CNDB-12484: Fix flaky VectorUpdateDeleteTest failures

[jkni]

What is the issue

VectorUpdateDeleteTest sees flaky failures where it fails to verify the checksums of all sstables.

What does this PR fix and why was it fixed

SAITester.verifyChecksum is swapped to consider canonical sstables rather than live sstables. If the SAITester.verifier overlaps with a compaction, it could previously see a preemptively opened SSTable. Validating the SAI components for this preemptively opened sstable fails, as they haven't yet been written. The verifier should only consider canonical sstables, which are expected to have valid SAI components.

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

[michaeljmarshall]

LGTM

[jbellis]

I'm going to scope creep you a bit here, sorry!

• Can you add javadoc to getLiveSSTables to point out what it does and when you need to consider early open (and by implication not use getLive)?
• Probably worth adding a getCanonicalSSTables method so getLive isn't the only "obvious" choice
• Almost all of the usage in the test package should probably be getCanonical instead of getLive, I toyed with the idea of asking an LLM to go through them on a case by case bases but I kind of think we'll be better served by just changing them all over to getCanonical and reverting any that break

[michaeljmarshall]

Probably worth adding a getCanonicalSSTables method so getLive isn't the only "obvious" choice

+1 great suggestion

[jkni]

Scope creep accepted. I'll update the PR shortly.

[jkni]

Ah, despite multiplexing this 200 times, the CI run here caught a real issue. We need a proper reference to the canonical sstables to prevent them from being removed from the filesystem during verification. I'll push a fix.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[cassci-bot]

❌ Build ds-cassandra-pr-gate/PR-1519 rejected by Butler

---

2 new test failure(s) in 3 builds
See build details here

---

Found 2 new test failures

Test	Explanation	Branch history	Upstream history
...tCorruptedSSTablesWithLeveledCompactionStrategy	regression	🔴🔴🔴	🔵🔵🔵🔵🔵🔵🔵
...ataset=uuid,wide=false,scenario=MEMTABLE_QUERY]	regression	🔴🔵🔵	🔵🔵🔵🔵🔵🔵🔵

---

Found 14 known test failures