Support NeoGAF

[old manatee]

It sounds like we are all lucky that his shitty code didn't delete the entire Internet.

 

[Gemüsepizza]

Probably fake. I've read that he tried to restore the data after the deletion and he claimed he made another mistake by using wrong arguments for a command, which resulted in a complete loss of the data. Nobody in such a position is that stupid.

 

[Mechanized]

This is why you have daily tape backups.

 

[ColtraineGF]

Oh man... rm -rf claims yet another victim (that slash was just asking for trouble. That script in general is just lol).

If you're backing up vital info always do it offsite, offline, and non-networked.

This applies even when you aren't deleting massive amounts of data, so that the effects of ransomware will also be limited.

 

[DOWN]

This made me cringe

 

[PillarEN]

Damn. Just like that one company that deleted their entire MMO. Devastating to the point of almost being comedic sans the real life implications.

 

[Qurupeke]

Okay, that's actually funny. Poor guy though.

 

[Goro Majima]

I imagine this to be the look on his face after realizing what he did

 

[industrian]

My reaction was literally this.

 

[Markster]

While the no backup situation is dumb, got kinda scared of "rm -rf {foo}/{bar}"

will keep it in mind to check they aren't empty next time I write a script that deletes stuff ahahaha

This is the appropriate way to feel about rm -rf
Also, test your code.

 

[mantidor]

That's some terrible backup setup if it got wiped by an rm -rf.

Yup.

I feel bad for the guy though.

 

[Nafai1123]

This is why you have daily tape backups.

Tape? What is this, the middle ages?

The fact they have no DR solution whatsoever is laughable.

 

[eddie4]

No backups? Lol.

It said the batch file connected to the network storage where the backups were and deleted that as well.

 

[Blue Submarine No. 6]

I could run that script on any box or network I manage and lose no more than 24 hours of information (which would be bad enough, but I wouldn't have that script running on my boxes in the first place)...

1500+ clients... good grief...

Frankly, if your livelihood is dependant on what is contained on your server or network, and you're too stupid or lazy to have a proper disaster recovery plan with a secured/off-site backup, you should look for another line of work... something in the food service or housekeeping industries, perhaps.

 

[Aureon]

I call fake.

 

[lordfuzzybutt]

I call fake.

Yep.

 

[Wowfunhappy]

Immediately power down everything and run data recovery programs, or get an expert to do so.

If he had redundant copies of all the data, and stopped using the drive immediately, he should be able to get everything back.

 

[Bob Coffee]

this is why you have at least 3 back ups for your professional stuff, one onsite, one tape, one offsite in da cloud.

 

[cityhunter]

.

 

[Dreavus]

How does this even happen accidentally? That seems like a very specific line of code that does one thing very well.

Was he talked into doing it by someone as a bad joke?

 

[LiK]

I keep multiple backups of files on separate external HDDs. Why don't more people do this on their own. Also, this guy is a dumbass and I'm sure those customers won't be back.

 

[Thorgal]

so you are telling me if i were to type * rm -rf Neogaf it would..... uh oh.

 

[Krejlooc]

isn't this just a quick format? As in, it sets a flag, but doesn't actually destroy the data. There are tools he should be able to use to recover that data from his harddrive.

 

[FortuneFaded]

 

[ElTorro]

How does this even happen accidentally?

That's not the implausible part. For instance, one could accidentally type

rm -rf /*

instead of

rm -rf ./*

Given that this guy constructed the path out of two variables, mistakes are even more likely.

 

[Pokemaniac]

Unless it's actually zeroing out all the data, this feels like something a data restoration company could handle pretty easily.

It's really easy to screw yourself over in a situation like this. Doing anything besides immediately cutting the power risks destroying data.

 

[FunkyDealer]

He dun goofed.

 

[thepotatoman]

isn't this just a quick format? As in, it sets a flag, but doesn't actually destroy the data. There are tools he should be able to use to recover that data from his harddrive.

