XPost: alt.os.linux.mint, alt.os.linux
From: dwhodgins@nomail.afraid.org
On Fri, 08 Sep 2023 12:28:04 -0400, Henry Crun wrote:
> My first and main objection is aesthetic. Now I know that is a
non-definable, highly personal quantity. It might be
> beautifully coded in it's internal working, but the overall efect is one of
an all-invading. all-encopmassing entity.
> Put it succintly. it breaks the tenet "Do one thing, and do it well."
> Systemd defintely does not "Do one thing". As for "do it well", it's
difficult to see through the maze of linked
> scripts, binary logs and what have you.
systemd is not just one thing. It contains many components, each of which do
one thing an does it well.
With initd it just handled scheduling of starting scripts, easily broken by
cyclic dependencies, especially in edge cases.
With systemd, it handles starting things, and restarting them when needed, not
just at boot time. Those scripts may or may not include functions other then
starting and stopping. The edge cases can be handled by custom service rules
that
override the defaults.
To do so properly, it also has components to handle things like ensuring file
systems are mounted when needed, as well as other resources such as network
access, etc.
While I disagree with the choice to use binary files for storing the journal,
I understand why the choice was made given that it simplifies the tools that
come with it.
The tools that come with systemd getting information about any given daemon
or the resources they use, including information from early in the boot process
that is not kept by initd.
Learning how to use systemd and all of it's components is a large learning
curve.
So was learning how to minimize cyclic dependencies with initd start up
scripts.
The more experience I have with it, the more I like it.
Regards, Dave Hodgins
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
