Emanuele Aina authored 5 years ago and Andrej Shadura committed 5 years ago

A long time ago, when Apertis was not yet Apertis and SAC was its name,
the LAVA workers had a pre-deployed image that people could use for quick
testing, and this job was used to run `apt dist-upgrade` on those images
to keep them up-to-date.

Now those pre-deployed images are long forgotten, so let's also drop their
vestigial remains.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Emanuele Aina authored 5 years ago and Andrej Shadura committed 5 years ago

The ADE tests have been put in their own group, which mean that a whole new
image gets deployed just to run a couple of tests, for no good reason.

Let's save some trouble and re-use the same instance set up for all the
other SDK tests.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

The `baseurl` variable is useful to load images to be tested from a
non-standard location, for instance when submitting tests from
branches that haven't been merged yet and that uploaded their
artifacts to https://images.apertis.org/test/ instead of
https://images.apertis.org/

.

However, to be used in that way the location of the nfs artifacts used in the
first stage should not change when setting `baseurl`.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

images.apertis.org removes files older than 1 day from test repo.
runtime and app repos has been moved to images.apertis.org/flatpak.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

Luis Araujo authored 5 years ago and Peter Senna Tschudin committed 5 years ago

The psdk-test is a script to semi-automate the following SDK persistent
tests:

https://qa.apertis.org/sdk-persistent-disk.html
https://qa.apertis.org/sdk-persistent-disk-home-user.html
https://qa.apertis.org/sdk-persistent-disk-etc.html



The script executes a prompt to enter a set of commands that will
automatically run most of the test setup and execution guiding the tester
through the different stages of the test.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

While converting the package to the new GitLab-to-OBS pipeline the project
was mistakenly forked under `pkg/` rather than transferred and now the
repositories went out of sync.

Fold the changes back so we can replace the forked repository with the
original one.

* packaging/apertis/v2021dev1:
  Releasing apertis version 0.2020.1.
  gitlab-ci: Link to the Apertis GitLab CI pipeline definition

This adds test to install the org.apertis.demo.curl application and test
it by retrieving https://www.apertis.org

 page.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

This commit adds a script to automate the apparmor-session-lockdown-no-deny
test case.

It also adds the test to run in LAVA jobs.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Remove setting terminal properties to show colored text.

This seems to break some tests in LAVA, and it is a feature not really
needed for LAVA test, so just remove it.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Signed-off-by: Andrew Lee (李健秋) <ajqlee@debian.org>

Commit bc9094f4 broke the first stage in many ways, and commit 3b9a1487
forgot to fix the path to the DTB files needed for the ARM64 Renesas R-Car
boards, resulting in the error below:

  Job error: Invalid job data: ["Resource unavailable at 'https://images.apertis.org/release/v2019/v2019.1/arm64/dtbs/renesas/r8a7796-m3ulcb.dtb

' (404)"]

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Commit bc9094f4 broke the first stage in many ways, the one
being fixed here is the change of the prompts to be expected
during autologin and when getting access to the shell.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Commit bc9094f4 broke the first stage, see for instance
https://lava.collabora.co.uk/scheduler/job/2166454

  Job error: Invalid job data: [
    "Resource unavailable at 'https://images.apertis.org/release/v2019/v2019.1/amd64/initrd.img' (404)",
    "Resource unavailable at 'https://images.apertis.org/release/v2019/v2019.1/amd64/apertis-nfs-amd64.tar.gz

' (404)"]

A `nfs/` component in the path was missing.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Emanuele Aina authored 5 years ago and Peter Senna Tschudin committed 5 years ago

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Emanuele Aina authored 5 years ago and Peter Senna Tschudin committed 5 years ago

Our main LAVA jobs have two stages:
1. boot a utility system via NFS whose only responsibility is to flash the
   real image to be tested on the SD card
2. boot from the newly-written SD card

Currently the first stage is picked up from a server location that is
maintained separatedly from the rest of the system. The version in the
URLs of the artifacts matches the release that will be loaded, not the
one used to build the artifacts: in fact, all of the current first stage
use artifacts built from the ancient 18.12 release. This is not an issue
because what we are actullay interested in testing here is the second
stage (the actual NFS artifacts use different jobs).

