I’ve done workstation maintenance in a previous job. Every part of the Linux centralized management was worse than Windows. We did it to support our coworker’s wishes, but SSSD constantly shits the bed, and having to code (config management) to write some pretty simple rules like default printers is super annoying compared to the Active Directory built ins.

Microsoft’s biggest strength is the Active Directory. Linux user and computer management is a huge PITA.

While it may be true that getting rid of SUID binary is ideal, widening systemd’s security surface area is much more concerning to me than the sudo binary.

Ink also runs when wet, so caveat emptor if you plan on your paper existing anywhere with water.

How would you deal with iowaits in a system like that? I can perfectly burn 100% of CPU time running a poll(), but that’s not useful work…

Tickless means it’s not based on the computer frequency and idle CPUs can stay idle rather than being annoyingly brought into high power mode ever 100 Hz, but it’s still firing interrupts based on scaling timed variables.

They’re now called “Dynticks”

SUSE wrote the vaguely more understandable write up that Linux foundation links to: https://www.suse.com/c/cpu-isolation-full-dynticks-part2/

BTW, the Linux RCU code is evil but interesting: https://www.p99conf.io/session/how-to-avoid-learning-the-linux-kernel-memory-model/

Quartz is the old macOS graphics framework, but the mouse shaking is probably just a cool show off feature of Core Animation. There’s uncontested Windows ports on GitHub, so I doubt Apple will throw any fits for Linux.

Yea, there’s no where near enough RAM in that system. Add RAM and I’d bet everything fixes itself for a bit.

META-C :wq!

ESC is all the way over there and my hand is already on the space bar.

fine Data Definition.

Disk Duplicator is a destroyer? Man, I used to image so many drives with DD back in my helpdesk days…

Most of the bolted on services are rather shitty. ResolveD in particular is straight up hot garbage unless the only thing you are comparing it to is nscd. I like shipping all my logs to a centralized service and journald just gets in the way. It feels like they took upstart, bolted on some really crappy daemons,

There’s a chain of reporting issue. Generally, the site you’re on (outside the social media shitholes) will keep track of some identifier of who you are, and which sites you visited. Then they will share your identifier with an advertising partner to deliver you advertisements. The advertising partner will take your data and pass it off to a data management platform (Hi Oracle!) who will then attempt to link you on that site to literally everything you have ever done. They have deals with Credit card companies, TV vendors, car manufacturers, cell service providers, public databases. That’s where the sketchiness happens. The worst part of it all is that realistically, the advertisers don’t care about all that data. They almost always want some very basic demographic data that fits in the old Nielsen family demographic data: It doesn’t make any sense to advertise a Lexus or investment advice to someone making minimum wage. Politicians want to know who likely voters are. Macy’s wants to advertise at people who shop at malls. They also want “Lives within 50 miles of my business”

The biggest worry is that the data platforms collect a lot of exact data that is not used except for super suspect Cambridge Analytica level targeted political advertising, and to add to it: They are reallllllly crappy at their jobs. I work in AdTech, so I can poke around at what they think I am and I’ve had things from “Salad Dressing lover” to being both unemployed and making $1 million/year in the same profile.

Hash the image, then assign hash ranges to servers that are part of the ring. You’d use RAFT to get consensus about who is responsible for which ranges. I’m largely just envisioning the Scylla gossip replacement as the underlying communications protocol.

Why? Use something like RAFT, elect the leader, have the leader run the AI tool, then exchange results, with each node running it’s own subset of image hashes.

That does mean you need a trust system, though.