A kubernetes node running a DaemonSet can no longer reach kubernetes master due to network issues or maintenance. The pod is still running until the host is restarted.

Is there a possibility to restart the pod if the node is rebooted and still not able to reach the master?

The image comes from a docker registry that is running on the kubernetes master, so connection to the registry is lost too in that case. The software running in the pod is able to record data when offline and resend when network is available, but this cannot happen if a restart occurs as the docker process is not started.

Edit: The example use case is: The cluster master node goes offline. The slave node is still running. Then the slave node is restarted due to power reset. The pods are not started on the slave after reboot, but daemon docker is running(docker ps -a reveals the docker ps of the pods but they're not running).

Goals: Have the docker process of the pods started on reboot. And still be able to manage the pods after the cluster master node gets back up.