd[1]: Failed to start Open vSwitch Forwarding Unit.
  

• Other ovs related systemd services could not run as well due to the dependency on ovs-vswitchd.service.
• Dynatrace agents deployed by Dynatrace Operator adding high priority SELinux modules to the cluster nodes.