Devuan bug report logs - #723
init-system-helpers: update-rc.d calls systemctl on a SysVinit system, creating a recursion

Full log

Message #25 received at 723@bugs.devuan.org (full text, mbox, reply):

Received: (at 723) by bugs.devuan.org; 14 Nov 2022 10:52:36 +0000
Return-Path: <wolfgang@leila.ping.de>
Delivered-To: bugs@devuan.org
Received: from email.devuan.org [2001:41d0:2:d06e::5c4:2612]
	by doc.devuan.org with IMAP (fetchmail-6.4.16)
	for <debbugs@localhost> (single-drop); Mon, 14 Nov 2022 10:52:36 +0000 (UTC)
Received: from email.devuan.org
	by email.devuan.org with LMTP
	id xR/UB8YdcmOSZwAAmSBk0A
	(envelope-from <wolfgang@leila.ping.de>)
	for <bugs@devuan.org>; Mon, 14 Nov 2022 10:51:50 +0000
Received: by email.devuan.org (Postfix, from userid 109)
	id 12CC0396A; Mon, 14 Nov 2022 10:51:50 +0000 (UTC)
X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on email.devuan.org
X-Spam-Level: 
X-Spam-Status: No, score=-0.0 required=5.0 tests=RCVD_IN_DNSWL_NONE,SPF_PASS
	autolearn=ham autolearn_force=no version=3.4.6
Received-SPF: Pass (mailfrom) identity=mailfrom; client-ip=83.97.42.2; helo=lilly.ping.de; envelope-from=wolfgang@leila.ping.de; receiver=<UNKNOWN> 
Received: from lilly.ping.de (lilly.ping.de [83.97.42.2])
	by email.devuan.org (Postfix) with ESMTPS id D728E4B
	for <723@bugs.devuan.org>; Mon, 14 Nov 2022 10:51:44 +0000 (UTC)
Received: (qmail 21370 invoked by uid 10); 14 Nov 2022 10:51:31 -0000
Received: from leila.ping.de by lilly.ping.de with UUCP (rmail-0.2-fdc); 14 Nov 2022 10:51:31 -0000
Received: by leila (Postfix, from userid 11341)
	id B2A297EC3; Mon, 14 Nov 2022 11:51:17 +0100 (CET)
Date: Mon, 14 Nov 2022 11:51:17 +0100
To: Mark Hindley <mark@hindley.org.uk>
Cc: 723@bugs.devuan.org
Subject: Re: bug#723: init-system-helpers: update-rc.d calls systemctl on a SysVinit system, creating a recursion
Message-ID: <20221114105117.GE7301@leila.ping.de>
References: <20221112202017.GC7301@leila.ping.de> <Y3E5pEc/vLJZHT5r@hindley.org.uk> <20221113210005.GD7301@leila.ping.de> <Y3IXZ9kz7gxOcgvI@hindley.org.uk>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Y3IXZ9kz7gxOcgvI@hindley.org.uk>
User-Agent: Mutt/1.5.13 (2006-08-11)
From: wolfgang@leila.ping.de (Wolfgang Wegner)

On Mon, Nov 14, 2022 at 10:24:39AM +0000, Mark Hindley wrote:
[...]
> I notice you have dummy-systemd-dev installed. You really don't want that on a
> production system. If is a hack for building sources that Build-Dep: systemd. It
> doesn't provide any systemd functionality.
> 
> If you remove it does it make any difference?

Thanks - I removed it now, but that does not make any difference.

> > > > Calling update-rc.d from the systemctl "abstraction" (surrogate?) seems
> > > > to be correct, but I would have expected update-rc.d to detect the system's
> > > > init system and - as I am using sysVinit on Devuan - just modify the links.
> > > > 
> > > > update-rc.d behaviour can also be reproduced when calling it for an already
> > > > installed service.
> > > 
> > > Could you give a specific example here. I can't reproduce it.
> > 
> > root@<host>:~# update-rc.d rsyslog enable
> > update-rc.d rsyslog enable
> > update-rc.d rsyslog enable
> > update-rc.d rsyslog enable
> 
> Hmmm, I get completely different behaviour with that command:-
> 
> mark@apollo:~% sudo update-rc.d enable rsyslog

Is the switched basename/command pair intended here?

> usage: update-rc.d [-f] <basename> remove
>        update-rc.d [-f] <basename> defaults
>        update-rc.d [-f] <basename> defaults-disabled
>        update-rc.d <basename> disable|enable [S|2|3|4|5]
>                 -f: force
> 
> The disable|enable API is not stable and might change in the future.
> 
> What systemd-requiring package are you trying to install? What other changes
> have you made?

The package I try to install is *cough* "cybereason-sensor", which is now
required to install from our IT department. Their post-install script
relies on either upstart or systemd being present. As in consequence, no
LSB init script is provided with the package, I created a basic one and
also added the standard links in /etc/rc?.d/.

I am not aware of other changes I did to the system - however, the
installation is originating from a Devuan 1.0 install and has only been
updated to the current state, which has been quite cumbersome with respect
to some in-between incompatibilities, especially the usual NVIDIA mess
but also some things concerning qemu/KVM and wine.
(This is to say: I do not remember any changes in the area of the init
system but there has been some fiddling around trying to get some updates
installed.)

As another test, I just tried to remove systemctl-service-shim, and
immediately the weird behaviour of update-rc.d is gone:

root@<host>:~# dpkg -r systemctl-service-shim
(Reading database ... 836967 files and directories currently installed.)
Removing systemctl-service-shim (0.0.4-1) ...
Processing triggers for man-db (2.11.0-1+b1) ...
root@<host>:~# update-rc.d rsyslog enable
root@<host>:~#

So to me it really looks like an unwanted interaction just between
update-rc.d and systemctl-service-shim.

> Mark

Best regards
Wolfgang

Send a report that this bug log contains spam.