[Bugfix for #5285] Expose flags for storage and untar images

[tkrishtop]

Description of the change:

Exposes two tags to provide a possibility to overwrite hardcoded images:

--storage-image busybox 
--untar-image registry.access.redhat.com/ubi8/ubi:8.4

Motivation for the change:

Closes #5285

Operator-sdk scorecard pulls two hardcoded images during the job execution:

operator-sdk/internal/scorecard/storage.go

Line 152 in 82e8ebb

Image: "busybox",

operator-sdk/internal/scorecard/testpod.go

Line 36 in 50caa74

scorecardUntarImage = "registry.access.redhat.com/ubi8/ubi:8.4"

In the disconnected environment that pull doesn't work because the images are pinned by tags and not by digests.
As discussed in #5285 with @jmccormick2001 and @estroz , the solution could be to expose two tags to provide a possibility to overwrite these images with their local-registry copies.

Tests in disconnected environment

The changes from this PR were installed as operator-sdk-flag and tested by running in the disconnected environment

operator-sdk-flag scorecard \
    --output json \
    --selector=test=basic-check-spec-test \
    --kubeconfig "{{ kubeconfig_path }}" \
    --namespace scorecard-testing \
    --service-account default \
    --config {{ scorecard_config_path }} \
    --verbose \
    --wait-time 300s \
    --untar-image registry.access.redhat.com/ubi8@sha256:910f6bc0b5ae9b555eb91b88d28d568099b060088616eba2867b07ab6ea457c7 \
    --storage-image docker.io/library/busybox@sha256:c71cb4f7e8ececaffb34037c2637dc86820e4185100e18b4d02d613a9bd772af \
    "{{ scorecard_operator_dir }}"

and

operator-sdk-flag scorecard \
    --output json \
    --selector=suite=olm \
    --kubeconfig "{{ kubeconfig_path }}" \
    --namespace scorecard-testing \
    --service-account default \
    --config {{ scorecard_config_path }} \
    --verbose \
    --wait-time 300s \
    "{{ scorecard_operator_dir }}"

Here are scorecard logs of the above tests

Checklist

If the pull request includes user-facing changes, extra documentation is required:

• Add a new changelog fragment in changelog/fragments (see changelog/fragments/00-template.yaml)
• Add or update relevant sections of the docs website in website/content/en/docs

[tkrishtop]

Hi @camilamacedo86, thank you for the review! I've addressed all the documentation issues and replied concerning SHA usage.

[estroz]

/lgtm

[jmrodri]

The test-sanity target is failing. It doesn't like something with the generated help text.

[openshift-ci]

New changes are detected. LGTM label has been removed.

[tkrishtop]

The test-sanity target is failing. It doesn't like something with the generated help text.

@jmrodri sorry for the issue, that was an uncommitted change for generated help text in cmd.go, don't know really how it was detected o_O. I've fixed that, and re-tested in local, make test-sanity works fine now.

[tkrishtop]

Hi @estroz and @jmccormick2001, could you please approve CI runs? Everything should work now.

[tkrishtop]

Hi @jmccormick2001, thank you for re-running the tests!

Yesterday's go/e2e tests were passing (the only problem was with the sanity tests), and today there is a timeout.
Since yesterday I only changed two lines in cmd.go in the text description of the flags. Normally, that should not impact go/e2e tests. Could you please re-run the tests once again?