Lessons · IT support · one change at a time
One change at a time, or you will never know which one worked
Change one thing, test, and only then change the next. If you change three things and it works, you have three suspects and no answer; if you change one and it works, you have a cause, a fix note and a knowledge article.
Hone is a place to practise a career, one idea a day. This is one of its lessons, written out in full and free to read without an account.
What it is for
A laptop will not print. You reinstall the driver, restart the spooler, and reboot, all at once. It prints. Next week the same thing happens on another laptop and you have to do all three again, for ten minutes, because you do not know which one was the fix. The ticket that says 'restarted the spooler; printed' is a thirty-second fix forever.
How to think about it
Write down the state before you touch anything. Make one change. Test with the user's own task. If it is not fixed, undo the change before trying the next one, so the machine is never carrying two of your experiments at once. When it is fixed, the last change is the fix, and it goes in the note by name.
Worked example
State: printer shows offline, one job stuck in the queueWritten before anything is touched. This is what you are comparing against.
Change 1: cancel the stuck job. Test: still offlineOne change, one test. Not the fix. Nothing else was touched, so this result is clean.
Change 2: restart the print spooler service. Test: printer online, test page printsThe second change, on its own, fixed it. That is a fact you can write down.
Note: 'stuck job cancelled (no effect); spooler restarted (fixed)'Both changes named, and which one worked. The next technician skips straight to the spooler.
Your turn
A change did not fix it. Write what you do to that change before trying the next one.
Not fixed → the change → try the next
Solve one, graded on the server
The trap
Leaving a failed change in place because 'it did no harm'. It did not fix this fault, and now it is a second difference between this machine and a working one, waiting to confuse the next fault.