Re: What's the best linux distro?

[F Russell]

On Mon, 23 Aug 2021 17:15:30 -0500, chrisv wrote:

>
> The source is available for all of them. Total choice exists for all
> of them.
>

Totally wrong.

If one desires to install an alternative init system on Fedora
it will be impossible. The distro is locked into systemd
as are many others. Choice does not exist.

Unless a distro specifically allows allows alternative options
for an init system and other systems, as do Gentoo and LFS,
then it is NOT a free distro in the spirit of FOSS.

Again: most distros are not free and violate FOSS principles.

Why is this so?

Because distro maintainers are wannabe gods and do not
want to work for the people but want to work only to glorify
their self-image.

Anyone can create a distro. But creating a distro that maintains
freedom and allows choice is a great labor that few desire to
undertake.

The case is closed.


--

Systemd free. D.E. free.

Always and forever.

[Carlos E.R.]

On 24/08/2021 01.04, F Russell wrote:
> On Mon, 23 Aug 2021 17:15:30 -0500, chrisv wrote:
>
>>
>> The source is available for all of them. Total choice exists for all
>> of them.
>>
>
> Totally wrong.
>
> If one desires to install an alternative init system on Fedora
> it will be impossible. The distro is locked into systemd
> as are many others. Choice does not exist.
>
> Unless a distro specifically allows allows alternative options
> for an init system and other systems, as do Gentoo and LFS,
> then it is NOT a free distro in the spirit of FOSS.
>
> Again: most distros are not free and violate FOSS principles.

Sue them.

>
> Why is this so?
>
> Because distro maintainers are wannabe gods and do not
> want to work for the people but want to work only to glorify
> their self-image.
>
> Anyone can create a distro. But creating a distro that maintains
> freedom and allows choice is a great labor that few desire to
> undertake.

Then why don't you create your's, and see how many want it?

>
> The case is closed.
>
>


--

Cheers, Carlos.

[F Russell]

On Tue, 24 Aug 2021 08:10:49 +0200, Joerg Lorenz wrote:

>
> Your posting is utter bullshit.
>

Where is your proof or example to the contrary? Without
it your statement is purely empty.

But my claims remain unassailable -- except to shallow and
deluded idiots like you.

>
> *You* are the dogmatic Troll here.
>

"Troll" is one of the three words in your entire pathetic
vocabulary which in turn stems from your totally reflexive
intelligence.

You are the the blind that leads the lame.

[The Natural Philosopher]

Exactly.

Linux runs my desktop and my servers. Its a bit like windows, only
better, more stable, and cost nothing more than the effort of installing
it .

I dont give a rats ass about whether its FOSS or not. I care that it
works, it's as secure as the people who do pick over the source code can
make it, and well supported, up to and including porting third party
apps to it.

I got bored with the niceties of the moralities of software development
sometime in the 1990s

>>
>> The case is closed.
>>
>>
>
>


--

Climate Change: Socialism wearing a lab coat.

[SixOverFive]

Then go back to Debian 1.0 ....

Sorry, but reality is *complicated* ....

[SixOverFive]

On 08/24/2021 08:53 AM, F Russell wrote:
> On Tue, 24 Aug 2021 08:10:49 +0200, Joerg Lorenz wrote:
>
>>
>> Your posting is utter bullshit.
>>
>

> Where is your proof or example to the contrary? Without...


PEOPLE !!! What's going on here ???

There IS NO "best" Linux distro. What's "best" depends
ENTIRELY on what YOU plan to DO with it.

And there's also all the BSDs too ... resurrected
Amiga-OS, Plan-9, DOS, CP/M, emulated C-64s, OS-9...
it's ALL out there.

What do YOU want to do with it ?

My office PC has long been vanilla Debian, but I put
UServer on my servers, I *like* OpenSuse, my notebooks
are MX Linux, my favorite VM is Manjaro right now but
I spent a day and a half making FreeBSD run in Virtualbox
with XFCE and LXDE. It JUST DEPENDS. "Best" is a total
illusion.

OK ... DO note NONE of the abovementioned included
Winders ... while there may not be a "best" there
may indeed be a "WORST" :-)

Well, SOMEWHERE, I have a Winders 1.0 floppy ....
I even have a BYTE Magazine issue that reviews
it ! (clue, C64/128 windowing systems were BETTER
at the time).

[The Natural Philosopher]

On 25/08/2021 05:41, SixOverFive wrote:
> My office PC has long been vanilla Debian, but I put
> UServer on my servers, I *like* OpenSuse, my notebooks
> are MX Linux, my favorite VM is Manjaro right now but
> I spent a day and a half making FreeBSD run in Virtualbox
> with XFCE and LXDE. It JUST DEPENDS. "Best" is a total
> illusion.

I stick with one distro - Mint - because after messing around with many
others, that proved optimal for *my desktop needs.
It also runs on my local server. There is no need, but i am used to its
foibles, and I likewise have Ubuntu on my main vps.
Its popular and well supported.

I haven't got time to be fannying around learning a hundred different
arrangements of this or that

--
Future generations will wonder in bemused amazement that the early
twenty-first century’s developed world went into hysterical panic over a
globally average temperature increase of a few tenths of a degree, and,
on the basis of gross exaggerations of highly uncertain computer
projections combined into implausible chains of inference, proceeded to
contemplate a rollback of the industrial age.

Richard Lindzen

[Carlos E.R.]

On 25/08/2021 06.41, SixOverFive wrote:
> On 08/24/2021 08:53 AM, F Russell wrote:
>> On Tue, 24 Aug 2021 08:10:49 +0200, Joerg Lorenz wrote:
>>
>>>
>>> Your posting is utter bullshit.
>>>
>>
>> Where is your proof or example to the contrary?  Without...
>
>
> PEOPLE !!! What's going on here ???
>
> There IS NO "best" Linux distro. What's "best" depends
> ENTIRELY on what YOU plan to DO with it.

Oh, you must be new in these parts :-D

F Russell is one of the local trolls, a fanatic. Linux has to be his way
or it is evil. Systemd is evil. He likes a linux with basically no
services, no init system, tailored solely for him. What the
distributions do, a system that works as is for many people is evil.
Linux for the masses is evil, it has to be for hackers only and must be
difficult to use.

He is impossible.

Ah, and he hates personally the people that don't agree with him.

He now and then posts something like that here, and sets the follow up
to the advocacy group, to have a heated mad discussion there, I suppose.

--
Cheers, Carlos.

[Richard Kettlewell]

"Carlos E.R." <robin_...@es.invalid> writes:
> He now and then posts something like that here, and sets the follow up
> to the advocacy group, to have a heated mad discussion there, I
> suppose.

And you indulge him.

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

[The Natural Philosopher]

On 25/08/2021 10:56, Carlos E.R. wrote:
> Systemd is evil.

Well on that I do agree.

Goes completely against the principles of simplicity, transparency, and
not doing stuff that was working perfectly well already


--
For in reason, all government without the consent of the governed is the
very definition of slavery.

Jonathan Swift

[Carlos E.R.]

On 25/08/2021 14.10, The Natural Philosopher wrote:
> On 25/08/2021 10:56, Carlos E.R. wrote:
>> Systemd is evil.
>
> Well on that I do agree.
>
> Goes completely against the principles of simplicity, transparency, and
> not doing stuff that was working perfectly well already

Well, the later is arguable. Many were not happy, and initd had no
maintenance or hopes of improvements, so they were looking for a
replacement. It could have been systemd or something else, but there was
nothing else suitable. So many distributions changed.

To me, it works very well, I have no complains.


--
Cheers, Carlos.

[Dan Espen]

Systemd is unarguably faster.
The user interface is way simpler than initd.
The design of systemd makes a whole bunch of things possible that were
never possible with initd. (Like actually being able to access a
description of each service or getting a list of all services.)
IMO it's a bunch of know-nothings that continue with this systemd
bashing.

--
Dan Espen

[Carlos E.R.]

Or easily modifying a distribution provided service file in a manner
that it survives updates.

--
Cheers, Carlos.

[The Natural Philosopher]

The problem is that it obsoleted a lot of tried and tested ways of
bringing stuff online.

For example, before systemd my laptops had hard wired NFS mounts to my
server in /etc/fstab.

The nfs layer came up after networking by wifi, the mounts were dine
later and all was well.

To years later systemd had broken all the scripts and Poeterring just
shrugged and said 'not my problem' and a further two years elapsed
before people actually managed to rewrite all their initialization to
conform to HIS opaque standards. And there are still messes around -
stuff that boots up 'oddly'.

Having scripts execute in an orderly fashion according to alphanumerical
order was perfectly OK

You looked at what systems your layer depended on, amd made sure you
were later in the list.


--
"In our post-modern world, climate science is not powerful because it is
true: it is true because it is powerful."

Lucas Bergkamp

[The Natural Philosopher]

On 25/08/2021 13:53, Dan Espen wrote:
> "Carlos E.R." <robin_...@es.invalid> writes:
>
>> On 25/08/2021 14.10, The Natural Philosopher wrote:
>>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>>> Systemd is evil.
>>> Well on that I do agree.
>>> Goes completely against the principles of simplicity, transparency,
>>> and not doing stuff that was working perfectly well already
>>
>> Well, the later is arguable. Many were not happy, and initd had no
>> maintenance or hopes of improvements, so they were looking for a
>> replacement. It could have been systemd or something else, but there
>> was nothing else suitable. So many distributions changed.
>>
>> To me, it works very well, I have no complains.
>
> Systemd is unarguably faster.

who needs to boot a server faster? when you only take it doiwn every few
months and its down for at least a few hours while you mess aroujnd with it.

