Saturday April 16, 2016

rm -rf *

Man accidentally 'deletes his entire company' with one line of bad code

article in The Independent, 15 April 2016, /via Gene Spafford in Twitter.

ScreenShot2016 04 16 13 56 10

"The problem command was "rm -rf": a basic piece of code that will delete everything it is told to."

I love this stuff. It proves we're human. (And also, EVERYTHING needs multiple off-site backups!)

I think this is almost a rite of passage for some of us. At least it tends to stop (or slow down) after it removes /bin/rm... and most of us only wipe out our own disk. (Extra credit if you ran it from a script and didn't test it sufficiently first with printf or echo.)

A few personal stories:

1. I managed to delete part of a company bug database when I issued a delete command on what I believed was my copy. It wasn't (and why did I, a lowly quality lead, have access permissions to run this on the real database???) I stopped it by hitting the shutdown button on my desktop Mac.

2. We also came into work one day at the same company (a pattern?) to find a notice on all of the doors that said that much of the CVS repository had been "accidentally" deleted and they were restoring from backups, please don;t do any work until we tell you it's OK".

3. At a previous company, our "tech lead" decided to re-use the disk that served as /usr/local on the group server because he "needed disk space". (He wasn't fired, but he was sent on "as much vacation as you have time earned" that afternoon.

Then there are the stories I've heard:

- the sys admin who decided to "test" the new freon system on the server room (it worked).

- the sys admin who was giving a tour of a secure server installation to a bunch of military higher-ups and becoming increasingly impatient with their questions. When one general expressed some distrust in the ability of the thermite charges attached to each Vax to render the data inaccessible, his reply was, apparently "Let's. Find. Out."

- A friend told me a story about how the University she worked at as a sysadmin would swap out certain disk packs at night so that people could use different OSs. One night, something went wrong, the system booted up and started running fsck on what was _not_ a Unix file system. The late night operator kept hitting return ('yes') until a final prompt announced "It's your funeral", at which point he stopped and picked up the phone.


Edited: According to PC world, the initial story is a bad "marketing effort" (a hoax). Sure. Because nothing says "Let's promote our company!" quite like publicly stating the owner is so stupid that he would delete all of the code by accident".

The truth or falsity of the story is not the point. This particular "cautionary tale" may have been a "hoax" (seriously? You post a death notice for a server and think it's a JOKE? What kind of cruel excuse for a human are you????) but it was too close for comfort for many of us. We've issued the

rm -rf * OMG OMG ^C ^C ^C 

I'm a bit worried about the technical acumen of the author of that story, who wrote "The most surprising thing might be that so many people believed him, including those on a forum for technology experts." It's not "surprising" at all, because we "technology experts" know all too well how plausible this is.

It happens. It happens too often. And when it happens, it is not funny.

"The moderators on Server Fault have been in contact with the author about this, and as you can imagine, they're not particularly amused by it," Stack Overflow said in a statement.

I should hope so.

rm -rf * ( in category Noteworthy , Random Thoughts ) - posted at Sat, 16 Apr, 14:05 Pacific | «e»

Post a comment

Any posted comments will be viewable by all visitors. Please try to stay relevant ;-) If you simply want to say something to me, please send me email.

All comments will be moderated. Thank you for your consideration.