autopkgtest success rate in Bionic reached an unusable amount

Bug #1817721 reported by Christian Ehrhardt  on 2019-02-26
This bug affects 2 people
Affects Status Importance Assigned to Milestone
systemd (Ubuntu)

Bug Description

Sorry to bother you - it seems we have this discussion every few months or so.

I was looking at this due to an SRU and I realized that the last 50-100 tests on all architectures are currently failing.

On one hand I see the known flaky boot-and-smoke but that is fine.
On the other hand I see the test "upstream" failing always and it seems not related to the actual proposed packages.

Please help to resolve this mid-term.
Short term I'll submit a force-badtest for this as it seems to hold up more and more with no gain of seeing plenty of peopl just hitting retry every now and then.

If test "upstream" is utterly broken consider skipping it in the next upload that you do for regular service. That would bring back at least the remaining tests coverage.

Related branches

Dan Streetman (ddstreet) wrote :

I'm fixing this.

Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in systemd (Ubuntu):
status: New → Confirmed

So after some investigation on systemd failures on ppc64el triggered by the kernel on eoan, I noticed some coredumps on processes when running under qemu inside the ADT environment.

As one of these processes is sleep, I am pretty confident this is caused by the lack of emulation or rather the failure to ignore mffsl extra bits, aka LP: #1847806.

I am not sure this is the right bug as it mentions lots of other upstream failures across all arches. But it looks like the ones we are seeing, and as we are referring this bug in our results, I thought it was worht mentioning.


To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers