XPost: alt.os.linux.mint, alt.os.linux

On Fri, 08 Sep 2023 08:18:49 -0400, Henry Crun <mike@rechtman.com> wrote:

I am gradually moving laptop...netbook...eventully desktop from Ub. to MX. Very smooth learning curve, some advantages, some disadvantages.
On Ubuntu I have been running for about year with no snap or flatpak. If only I could avoid the horror that is systemd.
We are supposed to have choice, but to choose 'no systemd' immediately implies 'not Ubuntu'
Pity 'bout that.

When Mageia first implemented systemd, I was not in favour of it at first, but decided to give it a fair chance and dig into it. After skimming through a lot of documentation and experimenting with it, I changed my mind.

As leader of the qa team for Mageia linux, I had a lot more exposure than most people to the problems with initd, where a change in one package would stop other packages from starting and in many cases stop the system from getting
to a state where anyone could log in, except in run level 1. Most of those mistakes were never seen by anyone other then the members of the qa team
and the packagers fixing the problems.

Similar mistakes in a package running under systemd are much less likely to stop the system getting to a point where login is possible, and even in those cases, they are much easier to fix.

There are things I don't like about systemd, such as the defaults for the max size of the journal, but they are minor things that are easy to fix once you understand how things work.

It's much more consistent in how things work. Instead of every package having it's own custom start up script that can do things in a wide variety of ways, each package has rules that systemd uses in a consistent way.

Like any init system, it has a learning curve in order to set things up properly
both for the packagers, and for the admin of the systems using it. It does much more than just replace initd, and does so in consistent logical ways.

The learning curve in order to set things up properly both for the packagers, and
for the admin of the systems using it is steep, but so was the learning curve for
initd when I first started using it and debugging packages that use it.

Regards, Dave Hodgins

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

David W. Hodgins wrote:

It does much
more than just replace initd, and does so in consistent logical ways.

What is your 'position' on the arguments completely separate from the
init improvement; that many don't like that systemd likes to take
'control' of so many non-init responsibilities?

I guess the systemd idea is that just like init, if systemd can do it
'better', then that should be just fine.

However, it does make it progressively more difficult to 'do without'
systemd, since it is SO much more than an init.

--
Mike Easter

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

On Fri, 08 Sep 2023 11:25:51 -0400, David W. Hodgins wrote:

When Mageia first implemented systemd, I was not in favour of it at first, but
decided to give it a fair chance and dig into it.

There is nothing wrong with systemd per se.

The problem is that distro maintainers see it totally as an
"either/or" issue.

GNU/Linux is all about CHOICE, and, to my knowledge, there are
only two distros that offer users a choice regarding systemd:
Gentoo and Linux From Scratch (LFS).

Why has systemd gained such a near-complete hegemony? Why
is there this egregious lack of choice?

Could it be that offering a choice would be too much work
for the distro maintainers? I strongly believe that this
is perhaps the major, if not only, factor.

My opinion is simple and clear:

If a GNU/Linux distro cannot offer CHOICE then it should
quietly step out of the game.

GNU/Linux is based in CHOICE, and the laziness/incompetence
of distro maintainers is no excuse to allow systemd to run
rampant.

I tremendously appreciate the fortitude and competence
of the Gentoo and LFS maintainers for continuing to offer
true and viable CHOICE.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

On 9/8/23 14:35, Mike Easter wrote:

David W. Hodgins wrote:

It does much
more than just replace initd, and does so in consistent logical ways.

What is your 'position' on the arguments completely separate from the
init improvement; that many don't like that systemd likes to take
'control' of so many non-init responsibilities?

I guess the systemd idea is that just like init, if systemd can do it 'better', then that should be just fine.

However, it does make it progressively more difficult to 'do without' systemd, since it is SO much more than an init.

The kernel has a sort of monolpoly in linuxland; nothing else should
ever be allowed to come close to that status or it will start dictating
to kernel development. That's my view of it or of anything else with
like attributes (or aims), I use eight different web navigators.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

On Fri, 08 Sep 2023 14:35:18 -0400, Mike Easter <MikeE@ster.invalid> wrote:

David W. Hodgins wrote:

It does much
more than just replace initd, and does so in consistent logical ways.

What is your 'position' on the arguments completely separate from the
init improvement; that many don't like that systemd likes to take
'control' of so many non-init responsibilities?

I guess the systemd idea is that just like init, if systemd can do it 'better', then that should be just fine.

