your backup strategy

Most amazing disk data recovery ever
Columbia's fragments were painstakingly and exhaustively collected. Amongst them was a 400MB Seagate hard drive which was in the sort of shape you think it would be in after being in an explosive fire and then hurled to earth from several miles up with a ferocious impact.

The Johnson Space Centre workers analysing the shuttle crash sent it off the CVX-2 (Critical Viscosity of Xenon) experiment engineers, who sent it on to Kroll Ontrack in Minneapolis, Minnesota, to see if the data, any data, could be recovered. For researcher Robert Berg and his team it was the only hope, a terribly slim hope, of salvaging significant data from the experiment looking at Xenon gas flows in microgravity.

The Kroll people managed to recover 90 percent or so of the 400MB of data from the drive with its cracked and burned casing.

Tags: ,

31 Responses:

  1. topbit says:

    And if the site ( is still down when you go to look at it, google the URL, and look at the cached page.

  2. skreidle says:

    Kroll Ontrack's got an office in Herndon, VA as well. They have a reputation as the go-to guys for all the -other- backup companies.

  3. Sure, they can do this bullshit, but they can't fix my drive that I just knocked off a table. Of course the heads crashed. If the heads hadn't crashed I could fix it myself.

    Maybe if I set fire to it that would help. But then I'd have to pay another $100...

  4. waider says:

    In 2005 I attended a security presentation from someone whose name I can't recall; a lot of what she talked about was in the "eh, I'll believe that if you show it to me but right now you sound like a raving and gullible lunatic", but she did have some interesting (and independently verifiable) stuff to say about Kroll and their data recovery abilities. In particular, I believe they either came up with or made practical a sort of statistical analysis technique which allows recovery of diskwiped drives, relying on the fact that the disk heads never quite travel the same path across the surface of the disk even if you're writing the same cylinder over and over again.

    Also, waiting for the "Columbia" line of Seagate drives. "Crash-tested FROM SPACE"

    • taffer says:

      That'd be way more impressive if they just soldered on a new controller board and plugged it in to "recover" the data...

  5. cara_chapel says:



  6. eqe says:

    I was hoping for a first-person account of how one goes about recovering data from a device in that condition, but alas, your quote contains all of the actual information about the drive recovery. The rest of the article is about xenon and so on; delightful, but not nearly so evocative as that photo.

    • ultranurd says:

      Well, if you'd been studying the shear fluid properties of xenon in low gravity for 20 years, you'd want your hard drive back too.

  7. hadlock says:

    I couldn't look at the picture and help but think after NASA's metric conversion follies, is that ruler in inches or cm?

    • wikkit42 says:

      Considering that it is subdivided into sixteenths, it is inches.

      • ultranurd says:

        Clearly you've never used Imperial Centimeters to determine the winner of a caber toss.

        • wikkit42 says:

          Point the second: how big was the last 400MB drive you saw? The gov uses ancient computer hardware in space because getting mil-spec space rating on new hardware is expensive and fairly useless for the manufacturers.

          • eqe says:

            Additional factors:
            1. a 400MB drive has much larger magnetic domains than a 400GB drive, so I'd imagine a single cosmic ray event is less likely to destroy data.
            2. The experiment was designed long long ago, when a 400MB drive on a DOS machine was perhaps more reasonable than today.

  8. travisd says:

    I'm more intrigued to figure out exactly which shuttle accident this is from, given the filename of the picture: Challenger_drive.jpg

    • martling says:

      Challenger was destroyed on launch, so wouldn't have had any scientific data.

      • quercus says:

        Challenger contained useful log data that was important in analysing the accident itself (AFAIR it was looking at control positions being moved, in consequence of the thrust having gone asymmetric). Some effort went into recovering it.

        As this was 20 years ago, the whole log was probably 5kb and a couple of punchcards.

        • martling says:

          By "scientific data" I meant "experimental data obtained in orbit like the stuff described in the article". I presumed they'd have retrieved what they could of the black box type stuff.

          • quercus says:

            This was more than black box though - it was stuff that was "aerospace" grade hardware (probably yet another Grid laptop!) but never deliberately intended to survive crashes, as their black boxen are.

  9. insomnia says:

    Here's what we were able to recover from your 100 MB of data. That will be $12,839,328 please!

    I know that you tend to avoid anything that even remotely smacks of politics like the plague, and would completely understand why you might send a two word reply to this comment, but I'll send it anyway.

    For a long, long time now, many at LiveJournal, including people like myself who helped create it long ago, have been extremely frustrated at the management of the site, their policies, their occasional censorship, their weak response to the Russian government's prosecution of a LJer for leaving a comment suggesting that corrupt cops should be publically burned at the stake, etc.

    Well, I've created an LJ community called ljunited, where our goal is to form a large, centralized group of LJ members to fight that.

    Specifically, our goal is to work towards defending and restoring the original promises made to LiveJournal's members by LiveJournal itself.

