Cirrus: Build multi-arch images + manifests

[cevich]

Github-actions for large/complex tasks is hard to read and maintain.
Reimplement the multi-arch image build workflow into a set of bash
scripts that use all native contrainer-org tooling. This requires
a special VM image setup with emulation to build foreign architectures.
It also requires renaming the helloimage directory, because the build
script uses the directory name in the image FQIN.

[edsantiago]

Mostly requests for clarification

[cevich]

Thanks for taking a peek @edsantiago. I'll definitely fix those FIXMEs and *standardvm thing. Until containers/automation_images#97 is ready, this can stay a 'draft' / POC.

[cevich]

Okay, this looks like it's working overall, building images for amd64, arm64, s390x and ppc64le:

• quay.io/podman/hello:latest took 45 seconds (somehow)!
• quay.io/podman/testing:latest tool 32 minutes
• quay.io/podman/upstream:latest took nearly 86 minutes
• quay.io/podman/stable:3.44 & latest (log) quay.io/containers/podman:3.4.4 & latest took 30 minutes

Before this merges, I intend to resolving the .cirrus.yml fixme to use only_if: $CIRRUS_CRON == 'multiarch' on the task. I'll then add a new Cirrus-cron job called "multiarch" set to run daily or weekly or whatever. So no PR will ever see this task run, but we'll pick up any failures by the daily notification e-mails.

@TomSweeneyRedHat For now, I'm just intending to replace the github-actions based workflow. But can you confirm those are all the right image FQINs we want?

@edsantiago when you have a moment, mind taking a final peek? I intend to add a similar task + cirrus-cron job, to both the Skopeo and Buildah repos.

[cevich]

force-push: Rebased, did final FIXME replacements, added push credentials, and removed github-action workflow.

[edsantiago]

I was halfway through looking at the logs from the previous run, and they seem fine (although they hurt to read). One question, otherwise LGTM.

[cevich]

although they hurt to read

I agree, though unfortunately this is somewhat the "nature of the beast", and not all that different from the github-actions output. It's really hard to present parallel stdout/err without resorting to eye-melting colors or heavy buffering then reformatting 😞

[cevich]

force-push: had to deal with github-actions YAML conflict (file was updated in another PR)

[cevich]

force-push: Since I got to re-run all the tests anyway, might as well remove the unnecessary skip Ed noticed.

[edsantiago]

There won't be logs to review this time, but LGTM based on earlier logs and on git range-diff

[edsantiago]

/lgtm
/hold

New task is in black, at 8 o'clock:

[cevich]

Thanks @edsantiago

[TomSweeneyRedHat]

LGTM
and thanks @cevich!

[TomSweeneyRedHat]

/hold cancel

[TomSweeneyRedHat]

/approved
/lgtm

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

Approval requirements bypassed by manually added approval.

This pull-request has been approved by: cevich

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment