linux-aws-5.15 ADT test MISS because it's unable to find package

[Bethany Jamison]

SRU cycle 2024.03.04 Focal aws-5.15 ADT test linux-aws-5.15 for both amd64 and arm64 results in a MISS with the error message below. It seems like the test was unable to locate the kernel-testing--linux-aws-5.15--modules-extra--preferred$ package which led to other missing packages and the test erroring out with exit code 1. This test has been SKIPPED before but with seemingly different reasons.

I have also attached the whole log amd64 output to this bug report.

"339s Reading state information...
339s E: Unable to locate package ^kernel-testing--linux-aws-5.15--modules-extra--preferred$
339s E: Couldn't find any package by glob '^kernel-testing--linux-aws-5.15--modules-extra--preferred$'
339s E: Couldn't find any package by regex '^kernel-testing--linux-aws-5.15--modules-extra--preferred$'
339s Reading package lists...
339s Building dependency tree...
339s Reading state information...
339s E: Unable to locate package ^linux-modules-extra-aws-5.15$
339s E: Couldn't find any package by glob '^linux-modules-extra-aws-5.15$'
339s E: Couldn't find any package by regex '^linux-modules-extra-aws-5.15$'
339s autopkgtest [16:53:45]: rebooting testbed after setup commands that affected boot
363s autopkgtest [16:54:09]: testbed running kernel: Linux 5.15.0-1057-aws #63~20.04.1-Ubuntu SMP Mon Mar 25 10:28:36 UTC 2024
363s autopkgtest [16:54:09]: @@@@@@@@@@@@@@@@@@@@ apt-source linux-aws-5.15
364s blame: linux-aws-5.15
364s badpkg: rules extract failed with exit code 1
364s autopkgtest [16:54:10]: ERROR: erroneous package: rules extract failed with exit code 1"

[Brian Murray (brian-murray)]

This also happens with linux-aws-6.5 on jammy amd64. Sorry for the bad formatting:

                                                                                                                         767s 0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
                                                                                                                         768s autopkgtest [15:06:12]: rebooting testbed after setup commands that affected boot
                                                                                                                         784s autopkgtest [15:06:28]: testbed running kernel: Linux 5.15.0-102-generic #112-Ubuntu SMP Tue Mar 5 16:50:32 UTC 2024
                                                                                                                         785s autopkgtest [15:06:29]: @@@@@@@@@@@@@@@@@@@@ apt-source linux-aws-6.5
                                                                                                                         786s blame: linux-aws-6.5
                                                                                                                         786s badpkg: rules extract failed with exit code 1
                                                                                                                         786s autopkgtest [15:06:30]: ERROR: erroneous package: rules extract failed with exit code 1
                                                                                                                         796s Creating nova instance adt-jammy-amd64-linux-aws-6.5-20240404-145324-juju-7f2275-prod-proposed-migration-environment-2-e4e1e79e-1749-4ae4-8386-7f19e3be6982 from image adt/ubuntu-jammy-amd64-server-20240404.img (UUID dc235da2-b119-4ee4-b09e-7c7ec6013d48)

[Paride Legovini (paride)]

Reproducer:

autopkgtest -U --apt-pocket=proposed linux-aws-5.15 -- qemu autopkgtest-focal-amd64.img

[Paride Legovini (paride)]

The issue is likely somewhere in

https://salsa.debian.org/ci-team/autopkgtest/-/blob/933bd29bb628bad8a66b877298ebcfaeb4aa41a6/runner/autopkgtest#L348

[Paride Legovini (paride)]

I now have an idea of where the problem is. This will be an interesting one.

@ubuntu-qa people: this may be related to the changes in 30d5aa7bd48ac15141f106dafb5930a1b5ac942f which we put in production with the switch to 5.32ish. So I expect this issue to be "new" but not super new (i.e. fist hit in the past month or so). I hope this matches observations.

@fginther: IIRC you also hit this, right? MM is down so I can't check the backlog.

[Francis Ginther (fginther)]

@paride: Yes, I've seen this with other kernels, mostly with the nvidia drivers. I think all of the runs of the following since March 20 show this problem:

https://autopkgtest.ubuntu.com/packages/n/nvidia-graphics-drivers-510-server/focal/amd64
https://autopkgtest.ubuntu.com/packages/n/nvidia-graphics-drivers-515/focal/amd64
https://autopkgtest.ubuntu.com/packages/n/nvidia-graphics-drivers-460-server/focal/amd64

As mentioned in MM, all of these started failing between March 11 and March 20. All of these drivers are transitional packages (they don't really do anything other then depend on the next driver in the series to keep users on a supported driver), but I don't know if this is of interest to the problem as I've seen this with other packages too.

[Paride Legovini (paride)]

This should fix it:

https://salsa.debian.org/ci-team/autopkgtest/-/merge_requests/317

[Paride Legovini (paride)]

That is now merged upstream. I cherry-picked the commit to the branch we use in the infrastructure, so I'm marking that task as Fix Released.