Re: Debian Stretch SELinux enforcing causes systemd --user unit to fail

[Laurent Bigonville]

C J du Preez wrote:


Good day,


Hi,


I would like to report a bug, but I am not sure which package to report it 
against.

I have SELinux enabled and enforcing on Debian Stretch (commandline via SSH 
only, no GUI is installed at all). I am trying to start a systemd --user unit 
(which I know is correct, because it works without SELinux enabled). When I try 
to start the service (using systemctl --user start ssh-agent) I get:

Failed to connect to bus: No such file or directory.

With SELinux enabled, DBUS_SESSION_BUS_ADDRESS is undefined (with SELinux 
disabled it is defined as unix:path=/run/user/1000/bus). With SELinux disabled 
that path exists, with SELinux enabled, it does not.
Could you also please try in permissive mode as well? That should fix 
the creation of the file on disk, if that problem is fixed I think you 
could open a bug against the selinux-policy-default package.


Regarding the environment variable not being set, I think it's this 
bug[0]. Would you be able to rebuild systemd with this patch[1] applied? 
If that patch fixes the problem, we could maybe try to include that in a 
stable point release.


Otherwise, you should try to upgrade systemd to an higher version using 
the backports.


Kind regards,

Laurent Bigonville


[0] https://github.com/systemd/systemd/issues/6120
[1] 
https://github.com/systemd/systemd/pull/6201/commits/51c7d5aa36e9ac0ec8ca6fef811a5f9deb7e4fd2

Re: Re: Systemd services (was Re: If Linux Is About Choice, Why Then ...)

[Laurent Bigonville]

Greg Wooledge mailto:wooledg%40eeg.ccf.org>> wrote:

On Fri, Apr 14, 2017 at 03:17:00PM +0200, Nicolas George wrote:
> Le quintidi 25 germinal, an CCXXV, Greg Wooledge a écrit :
> > Some day there will be actual end-user-friendly systemd documentation
> > somewhere, consolidating all of these pieces of wisdom together.  I hope.
>
> Note: systemd is not for end-users, it is for system administrator and
> distribution authors.

The end users of systemd are Linux system administrators.  You and me.
The people on this mailing list.  That's us, the users.  That's why
it's called "debian-user".

If you'd prefer "Some day there will be a system administrator's guide
for systemd", that's an acceptable wording.

There is already extensive documentation about how to administrate systemd:

https://www.freedesktop.org/wiki/Software/systemd/#manualsanddocumentationforusersandadministrators

"The systemd for Administrators Blog Series" worth reading.

Re: fprintd-enroll

[Laurent Bigonville]

Shahryar Afifi > wrote:

after i enroll as super user i get the message " enroll resault: enroll 
compeleted" but when i log in as super user again, shell asks for password not 
finger authentication.


Did you install the libpam-fprintd? It should configure the PAM stack 
for you.

Re: gdm3 doesn't work any more after the upgrade from Wheezy to Jessie 8.5

[Laurent Bigonville]

Jean-Paul Bouchet wrote:
> [...]
> Check that logind is properly installed and pam_systemd is getting 
used at login.

> [...]

Could you check if you have libpam-systemd package installed? And also 
please check if "loginctl" shows sessions.


Cheers,

Laurent Bigonville

Re: Re: System in broken state after dpkg upgrade

[Laurent Bigonville]

Don Armstrong wrote:

That's basically because the policy wasn't fixed in time for the jessie
release (see #756729 and #771484). If you're using selinux on Debian, it
would probably be good to participate in the development of the default
policy and refpolicy packages.


Yes please

Re: Can't load X

[Laurent Bigonville]

Le 12/10/15 17:07, Mitt a écrit :

Hi,

Hi,

I installed xserver-xorg-legacy but it changed nothing.

I have Intel HD and Radeon, I don't use logind but ConsoleKit instead, 
I don't have anything related to systemd, except for udev.


You need a logind session to be registered. If it's not the case, you'll 
have to add "needs_root_rights = yes" in /etc/X11/Xwrapper.config



If server is supposed to run with root, how should a regular user use X?


X has always been running as root via a setuid wrapper, now with logind 
it gained the ability to delegate the things that require root 
privileges to it.


Cheers,

Laurent Bigonville

Re: Can't load X

[Laurent Bigonville]

Hello,

With the latest upload of the xserver-xorg package in unstable 
(2:1.17.2-3), the server is now supposed to run with root privileges.


Do you have a logind session registered? What kind of graphic card do 
you have?


Try to install xserver-xorg-legacy package and see if that helps.

Cheers,

Laurent Bigonville

Re: Okay, that's too much now!

[Laurent Bigonville]

Erik Lauritsen wrote:


I have been a Debian user for more than 15 years, when the "war"
about systemd broke out I mostly ignored it, I just removed systemd
from my systems because I don't like the implementation.

Today I was setting up a new Debian system and wanted to remove
systemd only to find our that the old tools "bsdutils" has been made
dependent upon libsystemd0

"This package contains the bare minimum of BSD utilities needed for a
Debian system: logger, renice, script, scriptreplay, and wall. The
remaining standard BSD utilities are provided by bsdmainutils."

What the freaking !#¤"#¤"¤#"#%" are people doing!?

Why the hell has this collections of utilities from FreeBSD been made
dependent upon libsystemd0!?!?!?


The bsdutils package (which is built from the *util-linux* source
package) contains reimplementation of tools that were initially existing
on BSD, not the BSD utilities themselves.

And to answer the question why that package has a dependency against
_libsystemd0_ package, it's because the logger utility is also writing
to the systemd journal if it's running in addition to good old syslog.
Not enabling the feature would mean functionally loss for people who
want to use systemd functionalities.


Freedom of choice my ass!


Well the fact that you have a dependency against _libsystemd0_ doesn't
say anything about running systemd running as PID1 as it turns itself to
a noop is PID1 is not systemd.

I personally still have troubles to understand why this is even an
issue. There are tenth of libraries that are installed on a debian
system that are only useful for some limited use-case (libselinux or
libaudit for example), Debian has a policy of enabling most (if not all)
the features in its packages, why is libsystemd suddenly a problem?

But if having libsystemd0 installed on your system is a problem for you,
you still have the freedom to rebuild the packages to remove it

Cheers,

Laurent Bigonville

Re: Debian fork: 'Devuan', Debian without Systemd

[Laurent Bigonville]

Le Wed, 03 Dec 2014 10:18:36 +0100,
Martin Steigerwald  a écrit :

> Am Mittwoch, 3. Dezember 2014, 08:35:00 schrieb Erwan David:
> > Le 02/12/2014 23:15, Martin Steigerwald a écrit :
> > > Am Dienstag, 2. Dezember 2014, 18:47:38 schrieb Renaud OLGIATI:
> > >> On Tue, 2 Dec 2014 14:22:13 -0700
> > >> 
> > >> Aaron Toponce  wrote:
> >  It's a waste. They shouldn't have left. I'm pretty neutral
> >  about systemd as I'm only an end user but I disklike having it
> >  forced upon me this way.
> > >>> 
> > >>> # apt-get install upstart
> > >>> # apt-get install sysvinit-core
> > >>> # apt-get install openrc
> > >>> No one is forcing you to stick with systemd. The "fork" is just
> > >>> silly.
> > >> 
> > >> Another way to look at it is "forward planning for the release
> > >> after Jessie, when systemd may well become compulsory..."
> > > 
> > > Or going beyond what is offered in Debian… like making GNOME
> > > installable without having any systemd related package installed.
> > 
> > The systemd package is just a small part of systemd. I'd like to
> > remove systemd-logind and lbpam-systemd, sinc I have no clue at all
> > that logind is better deisgned and programmed than resolved, which
> > showed it was designed without any care for well known attacks.
> 
> I explicetely wrote "any systemd related package".
> [...]
> 
> So you can still choose to what init system to use, but running
> completely without any systemd related packages gives you a really
> crippled system.

