Install oc

[cgwalters]

The immediate motivation is https://github.com/openshift/release/blob/master/ci-operator/config/openshift/os/openshift-os-master.yaml#L77

For RHEL CoreOS related stuff in particular it's just really useful to have
the oc binary in coreos-assembler to streamline things like
"build a new release image that overrides the OS" etc.

I suspect this would help OKD too.

---

[cgwalters]

(Feel free to bikeshed on naming...rhcos-assembler:latest? Doesn't quite fit since OKD could want this too. Maybe ocp-coreos-assembler:latest?)

[jlebon]

Or should we just ship oc in cosa purely for practicality? Even if nothing in cosa actually uses it (though cosa oc-adm-release could).

[cgwalters]

Or should we just ship oc in cosa purely for practicality? Even if nothing in cosa actually uses it

I'm a bit uncertain. One thought I have here is that in the future, the version of oc we use should actually be driven by OCP branches which we already have (except they're called "rhcos-"). And mechanically, I was thinking we may find ourselves wanting to add more into coreos-assembler that is actually OpenShift specific. (Although OTOH if we do a lot of it, it could become a separate git repository)

(though cosa oc-adm-release could).

Ah yeah, I should have called that out in this PR. Hmm. That code almost seems like it should live in github.com/openshift/coreos-assembler or so. If we did that, then we could have this oc injection live there too, and cleanly use all the same Prow tooling there, e.g. auto-create release-4.X git branches, etc.

A lot of tradeoffs (advantages/disadvantages).

Just adding oc into the existing image is definitely easier.

So...hmm. I dunno. I guess, does anyone object to just adding in $current oc into the image?

(Side note, looks like https://src.fedoraproject.org/rpms/origin/blob/rawhide/f/origin.spec hasn't been updated in nearly 2 years)

[cgwalters]

Just adding oc into the existing image is definitely easier.

Done!

[jlebon]

LGTM! Would be good to have at least one other person chime in.

[miabbott]

I like this approach. LGTM!

[jlebon]

[2022-03-28T22:36:11.235Z] src/cmd-oc-adm-release:18:1: F401 'shutil' imported but unused
[2022-03-28T22:36:11.235Z] src/cmd-oc-adm-release:20:1: F401 'tarfile' imported but unused
[2022-03-28T22:36:11.235Z] src/cmd-oc-adm-release:21:1: F401 'tempfile' imported but unused
[2022-03-28T22:36:11.235Z] src/cmd-oc-adm-release:22:1: F401 'urllib.request' imported but unused
[2022-03-28T22:36:11.235Z] make: *** [Makefile:44: flake8] Error 1
script returned exit code 2

[cgwalters]

[2022-03-28T22:36:11.235Z] src/cmd-oc-adm-release:18:1: F401 'shutil' imported but unused

I am so done with dynamic programming languages 😦

[jlebon]

/cherrypick rhcos-4.9-new

[openshift-cherrypick-robot]

@jlebon: #2777 failed to apply on top of branch "rhcos-4.9-new":

Applying: Install `oc`
Using index info to reconstruct a base tree...
M	build.sh
Falling back to patching base and 3-way merge...
Auto-merging build.sh
CONFLICT (content): Merge conflict in build.sh
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Patch failed at 0001 Install `oc`
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".

Details

In response to this:

/cherrypick rhcos-4.9-new

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.