Configure tcp_retries2 for all the pods in the OpenShift cluster

Issue

• After customizing the value of net.ipv4.tcp_retries2 in /etc/sysctl.d/openshift_sysctl.conf file, all the pods in the node do not reflect the updated value:

// configured on all the worker nodes via machine-config, loaded the sysctls and then the nodes underwent reboot
$ sudo cat /etc/sysctl.d/openshift_sysctl.conf; done
# The maximum number of times a TCP  packet  is  retransmitted  in established  state  before  giving up
net.ipv4.tcp_retries2 = 5

$ cat /proc/sys/net/ipv4/tcp_retries2
5

// Does not reflect in the application pods (default value 15 still persists):
$ for a in $(oc get pod --no-headers|grep Running |awk '{print $1}'); do oc rsh $a more /proc/sys/net/ipv4/tcp_retries2; done
15
15
15
15
15

• Changes to unsafe sysctl parameters are not reflected for specific pods like router, and not for all the pods scheduled on the specific nodes.

• How to configure namespaced sysctl at node-level instead of per pod?