As explained several times on this ML, depending against libsystemd0
package doesn't mean anything about requiring systemd to be used as
PID1 or not. Even Ian's GR was not taking the "I don't want any systemd
package on my machine" use case into account you know.

But if you have that special concern, you'll have to start recompiling
the packages I'm afraid. Start with policykit and network-manager (and
other package defining a dependency against libpam-systemd) to make
them use ConsoleKit again, you would at least be able to remove the
systemd package completely.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141203123926.4fb34...@soldur.bigon.be

Re: Replacing systemd in Jessie

[Laurent Bigonville]

Le Mon, 1 Dec 2014 23:05:09 -0800,
Patrick Bartek  a écrit :

> On Tue, 02 Dec 2014, Scott Ferguson wrote:
> 
> > On 2 December 2014 at 08:18, Patrick Bartek 
> > wrote:
> > > On Mon, 01 Dec 2014, Ric Moore wrote:
> > >
> > >> On 11/30/2014 11:27 PM, Patrick Bartek wrote:
> > >>
> > [snip]
> > 
> > > > > as the
> > >> > default init more distros will follow suit,
> > 
> > Very few do not include systemd. I'd welcome a definitive list of
> > those that don't.
> 
> As as option at install time or during an upgrade?  Don't know of any.
> 
> As far as I've read, I believe only Slackware absolutely refuses to
> use systemd.  I don't even think it's in the repo.  I don't know if
> systemd will even work with Slackware.


Well according to the following wikipedia page, Patrick Volkerding
(Slackware founder) has not completely ruled you systemd:
https://en.wikipedia.org/wiki/Systemd#Adoption


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141202130423.3b09a...@soldur.bigon.be

Re: Replacing systemd in Jessie

[Laurent Bigonville]

Le Mon, 01 Dec 2014 13:45:12 -0500,
Miles Fidelman  a écrit :

> Ric Moore wrote:
> > On 11/30/2014 11:27 PM, Patrick Bartek wrote:
> >
> >> I fear that once systemd is firmly entrenched in Debian as the
> >> default init more distros will follow suit, and more and more
> >> developers will start writing apps with systemd, or parts of it,
> >> as a dependency for the "features" it offers.
> >
> > Every other distro of merit has long since made the switch. We're
> > just late to the party.  Are you just figuring it out now? Ric
> >
> 
> Just to be clear... you're saying that Slackware, Gentoo, and their 
> derivatives are not "distros of merit?"  Or, for that matter, BSD and 
> illumos derivatives?

Did you saw that a co-founder of FreeBSD is proposing to switch to a
system very similar to systemd?
http://www.phoronix.com/scan.php?page=news_item&px=MTg0ODE


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141202114915.5b7c7...@soldur.bigon.be

Re: udev memory problem when trying to plug a disk with corrupted partition table

[Laurent Bigonville]

Le Mon, 17 Nov 2014 16:56:51 +0100,
berenger.mo...@neutralite.org a écrit :

> Now, fact is that the hard-disk partition table is no longer correct, 
> and when I plug it (it is an USB HD) into a Debian system, it makes
> udev eating all my memory, and more.

Could you please open a bugreport against udev with all the
information like the fdisk output and such?

I feel that even if the partition table is invalid, it shouldn't DoS
udev.

Thanks!

Laurent Bigonville


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141120100825.562a7...@fornost.bigon.be

Re: If Not Systemd, then What?

[Laurent Bigonville]

Le Sun, 16 Nov 2014 18:08:48 +,
Martin Read  a écrit :

> On 16/11/14 17:33, Laurent Bigonville wrote:
> > Are you aware that this is the approach that systemd and upstart
> > have taken, right?
> >
> > 1) Both systemd (PID1) and upstart are drop-in replacement for the
> > good old SysVinit as they both support the common "standard" that
> > are LSB scripts (A really good share of the existing LSB
> > initscripts in the debian archive are just working out of the box).
> 
> Well. They're (mostly) a drop-in replacement for sysvrc and its 
> supporting tools. They're certainly not a *drop-in* replacement for 
> *sysvinit*, because they don't support all of sysvinit's interfaces; 
> specifically, they don't support /etc/inittab.
> 
> Luckily (for some values of lucky), /etc/inittab was such a terrible 
> interface (it's unpleasantly reminiscent of Angband's monster, item, 
> etc. databases) that it seems even most people who prefer sysvinit to 
> systemd or upstart were using a factory-default /etc/inittab.

Note that there were plans to either abort systemd-sysv installation or
at least display a big fat warning in case /etc/inittab was modified on
the machine.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141116194032.38ec5...@fornost.bigon.be

Re: init scripts [was: If Not Systemd, then What?]

[Laurent Bigonville]

Le Sun, 16 Nov 2014 11:50:25 -0500,
Miles Fidelman  a écrit :

[...]
> So... with systemd, one has to:
> - rely on packagers to generate systemd service files, and/or,
> - rely on systemd's support for sysvinit scripts, which
> 
> In the later case, one just has to read:
> http://www.freedesktop.org/wiki/Software/systemd/Incompatibilities/
> to get very, very scared

Do you have any bug reports at hand that show that there are issues
with the initscripts that are shipped in the debian archive? Or are you
spreading FUD again?

> Among the implications of this, the old standby of installing
> software from upstream (bypassing packaging), has just gotten a lot
> riskier.

That's indeed true, but usually 3rd party vendors of (commercial)
software are certifying their software for a limited set of
distributions/version anyway.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141116193635.242fc...@fornost.bigon.be

Re: If Not Systemd, then What?

[Laurent Bigonville]

Le Sun, 16 Nov 2014 13:53:24 +,
Nuno Magalhães  a écrit :

> On 2014-11-16 11:40, Klistvud wrote:
> > 1. Reviving the existing init systems. Modernizing them, making them
> > into true, interchangeable drop-in replacements of each other,
> > which do the task assigned, and do it well. Each of them
> > accomplishing at least the common subset of tasks an init system is
> > supposed to provide.
> > 
> > 2. Complementing them with existing or new tools (again, drop-in
> > interchangeable replacements of each other) which build on them and
> > provide the next layer. For example, the kernel autofs facility
> > provides very nice automounting and could be deployed to the
> > majority of desktop installs (instead of being just an optional
> > package, as it is now), thus making the various automount daemons
> > of the various desktop environments/file managers virtually
> > superfluous. As a further example, the former udev (prior to being
> > merged into systemd) has already been forked and could/will serve
> > us well for years to come. And so on.
> 
> +1 for being reasonable and making sense
> 
> It's an approach that would keep a lot of people happy and, more
> importantly (at least to me), it gives the user choice instead of
> taking it away. At least this way each user could choose the
> loosely-coupled components s/he wanted.

Are you aware that this is the approach that systemd and upstart have
taken, right?

1) Both systemd (PID1) and upstart are drop-in replacement for the good
old SysVinit as they both support the common "standard" that are LSB
scripts (A really good share of the existing LSB initscripts in the
debian archive are just working out of the box).

2) Again that's exactly what systemd and upstart are doing, they have
added extra features to PID1 like socket activation, process tracking
or the fact that the daemons are started in a clean environment. And
then to that, the systemd project (outside of PID1) has consolidated
services (some of them dead upstream for _years_) under the same
umbrella project. All of this without preventing the already existing
implementations to be used. journald is _not_ preventing a syslog
daemon to be used, the .timer unit files are _not_ preventing cron to
be used and so on...

But then you cannot blame the systemd project for 3rd party software
taking advantages of these new functionalities if they think they fit
their usecases.

Laurent Bigonville


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141116183351.3bd37...@fornost.bigon.be

Re: engineering management practices and systemd (Re: Installing an Alternative Init?)

