Skip to content
This repository has been archived by the owner on Sep 23, 2024. It is now read-only.

Latest commit

 

History

History
63 lines (53 loc) · 2.52 KB

README.md

File metadata and controls

63 lines (53 loc) · 2.52 KB
Raw

Beaker tasks used with skt runner

THIS PROJECT HAS MOVED TO https://gitlab.com/cki-project/kernel-tests/.

How to run tests

Here is a list of common prerequisites for all beaker tests. Test-specific dependencies and steps can be found in the README.md within each test's directory.

$ sudo wget -O /etc/yum.repos.d/beaker-client.repo https://beaker-project.org/yum/beaker-client-Fedora.repo
$ sudo wget -O /etc/yum.repos.d/beaker-harness.repo https://beaker-project.org/yum/beaker-harness-Fedora.repo
$ sudo dnf install -y beaker-client beakerlib restraint-rhts

Test onboarding

Currently, all onboarded tests must use the following combinations of result/status fields:

  • SKIP/COMPLETED if the test requirements aren't fulfilled (eg. test is running on incompatible architecture/hardware)
  • PASS/COMPLETED if the test finished successfully
  • WARN/ABORTED in case of infrastructure issues or other errors (eg. the test checks out a git repo and the git server is unavailable)
  • WARN/COMPLETED or FAIL/COMPLETED in case of any test failures, based on how serious they are (left to decide by test authors)

See examples below to properly abort or skip in beaker:

Abort task if infrastructure failure is task only related

if [ $? -ne 0 ]; then
    rlLog "Aborting test because $reason"
    rstrnt-abort --server $RSTRNT_RECIPE_URL/tasks/$TASKID/status
fi

Abort recipe if infrastructure failure affects the entire recipe

if [ $? -ne 0 ]; then
    rlLog "Aborting recipe because $reason"
    rstrnt-abort recipe
fi

Skip the task (e.g. testing with unsupported hardware)

if [ $? -ne 0 ]; then
    rlLog "Skipping test because $reason"
    rstrnt-report-result $TEST SKIP
    exit 0
fi

When onboarding a test, please check especially the point about infrastructure issues: a lot of tests simply report a warning if eg. external server can’t be reached and then continue. This kind of situation falls under infrastructure issues and the test must use the WARN/ABORTED combination, otherwise the infrastructure problem is reported to people as a bug in their code!

The order of Beaker tasks in the XML determines if the task is a preparation for the testing or a test. Anything after kpkginstall task is treated as a test and must follow the rules above. Anything except PASS before kpkginstall is treated as infrastructure issue. Machineinfo (to get HW specification) is ran before kpkginstall, so the same logic applies there. PANIC during kpkginstall means the kernel is bad (can't boot).