Why systemd?

I’ve been reading about how linux works and all the docs speak of init as a shell script. And I’ve noticed, by a little listening that everyone is switching to systemd as the init program and I wanted to know why? What does systemd have that init.sh lacks? Linux is world famous for multiple choices (ever counted the number of text editors?) and now all the sudden it seems that systemd is the only choice, which is strange. A web search turns up only information about a debate going on in ubuntu and debian and conflicts of interest thereof but no hard facts.

Not a request for help. Please do not reply. It will be moved to General Chit-Chat…

Moved from Applications. This thread is now open for posting.

On 2014-06-13 22:46, deano ferrari wrote:
>
> Moved from Applications. This thread is now open for posting.

Sigh. The moving notice did not appear in my Thunderbird till after I
posted. I guess my answer is now lost, or worse, appeared as a new
thread in applications. You can delete it if so. Sorry.

I’ll repost my answer here, then.

On 2014-06-13 21:56, ballsystemlord wrote:
>
> I’ve been reading about how linux works and all the docs speak of init
> as a shell script. And I’ve noticed, by a little listening that everyone
> is switching to systemd as the init program and I wanted to know why?

No.

openSUSE switched, several releases ago, to systemd. If you are using
openSUSE, you have no choice in the matter: we all use systemd and there
is no alternative (or yes: roll your own distro, or switch to a
different distro).

There have been tons of discussions about this in the openSUSE mail
lists and forums, over several years.

I don’t see this as a an application question, perhaps the thread should
be moved to chit-chat.


Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)

I’m not objecting. I’m curious. You don’t ask a stupid question you don’t get a stupid answer. And I did notice but I deemed myself to much a newbie at the time grasp the understanding behind the decision. And I’ve been thinking about it so I decided to research it and came up blank. So, I’ll ask.

Well, that’s not really true. There are a number of distros for which the systemd vs something else is still a live civil war. You mention Ubuntu and Debian (and one argument is whether, if Debian take one particular choice, does Ubuntu really have an independant choice, but that’s by the by). Slackware, for example, has quite a number of users who are virulently anti-systemd, but, in the end, most of their users are of the mind that ‘Our BDFL will take a wise and enlightened decision, and, as we like wise and enlightened decisions we will live with that, whatever it finally turns out to be’. IIRC, Gentoo is one of the distros that seems to be going a different way, but that might just be brain fail on my part.

Theoretically a new init system has several advantages. (A slightly different question is to what extent are the theoretical advantages delivered without counterbalancing disadvantages which render the whole project either worthless or of so minimal a worth that it isn’t worth the bother - the issue here is that you only find the answer to that, if you commit to the new init system, and that doesn’t render settling these religious wars any easier, because no one goes down all the different paths, assesses the results, and makes the logical decision that results from the evidence.)

In general, if you want to start lots of different processes in parallel, and you want a general solution to doing that, you’ll have to change from the traditional approach (and, presumably, break stuff until some tweaks have been made to adapt).

In theory, systemd can not only start stuff up in parallel, but, with a bit of thought, you don’t even have to do too much ‘fine tuning’ or ‘trial and error’ in order to get it to work. (You’ll have to ask someone else to what extent this part of the process is delivered in practice, and to what extent they still have to ‘tweak’, but I do get the impression that, somewhere in the world, there are people who have this expertise, that it isn’t necessarily that difficult for them and I’m not one of them.)

Now, the end result of all this cleverness ought to be a faster start up. One of the differences behind the religious wars is that some people rarely re-start their Linux boxes, and to them a faster start-up is worth almost exactly nothing, and it certainly isn’t worth the transition to some difficult-to-understand or partially documented process in order to get it. Another is that some people just can’t stand the people behind it…

Seems that some improvement in search-fu is required rotfl!

https://www.linux.com/learn/tutorials/524577-here-we-go-again-another-linux-init-intro-to-systemd
https://en.wikipedia.org/wiki/Systemd

were both in the first five of a simple search on ‘systemd’. I’m sure that I could do better, if I really tried.

On 2014-06-13 23:26, ballsystemlord wrote:
>
> I’m not objecting. I’m curious. You don’t ask a stupid question you
> don’t get a stupid answer. And I did notice but I deemed myself to much
> a newbie at the time grasp the understanding behind the decision. And
> I’ve been thinking about it so I decided to research it and came up
> blank. So, I’ll ask.

You can read here a recent thread about it.

<http://forums.opensuse.org/showthread.php?t=498290>


Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)

Here we go again :sigh

“And now for something completely different”… :D. You might not notice “faster start up” [e.g. on older laptops], once we have btrfs and Snapper as a default (13.2?).

In my opinion here is a good comparison vs upstart and sysinitv:
[http://0pointer.de/blog/projects/why.html

http://0pointer.de/blog/projects/the-biggest-myths.html](http://0pointer.de/blog/projects/why.html)

My options are always my own. Maybe not the best idea to refer to pages written and maintained by Lennart Poettering in this case.

Who are I’m to judge. None. Sorry if I named a person here in the forum.

Regards

Why not ? :slight_smile: He sure is one of the people that knows the most about systemd.

You are right! Why not? Gives a thought of the developers goal. Lennart P is not all evil. I like Pulse audio a lot.

Regards

Didn’t we just have this thread? :sarcastic:

http://www.networkworld.com/article/2175826/software/linus-torvalds-suspends-key-linux-developer.html

Sievers was banned by kernel maintainer Linus Torvalds on Wednesday for failing to address an issue that caused systemd to interact with the Linux kernel in negative ways. Specifically, the command line entry “debug” ran both the base kernel’s debugging routine and that of systemd, potentially flooding some systems.
Torvalds bashed Sievers for lacking responsibility in a mailing list message. “Key[sic], I’m [expletive] tired of the fact that you don’t fix problems in the code you write, so that the kernel then has to work around the problems you cause,” he wrote.
Torvalds went on to state that “this has been going on for years,” and said that he will refuse to accept patches from Sievers until Sievers cleans up his act.

This is like the gift that keeps on giving endless drama forever.

Thank you systemd

On 2014-06-17 13:16, BSDuser wrote:
>
> http://tinyurl.com/owfrdx3
>
> Sievers was banned by kernel maintainer Linus Torvalds on Wednesday

That was posted Apr 3, 2014 5:13, so it is not this Wednesday.


Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)

Since everyone is on their “let’s hate systemd” parade, let me ask you something:

How would you, without using external applications such as monit, implement a high availability system service - using sysv - that was able to recover from catastrophic application failure and restart itself?

I’m pretty sure that is what the “respawn” entries in “/etc/inittab” were for, though that facility was not much used.

It has literally no safety checks, it doesn’t handle .pids and has a multitude of other issues - in short, there’s a reason why it wasn’t used.

That is something systemd aims to fix. Which is good.