Picture a pair of training schools facing each other across one busy airfield. Both teach private pilot courses, both keep a fleet of trainers and a simulator, and both hold years of student files. Both would say they take their records seriously. Yet when a server failed at each of them in the same month, one reopened its training office the next morning and the other spent three weeks reconstructing paper trails. The difference was not budget or software. It was a habit, and that habit is what BackupOps is about. These two schools are invented for illustration, a composite drawn from common patterns, not any real organisation.
School A: the comfort of the green light
School A bought a backup package years ago and has not touched its settings since. Every night the system copies the records server to a disk in the same rack. A monitor in the dispatch office glows green. Staff feel safe, and the owner mentions it proudly to prospective students’ parents.
What School A does not know is how many ways that single disk shares the fate of the server. A power surge would damage both. A fire in the equipment room would take both. A staff member with administrator rights, or a malicious program running under that account, could erase the originals and the copy in the same afternoon. And nobody has ever tried to bring a file back.
School B: the habit of rehearsal
School B’s chief instructor handed records oversight to the training coordinator, with a name written next to every category of file. Each quarter the coordinator picks something random, a student’s training file from two years back or the simulator’s last configuration, and restores it onto a clean laptop. The date of that test goes onto a sheet pinned in the office. A second copy lives in a separate cloud account that the daily administrator login cannot delete from. If no good copy arrives for two days, a message goes to the coordinator’s personal phone.
School B spends a little more time and a little more money. The difference shows when things go wrong.
Side by side
- How success is measured. School A counts completed jobs. School B counts assets it has actually restored.
- Where the copies live. School A keeps them nearby, under the same login. School B keeps at least one copy where daily credentials cannot reach.
- Who is responsible. At School A it is “the IT person, probably”. At School B a named person owns each asset and its test date.
- What happens on silence. School A’s monitor only complains about errors. School B’s alert fires when success is absent.
- How long records are kept. School A keeps the last seven copies. School B keeps everything from the past year plus yearly snapshots, so one stretch of corrupted copies cannot push out the good ones.
The month it went wrong
At School A, the failed server was rebuilt by a contractor, who then discovered that the nightly copy had been silently skipping the training database for most of the year. The backup tool had been failing on a file that was open during the run, and reported success for everything else. The school recovered some information from paper logbooks, instructors’ personal notes and emailed schedules, and still ended up asking students to re-sign forms.
At School B, the coordinator opened the recovery sheet, pulled data from the separate account, and had the records system working by lunchtime on a spare machine. It was slower than normal, but nobody needed to guess what had been lost.
What the Atlas says about the gap
The contrast above is exactly what the Ops Atlas page on BackupOps tries to describe. Its core idea is that backup work should be judged by what you can restore. It also quotes a survey from early 2024 in which 94 percent of IT staff said attackers had targeted their backups and, by their own account, 57 percent of those attempts succeeded. Those are self-reported figures from a vendor survey, so read them as a warning light, not a calibrated gauge. The point stands: a copy that sits within an intruder’s reach is a copy that can be taken away.
Borrowing from the cockpit
Pilots already live by the habits School B practises. They run checklists, they log defects, and they rehearse engine failures in a simulator long before one happens at altitude. Take the simulator write-up about a new device for IFR procedure training and the choice of a simulator for instrument, proficiency and upset work show how much the industry values safe rehearsal. Data deserves the same: practise recovery where it is cheap, so you are not improvising where it is expensive. Every new aircraft, like the advanced twin with a full glass cockpit announced earlier, adds one more set of records to protect.
What School A could do on Monday
The good news is that School A’s path to School B is short. First, list the five record sets that would hurt most to lose, and add the missing training database to the backup scope. Second, open a free or low-cost storage account under a login used by nobody except the records owner, and send a weekly copy there. Third, pick a student file from last year, restore it onto a different computer, and watch for anything that does not open. Fourth, put the next test on the calendar. The Atlas calls this early stage “scheduled” and describes the next step up as “owned”, with each top-tier asset restored at least once. Neither step demands a large project; each simply replaces a guess with an observation.
Which school are you?
Most organisations are somewhere between the two. A quick self-check:
- When did you last restore a real file from your backup, onto a different machine?
- Can the account you use every day delete every copy you have?
- Would anyone be alerted if backups quietly stopped for a week?
- Does each important record category have a named owner?
If you hesitated on any of those, you have your first project. No new purchase is needed to become School B. You need a calendar entry, a person with a name, and the discipline to open a restored file and say, out loud, “yes, that is it”.