[Laurent Bigonville]

Le Sat, 15 Nov 2014 20:21:49 -0500,
Marty  a écrit :

> On 11/15/2014 06:49 AM, Andrei POPESCU wrote:
[...]
> >
> > At least some of people rejecting systemd demand that it be removed
> > completely, including libsystemd. How is it pro-choice to forbid me
> > from being able to use a software at its full potential?
> 
> For me it's more about being unable to remove it completely, because
> of vendor lock-in. There's no technical reason that I know of that
> anything in userspace cannot modular, and replaceable, so when
> something cannot be replaced then an alternative must be provided, or
> else my default assumption is that vendor lock-in is in effect.

Well, yes there are technical reasons why you cannot remove a library
package when you have symbols provided by this library used in an
executable. You can still recompile the package and remove some
configure flag if you want to remove this dependency.

OTHO there is no technical reasons that I can see, to completely remove
libsystemd package. You have tenth of other libraries on your system
that, like libsystemd, turn them self into a noop if some some
functionality or daemon are not enable. I'm thinking here for example
about libselinux and libaudit (both also written by Red Had and the
NSA, OMG!!!11), and yet I fail to see any outcry here...

So as long as you cannot _prove_ that having libsystemd installed on
your machine is causing any issues, I'll personally mentally classify
your request as "I don't want to see any packages containing the
"systemd" string on my machine" and redirect these to /dev/null.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141116112652.78d2a...@fornost.bigon.be

Re: engineering management practices and systemd (Re: Installing an Alternative Init?)

[Laurent Bigonville]

Le Fri, 14 Nov 2014 12:26:09 +0200,
Andrei POPESCU  a écrit :

> On Vi, 14 nov 14, 08:59:11, Joel Rees wrote:
[...]
> 
> > So Fedora is not, itself, really ready yet, except for two groups, a
> > certain group of workstation users who want and are willing to use
> > fairly new, relatively high-end hardware, including enough RAM and
> > processors to use VMs for certain things, and a certain group of
> > server-farm users who want and have budget for similarly recent,
> > relatively high-end hardware and lots of RAM and processors for lots
> > of VMs.
> > 
> > The rest of the Fedora users jumped ship.
> 
> Again, I can't comment on Fedora, but my Raspberry Pi runs systemd
> just fine. Also my laptop running is quite far from being a speed
> monster. 

Also, different embedded projects (the Jolla phone, car board
computers,...) are already using or planning to use systemd.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141114141021.435d4...@soldur.bigon.be

Re: what's the difference btw. libelf1 and libelfg0?

[Laurent Bigonville]

Le Wed, 12 Nov 2014 17:09:05 -0800,
Don Armstrong  a écrit :

> On Thu, 13 Nov 2014, Christian Groessler wrote:
> > Ignoring the freebsd version, I'm wondering what's the difference
> > between libelf1 and libelfg0. They seem to be built from the same
> > sources, 'apt-get source' retrieves the same files.
> 
> Two different sonames. The g0 version exists for compatibility with
> certain pre-packaged binaries which require it.
> 
> Most (all, hopefully) things in Debian should require the libelf1
> package instead of the libelfg0 package:
> 
>  This shared library may be needed by pre-packaged programs. To
> compile programs with this library, you will need to install the
> libelfg0-dev package as well.
> 

Thanks Don for this, there were still some packages in the archive
using libelfg0, I've opened bug reports against them.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141113174745.21555...@soldur.bigon.be

Re: systemd - so much energy wasted in quarreling

[Laurent Bigonville]

Le Wed, 12 Nov 2014 15:29:27 -0200,
Andre N Batista  a écrit :

> On Tue, Nov 11, 2014 at 06:23:26PM +0100, Laurent Bigonville wrote:
> > Le Tue, 11 Nov 2014 11:58:33 -0500,
> > Tanstaafl  a écrit :
> > 
> > > On 11/11/2014 9:26 AM, Didier 'OdyX' Raboud 
> > > wrote:
> > > > Blaming the Debian project for letting the Debian distribution
> > > > evolve in ways defined by its volunteers is unfair.
> > > 
> > > Eh? My understanding is that this systemd mess is due to a vote
> > > of the technical committee, a vote that was in fact tied and the
> > > chair had to cast the tie-breaker.
> > > 
> > > Hardly waht I would call an 'evolution defined by its
> > > volunteers'...
> > 
> > The members of the project have delegated the power to arbitrate
> > technical decisions to the the technical committee via the
> > constitution. Some people might not agree with this, but this how
> > the project is working today.
> 
> And that power which was delegated can be reclaimed anytime by it's
> true and legitimate owner if, for some reason - say the TC reached a
> decision with a Minerva vote by Project Leader and that decision has
> proven over the time to be highly controversial among debian users
> and developers -, he feels the need to do so.
> 
> As far as debian is related to software, very question can be
> expressed as a technical one. As far as it is a project commited to
> some political goals, some technical decisions can hinder, others can
> further project's goals.
> 
> Nobody is claiming that the TC shouldn't be allowed to decide on
> technical issues, but there are people who seem to think that this is
> equal to TC's decisions being final, undebatable and impossible to
> overrule. If this is true, however, we cannot claim that TC's power
> is delegated, we need to recognize that it is an unlimited de facto
> power.

The TC decisions can be overruled by a GR with a simple majority, see
the point 4.1.4 of the Constitution[0]

[0] https://www.debian.org/devel/constitution#item-4


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141112195508.113dc...@fornost.bigon.be

Re: Installing an Alternative Init?

[Laurent Bigonville]

Le Wed, 12 Nov 2014 06:10:55 -0500,
Tanstaafl  a écrit :

> On 11/11/2014 3:33 PM, Miles Fidelman 
> wrote:
> > Actually, there's a patch (thank you Kenshi).  It has not been
> > applied. Hence, to use it right now, one has to build a custom
> > version of the installer.  I hope, that post the initial Jessie
> > release, the deboostrap and installer maintainers will apply the
> > patch.
> 
> Since the bug is so old (dates back to wheezy), and a patch exists and
> still hasn't been applied, I think it is likely that they simply don't
> *want* to fix this bug, since that would negatively impact the desire
> to get as many people using systemd as possible, so they can be
> counted in the stats of 'satisfied systemd users', even if many/most
> don't even *know* they're running a different init system.

Any evidences of this? Because you know this also has an impact for
other packages that have OR statement in their dependencies.

Also are you aware that back to 2012, the changes needed to install
systemd-sysv package on the machine were way bigger than today due to
the fact that sysvinit was the only package that you could install on
your machine that was providing "/bin/init" because this package was
marked "Essential: yes". In the meantime, the Essential flag moved from
sysvinit to the "init" meta-package to allow more init system to be
installable. So like Michael said, Jessie will indeed be the first
version that allows you to have an alternate init without modifying the
kernel cmdline.

Now the bug is "just" debootstrap ignoring OR in the dependencies. And
as Kibi said on the bug, the patch arrived late in the d-i schedule
(patch proposed the 17/10, more than 10 days after the release of the
2nd beta of the installer).


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141112150228.3aa7a...@soldur.bigon.be

Re: systemd - so much energy wasted in quarreling

[Laurent Bigonville]

Le Tue, 11 Nov 2014 11:58:33 -0500,
Tanstaafl  a écrit :

> On 11/11/2014 9:26 AM, Didier 'OdyX' Raboud  wrote:
> > Blaming the Debian project for letting the Debian distribution
> > evolve in ways defined by its volunteers is unfair.
> 
> Eh? My understanding is that this systemd mess is due to a vote of the
> technical committee, a vote that was in fact tied and the chair had to
> cast the tie-breaker.
> 
> Hardly waht I would call an 'evolution defined by its volunteers'...

The members of the project have delegated the power to arbitrate
technical decisions to the the technical committee via the
constitution. Some people might not agree with this, but this how the
project is working today.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/2014182326.5294f...@fornost.bigon.be

Re: Installing an Alternative Init?