    We want to stand up for the free speech of ALL of LiveJournal's members, restore LJ's original strong support for community-based Open Source Software development, and restore community participation in the successful operations of LiveJournal... because frankly, the people managing the site are also slowly killing it due to their poor policies.

    The site is shrinking in activity about 1% a month currently, and it's within the management's power to do something about this trend, rather than contribute to it by further alienating its members.

    In the past week or so, I have attracted about 150 fairly motivated people to help out. We're also working with helping candidates for the upcoming LJ advisory council elections who embrace our views.

    We've endorsed rm on LiveJournal, who is a very knowledgeable, thoughtful individual with a lot of persuasive oomph who has been on LJ forever. A talented writer, longtime LJ member, very knowledgeable, very talented individual.

    We've also attracted several LJers in the Open Source community, who have their own issues relating to the open source side of LJ that need to be addressed, because LJ has turned its back on working with the open source community for its development over the years, shutting out needed improvements to the software. We want to make the case that they should reverse this policy.

    Despite current crap Russian management at LJ's parent company SUP, we have some reasons for hope. Recently, they brought in someone who previously oversaw the internet development at ITV to be their general manager, so there seems to be some recognition that they need a bit of a change. Our job is to basically do a far better job in making the case for it, and to lay out where they're going wrong and what the better alternatives would be.

    I'd be curious what your concerns are regarding LiveJournal and its future, because it's our goal to bring people into our group and have them positively effect our policies and ideas.

    Sure, it would be a wonderful thing to have you join our community, to have your support for LJ United, its goals, and its candidate... but failing that, I invite you to find out more about us.

    • Companies are not democracies, they exist to make a profit -- not for their users/citizens, but for themselves.

      Livejournal filled a niche that was unfilled. While you have explosive growth as a company, you can do the right thing and make a profit, especially when you have a stubborn benevolent dictator.

      Then, the internet catches up with you. When you've just bought a company on the decline, it's time to start acting like a douche because you want to make a return before the mass exodus -- a mass exodus that's going to happen WHETHER OR NOT you act like a douche because all the kids are into different things now.

      I mean goddamn, how many people who use Livejournal honestly give a shit whether or not they're cooperating with the Open Source community? How many people who use Livejournal even know what the fuck Open Source is?

      This entire reply was a waste of time. I know this.

      • insomnia says:

        I agree with you, actually. Companies do exist to make money, and I think LJ should make money...

        But what happened over time with LJ has been a case of the golden goose... a profitable business that clueless management has tried to cut into, in order to get at the eggs. They seemed to overlook the fact that the money that LJ made was directly connected to customer satisfaction.

        While I understand that some people will leave LJ because it's not the newest, shiniest toy out there, there is still the potential for continued -- albeit lessened --growth. Some of that growth is being cut into by users leaving LJ for other LJ code sites... InsaneJournal has had very significant growth lately, for instance.

        What I want to see is that LJ is viable for as long as possible... and from what I can see, that is a long time, as there is still considerable expansion of the site going on in other countries.

        "I mean goddamn, how many people who use Livejournal honestly give a shit whether or not they're cooperating with the Open Source community? How many people who use Livejournal even know what the fuck Open Source is?"

        I agree with you. More should care, frankly. But the point is, more than enough people on LJ *DO* care about these issues in order to get someone named to the advisory board who will stand up for them... even if they are wonky issues that an average 17-year-old female LJer won't care about.

        Don't get me wrong... there's *plenty* of reason for cynicism. I just think that doing something is better than doing nothing.

      • insomnia says:

        It should be noted, though, that SUP don't view themselves as buying a company on a decline. They view it, quite rightly, as one of the fastest growing online services in their country. The fact that the usage outside of Russia is on the decline is unfortunate, but not necessarily an impediment to the value they are paying for.

    • jwz says:

      A random post on my LJ is clearly not the place to discuss this.

      I wish you luck, but I think you're pissing up a flagpole. LJ has been in technical stagnation (I'm being generous here) since Brad disengaged from it shortly after the 6A sale, and all else follows from it having no leadership. If you think a bunch of angry users are going to make Sup change their tune, well, I envy your optimism.

      I'll miss LJ, but come on, man, the writing's on the wall here. I'll jump ship just as soon as there's a reasonable alternative. In the meantime, I'm just hoping they don't fuck it up too badly before that alternative appears.

      (I suspect that reasonable alternative will involve self-hosting and a usable, distributed social graph, but my magic 8-ball says "answer hazy".)

  10. korgmeister says:

    And that's why I buy Seagate!

  11. loosechanj says:

    The Kroll people managed to recover 90 percent or so

    That'd be the 90% that's crap, wouldn't it?

  12. lohphat says:

    $1800 in 2005 to fix a drive which had a simple actuator motor failure. Data was fully intact.

    RAID people, RAID. It's not a matter of if, but when.

    Oh, and RAID != backup. Make BD copies and get them off-site too.