121–130 of 1001 posts
We are a major CS client, with 50k windows-based endpoints or so. All down. There exists a workaround but CS does not make it clear whether this means running without protection or not. (The workaround does get the windows boxes unstuck from the boot loop, but they do appear offline in the CS host management console - which of course may have many reasons).
SLAs will be breached anyway
The biggest mistake here is running a global update on a Friday. Disrespect to every sysadmin worldwide.
Can someone explain what Crowdstrike actually is? Reading Wikipedia it seems to be some sort of anti-virus software?
On a related note, this also demonstrates the danger of centralized cloud services. I wish there were more players in this space and the governments would try their very best to prevent consolidation in this space. Alternatively, I really wish the CS did not have this centralized architecture that allows for such failure modes. Software industry should learn from great & age old engineering design principles. For example, a large ships have watertight doors that prevent compartments from flooding in case of a breach. It appears that CS didn't think the current scenario was not possible therefore didn't invest in anything meaningful to prevent this nightmare scenario.
1. Stop putting mission critical systems on Windows, it's not the reliable OS it once was since MS has cut off most of its QA
2. AV solutions are unnecessary if you properly harden your system, AV was needed pre-Vista because Windows was literally running everything as Administrator. AV was never a necessity on UNIX, whatever MS bundles in is usually enough
3. Do not install third party software that runs in kernel mode. This is just a recipe for disaster, no matter how much auditing is done beforehand by the OEM. Linux has taught multiple times that drivers should be developed and included with the OS. Shipping random binaries that rely on a stable ABI may work for printers, not for mission critical software.
We are a major CS client, with 50k windows-based endpoints or so. All down. There exists a workaround but CS does not make it clear whether this means running without protection or not. (The workaround does get the windows boxes unstuck from the boot loop, but they do appear offline in the CS host management console - which of course may have many reasons).