Systemd is more than an init system. Systemd was designed to be different from previous Unix-style single-/narrow-purpose services. Many distros making the switch seems to indicate that such a switch had significant enough upsides or necessities. No?
I read an article about why Systemd became what it is, and why it makes sense, and that made sense to me. Integration and a fully designed system has advantages over disconnected utilities and systems you have to connect and negotiate, especially on system- and boot-up level concerns.
Systemd is more than an init system. Systemd was designed to be different from previous Unix-style single-/narrow-purpose services. Many distros making the switch seems to indicate that such a switch had significant enough upsides or necessities. No?
I read an article about why Systemd became what it is, and why it makes sense, and that made sense to me. Integration and a fully designed system has advantages over disconnected utilities and systems you have to connect and negotiate, especially on system- and boot-up level concerns.
And therein lies one of its problems.
Other init systems are able to handle those issues without requiring the absolutely insane amount of scope creep that systemd exhibits though.
That comes with the price of lower reliability, highly non-linear behaviour and a central point of failure (or control). But, its thr user’s choice.
Is it though? If it’s an app you can (usually) replace it, but the init?
The choice there is, in most cases, to replace the whole distro.