However, it does make it progressively more difficult to 'do without' systemd, since it is SO much more than an init.

All of the features included in systemd logically belong together. A system level
application can not start if the resources it needs are not available. Ensuring all the resources needed are available, and starting them if need be makes sense.
Same with mounts, and starting other applications at user login or on demand.

It's becoming harder not to use systemd because it simplifies things for applications that don't have to duplicate code to make sure things they need are available.

It's a system resource and application start/stop manager. While it's much more than a boot time daemon starter, everything in it makes sense to integrate into it, in my opinion.

The biggest problem with the way it was introduced was with the impression given (though never stated that I can remember) that it was only intended as
a replacement for initd.

Regards, Dave Hodgins

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

David W. Hodgins wrote:

The biggest problem with the way it was introduced was with the
impression given (though never stated that I can remember) that it
was only intended as a replacement for initd.

That was the /purpose/ of its introduction in 2010, a significantly needed purpose.

wp

They sought to surpass the efficiency of the init daemon in several
ways. They wanted to improve the software framework for expressing dependencies, to allow more processing to be done concurrently or in
parallel during system booting, and to reduce the computational
overhead of the shell.

By 2014, Poeterring saw it/ his idea/ more broadly as:

A system and service manager (manages both the system, by applying
various configurations, and its services)

A software platform (serves as a basis for developing other
software)

The glue between applications and the kernel (provides various
interfaces that expose functionalities provided by the kernel)

... and thus the 'mission creep' that became so upsetting to anti-systemd.

The anti-/s see it as 'Hitlerian' - going from party intelligence agent to
Nazi party orator to leader of the Nazi party to Chancellor of a
coalition to dictator and suppression of the other parties. Absolute
dictator, and not a benevolent one.

I would say systemd's success at that is based on what you said;

DH

While it's much more than a boot time daemon starter, everything in
it makes sense to integrate into it, in my opinion.

--
Mike Easter

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

On 09/09/2023 4:39, Mike Easter wrote:

David W. Hodgins wrote:

...Snipped...

The anti-/s see it as 'Hitlerian' - going from party intelligence agent to Nazi party orator to leader of the Nazi party to Chancellor of a
coalition to dictator and suppression of the other parties. Absolute dictator, and not a benevolent one.

I would say systemd's success at that is based on what you said;

DH

While it's much more than a boot time daemon starter, everything in it makes sense to integrate into it, in my opinion.

I invoke Godwn's Law!!

--
No Micro$oft products were used in the URLs above, or in preparing this message.
Recommended reading: http://www.catb.org/~esr/faqs/smart-questions.html#befor

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

"David W. Hodgins" <dwhodgins@nomail.afraid.org> writes:

The biggest problem with the way it was introduced was with the impression given (though never stated that I can remember) that it was only intended as a replacement for initd.

I’m not sure that’s historical. In the original announcement in 2010[1]
it already extends beyond the sysvinit responsibilities to include
socket activation, filesystem mounting, cgroup management, logging and
login accounting, with plans already in place for swap management,
session management and job scheduling. Anyone who imagined it had a
narrower focus wasn’t paying attention.

The name is a bit of a hint, really.

[1] http://0pointer.de/blog/projects/systemd.html

--
https://www.greenend.org.uk/rjk/

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

Henry Crun wrote:

Mike Easter wrote:

The anti-/s see it as 'Hitlerian' - going from party intelligence
agent to Nazi party orator to leader of the Nazi party to
Chancellor of a coalition to dictator and suppression of the other
parties. Absolute dictator, and not a benevolent one.

I invoke Godwn's Law!!

:-) I knew /someone/ was going to say that :-)

In 1994, Mike Godwin wrote a fun article for Wired mag.

https://www.wired.com/1994/10/godwin-if-2/

He stated that he introduced Godwin's law in 1990 as an experiment in memetics

The/His Wired article also provides various corollaries.

And, I tend to agree w/ DWH's position, even while bringing up some anti-systemd arguments.

My everyday driver is a systemd distro and I'm also a 'fan of' MX Linux.

--
Mike Easter

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

On Fri, 8 Sep 2023 18:39:54 -0700, Mike Easter wrote:

The glue between applications and the kernel (provides various
interfaces that expose functionalities provided by the kernel)

... and thus the 'mission creep' that became so upsetting to anti-systemd.

There never was any "mission creep."

