• (cs) in reply to instigator
    instigator:
    Balr0g:
    I call bullshit on this one.
    1. They really had smb-mounts for the system directories - usr/bin or usr/local/bin/ - Why would any sysadmin do this?
    2. Why and how would the Windows management software then detect the Linux binaries?
    3. As the Windows management software is not root on the Linux machines it should not have permission to delete anything.
    If users don't have root access they might install binaries in ~/bin.

    "Why can't I install my internet to /dev/null?"

  • bluesman (unregistered)

    The real WTF is Linux, amirite?

  • mophobiac (unregistered) in reply to Ian
    Ian:
    The article title should have been something about cats and dogs sleeping together.

    No, it's great! War of the worlds leading to deadlock = Warlock. TRWTF is not implying use of the blessed hammer on the blessed software.

  • (cs) in reply to no laughing matter
    no laughing matter:
    So this was put in production without ever testing it?

    Yeah, they should have had some butter testing with it for sample group of users.

  • (cs) in reply to faoileag
    faoileag:
    muttleee:
    It took managers all day to work out that the changes which had just gone live caused the problem?
    Erm, no. Cheryl found out, after she met that guy in a bar.

    The managers remained clueless until Cheryl presented her findings.

    All good thing can be found in bars. So the moral of story is keep visiting bar during office hours. I will now speak with my manager.

  • (cs) in reply to $$ERR:get_name_fail
    $$ERR:get_name_fail:
    Boomer:
    the executables are not going to have the same names or signatures--is the Windows software looking for ELFs?
    They could have the same signature when they are Java applications. One Java application allowed on one OS and forbidden on the other would be enough to start that loop.
    From what I can gather, it would also uninstall anything not blessed (in other words, they were using a white list.) So, if you have a whitelist, if its not on it, it gets uninstalled. Then it installs the proper software that is found to be missing. You updated that software from microsofts website, oh well, we will uninstall it now cause it's sig isn't on the whitelist and reinstall the old version.

    I fear the solution was about that simple. Just 2 different whitelists competing back and forth.

  • mabinogi (unregistered) in reply to no laughing matter
    no laughing matter:
    So your saying that Android Smartphones are PCs?
    You're saying they're not?

    Let's see, 32GB storage, 1GB Ram, 1.4Ghz quad-core processor, multitasking, multiuser Operating System, supports keyboard, monitor and mouse, 3D video acceleration, USB support, networking.

    That's a more capable PC than half the "PCs" in my house, and has the bonus of only being 5" diagonal and less than a cm thick, with built in screen and pointing device when I don't have it docked.

    About the only difference I see, is the CPU is ARM, not x86 - but x86 isn't what makes something a PC.

  • (cs) in reply to Jazz
    Jazz:
    What I'm hearing you say is "All of my clients who dislike IE prefer Chrome. Therefore, I'd really like to give them Firefox."

    On the downside, that's the worst possible mindless-support-drone type solution to the problem. Listen to what your users are telling you, for chrissakes. They don't WANT Firefox any more than they want IE. Don't shove a solution down your users' throats.

    On the upside, with that sort of approach, you have a promising career in management ahead of you.

    I'm confused. Nothing Spenzer said implied the clients wanted Chrome.

    I'm in the same boat as him, and when people ask for assistance to install another browser instead of IE, they would generally prefer Firefox, which they can't really work with due to Policy issues. I'm not saying one is better then the other, but I end up having to drop Chrome instead because we can't support Firefox.

  • Luiz Felipe (unregistered) in reply to Ralph
    Ralph:
    foxyshadis:
    It's true, it's a pain in the ass and somewhat fragile, and most importantly doesn't integrate into group policy at all. You have to make policy _files_ rather than registry entries
    Sorry, that's not a bug, that's a feature, and a damn good one too. There are soooooo many nice things you can do with files: individually copy them, back them up, restore them, put them in version control, compare them (in version control or not) with previous to see what changed when, set permissions on them, print them, paste them into instruction documents, link to them from how to articles, check date last modified........... and all with the same standard familiar utilities you use to work with all your other files.

    By comparison (although you can do a few of those things, in a clunky way, with a whole different tool set) registry entries suck manure-coated fly-swarming donkey balls while bragging about how tasty they are.

    foxyshadis:
    occasionally the FF people change around about:config entries, although that's a lot more rare now.
    The inability of the FF people to stabilize their configuration options is not a reflection on where they store them. You think if they used registry entries that would stop them from FFing around with things every so often?

    All this you said can be done with registry subtree, you only need one tool to transform it to a text format. It behave like a database, its encapsuled, you need to use regular tools to access it, not that this prevents corruption. Windows registry is a database, not a text file. It has all burdens such thing has, but it has benefits. Like standard access routines, on linux config, every file has a fucking different sintax that you must learn, also, you cant easilly parse it. with Windows registry, no parse needed, just plain C compiler, include windows header and 4 function calls to recursive load keys/values. Also about firefox, i think you are correct, they cant stabilize config, it all sucks. But they can make some interface to control security, and like an external API, make that static, then througth it you change the fucked configuration of firefox.

