Details
Goals & objectives
Hardware & software
Solution overview
In this lab you take a finished bash tool and package it as a proper systemd service instead of a cron entry. You write the unit file from scratch, then reproduce a startup race between two dependent services and fix it with the right combination of Wants and After. You tune the restart settings so a crash recovers on its own but a broken service stops looping instead of failing forever, and you cap its memory and CPU so one runaway job can never take the host down. In the last module you look at how Ubuntu already socket-activates OpenSSH, switch it to one instance per connection, prove it wakes on demand, and apply the good-fit or poor-fit judgment from the videos to decide whether that startup model belongs on this host.