systemd-tmpfiles breaks daily builds

[zeha]

Setting up sudo (1.9.17p2-7) ...
/proc/ is not mounted, but required for successful operation of systemd-tmpfiles. Please mount /proc/. Alternatively, consider using the --root= or --image= switches.
dpkg: error processing package sudo (--configure):
 old sudo package postinst maintainer script subprocess failed with exit status 1

[zeha]

Currently we cannot mount /proc, as the builds run inside an OCI container (unprivileged of course).
We already tried switching docker for podman, but doesn't help a thing - the container still doesn't have CAP_SYS_ADMIN which is needed for mount.

[zeha]

Current plan: do away with the containers altogether, let grml-live run as a normal user, have it create its own userns as needed.

[mika]

Do we have a tracking bug in either Debian's BTS or systemd upstream that's related to this failure / behavior change? 🤔

[zeha]

Do we have a tracking bug in either Debian's BTS or systemd upstream that's related to this failure / behavior change? 🤔

No. The last time I reported something similar in another systemd binary, it was closed saying that our build host is broken, effectively claiming that CI with unprivileged containers is not a support target.

[mika]

Do we have a tracking bug in either Debian's BTS or systemd upstream that's related to this failure / behavior change? 🤔

No. The last time I reported something similar in another systemd binary, it was closed saying that our build host is broken, effectively claiming that CI with unprivileged containers is not a support target.

😢

I recall that, but couldn't find the related bug number or alike, do you recall the bug number or so where this happened? :)
(Would be nice to be able to point people to this? IMO this is a problem within systemd :-/)

[zeha]

systemd/systemd#38632 was back then