Unusual Data Disaster Horror Stories

Lucas123 writes "Computerworld has posted stories from a disaster recovery company that include a scientist who drilled into his hard drive in order to pour oil into the mechanism to stop the squeaking. It worked. Of course a dead drive makes no noise. And, then a guy in Thailand who, after discovering ants in his external hard drive, took the cover off in order to spray the interior with insect repellent. Both the ants and the drive died."

Re:If you want a good laugh, go into atom smashing

[forkazoo]

"There is literally no limit to the human inventiveness when it comes to breaking stuff."

Try breaking reality.

Try studying quantum mechanics. ;)

A website that collects these

[Anonymous Coward]

Someone may have posted it but I've been referring to this site every six months or so to check on updates since my highschool years back in the late 90's. It is a list of "computer stupidities", some of which are actually pretty funny.

http://www.rinkworks.com/stupid/ [rinkworks.com]

yawn

[adolf]

TFA reads like a press release for Kroll [kroll.com]. The whole thing is (almost) written like a short superhero story, with several paragraphs about Kroll saving the day in a small variety of mishaps which are neither very original nor particularly amusing.

These aren't disasters; all of these folks got their data back.

If this is the going rate for disaster articles these days, I might as well tell you all about the hard drive I recently rescued out of a Dell laptop after the Geek Squad had given up on it (big surprise, that). The Toshiba drive had either very bad spindle bearings or a failed head stack (or both), as when I powered it up it vibrated like crazy and made a very rapid thumping noise, but none of this was a big surprise given that it was a little over four years old.

In experimenting with it, I found a few interesting features:

Plugging it into a Windows box to try running Acronis against it immediately bluescreened the host machine.

When powered up, if the drive was slowly rotated, the nature of the thump would change, and something inside would emanate a horrible metal-on-metal grinding sound for as long as I kept rotating it (apparently due to the gyroscopic effect of the spinning platters along with the failed bearings).

The drive was totally unusable in its normal (label-side up) orientation; Linux wouldn't even read the partition table in that state.

But if I carefully propped the drive up, in a very particular, almost-vertical position resting on its connector, it worked. Not only that, but dd was able to recover every single sector of the disk, without error. I then dd'd that back to a new disk, reinstalled Windows (the theory is that Best Buy's fine Geek Squad managed to fuck up XP somehow) on it, did some shuffling of partitions in Acronis, and gave the customer back a working computer complete with their family photos and music library.

Total recovery of user data, much rejoicing, !=disaster.

Or, there was the 200GB Seagate desktop drive that was under six feet of water for about 48 hours. It worked just bloody fine after letting it dry for a week, and then removing the cover to dry out the innards a bit more. Despite the visible traces of river silt still laying on the platters, Windows Explorer was more than capable of retrieving all of the requested data.

Total recovery of user data, much rejoicing, !=disaster.

On the other hand, another (different model) Seagate drive which was also in the same flood failed miserably. Swapping controller boards did not help. Kroll's pricing for recovery was deemed too expensive, and it was therefore a total loss.

It was the hard drive from one of my boss's machines. Years worth of quotations and customer data that were stored in Outlook which he had been accustomed to referring to, all gone. This, of course, ==disaster. (But it was a minor disaster compared to the rest of the flood, which destroyed his office building, trashed the basement at his house, and ate enough of my own house that it is now condemned.)

He is still insistent on maintaining his own PCs, and has subsequently been given the standard-issue lecture about backups, which he'd already heard in the past. We'll see if it soaked in, this time.

But I seem to be digressing a lot, here. The point is, in a world stuffed [catb.org] full [essex.sch.uk] of stupid [rinkworks.com] and funny [theregister.co.uk] computer stories, TFA doesn't seem to include any. The absence of both well-written humor and real disasters factored with the total lack of technical details equates to this article being positively inane and simply as useless as common whitewash [wikipedia.org]. (Another example of this same PR tactic, not surprisingly from Kroll'

Paid articles?

[Kintanon]

Even the original article is essentially just an advertisement for OnTrak. WTF? Why is slashdot inserting ad content into the story sections now? Keep that shit in the banner ads.

Re:Backups...

[djh101010]

There is really nothing wrong with riding an old computer into the ground. Just make sure you plan an escape path. I.e. Test your software and configuration on new hardware. If you have multiple ancient boxes in your data centre and the testing is a routine matter then keeping just a few spares around to swap out whichever old box keels over is a cost saving measure.

Actually, I disagree. We've got almost 2000 Unix servers in our environment. The oldest 10% of them give us maybe 50% of our problems. In the case of a sparcstation 10, it has gone well past end of life, end of service life, and is into the "You're joking, right" phase of support from the vendor. So when something like this dies, someone on my team has to spend a day or two doing heroics to compensate for something that shouldn't have been in the data center in the first place.

I should know we. We just replaced an old monstrosity with 4 CPUs and dedicated external storage with a bare bones PE1950 and internal 250 GB SATA RAID1.

Not because the new box was faster or more reliable. But simply to save on electricity.

In addition to heat and power savings (same thing, really), another consideration is licensing cost. If you're running an app that is priced by CPU, then keeping that old 16 CPU Sun E4500, at 350MHz, is a pretty expensive cost savings. The license savings alone can pay for the hardware upgrade, because of processor improvements. I had a project a couple of years ago to "migrate" a business that we had acquired into our infrastructure. E3500s and similar stuff, really old big heavy servers, half a country away. Turned out to be considerably cheaper to scrap those and buy new here, mostly financed by license savings due to fewer processors.

Sometimes, saving money by keeping the old stuff around is _very_ expensive.