Systemd 249-rc1 Released With Many New Features

[phoronix]

Phoronix: Systemd 249-rc1 Released With Many New Features

The first release candidate of systemd 249 is now available for testing with yet more new and improved features...

https://www.phoronix.com/scan.php?pa...ystemd-249-rc1

[andyprough]

... more ... big ... ton ... lengthy ...

Are you describing an init system or a cargo ship that got stuck in the Suez Canal, Michael? Sounds almost "bloaty".

[jacob]

Are you describing an init system or a cargo ship that got stuck in the Suez Canal, Michael? Sounds almost "bloaty".

For what must be the 1000000000th time, systemd is not, has never been and was never meant to be "an init system". It's an umbrella project more analogous to GNU or BSD. The service manager was just the first component they released but from day 1, in fact from the very first post about systemd from Poettering, they announced that the goal of the project was to rebuild the userland, not to develop an "init". So yes, it makes perfect sense for systemd to implement all this and much much more in the future. Or are you complaining that GCC is so terribly bloated that it includes a shell, coreutils, an Emacs editor complete with a Lisp interpreter, a debugger and a complete UI framework (GNUstep)?

[andyprough]

For what must be the 1000000000th time, systemd is not, has never been and was never meant to be "an init system".

You are lecturing the wrong person.

Systemd 249 is going to be another big summertime update to this Linux init system.

[skeevy420]

Ha, ha, ha.

                                                                                              
>>>>>>>>>>>>>>>>>flying over>>>>>>>>>>>>>>>>>>>>>>>>andyprough's joke




            jacob's head

After the horrible day I've had I needed an unexpected laugh.

[aht0]

For what must be the 1000000000th time, systemd is not, has never been and was never meant to be "an init system". It's an umbrella project more analogous to GNU or BSD. The service manager was just the first component they released but from day 1, in fact from the very first post about systemd from Poettering, they announced that the goal of the project was to rebuild the userland, not to develop an "init". So yes, it makes perfect sense for systemd to implement all this and much much more in the future. Or are you complaining that GCC is so terribly bloated that it includes a shell, coreutils, an Emacs editor complete with a Lisp interpreter, a debugger and a complete UI framework (GNUstep)?

Sadly you are not quite correct. "new, better init" was hugely advertised MAJOR selling point of systemd by it's proponents at the time. Very little was spoken about it's other aspects, quite the opposite - other planned features were downplayed because changes that drastic would have made lots of people more hesitant rather than enthusiastic about it. It was akin to invasive malware attack - psychological carrot was dangled in front of people to get it installed and now it's present and you cant really get rid of it without making continous huge effort.

[jacob]

Sadly you are not quite correct. "new, better init" was hugely advertised MAJOR selling point of systemd by it's proponents at the time. Very little was spoken about it's other aspects, quite the opposite - other planned features were downplayed because changes that drastic would have made lots of people more hesitant rather than enthusiastic about it. It was akin to invasive malware attack - psychological carrot was dangled in front of people to get it installed and now it's present and you cant really get rid of it without making continous huge effort.

Well look it up yourself:

https://books.google.com/books?id=_h_lCwAAQBAJ:
"systemd defines itself as a system and service manager. The project was initiated in 2010 by Lennart Poettering and Kay Sievers to create an integrated set of tools for managing a Linux system including an init daemon." (emphasis mine)

And of course: https://www.freedesktop.org/wiki/Software/systemd/:
"systemd is a suite of basic building blocks for a Linux system. It provides a system and service manager that runs as PID 1 and starts the rest of the system. [...] Other parts include a logging daemon, utilities to control basic system configuration like the hostname, date, locale, maintain a list of logged-in users and running containers and virtual machines, system accounts, runtime directories and settings, and daemons to manage simple network configuration, network time synchronization, log forwarding, and name resolution." (emphasis mine)

The idea that systemd was supposed to be "init" that has somehow outgrown its initial goals is a myth. A widespread myth in some communities, but still a myth. They first released the service manager part (which btw is not at all an "init" in the traditional *nix sense, it's very different in its both purpose and operation) but claiming that they somehow promised "init" and then somehow cheated everyone by creating something different and larger is revisionism. GNU's ultimate purpose has never been to create a C compiler, that was just the first step toward their goal and, similarly, systemd created a service manager but that was also only the first step toward their goal which is to develop a base OS userland environment.

[egberts]

Too bad that systemd cannot be started by SysV init.

[Nocifer]

Well look it up yourself:

https://books.google.com/books?id=_h_lCwAAQBAJ:
"systemd defines itself as a system and service manager. The project was initiated in 2010 by Lennart Poettering and Kay Sievers to create an integrated set of tools for managing a Linux system including an init daemon." (emphasis mine)

And of course: https://www.freedesktop.org/wiki/Software/systemd/:
"systemd is a suite of basic building blocks for a Linux system. It provides a system and service manager that runs as PID 1 and starts the rest of the system. [...] Other parts include a logging daemon, utilities to control basic system configuration like the hostname, date, locale, maintain a list of logged-in users and running containers and virtual machines, system accounts, runtime directories and settings, and daemons to manage simple network configuration, network time synchronization, log forwarding, and name resolution." (emphasis mine)

The idea that systemd was supposed to be "init" that has somehow outgrown its initial goals is a myth. A widespread myth in some communities, but still a myth. They first released the service manager part (which btw is not at all an "init" in the traditional *nix sense, it's very different in its both purpose and operation) but claiming that they somehow promised "init" and then somehow cheated everyone by creating something different and larger is revisionism. GNU's ultimate purpose has never been to create a C compiler, that was just the first step toward their goal and, similarly, systemd created a service manager but that was also only the first step toward their goal which is to develop a base OS userland environment.

Not to mention that even if systemd had been supposed to be just an "init" that has somehow outgrown its initial goals, that's no crime in any sane person's book. As you say, systemd is an umbrella project that aims to reshape the whole of Linux's userland, part of which is of course the "init", but there is also a bootloader, a login manager, an OOM manager, a network manager, a log manager, a device manager, etc. Most of these are stuff that have always been part of the Linux userland but used to be made by separate parties and just glued on to each other and made to work together, until systemd came along and truly united them in true BSD fashion (which is usually praised for its model of tightly integrated userland development).

The really fun part (that has also been repeated more than 1000000000 times) is that almost all of these modules are optional and if you want you can use systemd as a pure init replacement (as most distros do, usually with some of the extras tacked on due to convenience). But who cares about the small details, SystemD is a cancer so f*ck you Nvidia, amirite.

[Danielsan]

I run daily Debian and Devuan Testing, but it is always Debian that gives me headache and always because systemd...

Last week I would use the cloud space my email provider gives me; you can connect to it through davfs, fine!

The Arch Wiki recommends to create a systemd unit (which is counter-intuitive) for the auto-mount but it also suggest you can add a much simplier line to fstab, and I went to the latter to avoid double work...

And guess what?

Debian doesn't auto-mount the davfs resource because systemd is complaining that is missing something, of course Devuan mounts the resource flawlessly.

Now I am evaluating to move all my desktop machines to Devuan since systemd is becoming utterly complicated and the headless ones to FreeBSD/OpenBSD, I prefer spending my time learning BSD rather than systemd.

Eventually it will happen the same as for pulseaudio, it will be replaced by something, same author, same issues and same destiny.