There's probably even a data recovery store somewhere near him that can do it for him.

Could probably do it yourself too, but paying the money for people that know what they're doing is worth it when it's your entire company on the line.

 

[sixteen-bit]

ctrl+z that shit

 

[GK86]

I feel for the poor bastard. I would cry if that ever happened to me.

 

[ElTorro]

Could probably do it yourself too, but paying the money for people that know what they're doing is worth it when it's your entire company on the line.

Maybe he could try to write a recovery script before spending good money for an external company.

 

[iPaul93]

Damn.I feel sorry for the guy.

 

[Kyuur]

Probably fake. I've read that he tried to restore the data after the deletion and he claimed he made another mistake by using wrong arguments for a command, which resulted in a complete loss of the data. Nobody in such a position is that stupid.

I agree, but lol @ bolded. That is unfortunately not always true.

 

[Faddy]

Should have used Windows...

Wait once Win10 gets Bash no server will be safe

 

[Jonnax]

Don't you need --no-preserve-root to do that?

 

[ElTorro]

Am I the only one is reading the title as "Florida Man" in his mind?

 

[Aureon]

How does this even happen accidentally? That seems like a very specific line of code that does one thing very well.

Was he talked into doing it by someone as a bad joke?

No, it can actually happen pretty easily.

you'll run a script with:
rm -rf {var1}/{var2}, which will delete the var1/var2 directory and all it's contents.
If by some mistake var1 and var2 are both null, it'll end up running:
rm -rf /
which will delete the entire drive.
This can only happen when you're root, and anyone worth their salt in sysadmin will input an additional check before rm -rf, AND will not be root if it's not necessary, AND will not keep the sole backup mounted, AND will run the script on a single test box, then on a real box, and only then on the entire network.
But idiots exist, even in system administrations. But still, it's fake. I don't think anyone really did rm -rf /.

Don't you need --no-preserve-root to do that?

Depends on version, now that i think of it.

http://www.ee.ryerson.ca/~elf/hack/recovery.html

 

[mclem]

Well, on the bright side, he didn't put the trailing slash in.

Edit: Oh, he did! The news report didn't mention that.

 

[blastprocessor]

Guy needs disksalv, saved me many a time on the Amiga.

 

[Princess Viola]

Pic of him after realizing what he did

 

[MJLord]

He should have physically unplugged the machines (if he was in control of the hardware, big if) straight away, maybe he could have recovered some of the data.

 

[ratcliffja]

Man now I don't feel so bad about accidentally doing that with college homework about 12 years ago.

 

[mugurumakensei]

Tape? What is this, the middle ages?

The fact they have no DR solution whatsoever is laughable.

Tapes last a really long time, take up very little space, are high capacity, and are cheap.

 

[TalesFromHoth]

That's awful to hear, but his back-up solution needs work.

Need to be cloud saving, protected external storage, tapes, whatever you can do!. Being the IT guy for a law firm, I'm scared as heck of loosing any file so redundancy is needed.

 

[Catdaddy]

 

[Joel Was Right]

Omg

 

[ApharmdX]

LOL. This is so avoidable, too. Original, disk-based backup, and cloud/tape/disk offsite backup.

Yeah I don't believe it.

I do.

 

[Lonestar]

I imagine this to be the look on his face after realizing what he did

Oh geez, I remember this game, and that image.

 

[Zepp Twofist]

How does this even happen accidentally? That seems like a very specific line of code that does one thing very well.

Was he talked into doing it by someone as a bad joke?

The explanation is that there were supposed to be arguments in his script that placed some variable in front of and after a / . So if those variables were "home" and "user", he would delete everything in a folder called home/user.

That being said, with the extra steps were he also messed up data recovery on his backup disk, I am inclined to not believe this story. It seems weird to have a script that would first mount a backup drive and then run a poorly written delete script on all servers. What non-destructive purpose could that serve?

 

[Jedi2016]

This is why you run backups to tape or otherwise off-site. It's his own damn fault.