lp:~fheimes/autopkgtest-cloud/+git/autopkgtest-package-configs

Author: Frank Heimes
Author Date: 2022-09-01 07:32:14 UTC

Add casync for ppc64el to big_packages, since it's likely that more resources are req.

Author: Frank Heimes
Author Date: 2022-08-31 21:12:27 UTC

Add linux-hwe-5.x armhf (and one arm64) to long_tests, since they time out.

Author: Mauricio Faria de Oliveira
Author Date: 2022-08-31 01:01:24 UTC

Add casper to big_packages on amd64

The change in 1.445.2 for more RAM in QEMU causes the
QEMU process to finish in the middle of `test_todisk`
in `debian/tests/boot`.

In 1.445.1 it reaches the 'reboot: Power down' point
before finishing, which suggests an unclean end/kill
on the newer version.

I could reproduce the failure locally with:

  $ autopkgtest casper -U --apt-pocket=proposed=src:casper \
      -- qemu --ram-size=1592 --cpus=1 ~/autopkgtest-focal-amd64.img

And have it pass with some more RAM:

  $ autopkgtest casper -U --apt-pocket=proposed=src:casper \
      -- qemu --ram-size=2048 --cpus=1 ~/autopkgtest-focal-amd64.img

Hence, let's move it to big_packages, so to give it more RAM.

Only amd64 is needed as `/d/t/boot` only runs on amd64/i386,
and the latter always fails for other reasons.

...

autopkgtests log snippets:

1.445.1:

  ### todisk does not mount any part of the boot media
  ...
  + timeout --foreground 10m qemu-system-x86_64 -m 768m ...
  ...
  [ 24.656985] EXT4-fs (sdb1): mounted filesystem with ordered data mode. Opts: (null)
  ...
  [ 83.371115] reboot: Power down
  + rm -rf result
  + mkdir result

1.445.3 (with autopkgtest change from 1.445.2):

  ### todisk does not mount any part of the boot media
  ...
  + timeout --foreground 10m qemu-system-x86_64 -m 1024m ...
  ...
  [ 25.696515] EXT4-fs (sdb1): mounted filesystem with ordered data mode. Opts: (null)
  ...
  [ 46.413507] clocksource: Switched to clocksource hpet
  + rm -rf result
  + mkdir result

Signed-off-by: Mauricio Faria de Oliveira <mfo@canonical.com>