I use systemd on all my servers. A couple of lines of declarative configure file to specify a daemon with its restart policy beats a messy shell script. For my

I use systemd on all my servers. A couple of lines of declarative configure file to specify a daemon with its restart policy beats a messy shell script. For my custom processes I've ripped out all the chroot, fork, and setuid code added in a couple lines for systemd watchdog and status support calls. I'm not going back. Granted, journalctl with a filter is longer to type than cat. Someday I'll add nice alias to my shell rc.