> The user interface is way simpler than initd.

Rubbish

> The design of systemd makes a whole bunch of things possible that were
> never possible with initd. (Like actually being able to access a
> description of each service or getting a list of all services.)

If you dont know what services are running on a machine or what they do,
you shouldn't be messing with them in the first place.

> IMO it's a bunch of know-nothings that continue with this systemd
> bashing.
>

No, it's people who have found stuff that worked perfectly well was
broken by systemd, and took ages to fix.




--
When plunder becomes a way of life for a group of men in a society, over
the course of time they create for themselves a legal system that
authorizes it and a moral code that glorifies it.

Frédéric Bastiat

[Bobbie Sellers]

Well certainly I suffer from vast areas of ignorance but GNU/Linux
without systemd was/is faster and is/was more reliable
than when I used systemd on Mageia 3.1. I started with Mandriva
2006. Before that I used Amiga computers from Commodore.

Systemd may be handy for people who must maintain a lot of
machines as in an Enterprise situation but for me on my laptops
other solutions are available.

PCLinuxOS 64 is my present choice, MX Linux is quite good, Devuan
is Debian without systemd, AntiX can be run without systemd and if
you have really old systems perhaps 4M Linux, or Slack by
Puppy drawing on the vast repositories of Slackware.

One simple job to start the computer but others to kick
off each service.

bliss - 'Nearly any fool can use a Linux computer. Many do.' After all
here I am...

[Robert Latest]

The Natural Philosopher wrote:
> For example, before systemd my laptops had hard wired NFS mounts to my
> server in /etc/fstab.
>
> The nfs layer came up after networking by wifi, the mounts were dine
> later and all was well.
>
> To years later systemd had broken all the scripts

I have a few NFS mounts to my local NAS on my systemd-running laptop without
problems. Just a couple of lines in /etc/fstab, no scripts required.

> Having scripts execute in an orderly fashion according to alphanumerical
> order was perfectly OK

I find writing custom services is also a lot easier in systemd. Until once when
I tried to write a timer unit. That was crap.

[Richard Kettlewell]

"Carlos E.R." <robin_...@es.invalid> writes:
> The Natural Philosopher wrote:
>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>> Systemd is evil.
>>
>> Well on that I do agree.
>>
>> Goes completely against the principles of simplicity, transparency,
>> and not doing stuff that was working perfectly well already
>
> Well, the later is arguable. Many were not happy, and initd had no
> maintenance or hopes of improvements, so they were looking for a
> replacement. It could have been systemd or something else, but there
> was nothing else suitable. So many distributions changed.

“Nothing else suitable” isn’t true at all. Disregarding the rather large
set of toy/experimental inits, Upstart was already deployed in
production Linux systems. Launchd would have needed porting for use on
Linux but was (and still is) happily managing large numbers of real
computers.

systemd as it is today has a lot of features they’re missing, certainly,
but applying the same effort to them that was in fact put into systemd
doesn’t seem like a totally unrealistic option, had anyone chosen to
pursue it.

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

[Dan Espen]

The Natural Philosopher <t...@invalid.invalid> writes:

> On 25/08/2021 13:36, Carlos E.R. wrote:
>> On 25/08/2021 14.10, The Natural Philosopher wrote:
>>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>>> Systemd is evil.
>>>
>>> Well on that I do agree.
>>>
>>> Goes completely against the principles of simplicity, transparency,
>>> and not doing stuff that was working perfectly well already
>> Well, the later is arguable. Many were not happy, and initd had no
>> maintenance or hopes of improvements, so they were looking for a
>> replacement. It could have been systemd or something else, but there
>> was nothing else suitable. So many distributions changed.
>> To me, it works very well, I have no complains.
>>
> The problem is that it obsoleted a lot of tried and tested ways of
> bringing stuff online.

initd still works on my Fedora system:

https://docs.fedoraproject.org/en-US/fedora/rawhide/system-administrators-guide/infrastructure-services/Services_and_Daemons/

--
Dan Espen

[Dan Espen]

The Natural Philosopher <t...@invalid.invalid> writes:

> Having scripts execute in an orderly fashion according to
> alphanumerical order was perfectly OK

Bull, it was a f***ing mess.

As I said, systemd faster, easier to understand, way more capable.

--
Dan Espen

[Dan Espen]

The Natural Philosopher <t...@invalid.invalid> writes:

> On 25/08/2021 13:53, Dan Espen wrote:
>> "Carlos E.R." <robin_...@es.invalid> writes:
>>
>>> On 25/08/2021 14.10, The Natural Philosopher wrote:
>>>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>>>> Systemd is evil.
>>>> Well on that I do agree.
>>>> Goes completely against the principles of simplicity, transparency,
>>>> and not doing stuff that was working perfectly well already
>>>
>>> Well, the later is arguable. Many were not happy, and initd had no
>>> maintenance or hopes of improvements, so they were looking for a
>>> replacement. It could have been systemd or something else, but there
>>> was nothing else suitable. So many distributions changed.
>>>
>>> To me, it works very well, I have no complains.
>> Systemd is unarguably faster.
>
> who needs to boot a server faster? when you only take it doiwn every
> few months and its down for at least a few hours while you mess
> aroujnd with it.

Oh here we go down that blind alley.
I said it was faster, whether you appreciate it or not is irrelevant.

>> The user interface is way simpler than initd.
>
> Rubbish

Oh really? Do you know anything about system design?
One file type, no forest of links, no utility to reconstruct the soft
link forest, a bunch of intuitive keywords in one file type.

>> The design of systemd makes a whole bunch of things possible that were
>> never possible with initd. (Like actually being able to access a
>> description of each service or getting a list of all services.)
>
> If you dont know what services are running on a machine or what they
> do, you shouldn't be messing with them in the first place.

That's the best you can come up with? You've never encountered a
service you are not running that you think you might need to run.
Pathetic.

>> IMO it's a bunch of know-nothings that continue with this systemd
>> bashing.
>>
> No, it's people who have found stuff that worked perfectly well was
> broken by systemd, and took ages to fix.

So sad.

--
Dan Espen

[jjb]

On 25-08-2021 17:47, Richard Kettlewell wrote:
> systemd as it is today has a lot of features they’re missing, certainly,
> but applying the same effort to them that was in fact put into systemd
> doesn’t seem like a totally unrealistic option, had anyone chosen to
> pursue it.
>

But they didn't...

[Dan Espen]

Bobbie Sellers <bl...@mouse-potato.com> writes:

> On 8/25/21 05:53, Dan Espen wrote:
>> "Carlos E.R." <robin_...@es.invalid> writes:
>>
>>> On 25/08/2021 14.10, The Natural Philosopher wrote:
>>>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>>>> Systemd is evil.
>>>> Well on that I do agree.
>>>> Goes completely against the principles of simplicity, transparency,
>>>> and not doing stuff that was working perfectly well already
>>>
>>> Well, the later is arguable. Many were not happy, and initd had no
>>> maintenance or hopes of improvements, so they were looking for a
>>> replacement. It could have been systemd or something else, but there
>>> was nothing else suitable. So many distributions changed.
>>>
>>> To me, it works very well, I have no complains.
>> Systemd is unarguably faster.
>> The user interface is way simpler than initd.
>> The design of systemd makes a whole bunch of things possible that were
>> never possible with initd. (Like actually being able to access a
>> description of each service or getting a list of all services.)
>> IMO it's a bunch of know-nothings that continue with this systemd
>> bashing.
>
> Well certainly I suffer from vast areas of ignorance but
> GNU/Linux without systemd was/is faster and is/was more
> reliable

There is nothing about the design of systemd that would make your
services run any different than they did before systemd, and there is no
doubt that startup and shutdown are faster with systemd. So I reject
your claim of slower.

There are facets of the systemd design that make it more reliable than
an init system that only covers startup and shutdown. If systemd had
teething problems, those are all behind us now.

> than when I used systemd on Mageia 3.1. I started with Mandriva
> 2006. Before that I used Amiga computers from Commodore.

An appeal to expertise? Back in the 90s I was using Solaris at home.
Somewhere in the 00s I transitioned to Linux, Mandriva was my first at
home install.

I believe I was on Fedora when systemd was introduced. I did an
upgrade, not an install. I was running for a few days before I became
aware of the new init system.

--
Dan Espen

[Dan Espen]

So now it's "some other init system could be better".

That's no way to bash systemd, you're supposed to attack the character
of the lead developer.

--
Dan Espen

[Richard Kettlewell]

I’m not sure how you get to that from the above. Just pointing out that
Carlos’s characterisation of the situation at the time decisions were
being made wasn’t accurate.

> That's no way to bash systemd, you're supposed to attack the character
> of the lead developer.

I’m not trying to bash anything here.

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

[Bobbie Sellers]

Because they see the featuritis of systemd as a complication
to be avoided rather than sought. One program one function
is the old GNU design criteria.

bliss - boots & runs a Pretty Cool Linux Operating System aka pclinuxos.

No to systemd.

