HDDS-14706. Make ozone admin om finalizeupgrade a no-op

[octachoron]

What changes were proposed in this pull request?

ZDU will use one command to finalize the whole cluster with its own status flag. The OM server's endpoint to start finalization should therefore trivially return success for compatibility. This PR makes that change, and adjusts tests accordingly.

As this change leaves OMs without a way to start finalization until the new mechanism is implemented, some tests are temporarily marked as unhealthy:

• OM finalization check after upgrade acceptance tests - Trivially depends on successful finalization.
• OM integration test TestOMUpgradeFinalization - Also trivially depends on successful finalization.
• Integration test TestOmSnapshot (all implementations) - Finalization after completing the pre-finalization checks has been adjusted to accept the unconditional ALREADY_FINALIZED. These checks would pass, and the finalization would not throw an error, but the actual snapshot-related tests would fail (as the finalization that enables the operations does not actually happen).
• OM integration tests TestOMBucketLayoutUpgrade and TestMultiTenantVolume - These also have pre-finalization and post-finalization checks, the latter of which would not pass without being able to finalize.

How much not being able to trigger OM finalization blocks other work is unclear. Status could be reported as successful or not required (as outlined in https://issues.apache.org/jira/browse/HDDS-14707), but operations that depend on it would still fail the way they do in TestOmSnapshot.

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-14706

How was this patch tested?

CI on fork: https://github.com/octachoron/ozone/actions/runs/23847445751

[sodonnel]

Is there more to do on this, or can it be moved from draft to ready for review?