Node Group Configmap changes may cause Node network traffic to cease flowing

Issue

Subsequent to editing the ConfigMap object cm/node-config-xxxxxxx in the openshift-node namespace, and adding or changing a parameter ....

Example
 system-reserved:
    - "cpu=500m,memory=1Gi"

Network traffic ceases to flow on those node_group members referencing the edited ConfigMap, when net.ipv4.ip_forward is set to 0 in /etc/sysctl.conf on these nodes, and the following error can be seen

Sep 01 08:16:05 thin-lizzy-comp-001 dockerd-current[1859]: time="2021-09-01T08:16:05.923135961+12:00" level=warning msg="IPv4 forwarding is disabled. Networking will not work"

The file /etc/sysctl.d/99-openshift.conf is often used to enable the net.ipv4.ip_forward parameter in RHOCP, opposed to being set in /etc/sysctl.conf

This issue manifests as a result of the update sequences that are followed when the 'sync' pods update node_group members. Namely that /etc/sysctl.conf is parsed AFTER /etc/sysctl.d/99-openshift.conf.

We do not see this issue manifest when Nodes are rebooted, or when docker or the kubelet is restarted.

This only occurs when the sync pod implements changes made to respective node_group ConfigMaps and propagates them to associated nodes.