[The Natural Philosopher]

On 25/08/2021 16:38, Robert Latest wrote:
> The Natural Philosopher wrote:
>> For example, before systemd my laptops had hard wired NFS mounts to my
>> server in /etc/fstab.
>>
>> The nfs layer came up after networking by wifi, the mounts were dine
>> later and all was well.
>>
>> To years later systemd had broken all the scripts
>
> I have a few NFS mounts to my local NAS on my systemd-running laptop without
> problems. Just a couple of lines in /etc/fstab, no scripts required.

now yes. two years ago, forget it

>
>> Having scripts execute in an orderly fashion according to alphanumerical
>> order was perfectly OK
>
> I find writing custom services is also a lot easier in systemd. Until once when
> I tried to write a timer unit. That was crap.
>

--
If you tell a lie big enough and keep repeating it, people will
eventually come to believe it. The lie can be maintained only for such
time as the State can shield the people from the political, economic
and/or military consequences of the lie. It thus becomes vitally
important for the State to use all of its powers to repress dissent, for
the truth is the mortal enemy of the lie, and thus by extension, the
truth is the greatest enemy of the State.

Joseph Goebbels

[Carlos E.R.]

On 25/08/2021 15.31, The Natural Philosopher wrote:
> On 25/08/2021 13:53, Dan Espen wrote:
>> "Carlos E.R." <robin_...@es.invalid> writes:
>>
>>> On 25/08/2021 14.10, The Natural Philosopher wrote:
>>>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>>>> Systemd is evil.
>>>> Well on that I do agree.
>>>> Goes completely against the principles of simplicity, transparency,
>>>> and not doing stuff that was working perfectly well already
>>>
>>> Well, the later is arguable. Many were not happy, and initd had no
>>> maintenance or hopes of improvements, so they were looking for a
>>> replacement. It could have been systemd or something else, but there
>>> was nothing else suitable. So many distributions changed.
>>>
>>> To me, it works very well, I have no complains.
>>
>> Systemd is unarguably faster.
>
> who needs to boot a server faster? when you only take it doiwn every few
> months and its down for at least a few hours while you mess aroujnd with
> it.

Me. I want booting my machines faster. Even the servers.

>
>> The user interface is way simpler than initd.
>
> Rubbish

In your opinion.



--
Cheers, Carlos.

[Carlos E.R.]

On 25/08/2021 17.38, Robert Latest wrote:
> The Natural Philosopher wrote:
>> For example, before systemd my laptops had hard wired NFS mounts to my
>> server in /etc/fstab.
>>
>> The nfs layer came up after networking by wifi, the mounts were dine
>> later and all was well.
>>
>> To years later systemd had broken all the scripts
>
> I have a few NFS mounts to my local NAS on my systemd-running laptop without
> problems. Just a couple of lines in /etc/fstab, no scripts required.

Same here.

>> Having scripts execute in an orderly fashion according to alphanumerical
>> order was perfectly OK
>
> I find writing custom services is also a lot easier in systemd. Until once when
> I tried to write a timer unit. That was crap.

You can use cron instead, you have both. At least I do.


--
Cheers, Carlos.

[Carlos E.R.]

On 25/08/2021 17.47, Richard Kettlewell wrote:
> "Carlos E.R." <robin_...@es.invalid> writes:
>> The Natural Philosopher wrote:
>>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>>> Systemd is evil.
>>>
>>> Well on that I do agree.
>>>
>>> Goes completely against the principles of simplicity, transparency,
>>> and not doing stuff that was working perfectly well already
>>
>> Well, the later is arguable. Many were not happy, and initd had no
>> maintenance or hopes of improvements, so they were looking for a
>> replacement. It could have been systemd or something else, but there
>> was nothing else suitable. So many distributions changed.
>
> “Nothing else suitable” isn’t true at all. Disregarding the rather large
> set of toy/experimental inits, Upstart was already deployed in
> production Linux systems. Launchd would have needed porting for use on
> Linux but was (and still is) happily managing large numbers of real
> computers.

Well, openSUSE was not about to use Ubuntu method :-p

> systemd as it is today has a lot of features they’re missing, certainly,
> but applying the same effort to them that was in fact put into systemd
> doesn’t seem like a totally unrealistic option, had anyone chosen to
> pursue it.
>

That's the point, no one chose to pursue it.

--
Cheers, Carlos.

[Aragorn]

On 25.08.2021 at 10:39, Bobbie Sellers scribbled:

> On 8/25/21 09:17, jjb wrote:
>
> > On 25-08-2021 17:47, Richard Kettlewell wrote:
> >
> >> systemd as it is today has a lot of features they’re missing,
> >> certainly, but applying the same effort to them that was in fact
> >> put into systemd doesn’t seem like a totally unrealistic option,
> >> had anyone chosen to pursue it.
> >
> > But they didn't...
>
> Because they see the featuritis of systemd as a complication
> to be avoided rather than sought.

Among other things that have already been addressed and that I shan't
get into anymore, the kernel was offering many advanced possibilities to
userspace that userspace wasn't making use of -- cgroups. And to the
best of my knowledge, even today, only one other init system beside
systemd offers support for that, i.e. Gentoo's openrc.

I won't contradict that there have been some ego issues involved --
e.g. Lennart asking the GNOME developers to make systemd's logind a
hard dependency of GDM -- but systemd really IS progress. It makes for
a consistent and transparent yet modular system daemon that replaces a
poorly maintained and archaic init system that blindly started services
with virtually no control over restarting them or managing them once
they were running, leaving disparate ways of managing them.

> One program one function is the old GNU design criteria.

Actually, that goes back to UNIX in general, and you might be sad to
hear that the GNU find utility sins badly against that principle, not
to mention GNU emacs.

> bliss - boots & runs a Pretty Cool Linux Operating System aka
> pclinuxos.
>
> No to systemd.

I've been running Manjaro -- which is also cool, and very clean under
the hood -- since late April 2019. Manjaro is based upon Arch and uses
systemd. The three main differences with Arch proper are...

1. It has a graphical installer.

2. It has many graphical utilities for installiong software and
configuring the system, including a Manjaro-specific package
manager that can be used as an alternative to pacman while still
maintaining full compatibility with it, and that can be used
graphically or from the command line.

3. It's a curated rolling release, so instead of there being updates
every day, updates are tested, bundled together and pushed out on
average twice a month, barring urgent updates to certain packages
for stability or security reasons.

My computer boots up in 10 seconds from pressing Enter at the GRUB menu
to the SDDM login screen, and from a successful login to a fully
loaded Plasma desktop takes about 2 seconds longer.

Given that kernels are updated with new patches without forcing one to
adopt a higher kernel version, it is necessary to reboot the machine at
least twice a month, or whenever a major update is rolled out. Updates
are always announced on dedicated threads at the forum, together with a
list of possible gotchas and how to deal with them.

(But of course, the n00bs won't look at those announcements. They're
even too lazy to do a forum search on whatever issue they may be
encountering with an update, and then we as moderators have to merge
their threads and point them at the announcement thread for a relief
from their oh-so-tragic burdens. <rolling eyes>)

--
With respect,
= Aragorn =

[Jimmy Johnson]

On 08/24/2021 06:19 AM, The Natural Philosopher wrote:

> I dont give a rats ass about whether its FOSS or not. I care that it
> works, it's as secure as the people who do pick over the source code can
> make it, and well supported, up to and including porting third party
> apps to it.
>
> I got bored with the niceties of the moralities of software development
> sometime in the 1990s

> Climate Change: Socialism wearing a lab coat.

I don't like you. There is nothing nice or good about Corporations or
the rich owning Open Source or murder, you're nothing less than a $atan
loving fool with no place to run and no place to hide, so don't be
surprised when your day comes. And saying you don't believe won't help,
you've been warned. LOL
--
James, Son of John

[Jimmy Johnson]

On 08/25/2021 02:56 AM, Carlos E.R. wrote:

> F Russell is one of the local trolls, a fanatic. Linux has to be his way
> or it is evil. Systemd is evil. He likes a linux with basically no
> services, no init system, tailored solely for him. What the
> distributions do, a system that works as is for many people is evil.
> Linux for the masses is evil, it has to be for hackers only and must be
> difficult to use.

Besides slowing a system down, spying and selling user info, what did
systemd bring to the table? Systemd was a lie from the very beginning
when they said over and over it was faster. Our $ystem is full of lie$.

[Robert Riches]

Jimmy Johnson, your post sounds like a threat of violence or
other harm against TNP. I hope he prints a copy of your post to
show to law enforcement if that becomes necessary.

--
Robert Riches
spamt...@jacob21819.net
(Yes, that is one of my email addresses.)

[Robert Latest]

Of course I ended up using cron for the timed thing. But for processes I want
to run permanently in the background, and that I want restarted upon a crash,
with proper logging, without having to bother how to properly demonize them,
systemd beats init.

[Robert Latest]

Jimmy Johnson wrote:
> Besides slowing a system down, spying and selling user info, what did
> systemd bring to the table?

Please elaborate on the "spying" bit. Never heard of that.

[Richard Kettlewell]

The Natural Philosopher <t...@invalid.invalid> writes:

> Having scripts execute in an orderly fashion according to
> alphanumerical order was perfectly OK
>
> You looked at what systems your layer depended on, amd made sure you
> were later in the list.

If you’re rolling your own, that’s fine, although doing it by hand
rather than getting a computer to work it out for you seems rather
strange.

