podvm: build rhel peer pod VM qcow2 image for s390x

[redobed]

This PR updates the podvm/Dockerfile.podvm_binaries.rhel and podvm/Dockerfile.podvm_builder.rhel to enable building a rhel based peer pod VM image qcow2 image for s390x. Packer binary is conditionally built from source s390x. The PR also allows building the Builder image either on a subscribed or an unsubscribed rhel host.

Signed-off-by: Obed Tetteh obed.tetteh@ibm.com

[redobed]

@jensfr
@bpradipt

[CarstenLeue]

LGTM

[snir911]

Thanks!
I added some comments, however, while review it seems to me like many changes should have been done in the builder container Dockerfile rather than here.
So I'd generally check what can be moved there (unless i missed something and for some reason you have to do it here)

[snir911]

I don't think it's needed in the binaries dockerfile, only in the podvm one, and it's already being set there, no?

[redobed]

This is required by this line in the binaries dockerfile

[snir911]

I think the right approch will be to add a default dummy link as done for x86 see here and here
(which is https://access.redhat.com/downloads/content/433/ver=/rhel---9/9.3%20Beta/s390x/product-software & d436ffeac7d468f75526ca8b8bab1cba64ba3ab2889ccd11fa107a7323b267db i suppose)

[redobed]

Thanks! I added some comments, however, while review it seems to me like many changes should have been done in the builder container Dockerfile rather than here. So I'd generally check what can be moved there (unless i missed something and for some reason you have to do it here)

@snir911 thanks for the review.
The changes were mostly done in that very file.
The binaries file only had:

ARG IMAGE_URL="/tmp/rhel.qcow2"
ARG IMAGE_CHECKSUM

ENV IMAGE_URL ${IMAGE_URL}
ENV IMAGE_CHECKSUM ${IMAGE_CHECKSUM}

which are required for the make target

[snir911]

Thanks! LGTM overall add one comment to address
also it seems the latter commit "podvm: align commit with other files" is empty, pls remove it

[snir911]

nit: typo

[snir911]

IIUC codeready expects x86_64 rather than amd64, how about treating the inconsistency using the following:

users will set only the ARCH value as done in other files:
ARG ARCH=x86_64 # set x86_64 or s390x
ENV ARCH=${ARCH}

then:
for protoc use: ${ARCH/s390x/s390_64}
for codeready use: ${ARCH}
for golang use: ${ARCH/x86_64/amd64}
for yq use: ${ARCH/x86_64/amd64}
for packer use: ${ARCH}

[snir911]

LGTM Thanks @redobed !
I think some of the commits needs to be squashed

[bpradipt]

@redobed please squash the commits and also remove the merge commits.
Here is a good tutorial on squashing the commits with git - https://www.git-tower.com/learn/git/faq/git-squash/

[tumberino]

@redobed I don't think the merge from main kept the correct changes, i.e. it is undoing changes done as part of #1448

[redobed]

@redobed please squash the commits and also remove the merge commits. Here is a good tutorial on squashing the commits with git - https://www.git-tower.com/learn/git/faq/git-squash/

@bpradipt commit are squashed

[redobed]

@redobed I don't think the merge from main kept the correct changes, i.e. it is undoing changes done as part of #1448

@jtumber-ibm resolved

[bpradipt]

/lgtm
Thanks @redobed