[sf-lug] Release of Debian (GNU/Linux) Buster Stable ... systemd?

Michael Paoli Michael.Paoli at cal.berkeley.edu
Fri Jul 5 20:07:50 PDT 2019


Yes, though Debian changed the *default* init system to
systemd with the release of Debian 8 "Jessie" on 2015-04-25,
it remains the case that one needn't use systemd, at least in general,
as the init system on Debian.
In fact even Debian's documentation with the Debian 8 "Jessie" release
on 2015-04-25 rather well covered that:
https://www.debian.org/releases/jessie/amd64/release-notes/ch-whats-new.en.html#systemd
https://wiki.debian.org/systemd#Installing_without_systemd
https://www.debian.org/releases/jessie/amd64/release-notes/ch-upgrading.en.html#recovery-shell-systemd

And, yes, absolutely possible to run Debian without systemd.
Heck, I've been doing if for *years* now!
$ lsb_release -d; uname -m; sudo ls -l /proc/1/exe
Description:    Debian GNU/Linux 9.9 (stretch)
x86_64
lrwxrwxrwx 1 root root 0 Jul  5 12:09 /proc/1/exe -> /lib/sysvinit/init
$ dpkg -S /lib/sysvinit/init
sysvinit: /lib/sysvinit/init
$

Debian is good/excellent on freedom and offering choices.  I see
zero evidence that Debian 10 "Buster" will at all force or coerce
anyone to using systemd if they don't wish to ... though it does
remain the *default*.

Continues to surprise/amaze me how many folks will jump distro just
because they don't like some default within a distro.  I mean, heck,
you did configure a non-default root password, did you not?  Uhm, ...
maybe I'm afraid to know the answer to that.

> From: "Rick Moen" <rick at linuxmafia.com>
> Subject: Re: [sf-lug] Release of Debian (GNU/Linux) Buster Stable
> Date: Fri, 5 Jul 2019 13:03:12 -0700

> Quoting aaronco36 (aaronco36 at SDF.ORG):
>
>> Lastly (in this particular posting), Rick M has his useful Debian
>> FAQ at [16], including a description of the means of converting
>> systemd-init into the "excellent dependency-based init system"
>> OpenRC [17] (IMNSHO, hope that this is all still completely valid
>> with Debian 10 Buster!)
>
> I'll probably find time to re-do the research that went into that page
> for Debian 10 Buster.  My expectation is that it'll work out the same.
>
> As context, that page was my non-rhetoric answer to a bit of rhetoric I
> kept hearing from frenetic systemd opponents:[1]  I kept hearing, over a
> couple of years, that it was utterly impossible to run Debian 8 'Jessie'
> (then the current 'stable' release) without the system package.  Every
> time I raised reasons to doubt the assertion, all that resulted was
> hearing the same thing back, asserted more firmly.  So, having gotten my
> stubborn up, I created a VirtualBox VM, installed Jessie into it,
> confirgured it to use WindowMaker (my preference in window manager), and
> then followed instructions on the without-systemd wiki to replace the
> systemd package with the openrc one including 'pinning' instructions to
> prevent systemd's reinstallation via dependencies -- and then checked to
> see what packaages or metapackages of interest were then uninstallable
> because of dependency chains including systemd.
>
> I simply included the exact results of that testing, including how to
> reporduce them, and a few musings about ways to contend with obstacles,
> and about other non-systemd init systems equally available in sundry
> ways.
>
> When Debian 9 'Stretch' became the new 'stable' branch, I did some
> spot-checking and updating the document, finding that its conclusions
> remain valid -- but I didn't rewrite it.
>
> Anyhow, it really takes only a few hours' work to check, and do a
> write-up.
>
>
> [1] This included my friend Steve Litt, who is like me a fan of Devuan
> Project but not a developer thereof.  Steve and the others kept citing
> the alleged total inability to run Jessie without system as a reason
> necessitating the fork.  I saw the fork as an on-balance reasonable step,
> but doubted the factual claim cited as justification.  Turns out, I was
> correct, and found just documenting my findings much more effective than
> arguing with those making what, in the final analysis, were just
> advocacy arguments with no factual basis.




More information about the sf-lug mailing list