If you’re building a distribution, with a wide range of possible end
user configurations, it’s not sustainable; you need something that will
work it out dynamically, based on the local configuration. IIRC Debian
had long standing bugs in this area until adopting systemd.

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

[The Natural Philosopher]

Well I must say that is very antisocial of you.
Did you get lost in someone else's Moral Maze?

--
The biggest threat to humanity comes from socialism, which has utterly
diverted our attention away from what really matters to our existential
survival, to indulging in navel gazing and faux moral investigations
into what the world ought to be, whilst we fail utterly to deal with
what it actually is.

[The Natural Philosopher]

Don't be silly. This is Usenet. Hate speech is all part of the fun, and
probably de rigeur.

Golly don't you remember the flame wars, way back when?
And the final solution - the Kill File.

[The Natural Philosopher]

And not a few due to adopting systemd...

[Carlos E.R.]

On 26/08/2021 10.49, The Natural Philosopher wrote:
> On 26/08/2021 09:09, Richard Kettlewell wrote:
>> The Natural Philosopher <t...@invalid.invalid> writes:
>>> Having scripts execute in an orderly fashion according to
>>> alphanumerical order was perfectly OK
>>>
>>> You looked at what systems your layer depended on, amd made sure you
>>> were later in the list.
>>
>> If you’re rolling your own, that’s fine, although doing it by hand
>> rather than getting a computer to work it out for you seems rather
>> strange.
>>
>> If you’re building a distribution, with a wide range of possible end
>> user configurations, it’s not sustainable; you need something that will
>> work it out dynamically, based on the local configuration. IIRC Debian
>> had long standing bugs in this area until adopting systemd.
>>
> And not a few due to adopting systemd...

Which are being worked on or at least have hopes, not as the previous
system.

--
Cheers, Carlos.

[Carlos E.R.]

Me neither. Prove it. The source code is available to any. It is just FUD.


--
Cheers, Carlos.

[Carlos E.R.]

On 26/08/2021 10.31, J.O. Aho wrote:
> On 26/08/2021 06.13, RonB wrote:
>> On 2021-08-26, jrg <jeff.g...@att.net> wrote:

>>> On 8/25/21 4:10 AM, Richard Kettlewell wrote:
>>>> "Carlos E.R." <robin_...@es.invalid> writes:

>>>>> He now and then posts something like that here, and sets the follow up
>>>>> to the advocacy group, to have a heated mad discussion there, I
>>>>> suppose.
>>>>
>>>> And you indulge him.
>>>>
>>> cross post deleted
>>>
>>> +1
>>
>> My slrn scorefile deletes all crossposts automatically. I find that
>> cleans
>> up a newsgroup faster than anything else.
>>
>
> Don't know which groups you subscribe for, but for me, most shit is
> multi posts and not cross posted, not sure it's possible with G2-web
> interface.

That way you have half the posts of the thread in one group and half in
the other, so that some people will not see yours.

--
Cheers, Carlos.

[Carlos E.R.]

I recently migrated a cron job to a systemd timer (for leafnode). I
didn't know how to do that, but fortunately the packager had already
done that, just that the timer was not enabled.

Then I wanted to change its timing, and I tried to do that, which
failed. So I asked in the openSUSE mail list, got an exact answer within
a day, and the issue was solved.

As I mentioned, the original file is intact, but my installation
contains a modification that will survive updates. It is a nice method
for distributions, easy to maintain.

--
Cheers, Carlos.

[Rich]

In comp.os.linux.misc The Natural Philosopher <t...@invalid.invalid> wrote:
> On 26/08/2021 04:47, Jimmy Johnson wrote:
>> On 08/24/2021 06:19 AM, The Natural Philosopher wrote:
>>
>>> I dont give a rats ass about whether its FOSS or not. I care that it
>>> works, it's as secure as the people who do pick over the source code can
>>> make it, and  well supported,  up to and including porting third party
>>> apps to it.
>>>
>>> I got bored with the niceties of the moralities of software development
>>> sometime in the 1990s
>>
>>> Climate Change: Socialism wearing a lab coat.
>>
>> I don't like you. There is nothing nice or good about Corporations or
>> the rich owning Open Source or murder, you're nothing less than a $atan
>> loving fool with no place to run and no place to hide, so don't be
>> surprised when your day comes. And saying you don't believe won't help,
>> you've been warned. LOL
>
> Well I must say that is very antisocial of you.
> Did you get lost in someone else's Moral Maze?

JJ's response would be the "hmm, you don't think like me, therefore you
must be _____" (pick your choice of "evil incarnate" for the blank).

Far too many recently have fallen into a trap of forgetting (or
possibly never understanding in the first place) that different people
can hold different opinions and it is not the end of the world. I.e.,
some have forgotten to "agree to disagree".

[Jasen Betts]

On 2021-08-26, Rich <ri...@example.invalid> wrote:
> In comp.os.linux.misc The Natural Philosopher <t...@invalid.invalid> wrote:
>> On 26/08/2021 04:47, Jimmy Johnson wrote:
>>> On 08/24/2021 06:19 AM, The Natural Philosopher wrote:
>>>
>>>> I dont give a rats ass about whether its FOSS or not. I care that it
>>>> works, it's as secure as the people who do pick over the source code can
>>>> make it, and  well supported,  up to and including porting third party
>>>> apps to it.
>>>>
>>>> I got bored with the niceties of the moralities of software development
>>>> sometime in the 1990s
>>>
>>>> Climate Change: Socialism wearing a lab coat.
>>>
>>> I don't like you. There is nothing nice or good about Corporations or
>>> the rich owning Open Source or murder, you're nothing less than a $atan
>>> loving fool with no place to run and no place to hide, so don't be
>>> surprised when your day comes. And saying you don't believe won't help,
>>> you've been warned. LOL
>>
>> Well I must say that is very antisocial of you.
>> Did you get lost in someone else's Moral Maze?
>
> JJ's response would be the "hmm, you don't think like me, therefore you
> must be _____" (pick your choice of "evil incarnate" for the blank).

It was a response in kind to the original post, and entirely justified.

Correct me ig I'm wrong, bufrom from your response it seems that you also
prefer superstition to science.

If you don't want your idocies mocked don't post them to usenet.

--
Jasen.

[The Natural Philosopher]

Well in the end some people - due mainly to insecurity one supposes -
move from purveyors of perspectives to Holders of the One True View.

Philosophy teaches one true fact, that there are no 'true' facts. There
are only inductive propositions, whose truth content is indecidable.

The problem with socialism per se, is that it is comprised of such
people. The tenets of Marxism and so called socialism are pronounced as
truth, not as informative perspectives.

It's as bad as science, which at least has the excuse that its
propositions are demonstrably not-false....yet anyway.


--
Truth welcomes investigation because truth knows investigation will lead
to converts. It is deception that uses all the other techniques.

[The Natural Philosopher]

No, I don't think he is Believer in Catastrophic Anthropogenic Global
Warming..

>
> If you don't want your idocies mocked don't post them to usenet.
>

Indeed.



--
To ban Christmas, simply give turkeys the vote.

[Aragorn]

On 25.08.2021 at 21:02, Jimmy Johnson scribbled:

> On 08/25/2021 02:56 AM, Carlos E.R. wrote:
>
> > F Russell is one of the local trolls, a fanatic. Linux has to be
> > his way or it is evil. Systemd is evil. He likes a linux with
> > basically no services, no init system, tailored solely for him.
> > What the distributions do, a system that works as is for many
> > people is evil. Linux for the masses is evil, it has to be for
> > hackers only and must be difficult to use.
>

> Besides slowing a system down, [...

My PCLinuxOS system with SysVinit booted up from hitting Enter at the
GRUB menu to a character-mode console login prompt in about 50 seconds.

My Manjaro system with systemd boots up from hitting Enter at the GRUB
menu to the SDDM login screen in 10 seconds.

Now, I could be grossly wrong, but I am still firmly convinced that 10
seconds is less than 50 seconds.

> ...] spying and selling user info [...

Links, please? To verifiable sources, please -- not to some "I hate
systemd" website.

>...], what did systemd bring to the table?

Uniformity and consistency in configuration, a failsafe mechanism for
the starting and managing of daemons, faster boot times, isolation and
containerization of processes, consistency in the naming convention of
attached peripherals -- e.g. drives, network interfaces -- as well as
autodetection of the EFI partition, the root partition and the swap
partition without needing any entries in /etc/fstab on drives with a
GUID partition table, et al.

AND... unlike with SysVinit, the code is actively being maintained.

[Dan Espen]

Still have me on ignore? Good post, finally some reason on the subject.

--
Dan Espen

[Aragorn]

On 26.08.2021 at 13:12, Dan Espen scribbled:

Gosh, no -- that was a long time (and two desktop computers) ago.

> Good post, finally some reason on the subject.

I used to be anti-systemd too, because I couldn't see its merits, and I
guess I was just jerking my knee, like so many others. (I am only
human (I think) and not above making mistakes. ;))

But the more I looked into it, the more systemd simply started making
sense. It's progress. And now that I am running a distribution that
uses systemd [*], I don't see myself ever going back to SysVinit.


[*] Manjaro started off as systemd-based-only, given that it's a
derivative of Arch, but it has in the past also also supported
openrc. The openrc support was however dropped a number of years
ago, which is why a fork of Manjaro was created, called Artix.

