Bug#923674: systemd: FTBFS (failing tests)

[Felipe Sateler]

Control: severity -1 important.

Resetting severity because this is not a problem in buildds, nor on my
own laptop.

On Sun, Mar 3, 2019 at 1:21 PM Santiago Vila <sanvila at debian.org> wrote:
>
> Package: src:systemd
> Version: 240-2
> Severity: serious
> Tags: ftbfs
>
> Dear maintainer:
>
> I tried to build this package in sid but it failed:
>
> The above is just how the build ends and not necessarily the most relevant part.
>
> I was able to build this package on buster just fine until version 239-15.
> Starting with version 240-2 it always fail for me.
>
> I'm using sbuild + schroot. I've tried so far:
>
> Self-hosted KVM machines with one CPU and 6 GB of RAM.
> 1-XS instances from Scaleway with one CPU and 1GB of RAM.
> 1-S instances from Scaleway with *two* CPUs and 2GB of RAM:
> n1-standard-1 instances from GCE with one CPU and 3.75 GB of RAM.
>
> You will see there is not a pattern here at all. Neither the amount of
> RAM or the number of CPUs seem to have any relation with the failure.
> (While monitoring RAM usage, I've never seen systemd build to require
> more than 500 MB to build anyway).
>
> It is possible that the package now requires an extraordinary amount
> of bogomips to build?

I checked a couple and they both fail on test-stat-util.  Any chance
you could get a backtrace of the failing test?

My guess is that sbuild setups something in /dev in a way that
test-stat-util is not liking. What is your sbuild setup? It's not
failing for me with sbuild either.

-- 

Saludos,
Felipe Sateler