chore(helm): add hook-delete-policy for Helm chart tests

[fty4]

WHAT

Adding Hook deletion policies for the Helm tests.

WHY

When running tests with Helm the pods will not be tidied up.
When the annotation "helm.sh/hook-delete-policy": hook-succeeded is added, after a successful run the pod will be deleted afterwards.
If the pod fails it will still be present.

Also the value before-hook-creation was added to be able to run the test again if it failed without manually deleting the pod.

---

Marco Lecheler [email protected] Mercedes-Benz Tech Innovation GmbH (ProviderInformation)

[fty4]

Test is failing because test-pod is cleaned up as expected when successful and not longer present:

tractusx-edc/.github/workflows/deployment-test.yaml

Lines 155 to 156 in 2f81217

	# execute the helm test
	helm test tx-prod --logs

[fty4]

The command should not be the problem - there is a bug that --logs will not act as expected:

helm/helm#10603

[paullatzelsperger]

The command should not be the problem - there is a bug that --logs will not act as expected:

helm/helm#10603

I tried this on my local machine, and there it works, but if removing the --logs flag solves it on CI, that would be fine. can you remove the flag in the workflow file, adding a comment about the helm bug and the link you posted?

[fty4]

Sure then i would revert the last commit which adds the tests.hookDeletePolicy value in 7a832c7 to keep the value file more clean, right?

I thought removing the --logs isn't an option to preserve potential error messages.

[paullatzelsperger]

Sure then i would revert the last commit which adds the tests.hookDeletePolicy value in 7a832c7 to keep the value file more clean, right?

I thought removing the --logs isn't an option to preserve potential error messages.

if i understand this right we have two options: to not delete the test pods, or to not have logs, right?

[fty4]

Right - here both options compared:

1. Option: persist test-pods for test workflow (current implementation of the PR)
With the last commit the workflow succeeds.
Here the Helm default will be overwritten to delete the test-pod after a successful run.
But because the workflow failed I have overwritten it again in test values back to the Helm defaults:

see: edc-tests/deployment/src/main/resources/helm/tractusx-connector-test.yaml

This would work and the pod will not be deleted in the test infrastructure which is no problem (will be deleted anyway).

+ Logs of failed pods in test workflow
- Additional values required in values.yaml and tractusx-connector-test.yaml

2. Option: no logs for test workflow
When removing the --logs flag the workflow would also succeed but in case of an error the log is not present in the workflow logs due the mentioned bug.

+ No additional values required
- No logs of failed pods in test workflow

---

I think I would prefer option 1.

[sonarcloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
0 Code Smells

No Coverage information
0.0% Duplication

[paullatzelsperger]

Additional values required in values.yaml and tractusx-connector-test.yaml

would that mean, if someone forgot to add those values to their xyz-values.yml, they'd still end up with zombie test pods?

[fty4]

would that mean, if someone forgot to add those values to their xyz-values.yml, they'd still end up with zombie test pods?

No the test-pods will be removed because in the default values.yaml
file the policy is set to "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded which removes succeeded test-pods.
Failed will still be available - because you want to see the error messages here therefore no deletion.

If you want all to persist test-pods you can adapt to the tractusx-connector-test.yaml test values file.

Without this PR the test-pods will not be removed after successful testing.

[paullatzelsperger]

Without this PR the test-pods will not be removed after successful testing.

ok, that sounds reasonable, thanks for the clarification.