systemd test failure with linux 4.8.0-11.12: scsi_debug missing in ppc64el, systemd-tmpfiles crash

[Andy Whitcroft]

Testing failed on:
    amd64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-yakkety-canonical-kernel-team-unstable/yakkety/amd64/s/systemd/20160918_093714@/log.gz
    ppc64el: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-yakkety-canonical-kernel-team-unstable/yakkety/ppc64el/s/systemd/20160918_231904@/log.gz

[Martin Pitt (pitti)]

amd64 failed in "upstream" tests during image preparation:

The file /dev/loop4p1 does not exist and no size was specified.
F: Failed to mkfs -t ext3

This might be just a rare race condition (I've never seen it). Does it reproduce on retry?

ppc64el:
 - One failure is due to linux-image-extra-4.8.0-11-generic losing scsi_debug (kernel bug)
 - The other is a regression in tmpfiles:
   Sep 18 22:23:56 autopkgtest systemd[1]: systemd-tmpfiles-setup.service: Main process exited, code=killed, status=11/SEGV
   That needs to be investigated more closely; the segfault is certainly a systemd bug, presumably a code path now triggered by different kernel behaviour.

[Brad Figg (brad-figg)]

This bug is missing log files that will aid in diagnosing the problem. From a terminal window please run:

apport-collect 1625100

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Martin Pitt (pitti)]

I tried to reproduce the full boot-and-services test log on scalingstack, and upon rebooting into the 4.8 kernel I got the attached kernel oops. This does everytime, though.

[Tim Gardner (timg-tpi)]

CONFIG_SCSI_DEBUG is now enabled for all arches in Ubuntu-4.8.0-12.13

[Martin Pitt (pitti)]

FTR, I ran the full boot-and-services test against the current PPA which has kernel -11.12 and could not reproduce the crash. Could it be that this was a regression in -10.11 (the test runs in the description) which got fixed in -11.12 again?

So the only concrete thing here that I can see is the missing scsi_debug module.

[Martin Pitt (pitti)]

> [kernel oops] This does everytime, though.

This was supposed to mean: This does *not happen* every time.

[Brad Figg (brad-figg)]

This bug is missing log files that will aid in diagnosing the problem. From a terminal window please run:

apport-collect 1625100

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Martin Pitt (pitti)]

This was fixed in -14, scsi_debug is back.

[Martin Pitt (pitti)]

http://autopkgtest.ubuntu.com/packages/s/systemd/yakkety/ppc64el passed a few times again, and that tmpfiles crash never reproduced.