Add Apache Kudu

[KazukiKandaKK]

Adds an impala-kudu/ entry. Kudu holds the data; Impala runs the SQL. This is separate from impala/, which queries the source Parquet file directly.

On c6a.2xlarge with Ubuntu 24.04, the full 100M-row run loaded 99,997,497 rows into Kudu, matching the source. Load took 3,800.123 s, and allocated persistent storage was 47,958,401,024 bytes. All 43 queries x 3 tries have timings.

The frozen validator passed 42 of 43 queries. Q4 was within the predeclared floating-point tolerance; exact checks also matched COUNT(*), COUNT(UserID), and SUM(CAST(UserID AS DECIMAL(38,0))). Original Q41 and Q42 selected different rows at tied boundaries. Q41 still passed under the frozen rules; Q42 was the only failure. Tie-breaker diagnostics matched both. The timed SQL was unchanged.

After check started verifying Kudu health, I reran the concurrent phase: 0.100 QPS, error ratio 0.016. The shared driver does not retain worker-level errors, so the failed concurrent query is unknown.

Tagged no-cold like the existing Impala entry: the driver drops the host page cache but does not restart the stack between tries. data-size reports allocated blocks for Kudu and Hive Metastore, excluding the source Parquet file.

Closes #1090.

[github-actions]

Results for impala-kudu are ready for: c6a.4xlarge.
The result files are committed as a6d958e.
Removed manually added result files: impala-kudu/results/20260927/c6a.2xlarge.json.

Logs:

• impala-kudu on c6a.4xlarge: benchmark log, cloud-init log