Control: tags -1 moreinfo

Hello Scott,

this really looks like an old bug that was fixed 9 months ago..

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=%231069075

but I fail to understand how this can happen in a fresh Trixie install.

when you experience the bug, where /etc/service/elogind link is
pointing?

On Sun, 2 Feb 2025 19:14:44 +0000
Andrew Bower <and...@bower.uk> wrote:
> It's probably worth adding the version of your runit packages to this
> bug report because a new version arrived in trixie TODAY with major
> changes (2.1.2-61)!
>
>  dpkg -l 'runit*'

yes, especially runit and runit-services package version are relevant
here.

On Sun, 02 Feb 2025 14:00:19 -0500
Scott Fitzgerald <sf@expat.email> wrote:

> Package: elogind
> Version: 255.17-1debian1
> Severity: important
> X-Debbugs-Cc: sf@expat.email
> 
> Dear Maintainer,
> 
> I did a fresh install of trixie on an old laptop, a "cli only"
> install to start, and chose a runit-based version.  Noticed a
> two-minute or so delay on every login.  Eventually I noticed that
> elogind was not running unless I manually started it from a terminal,
> so I looked at the init scripts for elogind, and noticed at the
> bottom it was referencing the wrong location, so I corrected it but
> this was not a fix.  Then I noticed that the run script was trying to
> use "invoke-run" which, according to the man page, attempted to start
> via an /etc/init.d script.

invoke-run may try to stop the SysV instance before starting the runit
service, it doesn't start a sysvinit script.

For reference, the current elogind runscript in runit-services is:

~$ tree /usr/share/runit/sv.current/elogind
/usr/share/runit/sv.current/elogind
├── check
├── control
│   └── t
├── env
│   └── SYSTEMD_LOG_TARGET
├── finish
├── log
│   ├── run -> /etc/sv/svlogd/run
│   └── supervise -> /run/runit/supervise/elogind.log
├── run
└── supervise -> /run/runit/supervise/elogind

run file and .meta/bin content are relevant
---------------------------------------
$ cat /usr/share/runit/sv.current/elogind/run 
#!/usr/bin/env /lib/runit/invoke-run
#Copyright: 2022-2024 Lorenzo Puliti <plore...@disroot.org>
#License: CC0-1.0

sv start dbus ||  exit 170
exec 2>&1

if [ -e /etc/runit/verbose ]; then
        echo "invoke-run: starting ${PWD##*/}"
fi
exec /usr/libexec/elogind

-------------------------
$ cat /usr/share/runit/sv.current/elogind/.meta/bin 
/usr/libexec/elogind


Best,
Lorenzo

>  So I tried things like
> "/etc/init.d/elogind start/status " etc, but got nothing.  Not even a
> "starting ...." message or some error about status.
> 
> I kinda gave up and just wrote my own simple runit run script using
> "exec chpst -P /usr/libexec/elogind" which, even though it throws one
> error while waiting for dbus to start, does start the daemon running
> in the background.  Now I have fast cli logins and can run loginctl.
> 
> Since I have your attention I wish to thank you and the boot
> diversity team because you have saved me from having to run a
> non-debian distribution just because I don't care for the systemd
> init system.
> 
> ---
> 
> Scotty Fitzgerald
> 
> -- System Information:
> Debian Release: trixie/sid
>   APT prefers testing
>   APT policy: (500, 'testing')
> Architecture: amd64 (x86_64)
> 
> Kernel: Linux 6.12.11-amd64 (SMP w/4 CPU threads; PREEMPT)
> Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8),
> LANGUAGE not set Shell: /bin/sh linked to /usr/bin/dash
> Init: runit (via /run/runit.stopit)
> LSM: AppArmor: enabled
> 
> Versions of packages elogind depends on:
> ii  dbus                 1.16.0-1
> ii  debconf              1.5.89
> ii  init-system-helpers  1.68
> ii  libacl1              2.3.2-2+b1
> ii  libc6                2.40-6
> ii  libcap2              1:2.66-5+b1
> ii  libmount1            2.40.4-1
> ii  libpam0g             1.5.3-7+b1
> ii  libselinux1          3.7-3.1
> ii  libsystemd0          257.2-3
> ii  libudev1             257.2-3
> 
> Versions of packages elogind recommends:
> ii  libpam-elogind  255.17-1debian1
> ii  polkitd         126-2
> 
> elogind suggests no packages.
> 
> -- no debconf information
> 

Reply via email to