Le 09/01/2024 à 03:37, Ludovic Bellière via Dng a écrit :
>
> There is fundamentally nothing wrong with the concept behind systemd.
> There is a
> fundamental issue with how the project is put together. When you
> install a
> distro with a systemd init system, you actually have the systemd
> "galaxy": a
> bunch of intertwined daemons that cannot be dissociated from systemd.

    I find several things wrong about the concept:

    The first thing is that it is not clear: it changes continuously
from accelerating init to encompassing the wole OS.

    The initial goal it was advertised for, accelerating init, is
futile since the init speed is mostly driven by the speed of mass
storage access, and, in addition, for servers, is negligible compared to
BIOS startup and tesing time.

    It ends up with replacing all Unix/GNU/Linux commands by systemd
commands, and, this way substitutes a new API between the OS and the
user: you buy Linux-GNU and you get Systemd.

    It makes everything more complex for actually no gain. But even if
there was a gain, the cost is too high.

>
> elogind, in itself, is a decent piece of software. Does what it needs
> to do,
> doesn't quite break, doesn't complain, relatively easy to troubleshoot
> (never
> had to personally). The way elogind works on devuan is the way it
> should work
> everywhere, except on systemd OS.
>
> The better question you should ask, though, is what does elogind can
> do for /you/?

    YES! I fully agree with you on this point; this is the question I
would like to get the answer to.

--     Didier