  • (cs) in reply to no laughing matter
    no laughing matter:
    lucidfox:
    TRWTF is that this article portrays PC and Linux as something mutually exclusive. PC doesn't just mean Windows, it means a computer using the PC architecture. Which includes most Linux computers.
    So your saying that Android Smartphones are PCs?

    I said computers. Last I checked, phones are not computers.

  • (cs)

    Linux is not a PC operating system? What, have the Linux geeks gone the way of the Mac elite?

  • (cs)

    So, anyway, with all this talk of "blessed" software... which side was the Catholics and which was the Protestants? :)

  • DB (unregistered) in reply to Anonymous Paranoiac
    Anonymous Paranoiac:
    Cheryl had lugged her laptop through the frigid weather to get to work early. She put it in the dock, hit the power button, and began to sip her coffee while the machine booted. After a couple of sips, the machine appeared to freeze up. After cursing under her breath, she rebooted and waited again. After several more sips, the laptop appeared to freeze up at the same point.

    Ah, I think there was a paragraph missing after that one:

    This was another blow in a series of unfortunate circumstances for Cheryl, the daughter of a university president. First she was sick, then that moron Andrew couldn't even manage to get his simple DNA sequencing done on time, now she's finally feeling well enough to return to work and this happens.

    Just made my day.

  • (cs) in reply to lucidfox
    lucidfox:
    I said computers. Last I checked, phones are not computers.
    Google:
    An electronic device for storing and processing data, typically in binary form, according to instructions given to it in a variable program.
    Check again.
  • (cs)

    TRWTF is nobody in the entire building disconnecting their machine from the network, booting it and then reconnecting to the network.

  • (cs) in reply to Luiz Felipe
    Luiz Felipe:
    [The Windows registry] behave like a database
    No, it doesn't really. It behaves like a database without consistency checks or multi-update transactions or anything that makes a database actually worthwhile.
  • IP-guru (unregistered) in reply to dkf

    Firefox stores its configuration entries in a json file so the order or parameters should not be an issue Provideing you are using a json parser & not just manually reading the file line by line

  • (cs) in reply to mabinogi
    mabinogi:
    no laughing matter:
    So your saying that Android Smartphones are PCs?
    You're saying they're not?

    Let's see, 32GB storage, 1GB Ram, 1.4Ghz quad-core processor, multitasking, multiuser Operating System, supports keyboard, monitor and mouse, 3D video acceleration, USB support, networking.

    That's a more capable PC than half the "PCs" in my house, and has the bonus of only being 5" diagonal and less than a cm thick, with built in screen and pointing device when I don't have it docked.

    About the only difference I see, is the CPU is ARM, not x86 - but x86 isn't what makes something a PC.

    It's only a PC if it comes with cargo-cult to adress 20-30-year-old-issues.

    Like x86 compatibility.

    Or A20-gate

  • Your Name (unregistered) in reply to $$ERR:get_name_fail
    $$ERR:get_name_fail:
    Aaron:
    I call shenanigans on this.

    Long ago, when I still used Windows, I used Firefox (and Netscape before that) because Internet Explore sucked balls. In fact, IIRC, Firefox was released for Windows long before it was released for Linux. I also tend to use Chrome on Linux when I'm not doing development (typing this in a Chrome window right now).

    In the company I work for we are using IE instead of Firefox on our windows machines mostly because IE can be administrated remotely through Windows group policies. Preventing the user from changing Firefox settings remotely is not that easy (at least according to our IT department).

    Why do you need to prevent the user from changing their browser settings? Why is this considered a desirable feature? I'm guessing approximately zero users have asked you to prevent them from changing their settings, correct?

  • PCGuy (unregistered)

    I saw a similar issue before at a just-merged company...only difference being that it was 2 different Windows programs fighting each other instead. Computers simply never got to the desktop after you logged in.

    The story was this (sorry it's so long, but there's probably enough WTF to make it worth reading):

    Company A had just merged with Company B a year earlier. Both were Windows networks. Initially, A & B's IT systems stayed separate. The few employees that needed access to both company's systems had two laptops, one A, one for B.

    The issues started when A & B finally decided to combine them.

    Company A had used an off-the-shelf network management product and ERP package. B used a home-grown system called RDL, something best thought of as an ERP-wannabe and the base framework on which almost everything at company B ran. The network & PC configuration manager was built in it, as was helpdesk ticketing, accounting system, manufacturing, inventory, sales, etc. Just about all (except payroll, outsourced) systems were in RDL.

