Build InstructLab RHEL AI container images

[lmilbaum]

Build the images listed in the following Makefile - https://github.com/containers/ai-lab-recipes/blob/main/training/instructlab/Makefile
Acceptance Criteria:

1. The images are built and pushed to quay.io/ai-lab once a day
2. They are built with automation. Preferably with GitHub Actions workflow.
3. All images are amd only. Multi-arch is NOT required at the moment.

[Gregory-Pereira]

Looking for acceptance criteria from @rhatdan and @sallyom. Many people have asks but no one has be able to clarify what the actual requirements are. From my understanding it would be something along the following lines

• Instructlab images get build and land in quay.io/ai-lab everyday
• Images are multi-arch

[rhatdan]

Instructlab+RHEL AI based images get in.

[Gregory-Pereira]

This has been addressed in a number of PRs that landed in main, but this successful action run shows that images are landing in quay.io.

[Gregory-Pereira]

This does work however it only tests 2 of the 3 images we want. We also want the intel based one. Therefore I will refactor this to use the make targets, tracked in #384

[lmilbaum]

Instructlab+RHEL AI based images get in.

@rhatdan Would you please map the images you have listed here with the ones listed in the Makefile. Which ones should be built by the end of this week?

[lmilbaum]

@Gregory-Pereira Isn't #384 somewhat a duplication of this issue? What would be different in the acceptance criteria from this issue?

[Gregory-Pereira]

my understanding was that #364 (this issue) tracks all the acceptance criteria and everything that needs to get done before summit for CI/CD and #384 was just tracking the 1 refactor from building from buildah to using make targets

[lmilbaum]

@Gregory-Pereira I'm aware that I'm nit picking. For the sake of the board hygiene let's carefully review the acceptance criteria this ticket has and create a new issue for outcomes which are not covered by the existing one.
Implementation details could be shared in this issue. Does that make sense?

[Gregory-Pereira]

As far as I can tell, at this point all 3 of our acceptance criteria have been met, but waiting for tomorrow for images to land to say for sure.

• The images are built and pushed to quay.io/ai-lab once a day (should be done but confirming tomorrow)
• They are built with automation. Preferably with GitHub Actions workflow.
• All images are amd only. Multi-arch is NOT required at the moment.

If our images show up tomorrow and I can check this box, im closing this issue