Autopkgtest system images include ifupdown which doesn't work
```irc 15:42 < arnaudr[m]> Hello there! I'm investigating some autopkgtest failures on debusine.debian.net, for bullseye LTS. In 2 of them, I noticed the same error. ifup is unhappy with ifup[183]: Cannot find device "eth0", which leads the systemd networking.service to fail. In one case it's clearly the cause of the autopkgtest failure. Is it a known issue, is there a bug already opened to track that? Thanks! 15:43 < jochensp> arnaudr[m]: that sounds like a bug in the test as it will likely fail soon on ci.d.n also 15:50 < arnaudr[m]> jochensp: I don't think so, the error happens during container startup (you have to dump the journal to see it). The error in itself doesn't necessary fail the test, well, it depends. 15:51 < jochensp> uh, then I don't know. Do you have a link? 15:51 < arnaudr[m]> Case 1: I see it in systemd autopkgtest 'boot-and-services', it tests that no service failed on startup IIUC. Avail at https://debusine.debian.net/debian/developers-arnaudr-libapache2-mod-python-2/artifact/3566332/, 15:52 < arnaudr[m]> Case 2: I see it in libapache2-mod-python autopkgtest. I modified the test for debug, and I can see the error as I dump the journal. However it's not certain that it's really the cause of this autopkgtest failure. Avail at https://debusine.debian.net/debian/developers-arnaudr-libapache2-mod-python-2/artifact/3566332/ 15:52 < arnaudr[m]> Sorry, the systemd link is this one: https://debusine.debian.net/debian/developers/work-request/566611/ 15:55 < arnaudr[m]> For libapache2-mod-python: I can't reproduce it locally, I tried with autopkgtest and incus backend. But in my local test, ifupdown is not installed in the testbed. 19:25 < tumbleweed> Case 1: I see ifupdown in the installed package list, that's presumably the problem 19:25 < tumbleweed> It's an issue in our base image 19:32 < tumbleweed> it is Priority: important, so it's included in the default mmdebstrap image variant (which is what we use for autopkgtests) 19:33 < tumbleweed> we should explicitly exclude it ```
issue