Use the Proper SLE15 SP2 JeOS GM image for testing on OpenStack

[dirkmueller]

Why is this PR needed?

The business case this fixes is that the CI is using the final SP2 image that we ship to customers rather than something older and unsupported.

As such it brings more relevance to what we test and how we test and that is considered a good thing.

What does this PR do?

It fixes the Image configuration to point to the right image.

Anything else a reviewer needs to know?

There is no manual testing needed.

Info for QA

There is no QA needed. if CI passes its good.

Related info

This PR can be merged independently.

Status BEFORE applying the patch

The wrong image is used.

Status AFTER applying the patch

The SLE15 SP2 GM JeOS image for OpenStack is being used for OpenStack testing.

Docs

The docs need to refer to the same image.

Merge restrictions

(Please do not edit this)

We are in v4-maintenance phase, so we will restrict what can be merged to prevent unexpected surprises:

What can be merged (merge criteria):
    2 approvals:
        1 developer: code is fine
        1 QA: QA is fine
    there is a PR for updating documentation (or a statement that this is not needed)

[evrardjp]

I am already using those images (but not the reboot fix), and that works for me. 👍

[evrardjp]

I think this should be placed on Z squad's radar. What do you think @jordimassaguerpla ?

[pablochacin]

There seem to be two unrelated changes in the PR: the image and the reboot logic. If you drop the second commit from this PR we can take the update of the image and test it in the CI.

The reboot logic, we prefer not to changed in the SAME PR unless it is broken, which is not the case as far as we know and JP reports.

[dirkmueller]

@pablochacin I removed the 2nd commit. I already tested it, and so did JP. I think a review would be enough.

[evrardjp]

That's what I am using.

[evrardjp]

Interesting that the pr-head chokes on a id_shared file... Should we edit the authorized_keys var to add the appropriate ci pubkey into ci/infra/openstack/terraform.tfvars.json.ci.example ? I suppose there is something else in our CI tooling dealing with this...

[pablochacin]

Interesting that the pr-head chokes on a id_shared file... Should we edit the authorized_keys var to add the appropriate ci pubkey into ci/infra/openstack/terraform.tfvars.json.ci.example ? I suppose there is something else in our CI tooling dealing with this...

I observed this same error in PR #1205 The problem seems to be related to the changes in PR #1168 (the location of files change) I'm looking at it.

[dirkmueller]

potentially this required a rebase. I've rebased on top of master.

[pablochacin]

I run an e2e test job with the PR and the cluster was deployed correctly. The deployment failed with an unrelated issue related to the image repository. So from my perspective, the new image solves the ssh issue.

I'm still investigating why the PR tests are failing. It also seems to be unrelated, as I mentioned in a previous comment.

[pablochacin]

I'm still investigating why the PR tests are failing. It also seems to be unrelated, as I mentioned in a previous comment.

Effectively, it is due to a problem with PR #1168

[pablochacin]

LGTM

[evrardjp]

It's weird we provide image_name in L4 instead of L3 on the example file, but it's good enough for me.

[kkaempf]

https://github.com/SUSE/avant-garde/issues/1739#issuecomment-651289097 - let's get this merged.