Comment on An angry admin shares the CrowdStrike outage experience

catloaf@lemm.ee ⁨1⁩ ⁨month⁩ ago

• Zron@lemmy.world ⁨1⁩ ⁨month⁩ ago

  I remember a few career changes ago, I was a back room kid working for an MSP.

  One day I get an email to build a computer for the company, cheap as hell. Basically just enough to boot Windows 7.

  I was to build it, put it online long enough to get all of the drivers installed, and then set it up in the server room, as physically far away from any network ports as possible. IIRC I was even given an IO shield that physically covered the network port for after it updated.

  It was our air-gapped encryption key backup.

  I feel like that shitty company was somehow prepared for this better than some of these companies today. In fact, I wonder if that computer is still running somewhere and just saved someone’s ass.

  source
• SapphironZA@sh.itjust.works ⁨1⁩ ⁨month⁩ ago

  We also backup our bitlocker keys with our RMM solution for this very reason.

  source

  • catloaf@lemm.ee ⁨1⁩ ⁨month⁩ ago

    I hope that system doesn’t have any dependencies on the systems it’s protecting (auth, mfa).

    source

    • SapphironZA@sh.itjust.works ⁨1⁩ ⁨month⁩ ago

      It’s outside the primary failure domain.

      source
• JasonDJ@lemmy.zip ⁨1⁩ ⁨month⁩ ago

  I get storing bitlocker keys in AD, but as a net admin and not a server admin…what do you do with the DCs keys? USB storage in a sealed envelope in a safe (or at worst, locked file cabinet drawer in the IT managers office)?

  Or do people forego running bitlocker on servers since encrypting data-at-rest can be compensated by physical security in the data center?

  source

  • catloaf@lemm.ee ⁨1⁩ ⁨month⁩ ago

    When I set it up at one company, the recovery keys were printed out and kept separately.

    source

    • nobleshift@lemmy.world ⁨1⁩ ⁨month⁩ ago

      Paper never goes out of style …

      source
  • Tankton@lemm.ee ⁨1⁩ ⁨month⁩ ago

    Paper print in a safe is what’s usual done.

    source

    • modeler@lemmy.world ⁨1⁩ ⁨month⁩ ago

      You need at least two copies in two different places - places that will not burn down/explode/flood/collapse/be locked down by the police at the same time.

      An enterprise is going to be commissioning new computers or reformatting existing ones at least once per day. This means the bitlocker key list would need printouts at least every day in two places.

      Given the above, it’s easy to see that this process will fail from time to time, in ways like accicentally leaking a document with all these keys.

      source

      • JasonDJ@lemmy.zip ⁨1⁩ ⁨month⁩ ago

        I think the idea is to store most of the keys in AD. Then you just have to worry about restoring your DCs.

        source

        • -> View More Comments
• ripcord@lemmy.world ⁨1⁩ ⁨month⁩ ago

  They also don’t seem to have a process for testing updates like these…?

  This seems like showing some really shitty testing practices at a ton of IT departments.

  source

  • USSEthernet@startrek.website ⁨1⁩ ⁨month⁩ ago

    Apparently from what I was reading these are forced updates from Crowdstrike, you don’t have a choice.

    source

    • ripcord@lemmy.world ⁨1⁩ ⁨month⁩ ago

      I’ve heard differently. But if it’s true, that should have been a non-starter for the product for exactly reasons like this. This is basic stuff.

      source

      • Entropywins@lemmy.world ⁨1⁩ ⁨month⁩ ago

        Companies use crowdstrike so they don’t need internal cybersecurity. Not having automatic updates for new cyber threats sorta defeats the purpose of outsourcing cybersecurity.

        source

        • -> View More Comments
  • catloaf@lemm.ee ⁨1⁩ ⁨month⁩ ago

    Unfortunately, the pace of attack development doesn’t really give much time for testing.

    source

    • ripcord@lemmy.world ⁨1⁩ ⁨month⁩ ago

      More time that the zero time than companies appear to have invested here.

      source

      • TonyOstrich@lemmy.world ⁨1⁩ ⁨month⁩ ago

        I was just thinking about something similar. I can understand wanting to get a security update as quickly as possible, but it still seems like some kind of rolling update could have mitigated something like this. When I say rolling, I mean for example split all of your customers into 24 groups and push the update once an hour to another group. If it causes a massive fuck up it’s only some or most, but not all.

        source

        • -> View More Comments