podman/tests
Ed Santiago 5b50b233ab Sigh. Put the reboot in a block.
Podman gating tests failed:

   ... back_again.changed): 'dict object' has no attribute 'changed'

Previous error (ignored) was a timeout in the reboot step. I have
no idea how to fix this, so let's try putting the reboot and wait
steps into a 'block' section and removing the '.changed'.

Tested with 1minutetip. Seems to work, but the reboot was quick
so it didn't trigger the same code.

Signed-off-by: Ed Santiago <santiago@redhat.com>
2020-05-11 14:12:45 -06:00
..
roles Sigh. Put the reboot in a block. 2020-05-11 14:12:45 -06:00
check_results.yml Complete rewrite of gating tests 2020-04-28 12:30:29 -06:00
README Complete rewrite of gating tests 2020-04-28 12:30:29 -06:00
test_podman_cgroups_vn.yml Complete rewrite of gating tests 2020-04-28 12:30:29 -06:00
test_podman.yml Complete rewrite of gating tests 2020-04-28 12:30:29 -06:00
tests.yml Quick simple tests for podman 2018-11-13 09:50:53 -07:00

I'm sorry. The playbooks here are a much-too-complicated way of saying:

   - test podman (root and rootless) under cgroups v2
   - reboot into cgroups v1
   - repeat the same podman tests

We can't use standard-test-basic any more because, tl;dr, that has to
be the last stanza in the playbook and it doesn't offer any mechanism
for running a reboot in the middle of tests. (I actually found a way
but it was even uglier than this approach).

The starting point is tests.yml . From there:

    tests.yml
      \- test_podman.yml
           |- roles/rootless_user_ready/
           \- test_podman_cgroups_vn.yml    (runs twice: cgroups v2, v1)
                |- roles/set_cgroups/
                \- roles/run_bats_tests/    (runs tests: root, rootless)

Principal result is the file 'artifacts/test.log'. It will contain
one line for each test run, format will be '(PASS|FAIL|ERROR) <test name>'

For each completed test there will also be a 'test.<name>.bats.log'
containing some setup blurbs (RPMs, environment) and the full BATS log.