[Stéphane CARPENTIER]

Le 25-08-2021, SixOverFive <hae274b.net> a écrit :
> On 08/24/2021 08:53 AM, F Russell wrote:
>
> What do YOU want to do with it ?

That part's easy to answer. The only thing he does is putting his name
in the source files before compiling the kernel. Like that he believe he
has his own kernel home made. It takes him a week: just in time to
repeat the process with the new version of the kernel. During the
compilation time he comes here to display his lack of knowledge.

--
Si vous avez du temps à perdre :
https://scarpet42.gitlab.io

[Stéphane CARPENTIER]

Le 25-08-2021, The Natural Philosopher <t...@invalid.invalid> a écrit :
>
> If you dont know what services are running on a machine or what they do,
> you shouldn't be messing with them in the first place.

OK, you don't know what you are talking about. There are legitimate
criticisms against systemd. This is not one of them. By design it's way
easier to know what services are running.

[Stéphane CARPENTIER]

Le 25-08-2021, Carlos E.R. <robin_...@es.invalid> a écrit :
>
> Me. I want booting my machines faster. Even the servers.

In fact, I'd say it's even more important for the servers. Because when
you need to restart your server, you know you have a lot of people
waiting for it to be available again.

[Stéphane CARPENTIER]

Le 25-08-2021, Dan Espen <dan1...@gmail.com> a écrit :
> The Natural Philosopher <t...@invalid.invalid> writes:
>
>> Having scripts execute in an orderly fashion according to
>> alphanumerical order was perfectly OK
>

> Bull, it was a f***ing mess.
>
> As I said, systemd faster, easier to understand, way more capable.

Agreed. I don't understand the need to have to chose the order of the
services. Why should I care to now if my printer must be started before
or after my webcam?

[Stéphane CARPENTIER]

Le 26-08-2021, Jimmy Johnson <Ji...@disposable.invalid> a écrit :
>
> Besides slowing a system down, spying and selling user info, what did
> systemd bring to the table? Systemd was a lie from the very beginning
> when they said over and over it was faster. Our $ystem is full of lie$.

s/Our Sytem/My message/

[Carlos E. R.]

On 26/08/2021 22.57, Stéphane CARPENTIER wrote:
> Le 25-08-2021, Carlos E.R. <robin_...@es.invalid> a écrit :
>>
>> Me. I want booting my machines faster. Even the servers.
>
> In fact, I'd say it's even more important for the servers. Because when
> you need to restart your server, you know you have a lot of people
> waiting for it to be available again.

True.

Equally important is not having to adjust everything on a new server to
have it working, but just install and go. More time available for the
real issues.

--
Cheers,
Carlos E.R.

[Carlos E. R.]

On 25/08/2021 18.21, Dan Espen wrote:
> Bobbie Sellers <bl...@mouse-potato.com> writes:

>> On 8/25/21 05:53, Dan Espen wrote:
>>> "Carlos E.R." <robin_...@es.invalid> writes:

>>>> On 25/08/2021 14.10, The Natural Philosopher wrote:
>>>>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>>>>> Systemd is evil.
>>>>> Well on that I do agree.
>>>>> Goes completely against the principles of simplicity, transparency,
>>>>> and not doing stuff that was working perfectly well already
>>>>
>>>> Well, the later is arguable. Many were not happy, and initd had no
>>>> maintenance or hopes of improvements, so they were looking for a
>>>> replacement. It could have been systemd or something else, but there
>>>> was nothing else suitable. So many distributions changed.
>>>>
>>>> To me, it works very well, I have no complains.
>>> Systemd is unarguably faster.
>>> The user interface is way simpler than initd.
>>> The design of systemd makes a whole bunch of things possible that were
>>> never possible with initd. (Like actually being able to access a
>>> description of each service or getting a list of all services.)
>>> IMO it's a bunch of know-nothings that continue with this systemd
>>> bashing.
>>
>> Well certainly I suffer from vast areas of ignorance but
>> GNU/Linux without systemd was/is faster and is/was more
>> reliable
>
> There is nothing about the design of systemd that would make your
> services run any different than they did before systemd, and there is no
> doubt that startup and shutdown are faster with systemd. So I reject
> your claim of slower.

Unless, as is often the case, the distribution also upgraded everything,
services and apps and libraries, with more features. Every upgrade of
software bring better features, more code, bigger and slower. Not
because of switching to systemd, but because that's the nature of
software in real life.

Besides that, systemd is faster.


--
Cheers,
Carlos E.R.

[The Natural Philosopher]

On 26/08/2021 21:53, Stéphane CARPENTIER wrote:
> Le 25-08-2021, The Natural Philosopher <t...@invalid.invalid> a écrit :
>>
>> If you dont know what services are running on a machine or what they do,
>> you shouldn't be messing with them in the first place.
>
> OK, you don't know what you are talking about. There are legitimate
> criticisms against systemd. This is not one of them. By design it's way
> easier to know what services are running.
>

It's a shame you cant tell a criticism from a refutation of a claimed
advantage

[The Natural Philosopher]

On 26/08/2021 21:57, Stéphane CARPENTIER wrote:
> Le 25-08-2021, Carlos E.R. <robin_...@es.invalid> a écrit :
>>
>> Me. I want booting my machines faster. Even the servers.
>
> In fact, I'd say it's even more important for the servers. Because when
> you need to restart your server, you know you have a lot of people
> waiting for it to be available again.
>

Since its already been down for 4 hours for upgrades, what difference
does 5 minutes make?

I mean all this is so ACADEMIC. So ArtStudent™ hand wavy.

IME the only thing that delays booting is the necessity of fsking the
file system and I dint think systemd will improve that.

[Carlos E. R.]

On 27/08/2021 03.04, The Natural Philosopher wrote:
> On 26/08/2021 21:57, Stéphane CARPENTIER wrote:
>> Le 25-08-2021, Carlos E.R. <robin_...@es.invalid> a écrit :
>>>
>>> Me. I want booting my machines faster. Even the servers.
>>
>> In fact, I'd say it's even more important for the servers. Because when
>> you need to restart your server, you know you have a lot of people
>> waiting for it to be available again.
>>
> Since its already been down for 4 hours for upgrades, what difference
> does 5 minutes make?

You get charged per second of downtime.

>
> I mean all this is so ACADEMIC. So ArtStudent™ hand wavy.
>
> IME the only thing that delays booting is the necessity of fsking the
> file system and I dint think systemd will improve that.

Not the only thing...

--
Cheers,
Carlos E.R.

[J.O. Aho]

On 26/08/2021 19.41, Aragorn wrote:

> I used to be anti-systemd too, because I couldn't see its merits, and I
> guess I was just jerking my knee, like so many others. (I am only
> human (I think) and not above making mistakes. ;))

Sure systemd has come with some merits, but not everything has been
great, as it do "eat up" other projects, which has lead to a number of
serious security issues with a high exportability as it's been the pid 1
that been using them.

Other init systems do handle the "benefit" that has been preached as the
main reason to use systemd, the boot up time. For me from Grub to Kodi
takes just 25 sec on OpenRC, would I want to monitor services, sure I
could use s6 and still get the same bootup time.

I do think that systemd has done a good job on possible service
isolation, but it's doable on other init systems too, but requires a lot
more hands on.

For me, who has work quite many years with systems that had to have high
security and prompt updates of high security vulnerabilities without too
much impact on down time, systemd has been a bad option compared with
say SysVinit.

> [*] Manjaro started off as systemd-based-only, given that it's a
> derivative of Arch, but it has in the past also also supported
> openrc. The openrc support was however dropped a number of years
> ago, which is why a fork of Manjaro was created, called Artix.

Artix claims to be a based on Arch and it do use the Arch repositories
and not the Manajro ones and you also notice that some packages are
released faster on Artix than on Manjaro.

--

//Aho

[Marc Haber]

The Natural Philosopher <t...@invalid.invalid> wrote:

>On 26/08/2021 21:57, Stéphane CARPENTIER wrote:
>> Le 25-08-2021, Carlos E.R. <robin_...@es.invalid> a écrit :
>>>
>>> Me. I want booting my machines faster. Even the servers.
>>
>> In fact, I'd say it's even more important for the servers. Because when
>> you need to restart your server, you know you have a lot of people
>> waiting for it to be available again.
>>
>Since its already been down for 4 hours for upgrades, what difference
>does 5 minutes make?

Wrong argumentation. With the right distribution, an upgrade doesn't
take four hours, and the services are not down the entire time.

But, if my server takes 5 minutes for its POST, it doesn't matter when
the OS boot takes. On the other hand, when I fire up containers that
only live for a minute, every second of bootup time counts.

I don't need to like systemd to see its advantages, hence I'm using
it. The pain to migrate to one of the systemd-less distributions is
way higher and the support surely worse.

Greetings
Marc
--
-------------------------------------- !! No courtesy copies, please !! -----
Marc Haber | " Questions are the | Mailadresse im Header
Mannheim, Germany | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 621 72739834

[Marc Haber]

Stéphane CARPENTIER <s...@fiat-linux.fr> wrote:
>Agreed. I don't understand the need to have to chose the order of the
>services. Why should I care to now if my printer must be started before
>or after my webcam?

