Exercise 1: Why a Year of Green Checkmarks Missed the Missing Config File — Possible Solution ==================================================================== WHAT THE COMPLETION STATUS WAS ACTUALLY MEASURING ------------------------------ Per this chapter, the backup job had been "'successfully completing' nightly database backups" for a year. A completion status, per Chapter 1's own central claim, only confirms that the copy process ran without error - it never checks what's actually inside the copy or whether it's sufficient to bring anything back online. The database files themselves were genuinely, correctly copied every night - the status was accurate about exactly what it measures. WHY THE MISSING FILE NEVER TRIGGERED AN ERROR ------------------------------ The backup job's scope simply never included the configuration file in question - it wasn't corrupted or lost, it was never being copied in the first place. A backup job has no way to flag "this thing you didn't tell me to back up is missing," because from the job's own perspective, nothing is missing - it copied everything it was configured to copy, successfully, every time. WHY ONLY AN ACTUAL RESTORE ATTEMPT COULD REVEAL THIS ------------------------------ The gap only becomes visible at the exact moment someone tries to use the restored data for its intended purpose - starting the database. Per this chapter, "the database restores - but won't start, because a configuration file the application depends on was never included in the backup scope." No amount of checking file counts, sizes, or completion logs could have surfaced this, because none of those checks attempt to actually operate the restored system the way starting the database does. WHY THIS IS A DIRECT ILLUSTRATION OF THIS CHAPTER'S OWN CORE CLAIM ------------------------------ This is precisely the "missing dependencies" failure mode this chapter lists among the reasons a "successful" backup can still fail to restore - files without the configuration needed to actually use them. The example makes that abstract category concrete: a real, specific gap invisible to every check except the one that actually matters. WHY THIS WORKS AS AN ANSWER ------------------------------ It explains what the completion status was and wasn't measuring, explains why a never-included file produces no error rather than a detectable failure, and explains why only the act of actually starting the restored database - the test restore itself - could reveal the gap.