    Because many employees from B (mostly sales, billing, and plant workers) liked the RDL system, it had been decided that RDL would be kept in those areas, and Company A's ERP would replace it in the others. Management decided they would use A's very strict, paranoid computer policies, enforced by A's management software, in place of B's lax policy (B gave employees full admin rights to their assigned laptop, and only mandated a provided AV, disk encryption product, and remote support program).

    The changes and new configurations were deployed over a weekend after months of testing, and would install on employee laptops on Monday at login.

    On that Monday, employees began arriving. They logged in, the new management software installed at logon, then began updating their system with approved software and policies. Both RDL and Company A's ERP were "approved", as was a pile of typical stuff (updates, Office, etc.)

    Since several departments were keeping RDL, it was still on their PCs (it had been removed from others where the ERP was replacing it). RDL was designed to remove incompatible programs and install updates to itself at startup. It saw the software installed by company A's management system, and removed part of A's ERP program. It also installed its own updates (which had also been deployed that weekend).

    Company A's software promptly noticed the removal and updates, then proceeded to reinstall the missing program and downgrade the updated components since they were unapproved. The PCs froze at a blank desktop (think when explorer dies) as the management software and RDL fought it out, each repeatedly reversing the other's action. 23 manufacturing plants, billing, and the entire sales department could do nothing. Helpdesk was flooded with calls.

    After about 2 hours, upper management finally decides to have a meeting with lower management and the deployment team. The end result was "Undo everything you did this past weekend, NOW!" Management in IT was never fast to make decisions (often a month or so for the most basic things), but we had approval to un-deploy that configuration in <5 hours that day.

    A week later it was learned that nobody on the deployment team (they were from A) knew anything about RDL (a product of B), so they decided to just leave it out of the test environment figuring it wouldn't be an issue. It was also discovered that the strict policies they deployed would've broken several features in RDL had the deployment succeeded.

    End result? The week before I left (I was an intern mostly doing helpdesk & desktop support), they deployed the opposite of what they originally planned: Put B's lax policies on A's systems and do away with A's management software. It worked perfectly. Devs had also fixed the incompatibility and got RDL and the ERP package to communicate and work together.

  • Check_your_facts (unregistered) in reply to Spenzer4Hire
    Spenzer4Hire:
    This. I'd love to deploy Firefox on all of our Windows machines (yes, they're all PCs as well) but it's a huge pain to manage. So, users who hate IE end up installing Chrome, which automagically installs into the user's profile if they're not an admin.

    If there are any Mozilla devs that read TDWTF, please stop hating on the enterprise. You're just pushing potential Firefox users to Chrome.

    Last time I checked Firefox can install in the user profile just as well, at least on W7 and WXP. Just click cancel when the 'please login as admin' message shows.

  • lolwtfbbq (unregistered) in reply to stew
    stew:
    I think I found the BOFH.
    FTFY
  • (cs) in reply to Ian
    Ian:
    The article title should have been something about cats and dogs sleeping together.

    Anyway, BS. What competent network admin lets Windows boxen connect?

    Oh, yeah, this was admin by management, not admin by competence. Never mind.

    One who wants to keep their job.

  • lollan (unregistered)

    Extraordinary. It's like we're living in the past ...

  • neminem (unregistered) in reply to Your Name
    Your Name:
    Why do you need to prevent the user from changing their browser settings? Why is this considered a desirable feature? I'm guessing approximately zero users have asked you to prevent them from changing their settings, correct?
    Not a desirable feature for the *users*... possibly a desirable feature for the users' *bosses*... if you can change your browser settings, you can do things like connect over a different network, thus totally sneaking past websense and everything else they set up to make sure you don't do anything you're not supposed to.

    Not that I would know anything about that.

  • (cs) in reply to Boomer
    Boomer:
    I agree with you on this. Apart from the question of mounting your bin directories, the executables are not going to have the same names or signatures--is the Windows software looking for ELFs?

    A more likely solution was that the developers actually used samba mounts to install windows software. A mounted drive works just as a normal drive, so you can install software on them just as easily as any other drive.

    If the software removal program only scanned C:\ it would be a nice workaround that enabled developers to install what they needed.

  • Latin captcha lover (unregistered) in reply to faoileag
    faoileag:
    snoofle:
    Then everyone was told to reboot ...just in time to go home.
    They were lucky! When our development server went down years ago, it took five days until we could log in again!!! But tell that to anybody today, and they would not believe you!
    Now that was quick!

    We once lost storage on our biggest repository server holding all the source code, but also the central repository of intermediate artefacts needed during builds. It turned out that a full backup was only taken once a quarter - and we were just before the end of the quarter. So after the day or two it took to restore the full backup, it took another two or three weeks to restore 10 weeks of daily incremental backups ...

    The outages costing only a full working day were typically something minor, like loss of home directories...

Leave a comment on “War of the Worlds”

Log In or post as a guest

Replying to comment #:

Log Out

« Return to Article