Imagine you have a DNS server for your IPv6 VPN users, and the VPN
server must be reachable over IPv6. That needs network, and the IPv6
setup to be finished (no more tentative addresses, that takes a few
seconds), before starting the VPN server (or it wouldnt be able to
bind the IPv6 address). Then the VPN server starts, a new interface
appears, the IPv6 address gets configured on the new interface (again
taking a few seconds for duplicate address detection to finish and the
address to lose its tentative status). Only if that's finished, the
DNS server can start.

It's all about order. And if the order isn't right, you either end up
without VPN, or without DNS server. That matters very much.

Took me half a day to get this right with systemd. But I was a newbie
back then, it would probably only be an hour today.

[Stéphane CARPENTIER]

Exactly. And you forget everything the user don't need and is not aware
of. But which are there in case he would need them. It's impressive the
number of packages installed on a popular distro just to avoid the
end-user the need to know what he needs or not.

> Besides that, systemd is faster.

Don't know about it. At home it's fast and doesn't use lot of
ressources. But I'm not able to compare with SysV.

[Stéphane CARPENTIER]

Le 27-08-2021, Marc Haber <mh+usene...@zugschl.us> a écrit :
> Stéphane CARPENTIER <s...@fiat-linux.fr> wrote:
>>Agreed. I don't understand the need to have to chose the order of the
>>services. Why should I care to now if my printer must be started before
>>or after my webcam?
>
> Imagine you have a DNS server for your IPv6 VPN users, and the VPN
> server must be reachable over IPv6. That needs network, and the IPv6
> setup to be finished (no more tentative addresses, that takes a few
> seconds), before starting the VPN server (or it wouldnt be able to
> bind the IPv6 address). Then the VPN server starts, a new interface
> appears, the IPv6 address gets configured on the new interface (again
> taking a few seconds for duplicate address detection to finish and the
> address to lose its tentative status). Only if that's finished, the
> DNS server can start.
>
> It's all about order. And if the order isn't right, you either end up
> without VPN, or without DNS server. That matters very much.

Of course some services must be started before others. And systemd
provide a way to do it. But the possibility to order some services
doesn't force me to chose the order every service. You can say the
logs must be started before everything else for example.

In some cases, as a socket is created, a service depending on another
one can be started before the first one because the information will
be cached and sent as soon as the service is available.

> Took me half a day to get this right with systemd. But I was a newbie
> back then, it would probably only be an hour today.

I never said systemd is easy to understand. It would be a lie. But
saying SysV is easy would be a lie too.

At first when something wrong happened and I discovered systemd, I
found the war against systemd. As I had issues with systemd I wanted
to join the war and looked harder. And the more I looked, the more I
found systemd interesting. There are real issues with systemd, it does
things it shouldn't. But nothing else does them and as it start the
process it makes sense to let it do them.

[Dan Espen]

Stéphane CARPENTIER <s...@fiat-linux.fr> writes:

> I never said systemd is easy to understand. It would be a lie. But
> saying SysV is easy would be a lie too.

Funny you should say systemd is not easy to understand.
I can't imagine a simpler system. One file type for configuration
with the simplest syntax possible, and just a few keywords to know about.

What could be eliminated or simplified?


--
Dan Espen

[Dan Espen]

You should know it's faster everywhere based on it's design.
It starts/stops services in parallel instead of serially.
That's ALWAYS going to be faster.

--
Dan Espen

[Stéphane CARPENTIER]

The boot process is faster, it's obvious. But the boot process takes
only the few first seconds of the time I'm using my computer. Once I'm
using my computer, I have no idea about the way to compare the
efficiency of SysV vs. systemd. Mostly when sytemd does stuff which where
lacking in SysV.

[Dan Espen]

Stéphane CARPENTIER <s...@fiat-linux.fr> writes:

> Le 28-08-2021, Dan Espen <dan1...@gmail.com> a écrit :
>> Stéphane CARPENTIER <s...@fiat-linux.fr> writes:
>>
>>> Le 26-08-2021, Carlos E. R. <robin_...@es.invalid> a écrit :
>>>> Besides that, systemd is faster.
>>>
>>> Don't know about it. At home it's fast and doesn't use lot of
>>> ressources. But I'm not able to compare with SysV.
>>
>> You should know it's faster everywhere based on it's design.
>> It starts/stops services in parallel instead of serially.
>> That's ALWAYS going to be faster.
>
> The boot process is faster, it's obvious. But the boot process takes
> only the few first seconds of the time I'm using my computer. Once I'm
> using my computer, I have no idea about the way to compare the
> efficiency of SysV vs. systemd. Mostly when sytemd does stuff which where
> lacking in SysV.

There should be no difference.
Both schemes start services. Those services will run at the same speed
regardless of how they are started.

--
Dan Espen

[Carlos E. R.]

The initd system used in openSUSE also started services in parallel.

--
Cheers,
Carlos E.R.

[Aragorn]

On 28.08.2021 at 23:08, Carlos E. R. scribbled:

So does the one in PCLinuxOS, but it's still SysVinit. It has simply
been patched with parallel execution, and while it's definitely a lot
faster than the original SysVinit, it doesn't come anywhere near
systemd.

[Richard Kettlewell]

Dan Espen <dan1...@gmail.com> writes:
> Stéphane CARPENTIER <s...@fiat-linux.fr> writes:
>
>> I never said systemd is easy to understand. It would be a lie. But
>> saying SysV is easy would be a lie too.
>
> Funny you should say systemd is not easy to understand. I can't
> imagine a simpler system. One file type for configuration with the
> simplest syntax possible, and just a few keywords to know about.

A simple config file syntax doesn’t make a system easy to understand.
There are thousands of lines of man pages containing over 160 options,
with an implementation exceeding half a million lines of C. It’s not
badly documented, but it’s a complex system and it’s silly to pretend
there’s no learning curve here.

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

[Dan Espen]

Where is the user interface? With initd, it's that pile of scripts,
each one a custom creation and a forest of symbolic links.
Where is the learning curve compared to systemd's simple config file
syntax?

No one said ANYTHING about "no learning curve". Systemd presents the
simplest user interface possible, especially compared to initd.

--
Dan Espen

[Richard Kettlewell]

Dan Espen <dan1...@gmail.com> writes:
> Richard Kettlewell <inv...@invalid.invalid> writes:
>> Dan Espen <dan1...@gmail.com> writes:
>>> Stéphane CARPENTIER <s...@fiat-linux.fr> writes:
>>>
>>>> I never said systemd is easy to understand. It would be a lie. But
>>>> saying SysV is easy would be a lie too.
>>>
>>> Funny you should say systemd is not easy to understand. I can't
>>> imagine a simpler system. One file type for configuration with the
>>> simplest syntax possible, and just a few keywords to know about.
>>
>> A simple config file syntax doesn’t make a system easy to understand.
>> There are thousands of lines of man pages containing over 160 options,
>> with an implementation exceeding half a million lines of C. It’s not
>> badly documented, but it’s a complex system and it’s silly to pretend
>> there’s no learning curve here.
>
> Where is the user interface? With initd, it's that pile of scripts,
> each one a custom creation and a forest of symbolic links.
> Where is the learning curve compared to systemd's simple config file
> syntax?

I have no idea why you’re focussing on the config file syntax, since
there’s a lot more to systemd (or any nontrivial init) than that.

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

[Carlos E. R.]

On 29/08/2021 12.52, Bud Frede wrote:
> "Carlos E.R." <robin_...@es.invalid> writes:
>

>> On 26/08/2021 10.31, J.O. Aho wrote:
>>> On 26/08/2021 06.13, RonB wrote:
>>>> On 2021-08-26, jrg <jeff.g...@att.net> wrote:

>>>>> On 8/25/21 4:10 AM, Richard Kettlewell wrote:
>>>>>> "Carlos E.R." <robin_...@es.invalid> writes:

>>>>>>> He now and then posts something like that here, and sets the follow up
>>>>>>> to the advocacy group, to have a heated mad discussion there, I
>>>>>>> suppose.
>>>>>>
>>>>>> And you indulge him.
>>>>>>
>>>>> cross post deleted
>>>>>
>>>>> +1
>>>>
>>>> My slrn scorefile deletes all crossposts automatically. I find that
>>>> cleans
>>>> up a newsgroup faster than anything else.
>>>>
>>> Don't know which groups you subscribe for, but for me, most shit is
>>> multi posts and not cross posted, not sure it's possible with G2-web
>>> interface.
>>
>> That way you have half the posts of the thread in one group and half
>> in the other, so that some people will not see yours.
>
> I tend to leave the groups alone when replying to a cross-posted
> article, although I do remove things like the advocacy cesspools.

Certainly, me too. The two groups listed in this post are precisely just
that.

--
Cheers,
Carlos E.R.

[Dan Espen]

It's the only visible user interface.

--
Dan Espen

[Richard Kettlewell]

systemd.syntax is tiny indeed, but if that’s all someone reads they will
still have no idea how to write or interpret a unit file.

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

[Carlos E. R.]

On 29/08/2021 13.11, Bud Frede wrote:

>
> I'm kind of surprised that some people are still whinging about
> it. Systemd is what most distros are using and that's that.
>
> It isn't perfect, but it works and it has useful features that SysV init
> did not - features that I wanted and still value.
>
> Actually, I lied. I'm not surprised that people are still whinging about
> Systemd. People love to argue and beat a drum. It's plastic-tip
> vs. metal-tip shoelaces all over again. :-)
>
> (When people were arguing about something that was irrelevant,
> meaningless, or already a forgone conclusion, my father used to say
> "plastic-tip vs. metal-tip" to signify that it wasn't worth discussing.)