[Laurent Bigonville]

Le Tue, 11 Nov 2014 07:42:33 -0500,
Tanstaafl  a écrit :

> On 11/10/2014 6:18 PM, Michael Biebl  wrote:
> > Am 11.11.2014 um 00:14 schrieb Miles Fidelman:
> >> Ok, then explain to me the procedure for running the installer in
> >> such a way that systemd is never installed, thus avoiding any
> >> potential problems that might result from later uninstallation all
> >> the dependencies that systemd brings in with it.
> 
> > Please be specific. What problems of of dependencies are you
> > talking about?
> 
> Please stop bring up irrelevant questions and address the question
> being asked.
> 
> This does require you to at least understand and acknowledge the
> difference between a *clean* install, and installing something one
> way, then having to uninstall a primary piece and replace it with
> something else.
> 
> The two are not the same, and no amount of you trying to act as if
> they are will change the fact that they are not.

There are no functional differences between an installation with
sysvinit-core out of the box or an install where sysvinit-core is
installed later, this is a fact.

Allowing the user to choose this at install time from the interface is
a "nice to have" feature (wishlist bug) not a RC bug like you were
claiming earlier.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/2014180749.7e240...@fornost.bigon.be

Re: Installing an Alternative Init?

[Laurent Bigonville]

Le Tue, 11 Nov 2014 10:47:35 +0200,
Andrei POPESCU  a écrit :

> On Lu, 10 nov 14, 21:12:10, Lee Winter wrote:
> > 
> > Of all the options available in the NON-expert installer, the
> > choice of init alternatives might not warrant a user selection
> > option, but all of the _consequences_ of that selection, i.e.,
> > things that get sucked in, mandate that users be offered a choice.
> 
> The systemd *package* would 'suck in' following packages (assuming 
> installation of Recommends, which is the default):

I want to add to this that I tried the following:

 - Install a wheezy system with debootstrap in a chroot
 - Change the source.list to jessie
 - Install sysvinit-core from jessie
 - Dist-upgrade the chroot to jessie
 - Try to install the systemd-sysv package

The extra dependencies being pulled are (with the recommends):

root@fornost:/# apt-get install systemd-sysv
Reading package lists... Done
Building dependency tree   
Reading state information... Done
The following extra packages will be installed:
  acl dbus dmsetup libcap-ng0 libcryptsetup4 libdbus-1-3 libdevmapper1.02.1 
libexpat1 libpam-systemd systemd
Suggested packages:
  dbus-x11 systemd-ui
The following packages will be REMOVED:
  sysvinit-core
The following NEW packages will be installed:
  acl dbus dmsetup libcap-ng0 libcryptsetup4 libdbus-1-3 libdevmapper1.02.1 
libexpat1 libpam-systemd systemd systemd-sysv
0 upgraded, 11 newly installed, 1 to remove and 0 not upgraded.
Need to get 3597 kB of archives.
After this operation, 14.8 MB of additional disk space will be used.

Same without the recommends:

root@fornost:/# apt-get install --no-install-recommends systemd-sysv
Reading package lists... Done
Building dependency tree   
Reading state information... Done
The following extra packages will be installed:
  acl dmsetup libcryptsetup4 libdevmapper1.02.1 systemd
Suggested packages:
  systemd-ui
Recommended packages:
  libpam-systemd dbus
The following packages will be REMOVED:
  sysvinit-core
The following NEW packages will be installed:
  acl dmsetup libcryptsetup4 libdevmapper1.02.1 systemd systemd-sysv
0 upgraded, 6 newly installed, 1 to remove and 0 not upgraded.
Need to get 2934 kB of archives.
After this operation, 12.8 MB of additional disk space will be used.


As you can see the number of new dependencies being pulled are minimal.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/2014130718.2ce34...@fornost.bigon.be

Re: "Lennart Poettering Linux" -- some real eye openers here ... don't be blindsided!

[Laurent Bigonville]

Le Mon, 10 Nov 2014 09:51:49 -0300,
Renaud (Ron) OLGIATI  a écrit :

[...]
> 
> If so, what happened to "Linux is all about freedom of choice" ?

GNU/Linux is maybe about "freedom of choice" but GNU/Linux
distributions have never been. You see the difference?

Distribution are integrators that sometimes have to take decisions to
ensure stability, reduce complexity and/or reduce exponential test case
scenarios.

If you want more scenarios to be supported, a distribution needs
testing, patches and manpower, if nobody is doing the job, well
nothing happens.

> 
> Will the only freedom left be that of not using Linux ?

As long as the upstream Linux kernel is not checking whether the PID 1
is a particular process, you have all the freedom you want to
fork/create a distribution. And even if it was the case, the code would
still be free to read and modify.

>  
> Cheers,
>  
> Ron.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141110142232.738c6...@fornost.bigon.be

Re: forks, derivatives, other distros - what are you thinking/doing

[Laurent Bigonville]

Le Mon, 10 Nov 2014 13:16:09 +0100,
berenger.mo...@neutralite.org a écrit :

> Le 09.11.2014 05:05, Hendrik Boom a écrit :
> > On Wed, 05 Nov 2014 09:32:57 -0500, Miles Fidelman wrote:
> >> 1. What are your issues, reasons for doing so - general and/or 
> >> specific?
> >
> > I've had trouble with passwords in the network-manager starting a
> > few months ago.  I tried a few other wifi connectivity tools, and
> > ended up
> > with wicd.  What was different about wicd was that (i) it worked,
> > and (ii) it was independent of systemd.  I don't know whether the
> > introduction of expansion of systemd had anything to do with my 
> > problems.
> 
> This might be a policykit-related issue. AFAIK, policykit has been 
> deprecated in favor of systemd-login0.
>  From what you describe, it seems the migration from policykit to 
> systemd-login0 (which is not systemd itself, but only a module. I
> hope I'm not wrong here, since I do not really understand the
> language used in systemd-world).

Not exactly, polkit is still used by DE, but it now uses logind
(instead of ConsoleKit) to track the users.

You need to be sure that you have a logind session registered (see
loginctl). If the session is not present, you need to check if
pam_systemd.so is called in the PAM configuration (/etc/pam.d).

> 
> > I've started to have trouble mounting the NTFS partition on my 
> > machine
> > from Linux.  No problem doing this in Windows, of course.  I used
> > to be
> > able to mount it from the file manager after entering the root
> > password.  Starting a month or so ago, the file manager would
> > tantalizingly show me the partition but refuse to let me mount it 
> > because
> > I didn't have the proveleges.  Finally, it stopped even showing me 
> > that
> > partition.  Of course I cann still log in as root and mount it from 
> > the
> > command line, copy any files from it, and chown them to myself.
> > But it
> > is unnecessarily awkward.  I understand systemd had involved itselg 
> > with
> > permissions.  Could this be relevant?  I have the same problem with 
> > usb
> > sticks -- having to be root to use them.  Again, I have no idea 
> > whether
> > the architecture changes caused by systemd has any relevance to
> > this, but
> > the general level of paranoia that is starting to exist makes me
> > suspicious, perhaps unjustly.
> 
> This could be udev-related. Udev is the part of the system which
> fills /dev.
> It does this at boot, and while your computer run.
> 
> The current man-page says that, "the udev daemon, 
> systemd-udevd.service(8), receives device uevents directly from the 
> kernel whenever a device is added or removed from the system, or it 
> changes its state".
> You can build rules in /etc/udev, which could eventually allow you to 
> fix your problem by yourself. I can't really help you about how to
> write the rule, since I have never tried to build some myself.
> Why things have stopped working, is a good question: maybe a change
> in /dev/disk/by-uuid? Does things works anew if you create a user to
> login with (and so, with a clear $HOME)?
> 
> It seems that your problems seems to not be systemd-related, since 
> systemd is only the PID1 process' stuff. They are only related to
> things which are parts of the systemd...hum... sorry, softwares which
> shares the same source-code repository that systemd uses.
> Yeah, some irony here, indeed. I won't argue about the fact that 
> udev/login0 are or are not parts of systemd. I do not mind, in facts.
> 

IMHO the lack of permissions on a device could also be related to the
lack of a logind session.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141110140716.10142...@fornost.bigon.be

Re: Mount order after systemd update

[Laurent Bigonville]

Le Thu, 6 Nov 2014 16:04:10 +,
Jonathan Dowland  a écrit :

> On Sun, Nov 02, 2014 at 07:31:06PM +0100, Sven Hartge wrote:
> > I know. The package description is lacking crucial information and
> > the web page isn't any better.
> > 
> > See
> > http://web.dodds.net/~vorlon/wiki/blog/Plymouth_is_not_a_bootsplash/
> > (Vorlon is Steve Langasek.)
> 
> You're absoutely right. Would anyone volunteer the beginnings of a
> re-word of the package description, that could be submitted as a
> wishlist bug against the package? I was about to file it, then I
> realised it would be worth submitting a patch for such a simple bug.

