CNDB-12434: Fix flaky CQLCompressionTest failures

[jkni]

What is the issue

CQLCompressionTest experiences flaky test failures when 3 sstables are detected rather than the expected 2.

What does this PR fix and why was it fixed

This PR switches back to inserting single rows before triggering flushes in the test. Multiple rows were needed due to UCS's previous maximal compaction behavior, but if an asynchronous CQLTester schema change dropped a table while rows were being inserted, an additional flush could be trigger in the middle of inserting multiple rows. This would generate more sstables than expected. Since multiple rows are no longer needed after #1342 due to changes in UCS behavior, we can revert to the previous behavior and avoid the flakiness without waiting for schema changes to complete between tests.

Checklist before you submit for review

• Make sure there is a PR in the CNDB project updating the Converged Cassandra version
• Use NoSpamLogger for log lines that may appear frequently in the logs
• Verify test results on Butler
• Test coverage for new/modified code is > 80%
• Proper code formatting
• Proper title for each commit staring with the project-issue number, like CNDB-1234
• Each commit has a meaningful description
• Each commit is not very long and contains related changes
• Renames, moves and reformatting are in distinct commits

[cassci-bot]

❌ Build ds-cassandra-pr-gate/PR-1512 rejected by Butler

---

2 new test failure(s) in 1 builds
See build details here

---

Found 2 new test failures

Test	Explanation	Branch history	Upstream history
o.a.c.i.s.c.VectorCompactionTest.testPQRefine[ca]	regression	🔴	🔵🔵🔵🔵
...et=map<int,int>,wide=true,scenario=MIXED_QUERY]	regression	🔴	🔵🔵🔵🔵🔵🔵🔵

---

Found 5 known test failures

[jtgrabowski]

@jkni we could merge this but @jacek-lewandowski has fixed CQLTester already https://github.com/datastax/cassandra/pull/1315/files#diff-822707a68576bd683b72fde39e19ab8a78ed5a1816d95f5d8a00b3f46ac78d03L452. Up to you.

[jkni]

@jtgrabowski yes, I saw that, and the CQLTester fix would also resolve this. I think we might as well merge as it reduces divergence from apache/cassandra in case we end up debugging this for some reason in the future.