News for hackers. Stuff that matters.
Re:Sad news today: systemd-resolved to be deployed in Ubuntu... by _nfin on Monday June 06, 2016 @06:36AM (#11845721)
systemd is a good thing (tm) - I recognize that and am grateful for it.

However.

In the beginning a lot of distros were "forced" to accept systemd because udev was merged into its codebase.

Systemd is great, but its group of devs are absolutely fucking terrible at public relations. Sometimes they make changes that upset users. Like the recent one that has (logind?) by default killing all user-created processes when the user logs out (this surprised people as it became a new default). Or the security incident before that where systemd made it possible for users to get access to framebuffer contents. Sometimes things happen.

It's just made worse when you go to the systemd group with your problem and they act like what they've done is correct and everybody should just live with it. Sometimes we're wrong, sometimes they're wrong. Poettering writes up these really long posts which are mostly good reads, but sometimes he calls us neckbeards who want to live in an unchanging, static past.

Great project, great code, great devs - but terrible attitudes [sometimes].

Re:Sad news today: systemd-resolved to be deployed in Ubuntu... by JdeBP on Monday June 06, 2016 @09:08AM (#11846475)
It was indeed logind.

* https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=825394#221

* https://news.ycombinator.com/item?id=11782364

The systemd developers' way of addressing that was to go and ask the people who wrote things like tmux to change tmux. The response of the tmux developers was to repeat the question asked and not answered 5 years ago.

* https://news.ycombinator.com/item?id=11798515

* https://news.ycombinator.com/item?id=11797075

The problem where the people who just switched to Ubuntu 16 are hitting problems with high-end database servers, middleware, containers, large builds, and so forth all running out of threads because of another systemd default that was switched by its developers, is still quietly on-going.

* https://news.ycombinator.com/item?id=11675129

Re:Sad news today: systemd-resolved to be deployed in Ubuntu... by nisa on Monday June 06, 2016 @09:56AM (#11846753)
> The problem where the people who just switched to Ubuntu 16 are hitting problems with high-end database servers, middleware, containers, large builds, and so forth all running out of threads because of another systemd default that was switched by its developers, is still quietly on-going.

That's evil. Thanks for pointing that out. Terrible to debug, poorly documented and funny side effects you don't immediately think off. That fork bomb reasoning for the change is also sketchy as if you have local privilege you can DoS the machine in a lot of other ways (cpuburn, filling the disk, hogging memory and causing swap of death). If you run a public SSH server you likely already used cgroups and other measures to prevent fork bombs.

Re:Sad news today: systemd-resolved to be deployed in Ubuntu... by Qantourisc on Monday June 06, 2016 @06:58AM (#11845809)
"Great code" -> I am sorry, but if the allegations are true it's not great. (Not saying every part of it is bad, I kinda like the idea of systemd-init-part, but I describe them as utterly inconsiderate to everyone else.)
Re:Sad news today: systemd-resolved to be deployed in Ubuntu... by rwmj on Monday June 06, 2016 @07:22AM (#11845893)
Gigantic thread on the KillUserProcesses setting: https://lists.fedoraproject.org/archives/list/[email protected]...
Re:Sad news today: systemd-resolved to be deployed in Ubuntu... by digi_owl on Monday June 06, 2016 @12:34PM (#11848042)