Yes please, I adopted plymouth a few months ago and as English is not
my first language I would be really happy if someone could provide me a
proper description.

Please file a bugreport with the "minor" severity.

Thanks

Laurent Bigonville


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141106184815.237ec...@soldur.bigon.be

Re: Mount order after systemd update

[Laurent Bigonville]

Le Thu, 6 Nov 2014 00:08:56 +0100,
Martin Manns  a écrit :

> On Mon, 03 Nov 2014 10:00:03 +0100
> Jonathan de Boyne Pollard 
> wrote:
> 
> > Plymouth is right there at the top of the list.
> 
> Since there seems to be little choice of password agents, I tried it
> out. Now update-initramfs fails:
> 
> # update-initramfs -u -v
> [...]
> Adding binary /bin/plymouth
> Adding binary /sbin/plymouthd
> Adding binary /usr/lib/i386-linux-gnu/plymouth//text.so
> Adding binary /usr/lib/i386-linux-gnu/plymouth//details.so
> E: /usr/share/initramfs-tools/hooks/plymouth failed with return 1.
> Removing /boot/initrd.img-3.16-3-686-pae.dpkg-bak
> update-initramfs: failed for /boot/initrd.img-3.16-3-686-pae with 1.
> 
> Any ideas how to fix this?

Which version of plymouth is that? Did you changed the default theme or
something?

In /usr/share/initramfs-tools/hooks/plymouth, could you add "set -x"
just after the already existing "set -e" and try again? the output
should be more verbose.

Laurent Bigonville


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141106044805.5e61e...@fornost.bigon.be

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

Le Tue, 04 Nov 2014 22:00:12 -0500,
Miles Fidelman  a écrit :

[...]

> - SUSE seems to be moving down the systemd path, but with less 
> aggregation (e.g., systemd w/o journald)

I personally didn't checked, but I can bet money on it if you want,
that SLES 12 is doing _exactly_ what debian IS doing, aka installing a
regular syslog daemon and not creating a /var/log/journal directory
meaning that the journal is in memory only and not saved across reboot.
But journald daemon will be still running.

The following URL seems to confirm that:
https://www.suse.com/documentation/sles-12/book_sle_admin/data/journalctl_persistent.html


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141105164149.5e5b7...@soldur.bigon.be

Re: How to use the network-manager-strongswan

[Laurent Bigonville]

Le Wed, 05 Nov 2014 21:41:04 +0800,
Gulfstrean Wang  a écrit :

> Hello,

Hello,

> I want to use the network-manager-strongswan to connect ikev2 vpn
> server, but I can not find how to configure the
> network-manager-strongswan via UI or configuration file. Could you
> tell me how to use it if you know? Thank you very much!
> 
> My debian version is testing, and the window manager is gnome 3.14.

You could try to use nm-connection-editor to configure the connection.

BUT, I'm not actually sure about the working state of nm-strongswan.
Last time I tried it, I end up loosing the loopback interface when
disconnecting. Not sure if anything changed.

Also note that it will be not release with jessie due to a RC bug (it
seems FTBFS with recent gnome-keyring versions).


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141105162224.09be9...@soldur.bigon.be

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

Le Tue, 04 Nov 2014 17:26:40 -0500,
Miles Fidelman  a écrit :

> Laurent Bigonville wrote:
> > Le Tue, 04 Nov 2014 16:57:20 -0500,
> > Jerry Stuckle  a écrit :
> >
> >> On 11/4/2014 4:41 PM, Laurent Bigonville wrote:
> >>> Le Tue, 04 Nov 2014 21:08:50 +0100,
> >>> Peter Nieman  a écrit :
> >>>
> >>>> On 04/11/14 19:04, Laurent Bigonville wrote:
> >>>>> Using the threat of forking to make people change their mind
> >>>> I didn't threaten anybody.
> >>>>
> >>>>> do not send 100 mails to ML's
> >>>> I didn't. I don't even know what "ML's" are.
> >>> That was not directed to you, but a more generic statement about
> >>> people talking about forking debian.
> >>>
> >>> (ML's == Mailing List's)
> >>>
> >>>
> >> I don't see that as a "threat".  I see it as an attempt to
> >> determine if there are enough others interested to make it a
> >> viable project.
> >>
> >> And lists like this one are a perfect place to find out of there
> >> are other users interested in such a project.
> > http://debianfork.org/:
> >
> > "If systemd will be substituting sysvinit in Debian, we will fork
> > the project and create a new distro. We hope this won't be
> > necessary, but we are well prepared for it."
> >
> > I call that a threat. And the same kind of message are all around
> > the debian mailing lists and other social media.
> >
> >
> 
> Does it really matter if it's a "threat," or a statement of intent,
> or even just a wish?  You seem to imply that there's something wrong
> with such a statement.
> 
> Given the number of folks who are saying "if you don't like systemd,
> go somewhere else," and the tradition of forking open source software
> (and linux distributions) over various differences of opinion -- what
> is it that you're objecting to?
> 
> Personally, I see it as a clear and strong statement of
> dissatisfaction that's likely to lead to significant defection from
> Debian - to other existing distros, and possibly to one or more new
> forks.
> 
> What, exactly, is it that you are objecting to here?  That folks
> aren't just happy to go along with a major shift in Debian?  That
> people are making their opinions clear?  That people won't adopt
> Jessie, and will desert Debian once Wheezy starts to reach EOL? That
> folks might go so far as to create a new fork?  Something else?

I'm objecting to the fact that they are doing announcement like that
instead of actually forking. That kind of message contributes to the
never ending discussion that is for a long time draining the energy of a
lot of developers (including me).

The situation for jessie will NOT change (the freeze is the Thu 6/11 at
00:00 UTC around 25h from now), systemd will stay the default init
system for jessie as this decision is clearly not challenged by the GR.

The "funny" thing is that there is ATM NO software in the debian archive
that requires systemd as PID1 to run[0] (zero, null, nil, None). So even
if Ian's GR passes, nothing will in practice change. There is software
however requiring _components_ of systemd to run properly, but from my
understanding this is again NOT covered by the GR.

So people not happy with today's situation and calling to fork will not
be more happy tomorrow.

So to summarize:

 - Debian releases jessie
 - People that still want to fork do whatever they want
 - Everybody is happy
 - End of the story

[0] OK I lied there is systemd-cron

> 
> Miles Fidelman
> 
> 
> 
> 


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141105000853.5bd8f...@fornost.bigon.be

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

Le Tue, 04 Nov 2014 16:57:20 -0500,
Jerry Stuckle  a écrit :

> On 11/4/2014 4:41 PM, Laurent Bigonville wrote:
> > Le Tue, 04 Nov 2014 21:08:50 +0100,
> > Peter Nieman  a écrit :
> > 
> >> On 04/11/14 19:04, Laurent Bigonville wrote:
> >>> Using the threat of forking to make people change their mind
> >>
> >> I didn't threaten anybody.
> >>
> >>> do not send 100 mails to ML's
> >>
> >> I didn't. I don't even know what "ML's" are.
> > 
> > That was not directed to you, but a more generic statement about
> > people talking about forking debian.
> > 
> > (ML's == Mailing List's)
> > 
> > 
> 
> I don't see that as a "threat".  I see it as an attempt to determine
> if there are enough others interested to make it a viable project.
> 
> And lists like this one are a perfect place to find out of there are
> other users interested in such a project.

http://debianfork.org/:

"If systemd will be substituting sysvinit in Debian, we will fork the
project and create a new distro. We hope this won't be necessary, but
we are well prepared for it."

