The analysis revealed that the Debian Linux configuration was not included in their test matrix.
You might as well say you don't support Linux.
"Crowdstrike's model seems to be 'we push software to your machines any time we want, whether or not it's urgent, without testing it'," lamented the team member.
I wonder how this shit works on NixOS.
If I'm remembering right, RHEL is Crowdstrike's primary Linux target. And NixOS wouldn't even be a factor since it's basically just not enterprise grade.
That said, they need a serious revision of their QA processes.
RHEL, Ubuntu, & Debian cover the vast majority of enterprise installs I imagine, and provide a solid testing base for developers in the Linux business software space.
Maybe you add Gentoo, some post-CentOS clones/forks, or other more niche industry/workload specific distros, but how you do skip Debian?
RHEL, Ubuntu, & Debian cover the vast majority of enterprise installs I imagine, and provide a solid testing base for developers in the Linux business software space.
Enterprises I imagine are using RHEL, Ubuntu, SUSE's SLES and Oracle Linux and probably not Debian. But that's a guess. Where can statistics and numbers be found ?
Largish enterprise heavily using Debian, just 1 data point here but we do exist.
Nice.
consultant for large enterprises in australia, and i literally can’t say i’ve ever seen anyone running anything other than RHEL and amazon linux (so… RHEL) in production… unless we’re talking not for profits, and then that’s been a bit of a mixed bag
In the enterprise realm it is typically SUSE and RHEL.
I'm not an expert in any sense.
But it was always my impression that Ubuntu and Debian were what you use on personal machines, while RHEL is the baseline standard for professional servers.
Is that not accurate? CrowdStrike's target customer seems to be the sort of company that would insist on using RHEL for the enterprise features.
That is not accurate.
I've been using Linux professionally for 15 years. It's been Debian or Ubuntu almost everywhere I have been. Although that might be regional.
A lot of companies run Debian and Debian based distros, Google on their servers for a start
Canonical and Debian both target the professional server space. I've spent pretty much my entire career working on Debian-based distros.
Hell, the one company I worked for that I expected to use RHEL used Ubuntu for everything, so 🤷♂️.
This is accurate.
There is another reply that says “this is not accurate” that includes true information to back you up.
For infrastructure, RHEL is the gold standard for large companies with a budget. The RHEL customer-base probably overlaps almost completely with CrowdStrike.
RHEL imitators are popular with people that value cost savings more than the corporate backing ( beyond individual cases, this DOES NOT describe the enterprise space ).
Ubuntu is very popular with developers in companies of all sizes. Outside of maybe being the base for containers, this is not how “infrastructure” choices are made though.
Debian is popular with Linux enthusiasts and, where they have influence, businesses may use that. In enterprise environments, it is less likely this group is the one making the decisions. Again though, individual cases exist.
Users of Debian and Rocky Linux also experienced significant disruptions as a result of CrowdStrike updates, raising serious concerns about the company's software update and testing procedures. These occurrences highlight potential risks for customers who rely on their products daily.
Hot take: maybe bossware is a fucking drain on society, and people should stop buying it.
Yeah, but our leadership had a really nice lunch with their sales rep! Licenses for everyone!
It's sad how accurate this is.
After getting a referral from your cyber insurance rep right?
Well, if the executive leech class wants workers to have bossware, there's not all that much people can do about it. Can't just decide to not use it if your employer demands it
Worse, my employer doesn't care about this shit but our clients are demanding we have the bossware installed.
I didn't mean the average worker. I meant the "executive leech class," because downtime of this scale means lost profits, which is something they care deeply about.
which is something they care deeply about.
They care about quarterly profits. Preventing fuckups of this scale requires long-term effort which is not profitable by itself, it only prevents possible future fuckups, and this is why proper QC etc. aren't done. Short term profits over everything else.
In that case, it's time for the average workers to sabotage the bossware. Let the leech class solve the problem they create.
The software is not the problem. Software breaks all the time. The problem is monocultures and centralization. Building entire industry ecosystems all around a single point of failure. This is the just-in-time manufacturing supply chain disruptions and fragility all over again.
Who knew, a diverse ecosystem was a strength, not a weakness.
The software is the problem if it's produced with a corporate mentality of "ship first, fix later".
Yep, at this point the "security" companies can do with imitating malware development practices.
Everyone got an MBA and failed to realize it was just corporate brainwashing.
Nature has been telling us all long, but we don't listen!
"I don't test often but when I do I test on the entire planet"
rootkit doing rootkit things
There's a concept in this industry where you eat your own dog food.
Deploying these updates to your own people could have avoided this mess.
Oh but they did. Turns out that this is specifically caused by one driver expecting another to be installed, the other one being for another of their products. If you have the other product installed, it doesn't crash, so it didn't crash on their machines because they have all their products installed and apparently not a single element of their test matrix has the single most common configuration they service
Do you have a source for that? I'm intrigued. Their own blog post is only talking about a "logic error".
I heard a different rumor, that the driver file they pushed was all zeros. I'm inclined to believe that one.
It's a very educated guess based on the following:
The crash is a null pointer dereference, which a linter ought to catch.
The crash does not happen if you have crowdstrike sensor installed, which is weird because crowdstrike sensor's job is not to prevent any crashes.
Hence the guess: the update the pushed tries accessing memory in sensor, but if it's not installed the pointer is null and that's Bye-Bye.
It's a well assembled article, but mostly based on a few comments in a hackernews post from yesterday. I would like to know how widespread it was.
The article implies that CrowdStrike issue impacted only Debian and Rocky 9.4. Debian I can see. But how did something impact Rocky but not RHEL itself or Alma or Oracle?
Is Rocky actually different from RHEL now? Their entire brand promise is that they are the same.
It also broke windows the other day…
😂
Believe me, people noticed that one.
Nobody breaks Rocky ok (spoken in a gruff, low, mumbling voice)
@lemmee_in I can't find any news about this. Just a statement in a forum and everyone basing subsequent articles on that. It appears to have been limited to a single company? Is there any support for this claim?
A community for everything relating to the linux operating system
Also check out !linux_memes@programming.dev
Original icon base courtesy of lewing@isc.tamu.edu and The GIMP