The RedHat/IBM lackey Poettering had stated from the very outset that
systemd was/is intended to be the one-and-only interface between the
Linux kernel and user space.

Fortunately, after more than a DECADE, he and his cronies have failed
to achieve that goal. Note the great failure of their "bus1" initiative.

But the attempt will continue, and all those fools who support systemd
are complicit in this willful expropriation of GNU/Linux by a corporate
agenda.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

On Fri, 08 Sep 2023 18:47:30 -0400, David W. Hodgins wrote:

It's becoming harder not to use systemd because it simplifies things for applications that don't have to duplicate code to make sure things they need are available.

Oh, is it?

I have never ever used that P.O.S. known as "systemd" and have never ever experienced any issues whatsoever regarding software usage.

Would you care to give some specific examples where systemd is essential?

In my opinion, any GNU/Linux software developer that creates software that
is dependent on systemd should be ostracized from the FOSS community.

But that will potentially implicate all the RedHat/IBM lackeys at freedesktop.org.

People stand up! Purge these commercial influences from FOSS.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

On Sat, 09 Sep 2023 08:29:24 +0100, Richard Kettlewell wrote:

I’m not sure that’s historical. In the original announcement in 2010[1] it already extends beyond the sysvinit responsibilities to include
socket activation, filesystem mounting, cgroup management, logging and
login accounting, with plans already in place for swap management,
session management and job scheduling. Anyone who imagined it had a
narrower focus wasn’t paying attention.

Exactly! Great point!

But there are multitudes of fools and dupes who are NOT paying attention.

The stated goal of systemd, as has been explicated in formal documents,
is the takeover and control of all user-space possibilities so that
COMMERCIAL SOFTWARE, ala RedHat/IBM, will have an easy time.

Richard Stallman and Linus Torvalds did not create their masterpieces
so that grubbing commercial enterprises could easily benefit.

The sooner systemd is rejected by all the better.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

On Sat, 09 Sep 2023 11:56:06 -0400, Nux Vomica <nv@linux.rocks> wrote:

On Fri, 08 Sep 2023 18:47:30 -0400, David W. Hodgins wrote:

It's becoming harder not to use systemd because it simplifies things for
applications that don't have to duplicate code to make sure things they need >> are available.

Oh, is it?

I have never ever used that P.O.S. known as "systemd" and have never ever experienced any issues whatsoever regarding software usage.

Would you care to give some specific examples where systemd is essential?

In my opinion, any GNU/Linux software developer that creates software that
is dependent on systemd should be ostracized from the FOSS community.

But that will potentially implicate all the RedHat/IBM lackeys at freedesktop.org.

People stand up! Purge these commercial influences from FOSS.

I am not saying it's essential. This is linux so there are always other ways of doing things. It's easier for developers and distribution providers to use and debug, and provides tools for them that simplify things. Those tools are now used by some desktop environments such as gnome. While gnome can be used without
it, it's more difficult.

Where systems like sysv init are more difficult to work with are the edge cases such as having /usr/local on a remote file system and it containing things to be started at boot. That type of setup is used in some organizations. It can be done, but is not easy, and since sysv init doesn't support having overrides in a location that is not part of the package, updates must be done with more caution.

If you just use a linux install, the learning curve of the switch will be annoying
because the benefits are mostly not visible to end users. They do benefit from fewer bugs, and having easier ways to customize things, but if they don't need to customize things they will not see those as benefits.

For the package developers and distribution creators the benefits are much more obvious.

If redhat and/or ibm ever decide to make changes to systemd that annoy enough developers and distribution creators, then like any gpl licensed software, it will get forked.

Regards, Dave Hodgins

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: alt.os.linux.mint, alt.os.linux

On Sat, 09 Sep 2023 15:36:57 -0400, David W. Hodgins wrote:

If redhat and/or ibm ever decide to make changes to systemd that annoy enough developers and distribution creators, then like any gpl licensed software, it will get forked.

There is much more to GNU/Linux than developers and distro creators. There
are also USERS and a LOT of users do not desire systemd, GNOME, KDE, etc.

But the voice of these users is totally disregarded because the developers
and distro maintainers have the "we know best" attitude.

This is no delusion. The "we know best and let the user be damned" attitude
is perhaps the norm among developers and distro maintainers.

There are some few exceptions, with Gentoo being the most significant.

It it were not for Gentoo I would not be able to maintain my custom GNU/Linux system as I desire.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)