Bug#1095025: elogind: Initialization Scripts fail to start elogind [Trixie]
Lorenzo
plorenzo at disroot.org
Sun Feb 2 23:37:21 GMT 2025
Control: tags -1 -moreinfo
Control: retitle -1 elogind service broken in bookworm-trixie upgrade
Control: reassign -1 runit-services 0.7.2
On Sun, 2 Feb 2025 16:05:06 -0500
Scotty Fitzgerald <sf at expat.email> wrote:
> Greetings Lorenzo!
>
> runit is at version 2.1.2.-60
> runit-services is at version 0.7.2
>
> /etc/service/elogind does not exist, however I note that
> /etc/service points to /etc/runit/runsvdir/current and
> /etc/runit/runsvdir/current points to /etc/runit/runsvdir/default and
> /etc/runit/runsvdir/default/elogind points to /etc/sv/elogind and
> /etc/sv/elogind/run refers to /lib/elogind/elogind not the correct
> /usr/libexec/elogind
>
> Interesting, it does look like the same bug.
>
> My method for getting my fresh trixie install is taken from the
> webpage at....
> https://wiki.debian.org/DebianTesting#How_to_install_Debian_.28next-stable.29_Testing
>
> Which gave the following method..
> 1) Use stable installer to install stable
> 2) modify /etc/apt/sources.list to use "trixie" instead of stable.
> 3) Use apt pinning to lock out systemd-init programs
> 4) perform dist-upgrade
>
> Could the upgrade process leave behind old init scripts?
I can see how this happens now: if in point 1) runit and runit-services
are installed from stable, then a copy of the elogind runscript with the
wrong elogind path is left in /etc/sv/elogind, triggering this bug.
It won't be a problem for fresh trixie installs, but systems
upgrading from bookworm to trixie are affected:
Scott, thanks for reporting this!
This is a bug in runit-services and need to be fixed there, reassigning.
>
> Thank you for your attention!
> ---
> Scotty Fitzgerald
>
More information about the Debian-init-diversity
mailing list