I work in IT. We don’t run CS here, but a bunch of clients of my old employer do. I had just shitposted about this fiasco in a discord server populated with fellow techs from that company, fired up lemmy, and this was the first post. What a fucking banger! I immediately shitposted again pasting this image. You’ve made my day. I need to stop using the internet because it’s all downhill from here. Thank you.
Fun story, my company just kicked off a PoC with crowdstrike 2 days ago. So far my computer was the only one that the agent was on as we had other work that needed to be done and we paused the rollout to the rest of my team. I woke up to boot loop hell today. Got it fixed right away, but so glad we didn’t roll it out any further. Not a good look to be starting a PoC with.
Sophos endpoint for years. We had an issue like this when we installed their software on one of Microsoft surface that use MS CPU we bricked every one of them. Kept Sophos and got rid of all the surfaces with Microsoft CPU.
If it’s any consolation, this is the first issue of its kind in the multiple years we’ve been using CS. Still unacceptable, but historically the program has been stable and effective for us. Hopefully this reminds higher ups the importance of proper testing before releases
For sure, my previous company I left last August ran CS for 3 years and we had no issues. Hopefully they hire a bunch of QA folks that were probably part of the layoffs earlier this year.
I mean it seems almost impossible that they either didn’t have a staging process to test N number of machines either virtual or otherwise or it passed through but it just seems insane that a problem of this magnitude went out to this many people.
Aw, you just reminded me of something. My grandma used to wash out soup cans, then use them to bake small raisin breads. She would make several at once and you could freeze them. I don’t know where she got this idea but it was awesome always having these tiny raisin breads available :) especially if you don’t want to commit to a whole full-sized loaf!
What amazes me is that so many big companies still use windows in critical core infrastructure.
Windows endpoints is one thing, but anyone using windows servers and MSSQL for mission critical application stacks need to be hit with the modernization hammer.
And then on top of that, they do not have a test rollout of any changes in a test environment, before rolling it out in the production stack.
Good luck to all the engineers in the trenches, having to fix the mistakes of their leadership.
I’ve not used crowdstrike, but looks like a part of the pitch is “cloud managed”, which often implies that the vendor takes care of everything, including updates. Particularly since they market it as a security solution, they weld likely emphasize that they can update rapidly enough to keep up with security attacks that move very quickly because they don’t care about “risk”.
The mistake was pushing it on Friday morning like a bunch of amateurs, they’re supposed to push it out scheduled on Friday at 5:03PM so you have enough time to get to your car and off the parking lot
AFAIK it was a Thursday night push for people in US mountain time / pacific time. But, that ends up being Friday early morning in Europe and Friday mid-day in Asia.
A car is parked at the far end of the street. Hidden by the shadow of an old elm, and a reflection of the blue sky on the windshield, an agent patiently writes out his notes:
8:15 am A leaves house on foot 8:17 am B arrives driving and parks car (license plate: GYX 455), walks away 8:40 am B arrives, enters house 9:20 am A arrives on foot, leaves in car
This is called a stakeout. A form of surveillance.
IT folks will also recognize this as analytics data. You can almost see the json: timestamp, event name, metadata
As analytics data gets tagged to individuals, it becomes targeted surveillance.
Regular analytics is like a surveillance camera: you just see each person in a snapshot, all in the same place. You’re seeing the story of the place. Like a 7-Eleven, tracking when its customers come to decide when to make the coffee.
But modern analytics is more and more all the events about a person or cluster of people. That’s a lot more like the FBI following Hemingway, keeping a log of all his activities to build a profile.
Quite the opposite, bigger grids are much more stable. When faults happen, tiny subsets of the grid get disconnected from the rest, it does not take the whole thing down at all…
lemmy.world
Oldest