I call that a threat. And the same kind of message are all around the
debian mailing lists and other social media.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141104230955.29263...@fornost.bigon.be

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

Le Tue, 04 Nov 2014 21:08:50 +0100,
Peter Nieman  a écrit :

> On 04/11/14 19:04, Laurent Bigonville wrote:
> > Using the threat of forking to make people change their mind
> 
> I didn't threaten anybody.
> 
> > do not send 100 mails to ML's
> 
> I didn't. I don't even know what "ML's" are.

That was not directed to you, but a more generic statement about people
talking about forking debian.

(ML's == Mailing List's)


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141104224116.62c2b...@fornost.bigon.be

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

Le Tue, 04 Nov 2014 15:22:32 +0100,
Peter Nieman  a écrit :

> On 04/11/14 03:53, Jerry Stuckle wrote:
> > On 11/3/2014 8:36 PM, Joel Rees wrote:
> > 
> >> I suppose it may be polemic to assert that forking debian and
> >> setting up a new community would be labor-intensive, fractious,
> >> divisive, and general not a wise use of precious free/libre/open
> >> community resources, in short, "dumb".
> >>
> >
> > But just the fact there are people who consider systemd to be
> > problematic enough to consider forking Debian should not be ignored.
> >
> > I agree it would be labor-intensive, fractious and divisive.
> > However, if the people feel it is that important, I think it would
> > be a wise use of community resources.
> 
> Forking often makes things worse (classical example: libav) and it 
> should always be the last resort. But when two entirely different 
> "philosophies" exist inside a project and the two parts of a whole
> start moving in opposite directions and keep doing so for some time
> it might be a natural and perhaps the only sensible thing to do, and
> in that case I would call it "dumb" to simply call people suggesting
> a fork "dumb".

That probably the only part of your mail that I'll agree with you.

Using the threat of forking to make people change their mind or change
a policy that has been decided by the project (the CTTE has been
delegated by the project the power to take such decisions) is bad. If
you want to fork just do it, do not send 100 mails to ML's, fork and
start working.

It's free software after all.

> In Debian we have two different groups of people with entirely
> different visions. One that tries to stick to the "traditional" Linux
> (or Unix) way of doing things and one that tries to create something
> that I would call a copy of MS Windows, something that the first
> group ran away from. The latter group is backed by powerful
> commercial companies and paid developers, which brings the first
> group into a situation where it increasingly feels compelled to fight
> in order not to lose what it has learned to love. That's my
> experience with Debian over the last few years at least.

Not sure what you are implying here. That people that are supporting
systemd are somehow related to Red Hat or that they want to transform
Debian into a MS/Windows clone?

Having unified plumbing across distribution is good, it reduces the
maintenance burden, it allows to leverage other people work and
knowledge. I don't really see where the problem is.

> Now, my impression is that some people advocating things like Gnome
> and systemd here are so much "experts" and concerned with technical
> details that they fail to see the overall picture and the fundamental 
> differences. They essentially ask the critics (of systemd in this
> case) to just give in unless they have precise technical problems, in
> which case they should start coding or at least report these problems
> because they can all be solved, and in the end everyone can
> reasonably be expected to be happy. I guess these people would also
> deny that Gnome is intransparent and that it resembles MS Windows,
> because everyone can look at the code, after all.
> 
> But to the user, things look quite different. Even if the user hasn't 
> consciously installed anything like Gnome, he will see that more and 
> more things are happening on his computer that he doesn't want and 
> doesn't understand, and that he has to spend more and more time
> looking for ways to understand things, disable things and restore the
> way things were done before, if that is still possible at all. And
> when he digs deep enough he will often find that the reason why
> something has changed for the worse is that "it's the Gnome way".

The change of initsystem has nothing to do with GNOME (even if GNOME is
using some features extensively). systemd (or upstart) is solving long
standing issues regarding starting of daemon (clean environment,
selinux context, loginuid attribute or prevents other stuff that can
leak from the user session) and daemon life cycle management (being
sure that when a service is stopped all the processes are effectively
stopped). Then systemd add other features like private /tmp directory
using namespace or socket activation. All of these features are for
servers, again nothing to do with GNOME. AND in addition to these, it
gives DE an unique API to interact with the power state of the machine
(inhibition, notification about power state changes,...) and user
session management via logind.

> So to my mind the fundamental question is if you want to keep control
> of your computer or if you prefer eye candy and things happening 
> "automagically". And there is no middle way (only "extremist" ones).

I fail to see how you are loosing control of your computer as systemd
provides more logging and more information about the state of the
services running on your machine. Without talking about the descriptive
language used to describe to .se

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

Le Tue, 04 Nov 2014 01:29:52 +,
Jonathan de Boyne Pollard  a
écrit :

> Laurent Bigonville:
> > The systemd umbrella project is  made of 10+ different executables
>  > that have all a specific scope (systemd PID1 used to manage the
>  > life cycles of the daemons, systemd-logind manage the user
>  > sessions, systemd-journald a logging system,...) and that are all
>  > communicating using well defined, stable and documented dbus
>  > interfaces that allow one to reimplement the functionalities as
>  > long as it exposes the same interfaces (ie. this is what
>  > systemd-shim is doing).
> 
> Not correct.  The "systemd process" D-Bus API is not stable and not 
> covered by the interface guarantee.  It's exactly this that is part
> of the hoo-hah and part of the problem with systemd-shim over the
> past year or so.
> 
> * 
> http://homepage.ntlworld.com./jonathan.deboynepollard/FGA/debian-systemd-packaging-hoo-hah.html

These are the "internal" API if I'm not wrong. The external ones are
stable as explained:

http://www.freedesktop.org/wiki/Software/systemd/InterfacePortabilityAndStabilityChart/
http://www.freedesktop.org/wiki/Software/systemd/InterfaceStabilityPromise/

But indeed you are correct, systemd PID1 and logind are tightly coupled
(the "Reimplementable Independently" column) and a project trying to
reimplement only one of them will have to play catch and run with
upstream. But this is NOT preventing to have reimplementations at all
as the external API used by 3rd party are stable. systemd is not alone
when providing stable external API and unstable internal ones.

Also note that this is not a huge issue for distribution like debian
that provides frozen in time stable release as we have a way to
synchronize the version (this is the job of a distribution) shipped to
our users.

Well at least this is my understanding of the matter.

Laurent Bigonville


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141104165653.4df3a...@soldur.bigon.be

Re: Mount order after systemd update

[Laurent Bigonville]

Le Sun, 2 Nov 2014 16:17:05 +0100,
Martin Manns  a écrit :

> Dear all

Hello,
 
> After switching to systemd, I would like to get back the following
> behavior:
> 
> Mount multiple lvm-crypt volumes with password entry on startup.
> Mount several loopback devices from files within these volumes.
> 
> With sysvinit, I had put the mount order into /etc/fstab and
> everything worked as expected.

Wild guess here, you should try to set the "use_lvmetad" to 1
in /etc/lvm/lvm.conf and see if it helps.

> 
> After switching to systemd, mount operations seem to be spawned in
> parallel. This has the following consequences:
> 
> 1) I am never sure which device requests password entry first.
> Therefore, password choice is a gamble. Furthermore, password entry on
> startup looks weird because some weird red moving stars are shown
> instead of a prompt.

For this issue I would suggest to install plymouth, it should
serialized the output on the console and allow you to have a better
idea of what's happening on the console.