However, this means that the hardlinks on the server side have to be set
up every time we branch, and it's not obvious what we are actually using
for the first stage.

This commit explicitly loads the v2019.1 NFS artifacts for the first stage
instead, so the job definition is in full control and does not rely on
some server state.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Emanuele Aina authored 5 years ago and Ritesh Raj Sarraf committed 5 years ago

Drop an example job that was used for experimentation.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Frédéric Dalleau authored 6 years ago and Emanuele Aina committed 5 years ago

It is now common with common-boot-tpl.yaml

Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>

Frédéric Dalleau authored 6 years ago and Emanuele Aina committed 5 years ago

Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>

Frédéric Dalleau authored 6 years ago and Emanuele Aina committed 5 years ago

Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>

Frédéric Dalleau authored 6 years ago and Emanuele Aina committed 5 years ago

Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>

Signed-off-by: Ritesh Raj Sarraf <rrs@debian.org>

Use specific prompts to match instead of just a single `#`.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

This commit changes the prompt that LAVA needs to match in the images
during tests execution.

For SDK images the 'user@apertis:' prompt is the initial prompt after
login and 'root@apertis:' is used after becoming root.

And for the rest of the images '\$ ' is the initial prompt after login,
the '/home/user #' is used right after becoming root, and '/tmp #' is
used during test execution.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Signed-off-by: Ritesh Raj Sarraf <rrs@debian.org>

With latest 5.4 kernel there's a message with a '# ' in it
(pci 0000:00:00.0: PME# supported from D0 D1 D3hot D3cold) which is
detected by Lava as a prompt and triggers send of the username before login
prompt appeared.
Adding a space in front of # fixes it.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Apertis-tests for geoclue fails with the following error:
/usr/share/chaiwala-tests/geoclue/automated/run-test.sh: 260: cd: can't cd to /usr/share/chaiwala-tests/geoclue/automated/../../resources

In this case /usr/share/chaiwala-tests/geoclue/automated is a link to
../../../lib/apertis-tests/geoclue/automated.

Force 'cd' to resolve symbolic links to their values.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

Denis Pynkin authored 5 years ago and Emanuele Aina committed 5 years ago

The board name is used in ostree-based images for pointing
the correct ostree branch name.

Signed-off-by: Denis Pynkin <denis.pynkin@collabora.com>

This reverts commit d8db26c9 which was
introduced to see how things would work. Apparently we still have some road to
do before being able to properly turn on the tests on all the platforms
since they were all failing.

Among others, https://phabricator.apertis.org/T6446 and
https://phabricator.apertis.org/T6444

 have been identified during this
experiment.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

This flag is no longer available for the timeout subcommand, making the
dbus test to fail, so remove it.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Signed-off-by: Ritesh Raj Sarraf <ritesh.sarraf@collabora.com>

Signed-off-by: Ritesh Raj Sarraf <rrs@debian.org>

Use correct boot commands for Amd64 since they are used
in all AUM tests during reboot.

Signed-off-by: Denis Pynkin <denis.pynkin@collabora.com>

Enable tests for ostree and Apertis Update Manager for all supported
architectures.

Signed-off-by: Denis Pynkin <denis.pynkin@collabora.com>

Add testing of signature verification of commits in offline updates.

Signed-off-by: Denis Pynkin <denis.pynkin@collabora.com>

Recent ARM internal jobs fails with a timeout on unpacking the overlay
into the image.

This commit increases the default timeout to the action transfering the
overlay in the LAVA job to 5 minutes for all jobs types.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

This test case has been failing for a long a time since the required
telepathy packages are missing from the images.

The test case is also not so relevant anymore and makes current SDK lava
tests jobs to fail, so this commit removes the test case.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Except for the SDK image, this test is failing in the remaining images.

This test is also not relevant anymore, so this commit removes it along
with the group-folks template file, so LAVA won't deploy and boot an
image in order to execute just one single test.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

All telepathy folks tests currently fail to execute because the
telepathy-gabble package is not available in the images.

Support for folks and telepathy libs also needs to be revised, so
this commit disables all the folks telepathy tests in the meantime.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>