[Cialug] The systemd Init System

Todd E Thomas todd.dsm at gmail.com
Wed Dec 9 20:26:55 CST 2015


That's a valid point but I don't think the purpose of it is as dry as "it's
broken, let's fix it". It seems more a matter of, "could it be better?";
can we, as admins, get more usable information with a few more key strokes?
As software people, this question never goes away.

As a young man I made one (of many) mistakes by saying "it's done - it's
soooo done"; that script(s) plagued me. My boss looked at me and realized I
was missing a larger piece of the puzzle. Patient as she was, sat me down
and asked me a few questions, essentially:

When is software ever done being written?
How does "done" leave room for the one constant, change?

Playing devil's advocate, she also asked me where the room for improvement
was built in(?).

The answer is, of course, that software is never done; ever. It's either
updated, in the case it's written well and will scale, or it's replaced in
the case where it's not. But the purpose for the software rarely goes away
and will forever require modification.
---

The benefit you are looking for probably isn't there for you yet. Perhaps
the problems you see on a daily basis are relatively easy to deal with.
Sometimes I envy that.

Consider the work of Fedora/Debian engineers. Their work trickles down in a
titanic way: Fedora > RHEL > CentOS / Debian > Ubuntu > Linux Mint, on and
on - seemingly forever. But there's a bigger picture. Given these OSs will
be deployed in many more scenarios than were even possible 5 years ago, how
can each system be more maintainable as individual parts of a greater whole?

System V UNIX was brilliant, no question, but it was built for a different
time. Someday you will need answers from many different types of systems,
both in-house and outdoors (in the cloud somewhere) and at that point you
will realize the beauty of it; it's just a common interface that's
scaleable.

In the case of journald - you now have the option to:
  1) ask very specific questions
  2) in the same way
  3) for every server; Fedora, Debian and all derivatives

That option just wasn't there yesterday; today it is.

Also, I can't tell you how happy I am not to be modifying init scripts any
more. Some customers paint themselves into a corner where that becomes a
necessity; and it sucks.

But I get it, I don't have all the time in the world either and learning
the new stuff just takes time. But, with a little project management
*magic* that time can be created. Then some of your job IS learning. And
there's really nothing better than being paid to discover something new.

TT



On Wed, Dec 9, 2015 at 4:13 PM, Scott Yates <Scott at yatesframe.com> wrote:

> Specifically, systemd seems to have the idea that all other init systems
> are broken, and they have taken it on themselves to "fix" them.  The idea
> of a monolithic init system makes me nervous because it throws away 30+
> years of proven ideas.
>
> Binary logging similarly bothers me.  Logging systems have been in place
> for years and have been battle tested and proven.
>
> The whole mind-set of fixing things that are not broken bothers me.
> Anointed Solution, does not mean it is the best solution.
>
> Obviously I will wait it out and see how things develop, as I am not nearly
> as smart as some of the guys working on systemd, but from what I have seen
> so far, I just don't see the benefit over the existing infrastructure that
> is already working and tested.
>
>
> On Wed, Dec 9, 2015 at 4:05 PM, Todd E Thomas <todd.dsm at gmail.com> wrote:
>
> > Scott, "things like this"? *What in specific?*
> >
> > The community doesn't leave much room for like or dislike; only puzzle
> > pieces that need to be fit together. Or, another way of saying it might
> > be...
> >
> > It is the anointed solution. When all older (non-systemd) OS versions
> have
> > been deprecated (no longer supported) the only choices left will be:
> >   *Linux (systemd),
> >   *Linux (non-systemd variant), or
> >   *Unix  (still Sytem V AFIK)
> >
> > Changing distros to avoid the future seems a long way to go, especially
> for
> > organizations with 100s of machines.
> >
> > So, what specifically is a hangup? Maybe we can setup a test environment
> > and smooth some wrinkles.
> >
> > TT
> > _______________________________________________
> > Cialug mailing list
> > Cialug at cialug.org
> > http://cialug.org/mailman/listinfo/cialug
> >
> _______________________________________________
> Cialug mailing list
> Cialug at cialug.org
> http://cialug.org/mailman/listinfo/cialug
>


More information about the Cialug mailing list