HDDS-15493. Limit checkpoint-format parameterization in TestOMRatisSnapshots to transfer tests

[chihsuan]

What changes were proposed in this pull request?

Test-only change. HDDS-14721 made the whole TestOMRatisSnapshots class parameterized (@ParameterizedClass) over ozone.om.db.checkpoint.use.inode.based.transfer (false/true), so every test runs twice even though the parameter only affects the checkpoint transfer path.

This PR keeps the v1/v2 parameterization only for the tests that actually exercise the transfer to speed up tests, by splitting the class:

Test	Lands in	Runs	Why
testInstallSnapshot	TestOMRatisSnapshotTransfer (new)	v1 + v2	only test of the multi-tarball path (max.total.sst.size + SnapshotMaxSizeInjector); not covered by TestOMBootstrap
2× @Unhealthy("HDDS-13300") incremental tests	TestOMRatisSnapshotTransfer (new)	v1 + v2	keep both-format coverage for when they're re-enabled
testInstallOldCheckpointFailure, testInstallCorruptedCheckpointFailure	TestOMRatisSnapshots	once	call installCheckpoint() directly, never download a tarball — parameter added no coverage
testInstallSnapshotWithClientWrite / ClientRead	TestOMRatisSnapshots	once	v1/v2 endpoint coverage duplicates TestOMBootstrap; their unique value (concurrent client IO during install) is format-agnostic
testInstallSnapshotFromLeaderFailedDownloadCleanupSucceeds	TestOMRatisSnapshots	once	cleanup behavior is format-agnostic

Net: both checkpoint formats stay covered on every transfer path (multi-tarball by the new class, endpoint selection by TestOMBootstrap), while the format-agnostic tests stop paying the 2× cluster-setup cost.

The moved test bodies are byte-identical to the originals; SnapshotMaxSizeInjector and the incremental-tarball helpers move with them. The new class reuses the static helpers (checkSnapshot, createOzoneSnapshot, writeKeys) from TestOMRatisSnapshots, following the existing TestOMBootstrap pattern.

For review: every retained line in TestOMRatisSnapshotTransfer matches the master TestOMRatisSnapshots source except two mechanical adjustments the split requires: (1) the checkSnapshot/createOzoneSnapshot/writeKeys wrappers call the shared statics via a TestOMRatisSnapshots. qualifier, and (2) conf is a local in init() rather than a field (the new class never reads it outside setup).

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-15493

How was this patch tested?

Full runs of both classes locally.

mvn -pl hadoop-ozone/integration-test test -Dtest='TestOMRatisSnapshots,TestOMRatisSnapshotTransfer'
TestOMRatisSnapshots:        Tests run: 5, Failures: 0, Errors: 0, Skipped: 0
TestOMRatisSnapshotTransfer: Tests run: 2, Failures: 0, Errors: 0, Skipped: 0

checkstyle passes on the module. Fork CI (build-branch) green: https://github.com/chihsuan/ozone/actions/runs/27064650428

Speedup

From the integration-om CI job, both on top of HDDS-10310 (so this isolates the effect of this change):

	Tests run	Time
master — TestOMRatisSnapshots (fully parameterized, v1 + v2)	6 + 6	402.6s
this PR — TestOMRatisSnapshots (single run)	5	131.4s
this PR — TestOMRatisSnapshotTransfer (v1 + v2)	1 + 1	153.0s
this PR total		284.4s

~118s faster (~29%), by no longer running the four format-agnostic tests twice.