> 
> 2) If I wait for some time before entering all passwords then some
> kind of timeout seems to kick in and the volume is not mounted at all.
> It is actually hard to enter everything without any timeout.

See systemd.mount(5) manpage, there is a way to increase the timeout.

> 3) If a loopback device is mounted before a preceding encrypted
> device then it obviously fails. However, this prevents correct startup
> and gives me an emergency console.

I guess you could write a .mount unit file if lvmetad is not helping,
you should be able to express dependencies there.

Cheers,

Laurent Bigonville


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141102175125.6519a...@fornost.bigon.be

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

Le Sat, 01 Nov 2014 14:24:28 -0400,
Miles Fidelman  a écrit :

> Laurent Bigonville wrote:
> > Le Sat, 01 Nov 2014 07:56:30 -0400,
> > Miles Fidelman  a écrit :
> >
> >> Steve McIntyre wrote:
> >>> Miles Fidelman wrote:
> >>>> Martin Read wrote:
> >>>>> On 01/11/14 01:53, lee wrote:
> >>>>>> It doesn't need these code paths.  The library doesn't do
> >>>>>> anything unless you do have the software actually running which
> >>>>>> the library makes useable --- at least that's what was said.
> >>>>>>
> >>>>>> Of course, not all cases are the same, yet in this case, the
> >>>>>> library shouldn't be installed unless the software it is for is
> >>>>>> installed.
> >>>>> Gentoo and Funtoo are > over there, just like they were
> >>>>> months ago when you first started complaining about systemd on
> >>>>> debian-user.
> >>>>>
> >>>>> If you move over to using them instead of Debian, you'll
> >>>>> probably be happier (because you'll have more control over what
> >>>>> software runs on your systems and how it's configured) and the
> >>>>> Debian maintainers will probably be happier (because there will
> >>>>> be one fewer person haranguing them for refusing to embrace
> >>>>> combinatorial explosion).
> >>>>>
> >>>>> Everyone wins.
> >>>> Right.  This sounds more and more like "we're going to rewrite
> >>>> the rules, and if you don't like it, we're taking our ball and
> >>>> going home."
> >>> Various people have tried to explain how a binary distribution
> >>> like Debian works (build packages with all options included by
> >>> defauls) and how shared libraries work on Linux (all the
> >>> libraries need to be there to satisfy symbol resolution at run
> >>> time, even if none of the code is ever used). When those
> >>> explanations fell on deaf ears, people have resorted to analogy.
> >>> That was clearly a waste of time too.
> >>>
> >>> These are standard "rules" that have existed for many years, there
> >>> is no rewriting going on at all. Instead, it seems there are
> >>> people who won't, or don't want to, understand explanations when
> >>> given. For people who claim to have technical backgrounds, that's
> >>> a surprising (and very frustrating) problem.
> >>>
> >> Yeah... the Unix way... which systemd and it's pieces violate in so
> >> many ways.
> > Surprisingly 10th of different executables talking to each other
> > using a common IPC mechanism (dbus here) seems to be really "unixy"
> > to me...
> >
> >
> 
> First off, we're talking about the hairball that is systemd, not the
> one specific piece of the ecosystem that is DBUS.

I was talking about systemd.

> Second, we're not talking about vaguely "unixy" - we're talking about
> a well developed philosophy of designing things that dates back to
> Ken Thompson, et. al (c.f., "The UNIX Programming Environment,"or 
> http://en.wikipedia.org/wiki/Unix_philosophy).

The systemd umbrella project is made of 10+ different executables that
have all a specific scope (systemd PID1 used to manage the life cycles
of the daemons, systemd-logind manage the user sessions,
systemd-journald a logging system,...) and that are all communicating
using well defined, stable and documented dbus interfaces that allow
one to reimplement the functionalities as long as it exposes the same
interfaces (ie. this is what systemd-shim is doing).


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141102005704.43119...@fornost.bigon.be

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

Le Sat, 01 Nov 2014 07:56:30 -0400,
Miles Fidelman  a écrit :

> Steve McIntyre wrote:
> > Miles Fidelman wrote:
> >> Martin Read wrote:
> >>> On 01/11/14 01:53, lee wrote:
>  It doesn't need these code paths.  The library doesn't do
>  anything unless you do have the software actually running which
>  the library makes useable --- at least that's what was said.
> 
>  Of course, not all cases are the same, yet in this case, the
>  library shouldn't be installed unless the software it is for is
>  installed.
> >>> Gentoo and Funtoo are > over there, just like they were
> >>> months ago when you first started complaining about systemd on
> >>> debian-user.
> >>>
> >>> If you move over to using them instead of Debian, you'll probably
> >>> be happier (because you'll have more control over what software
> >>> runs on your systems and how it's configured) and the Debian
> >>> maintainers will probably be happier (because there will be one
> >>> fewer person haranguing them for refusing to embrace
> >>> combinatorial explosion).
> >>>
> >>> Everyone wins.
> >> Right.  This sounds more and more like "we're going to rewrite the
> >> rules, and if you don't like it, we're taking our ball and going
> >> home."
> > Various people have tried to explain how a binary distribution like
> > Debian works (build packages with all options included by defauls)
> > and how shared libraries work on Linux (all the libraries need to
> > be there to satisfy symbol resolution at run time, even if none of
> > the code is ever used). When those explanations fell on deaf ears,
> > people have resorted to analogy. That was clearly a waste of time
> > too.
> >
> > These are standard "rules" that have existed for many years, there
> > is no rewriting going on at all. Instead, it seems there are people
> > who won't, or don't want to, understand explanations when given. For
> > people who claim to have technical backgrounds, that's a surprising
> > (and very frustrating) problem.
> >
> 
> Yeah... the Unix way... which systemd and it's pieces violate in so
> many ways.

Surprisingly 10th of different executables talking to each other using
a common IPC mechanism (dbus here) seems to be really "unixy" to me...


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141101175853.408f5...@fornost.bigon.be

Re: Preventing the computer from shutting down.

[Laurent Bigonville]

Le Thu, 30 Oct 2014 12:11:25 -0600,
Mario Castelán Castro  a écrit :

> Hello.

Hello,

> I can set up a script for backup with cron or anacron, but how can I 
> prevent the computer from shutting down while the backup is being 
> performed so as to not to leave it incomplete?.

If you are using systemd, you can use systemd-inhibit command to do
this.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141030200816.6d780...@fornost.bigon.be

Re: Suggestion for systemd and /usr on seperate partition

[Laurent Bigonville]

Le Thu, 30 Oct 2014 10:27:50 +0100,
Hans  a écrit :

> Dear maintainers,
> 
> completely without starting any flamewars: 
> 
> I am using systemd and I have /usr mounted on a separate partition as
> well as /var, /home, /boot and /.
> 
> Additionally /usr, /var and /home are luks encrypted. 
> 
> Due to this profile, I get a lot of annoying errors, as systemd does
> not find /usr when it is started, because it produces an error and
> then switches to verbose mode. This is very annoying!
> 
> For a new installation it might be ok, to put /usr on the root
> partition, but I guess, there are a lot of systems in the world
> running a partition profile like mine.
> 
> Besides of the mentioned problem systemd is running well.
> 
> I thought about this problem. Might it be possible, to change systemd
> in that way, that it will start after all partitions are mounted? I
> know, it must be done in the source code, but as I am no coder, I
> cannot do it myself.
> 
> So I ask the developers hereby, maybe it wil be possible to do that.

I bet you are running jessie/testing?

This issue should be fixed in initramfs-tools >= 0.117 that is
currently in unstable (and unfortunately blocked by some RC bugs).

The newer version include the following change:

  * Mount /usr if present in the /etc/fstab on the mounted rootfs
(Closes: #652459)

I personally really hope this change will make it into jessie.

