You can connect to a running pod with oc/kubectl rsh pod_name, or start a copy of a running pod with oc debug pod_name, but as best practises recommend unprivileged, slim container images, where do you get sosreport, kdump, dig and nmap for troubleshooting ?
Fortunately you can start either a transient Debian troubleshooting container with:
oc run troubleshooting-pod --stdin --tty --rm --image=docker.io/library/debian:buster
or a Red Hat Entreprise Linux:
oc run troubleshooting-pod --stdin --tty --rm --image=registry.access.redhat.com/rhel7/rhel-tools
