killing microshift-etcd with a normally non-fatal signal like SIGHUP or SIGINT the process becomes defunct requiring a microshift restart.   Normally systemd managed processes automatically restart on signals like this.

4.13

Always

1. Get the pid of microshift-etcd and kill -1 on it

1. ps -ef on microshift-etcd shows it as defunct and still parented by microshift
2. Nobody home on port 2379 and the cluster is unresponsive until microshift is restarted.

Cleanly handle signals and restart