You are right :-)


--
Cheers,
Carlos E.R.

[Carlos E. R.]

On 29/08/2021 13.26, Bud Frede wrote:
> "Carlos E.R." <robin_...@es.invalid> writes:

>> On 25/08/2021 15.31, The Natural Philosopher wrote:

>>> On 25/08/2021 13:53, Dan Espen wrote:
>>>> "Carlos E.R." <robin_...@es.invalid> writes:
>>>>

>>>>> On 25/08/2021 14.10, The Natural Philosopher wrote:
>>>>>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>>>>>> Systemd is evil.
>>>>>> Well on that I do agree.
>>>>>> Goes completely against the principles of simplicity, transparency,
>>>>>> and not doing stuff that was working perfectly well already
>>>>>
>>>>> Well, the later is arguable. Many were not happy, and initd had no
>>>>> maintenance or hopes of improvements, so they were looking for a
>>>>> replacement. It could have been systemd or something else, but there
>>>>> was nothing else suitable. So many distributions changed.
>>>>>
>>>>> To me, it works very well, I have no complains.
>>>>
>>>> Systemd is unarguably faster.

>>> who needs to boot a server faster? when you only take it doiwn every
>>> few months and its down for at least a few hours while you mess
>>> aroujnd with it.

>>
>> Me. I want booting my machines faster. Even the servers.
>>
>

> Especially the servers sometimes. If something is down, or if load
> skyrockets and it needs to autoscale, I very much want servers that boot
> fast.

>
>>>
>>>> The user interface is way simpler than initd.

>>> Rubbish
>>
>> In your opinion.
>
> SysV init can be simple to use, for simple tasks. However, when you need
> something complex, it gets rather baroque, and no two people set it up
> the same way. That makes for a whole lot of complexity for a typical
> computer.
>
> Systemd involves some more complex upfront design (by the systemd
> developers), but in operation it is simple and consistent. It remains
> simple and consistent even when your needs are more complex.
>

That's it, absolutely.

--
Cheers,
Carlos E.R.

[Carlos E. R.]

On 29/08/2021 13.32, Bud Frede wrote:
> "Carlos E.R." <robin_...@es.invalid> writes:

>> On 25/08/2021 17.47, Richard Kettlewell wrote:
>>> "Carlos E.R." <robin_...@es.invalid> writes:
>>>> The Natural Philosopher wrote:
>>>>> On 25/08/2021 10:56, Carlos E.R. wrote:
>>>>>> Systemd is evil.
>>>>>
>>>>> Well on that I do agree.
>>>>>
>>>>> Goes completely against the principles of simplicity, transparency,
>>>>> and not doing stuff that was working perfectly well already
>>>>
>>>> Well, the later is arguable. Many were not happy, and initd had no
>>>> maintenance or hopes of improvements, so they were looking for a
>>>> replacement. It could have been systemd or something else, but there
>>>> was nothing else suitable. So many distributions changed.

>>> “Nothing else suitable” isn’t true at all. Disregarding the rather
>>> large
>>> set of toy/experimental inits, Upstart was already deployed in
>>> production Linux systems. Launchd would have needed porting for use on
>>> Linux but was (and still is) happily managing large numbers of real
>>> computers.
>>
>> Well, openSUSE was not about to use Ubuntu method :-p
>>
>
> Redhat was using upstart. That's arguably more persuasive than anything
> that SuSE was doing or not doing. :-)

I don't remember what was said at the time about upstart, but basically
it was "no" :-)

Me, at the time I preferred initd scripts. Easy to change /one/ script
that misbehaves. But of course, each script was different with its own
learning curve. I did not see the reason to change. But change came and
now I came to accept/like it.

Doesn't mean I /know/ systemd. When I have to do something I often have
to ask.

--
Cheers,
Carlos E.R.

[Carlos E. R.]

On 29/08/2021 14.09, Bud Frede wrote:
> Bobbie Sellers <bl...@mouse-potato.com> writes:

...
> I'm not going to claim that Systemd was as brilliant an improvement as
> the Amiga was in 1985. However, for some people and some uses, it's
> pretty brilliant. :-) I find it to be a very useful tool.
>
> https://opensource.com/article/20/10/cgroups
>
> This article talks about something that I find to be pretty brilliant
> about Systemd. It's may not be meaningful for a desktop system that
> only sees light duty, but for home lab or development use, or for
> servers, it's very intriguing.

cgroups, yes. I have used it and it is indeed useful and simple with
systemd.

>
>
>>
>> Systemd may be handy for people who must maintain a lot of
>> machines as in an Enterprise situation but for me on my laptops
>> other solutions are available.
>
> It's fine that you prefer other solutions on your laptops, and fine that
> you use them. Use whatever you prefer. :-)
>

Absolutely :-)

--
Cheers,
Carlos E.R.

[Stéphane CARPENTIER]

Le 29-08-2021, Dan Espen <dan1...@gmail.com> a écrit :

> Richard Kettlewell <inv...@invalid.invalid> writes:
>
>> I have no idea why you’re focussing on the config file syntax, since
>> there’s a lot more to systemd (or any nontrivial init) than that.
>
> It's the only visible user interface.

What about that?

journalctl systemctl systemd-cryptenroll systemd-hwdb systemd-nspawn
systemd-stdio-bridge systemd-analyze systemd-delta systemd-id128
systemd-path systemd-sysext systemd-ask-password systemd-detect-virt
systemd-inhibit systemd-repart systemd-sysusers systemd-cat
systemd-dissect systemd-machine-id-setup systemd-resolve
systemd-tmpfiles systemd-cgls systemd-escape systemd-mount systemd-run
systemd-tty-ask-password-agent systemd-cgtop systemd-firstboot
systemd-notify systemd-socket-activate systemd-umount

They don't count as visible user interface?

Systemd is not only about launching services, but managing them too.
Unlike SysV which was only starting the system and didn't care of the
processes once launched. The best things brought by systemd are not only
a few seconds at the startup. The few seconds are just a good side
effect.

For example, I'd say systemd was the first to use the cgroups, even if
they were available in the kernel a long time before. And to understand
the way the cgroups are managed by systemd, it's not enough to look at
the config file syntax.

Another example, some people complain because the logs are by default in
a binary format. But now the search in the logs is really great. Using
this new way isn't related with the config file syntax.

[J.O. Aho]

On 29/08/2021 13.42, Bud Frede wrote:
> Aragorn <thor...@telenet.be> writes:
>
>> On 25.08.2021 at 21:02, Jimmy Johnson scribbled:
>>
>>> On 08/25/2021 02:56 AM, Carlos E.R. wrote:
>>>
>>>> F Russell is one of the local trolls, a fanatic. Linux has to be
>>>> his way or it is evil. Systemd is evil. He likes a linux with
>>>> basically no services, no init system, tailored solely for him.
>>>> What the distributions do, a system that works as is for many
>>>> people is evil. Linux for the masses is evil, it has to be for
>>>> hackers only and must be difficult to use.
>>>
>>> Besides slowing a system down, [...
>>
>> My PCLinuxOS system with SysVinit booted up from hitting Enter at the
>> GRUB menu to a character-mode console login prompt in about 50 seconds.
>>
>> My Manjaro system with systemd boots up from hitting Enter at the GRUB
>> menu to the SDDM login screen in 10 seconds.
>>
>> Now, I could be grossly wrong, but I am still firmly convinced that 10
>> seconds is less than 50 seconds.
>>
>
> One of my friends did some testing way back when systemd was first being
> used, in Arch IIRC what he tested with. He found that systemd did indeed
> speed up the boot process vs. SysV init.

One thing that systemd does which none of the other init systems (as far
as I know) do set some things to be executed first when you need it, so
by default nfs mounts aren't mounted at boot up, they are mounted first
when you access the nfs mount, which makes the first access extremely
slow. In this case you save some time during the initial boot, but
detecting access and the mount will actually make it take longer time.

Sure you can reconfigure the unit file in systemd to mount the nfs at
boot time and you increase the boot up time, but on the other hand the
user will not get the long delay when first accessing the share.


Just compare the boot up time isn't enough, you have to take a look if
everything is really started by systemd or if it's left it to later
waiting for user action, as that extra wait time for the user waiting
for the service to be started up should be added to the boot up time to
make it a more comparable with other init systems.

For me the boot time ain't that important as I seldom reboot my machines
that thanks to live patching of the kernel that works fine in most cases
(for me).


The issue I have with systemd is the somewhat frequent security issues
and in most cases those tend to be able to give the attacker the most
privileged access of the system outside the kernel.

--

//Aho

[Carlos E. R.]

