CASMCMS-8903: Remove BOS v1 CLI commands & tests; add some BOS v2 tests

[mharding-hpe]

Summary and Scope

BOS v1 is finally going away in CSM 1.6.
It has already been removed from the cmsdev test tool and the barebones boot test.
This PR removes the BOS v1 command paths from the Cray CLI, along with the BOS v1 CLI tests.
It also adds some BOS v2 tests (although more should be added).

It does not replace the BOS swagger file because the CLI doesn't play nicely with anyOf/allOf/oneOf, which the current BOS API spec makes use of. Instead, I just modified the CLI generation for BOS to delete the v1 paths.

I ran the CLI test suite and it passed (including all of the new/updated BOS tests). I also installed the RPM on mug and verified that the BOS CLI command still work, and that the v1 paths are missing.

• Fixes: CASMCMS-8903

Issue Type

• RFE Pull Request

Prerequisites

• I have included documentation in my PR (or it is not required)
• I tested this on internal system (if yes, please include results or a description of the test)

[jsollom-hpe]

Is removing BOS V1 wise at this point? I know we said we were going to deprecate it; however, we're nearing the end of life of CSM. BOS V1 and BOS V2 can coexist. I don't think removing BOS V1 eliminates enough of a support burden to justify it. We still have to support it through CSM-1.5. I worry that people will either hate BOS V2 or there will be some insurmountable bug in it that would push them back to V1.

[mharding-hpe]

Is removing BOS V1 wise at this point? I know we said we were going to deprecate it; however, we're nearing the end of life of CSM. BOS V1 and BOS V2 can coexist. I don't think removing BOS V1 eliminates enough of a support burden to justify it. We still have to support it through CSM-1.5. I worry that people will either hate BOS V2 or there will be some insurmountable bug in it that would push them back to V1.

Well, we've already done a fair amount of work in CSM 1.6 toward removing it. If we had changed our plans on this, it would have been nice to find out before all of that. :) We'll also need to go back and revert a lot of that.

I don't know if it's more work now to finish removing v1, or to undo the work that has already been done.

We've already made BOS v2 the default, and tools like SAT and IUF have switched over to it. If we run into any BOS v2 bugs, we're going to have to fix them anyway. And I know that I still run into plenty of problems with BOS v1 when I'm on triage, often things that are just limitations of BOS v1. So barring some catastrophic unfixable BOS v2 bug, it seems to me like it makes our life easier to remove BOS v1.

Plus, while we may be stopping active development of new features for CSM, I wouldn't describe it as nearing end of life -- at least not based on the last info I got. I think we're obligated to support it for quite a while yet.

Oh, and BOA still uses CAPMC, which is also slated to be removed in CSM 1.6. @rfrost-hpe opened a Jira for that work just a couple of days ago (CASMHMS-6145), so I assume it is still the plan.