Late last week the story emerged of Marco Marsala inadvertantly deleting all his company data, including his customer's websites that he was hosting.

His problem, he claimed, was a mistake in an automated script which invoked rm -rf {foo}/{bar}, but a bug left the variables undefined, thus invoking the dreaded rm -rf / command.

Masala posted a request to the popular ServerFault forums asking for assistance, only to be soundly rebuked because such recovery was not trivial, and it transpired the only backups made were sitting on the same server, in the same disk hierarchy, and thus were deleted also.

Some on the forum questioned just how this could have happened. While it was always a macabre joke over the decades, and entirely possible to do this in times gone past, Linux has evolved the rm command to include a measure of safety precautions.

For the curious, Kyle Kelley decided to see what happened if you actually do run rm -rf / on your Linux system as did Mario Wolczko prior to him.

Now, Masala is now claiming his story was a farce, that he did no such destruction, and that it was all in fact a viral marketing attempt.

It's difficult to know what is and is not truthful in this situation. On the one hand, Linux does require a verbose command-line flag be added if you really want to delete files from the root partition and it seems possibly that Masala's original story was not true.

Yet, on the other hand, it is unclear what possible "marketing" could be gained by making yourself out to be someone who destroys all their client files. Is it a case of mistakingly believing any news is good publicity? Perhaps we could ask former Telstra Operations Chief Greg Winn who famously said Apple had no business in mobile phones and should "stick to its knitting", this quote now surfacing anytime anyone Googles him.

Nevertheless, whether Masala is a fraudster the first or second time, it is a timely reminder that backups must exist, and they must exist on an off-site storage device which is not directly accessible from the equipment it represents.

Consider the hapless case of Jeff Atwood, the very founder of the Server Fault (and Stack Overflow) Q and A site used by Masala who himself lost all his backups because they resided on the same disks as his faulty server.

Or, consider the increase threat of cryptographic trojan/virus attacks which encrypt all a company's file shares rendering them useless unless either a fee is paid for a decryption key, or a backup exists and is available. The best recovery mechanism from CryptoWall attacks is a good, reliable, recent and importantly, offline, backup.

Even if you're just an individual, and not an enterprise, backups still matter. Consider the varied cloud storage options which can help you out.

Backups matter. In the immortal words of Jeff Atwood - who perhaps should have listened to himself - "Shut up. I know things." Don't complain about backups. Do it.