On 29/08/2021 14.59, Stéphane CARPENTIER wrote:
> Le 29-08-2021, Dan Espen <dan1...@gmail.com> a écrit :
>> Richard Kettlewell <inv...@invalid.invalid> writes:
>>
>>> I have no idea why you’re focussing on the config file syntax, since
>>> there’s a lot more to systemd (or any nontrivial init) than that.
>>
>> It's the only visible user interface.
>
> What about that?
>
> journalctl systemctl systemd-cryptenroll systemd-hwdb systemd-nspawn
> systemd-stdio-bridge systemd-analyze systemd-delta systemd-id128
> systemd-path systemd-sysext systemd-ask-password systemd-detect-virt
> systemd-inhibit systemd-repart systemd-sysusers systemd-cat
> systemd-dissect systemd-machine-id-setup systemd-resolve
> systemd-tmpfiles systemd-cgls systemd-escape systemd-mount systemd-run
> systemd-tty-ask-password-agent systemd-cgtop systemd-firstboot
> systemd-notify systemd-socket-activate systemd-umount
>
> They don't count as visible user interface?

No :-)

I only use a handful of commands or maybe less. systemctl and
journalctl, that's it. :-)



--
Cheers,
Carlos E.R.

[Richard Kettlewell]

"J.O. Aho" <us...@example.net> writes:
> One thing that systemd does which none of the other init systems (as
> far as I know) do set some things to be executed first when you need
> it, so by default nfs mounts aren't mounted at boot up, they are
> mounted first when you access the nfs mount, which makes the first
> access extremely slow. In this case you save some time during the
> initial boot, but detecting access and the mount will actually make it
> take longer time.
>
> Sure you can reconfigure the unit file in systemd to mount the nfs at
> boot time and you increase the boot up time, but on the other hand the
> user will not get the long delay when first accessing the share.

That’s always been possible with automounting, since before Linux even
existed. The only new thing here is that it’s integrated into init.

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

[Aragorn]

On 29.08.2021 at 07:11, Bud Frede scribbled:

> Dan Espen <dan1...@gmail.com> writes:
>
> > Systemd is unarguably faster.

> > The user interface is way simpler than initd.

> > The design of systemd makes a whole bunch of things possible that
> > were never possible with initd. (Like actually being able to
> > access a description of each service or getting a list of all
> > services.) IMO it's a bunch of know-nothings that continue with
> > this systemd bashing.

>
> I'm kind of surprised that some people are still whinging about
> it. Systemd is what most distros are using and that's that.

Haters gonna hate, and some people simply have an irrational fear of
change, even when "change" equals "better". <shrug>

[The Natural Philosopher]

On 29/08/2021 13:36, Carlos E. R. wrote:
> I did not see the reason to change. But change came and
> now I came to accept/like it.
>

> Doesn't mean I/know/ systemd. When I have to do something I often have
> to ask.

Its like postscript and X windows. Designed to be able to do anything,
but finally tamed by putting in shims to ensue that there is a way
simpler API that only uses the useful 1% of the features...

But they are standards and we are stuck with them


--
"Women actually are capable of being far more than the feminists will
let them."

[Dan Espen]

"No idea"? Hardly.
Have you looked?
They seem pretty straight forward to me.

--
Dan Espen

[Dan Espen]

Stéphane CARPENTIER <s...@fiat-linux.fr> writes:

> Le 29-08-2021, Dan Espen <dan1...@gmail.com> a écrit :
>> Richard Kettlewell <inv...@invalid.invalid> writes:
>>
>>> I have no idea why you’re focussing on the config file syntax, since
>>> there’s a lot more to systemd (or any nontrivial init) than that.
>>
>> It's the only visible user interface.
>
> What about that?
>
> journalctl systemctl systemd-cryptenroll systemd-hwdb systemd-nspawn
> systemd-stdio-bridge systemd-analyze systemd-delta systemd-id128
> systemd-path systemd-sysext systemd-ask-password systemd-detect-virt
> systemd-inhibit systemd-repart systemd-sysusers systemd-cat
> systemd-dissect systemd-machine-id-setup systemd-resolve
> systemd-tmpfiles systemd-cgls systemd-escape systemd-mount systemd-run
> systemd-tty-ask-password-agent systemd-cgtop systemd-firstboot
> systemd-notify systemd-socket-activate systemd-umount
>
> They don't count as visible user interface?

Yes it does.
Of course that's a command line tool and not that hard to learn, but I'd
expect the distro to be loaded with tools making use of the command line
tool an exception rather than the rule.

> Systemd is not only about launching services, but managing them too.
> Unlike SysV which was only starting the system and didn't care of the
> processes once launched. The best things brought by systemd are not only
> a few seconds at the startup. The few seconds are just a good side
> effect.
>
> For example, I'd say systemd was the first to use the cgroups, even if
> they were available in the kernel a long time before. And to understand
> the way the cgroups are managed by systemd, it's not enough to look at
> the config file syntax.

True, but then I'd consider cgroups distinct from systemd.

> Another example, some people complain because the logs are by default in
> a binary format. But now the search in the logs is really great. Using
> this new way isn't related with the config file syntax.

The journal and systemd the init system are distinct.
The journal brings lots of new capabilities while leaving all the old
/var/log stuff in place.

--
Dan Espen

[Richard Kettlewell]

Yes, I have written a number.

> They seem pretty straight forward to me.

Go on then, explain how to write a unit file that actually does
something useful, using only information in `man systemd.syntax`.

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

[Carlos E. R.]

I don't. I look at existing files.

--
Cheers,
Carlos E.R.

[Dan Espen]

Here is the existing emacs.service file on my system:

[Unit]
Description=Emacs: the extensible, self-documenting text editor

[Service]
Type=forking
ExecStart=/usr/bin/emacs --daemon
ExecStop=/usr/bin/emacsclient --eval "(kill-emacs)"
Restart=always

[Install]
WantedBy=default.target

If I wanted to add arguments to emacs I'd change the ExecStart line.

No reference to the man page is required.

--
Dan Espen

[Richard Kettlewell]

Cargo-culting one example is a bit different to understanding the
system. If you wanted to do anything even slightly different you’d need
to read the manual. Which isn’t unreasonable, but it’s a long way beyond
knowing the config file surface syntax.

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

[Dan Espen]

I was asked for an example, I gave it.
Not enough for you?
I'm wasting my time.

Cargo cult:

A cargo cult is an indigenist millenarian belief system in which
adherents perform rituals which they believe will cause a more
technologically advanced society to deliver goods.

WTF?

--
Dan Espen

[The Natural Philosopher]

Well yes, I was a bit puzzled by the use of that term. I assumed it was
sort of like 'using something not invented or understood here, but
delivered by someone who knew what they were doing'
Which is stretching the definition of cargo cult past breaking part.

'cribbing' would have been simpler to say.

It's a tradition that goes back to Bletchley Park, cryptography and the
dawn of the computer age.

We used to call it hacking - copy everything and then hack it about
till it does what you want. A very fast way to develop similar code
since you only need to really understand the bit that needs changing.



--
Gun Control: The law that ensures that only criminals have guns.

[Richard Kettlewell]

You were asked for an example that used only knowledge of the syntax,
which is what you’ve spent the last few posts claiming is all you
need. Turns out what you really meant by “knowledge of the syntax” was
“an example to blindly copy, with no particular understanding of the
wider system”. Perhaps the conversation would have gone faster if you’d
said what you meant at the start.

> I'm wasting my time.
>
> Cargo cult:
>
> A cargo cult is an indigenist millenarian belief system in which
> adherents perform rituals which they believe will cause a more
> technologically advanced society to deliver goods.
>
> WTF?

The practice of copying things without understanding why they do what
they do. In my experience it’s a commonly used term among techies, at
least among the ones who do make the effort to fully understand the
technology they use, when faced with the output of people who guess and
bluff.

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

[The Natural Philosopher]

On 31/08/2021 10:13, Richard Kettlewell wrote:
> In my experience it’s a commonly used term among techies, at
> least among the ones who do make the effort to fully understand the
> technology they use, when faced with the output of people who guess and
> bluff.

Never heard it used that way, ever.


--
For in reason, all government without the consent of the governed is the
very definition of slavery.

Jonathan Swift

[Eli the Bearded]

In comp.os.linux.misc, The Natural Philosopher <t...@invalid.invalid> wrote:
> On 31/08/2021 10:13, Richard Kettlewell wrote:
> > In my experience it’s a commonly used term among techies, at
> > least among the ones who do make the effort to fully understand the
> > technology they use, when faced with the output of people who guess and
> > bluff.

Yes. Sometimes used in reference to people who lean on Stack Overflow
too much.

> Never heard it used that way, ever.

Congratuations, now you have.

Some further reading:

https://en.wikipedia.org/wiki/Cargo_cult_programming

https://en.wikipedia.org/wiki/Cargo_cult_science

Elijah
------
the "see also" sections of those two pages are also good reading

[Branimir Maksimovic]

On 2021-08-25, Carlos E.R. <robin_...@es.invalid> wrote:
> On 25/08/2021 06.41, SixOverFive wrote:
>> On 08/24/2021 08:53 AM, F Russell wrote:
>>> On Tue, 24 Aug 2021 08:10:49 +0200, Joerg Lorenz wrote:
>>>
>>>>
>>>> Your posting is utter bullshit.
>>>>
>>>
>>> Where is your proof or example to the contrary?  Without...
>>
>>
>> PEOPLE !!! What's going on here ???
>>
>> There IS NO "best" Linux distro. What's "best" depends
>> ENTIRELY on what YOU plan to DO with it.
>
> Oh, you must be new in these parts :-D

>
> F Russell is one of the local trolls, a fanatic. Linux has to be his way

he is either unemployed or he is payed for doing this.



--
bmaxa now listens rock.mp3