Cheers,

Laurent Bigonville


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141030145536.601a6...@soldur.bigon.be

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

Le Thu, 30 Oct 2014 02:24:27 -0400,
Marty  a écrit :

> On 10/29/2014 06:53 AM, Laurent Bigonville wrote:
> > Le Wed, 29 Oct 2014 00:52:54 -0400,
> > Marty  a écrit :
[...]
> >> By "problem" I meant what I consider the problem of not having an
> >> overlap between old and new solutions, and no deprecation period or
> >> warning. I don't argue that it should not be corrected. My point
> >> was more of a policy and design strategy issue.
> >
> > There is NO functional change at all, what was working before is
> > still working now. The only difference is that you have an extra
> > daemon running from the start (and running in a clean context).
> 
> The deprecation issue would not apply to Jessie, just some legacy
> code.

I'm not following you here
 
> > The daemon here on my machine is using 1224KiB of resident memory,
> > this is nothing on modern machines. I personally don't even see why
> > we are discussing this tbh.
> 
> I don't know about you, but I am here because I am trying to decide
> if this is a case of "accidental vendor lock-in" and more broadly if
> there is a backward compatibility policy issue that encourages it, or
> if this is an isolated instance.
> 
> There's also the matter of the missing init script, doubts about init 
> script support, the freeze, the GR, the vote, and questions about 
> deprecation policies in general, as background. At this point I'm 
> willing write it off as undecidable, at least by me. :)

There is a /etc/init.d/uuidd initscript currently in jessie, see:

https://packages.debian.org/jessie/amd64/uuid-runtime/filelist
http://sources.debian.net/src/util-linux/2.25.1-5/misc-utils/uuidd.rc.in/

And it has always been clear (at least to me) that all the packages must
continue to support sysvinit and not remove any LSB initscripts for the
jessie release.


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141030092935.3400d...@fornost.bigon.be

Re: Perfect Jessie is something like this...

[Laurent Bigonville]

 always starting uuidd would have probably
> > been the best solution.
> 
> By "problem" I meant what I consider the problem of not having an 
> overlap between old and new solutions, and no deprecation period or 
> warning. I don't argue that it should not be corrected. My point was 
> more of a policy and design strategy issue.

There is NO functional change at all, what was working before is still
working now. The only difference is that you have an extra daemon
running from the start (and running in a clean context).

The daemon here on my machine is using 1224KiB of resident memory, this
is nothing on modern machines. I personally don't even see why we are
discussing this tbh.

Laurent Bigonville


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141029115314.188b1...@soldur.bigon.be

Re: Understanding DNS, Create an "Failover"

[Laurent Bigonville]

basti wrote:
> Hello,

Hello,

> last weekend my primary DNS-Server goes down, and some of my server
> can't find each other.
> 
> I have a Primary and a Secondary DNS-Server using bind9.
> The resolv.conf file looks like:
> 
> nameserver 
> nameserver 
> nameserver <2'ndOfISP>
> 
> For understanding:
> Is the secondary DNS just a backup of the Master for "loadBalancing"?
> What does the secondary DNS do if master is down?
> 
> How can I fix this?
> 

According to the resolv.conf(5) manpage the nameservers are used in the
specified order. The extra nameservers are used as backup if the
previous one is not answering. There is no load balancing done here.

"The algorithm used is to try a name server, and if the query times
out, try the next, until out of name servers, then repeat trying all
the name servers until a maximum number of retries are made."

Cheers,

Laurent Bigonville


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141028110448.65cf2...@soldur.bigon.be

Re: Who's locking down the code?

[Laurent Bigonville]

Marty wrote:
> [...]
> I looked and did not find the script. I still don't know what the
> original problem was (although I didn't dig very deep). In any case I
> don't see how the script is more "elegant" than the original
> solution. Can anyone explain what's going on here?

You mean why is it better to either socket activate uuidd or run in from
the start instead of making the library start it on demand?

If so, I would suggest you to read Christian's mail:
https://lists.debian.org/debian-user/2014/10/msg02380.html

Cheers,

Laurent Bigonville


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141028092028.049ba...@fornost.bigon.be

Re: Who's locking down the code?

[Laurent Bigonville]

Andrew McGlashan wrote:
> On 26/10/2014 4:30 AM, goli...@riseup.net wrote:
> >> The fact that an executable is linked against a systemd library
> >> doesn't automatically mean you have to run systemd as PID1.
> >>
> >> This is especially true for the sd-daemon and sd-journal libraries
> >> in this case.
> >>
> >> Laurent Bigonville
> >>
> > 
> > 
> > I have heard that argument before.  I counter that it's about more
> > than PID1.  It seems that even having systemd libraries etc. is a
> > little like being somewhat pregnant - precursors to a little bundle
> > of joy to be delivered at a later date when the PTB see fit. In
> > other words, a trojan of sorts that will come to bite us. Sorry,
> > not much trust these days . . .  :(
> 
> That is 100% true, I couldn't give a rats if it is PID1 or not.  It IS
> systemd, that's more than enough for me to want it OUT -- it's a
> cancer that is spreading and it needs to be eradicated *before* it is
> nigh impossible.

It is an habit in debian to compile the packages with as many options
as possible as long as it's not adding pile of new dependencies or
causing issues to the other packages in the archive.

I don't see how a library that turns itself into a noop if PID1 is
not systemd fits into any of these 2 categories in the case of
util-linux (or probably any packages depending against libsystemd0).

IMHO, if you have the (non-technical?) requirement to not have any
systemd component on your system, you'll have to either start building
your own packages (you can have a look at apt-build) and maybe propose
sensible patches to make it easier for the debian users to opt-out when
rebuilding packages. Or switch to a distribution that allows you to
select which components are enabled at build time.

Laurent Bigonville


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141026122446.2d8f2...@fornost.bigon.be

Re: Who's locking down the code?

[Laurent Bigonville]

John Hasler wrote:
> Slavko writes:
> > IMO there can be two versions of these packages in the repositories,
> > one which depends on systemd and one which don't depends on it, but
> > i afraid, that there will be response: "not enough resources".
> 
> Any developer can package and upload anything that is legal to
> package. There is no central authority determining what there is and
> is not enough resources for.  Just convince one developer to sponsor
> your systemd-free packages and you're there.

You forgot about the ftp-masters, the release team and the security
team. They all have their word to say about which package can or cannot
be in the archive or part of a release. And usually they are not really
pro having the same code base twice in the archive.


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141025231523.26388...@fornost.bigon.be

Re: Who's locking down the code?

[Laurent Bigonville]

golinux wrote:
> Would appreciate comments on this observation:
> 
> "Look at the source code (while you're at it, note who are the
> upstream maintainers of util-linux). Even they (so far) allow
> compilation without systemd. It is Debian who are introducing systemd
> dependencies even where it is actually optional in the upstream
> source."
> 
> 
> If upstream is allowing choice, why is Debian cutting it off? Maybe
> I'm missing something . . .

The fact that an executable is linked against a systemd library doesn't
automatically mean you have to run systemd as PID1.

This is especially true for the sd-daemon and sd-journal libraries in
this case.

Laurent Bigonville

> 
> 
> golinux
> 


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141025191427.7cdd6...@fornost.bigon.be

Re: Re: Synaptic slow when Caribou is running after Gnome-shell update

[Laurent Bigonville]

Stephen Allen wrote:
> On Mon, Oct 13, 2014 at 03:35:51PM +0200, Michael Biebl wrote:
> > Am 13.10.2014 um 12:06 schrieb Stephen Allen:
> > > I too had experienced this, seems to be fixed as of yesterday.
> > > Thanks Michael.
> > 
> > Unfortunately, this is not fixed yet.
> > Apparently the start is not triggered via the xdg autostart file,
> > as I had originally suspected, but rather via D-Bus activation,
> > i.e.
> > via /usr/share/dbus-1/services/org.gnome.Caribou.Daemon.service.
> 
> Well, Synaptic isn't waiting 5 minutes to start anymore. :)

This has been fixed with the at-spi-* updates to 3.14


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141014170136.023ec...@soldur.bigon.be