add index creation to migration scripts

[jgao54]

and remove the one-off run_non_transactional_migrations.

For existing prod DBs this is a no-op; for fresh DBs the table is empty, so a non-concurrent build is fine.

Will add a comment in the next release note about this so anyone using this in prod can also create this index before performing the upgrade.

[claude]

Code review

No issues found. Checked for bugs and CLAUDE.md compliance.

[github-actions]

🔄 Flaky Test Detected

Analysis: The e2e test suite hit the hard 900s timeout with no individual test assertion failures, indicating a CI infrastructure slowdown or a hung test rather than a code regression.
Confidence: 0.85

✅ Automatically retrying the workflow

View workflow run

[github-actions]

🔄 Flaky Test Detected

Analysis: The e2e test suite in the maria/8.0 CI matrix job hit the hard 900s timeout limit (ran for 900.754s) with no assertion failures, indicating a flaky timeout rather than a code regression.
Confidence: 0.88

✅ Automatically retrying the workflow

View workflow run

[github-actions]

❌ Test Failure

Analysis: All four e2e ClickHouse test suites fail Test_PartitionBy with a deterministic value mismatch (expected "num", got "(num)"` at clickhouse_test.go:3028), indicating a real regression in how the ClickHouse PARTITION BY key is extracted rather than a flaky/intermittent failure.
Confidence: 0.85

⚠️ This appears to be a real bug - manual intervention needed

View workflow run

[github-actions]

❌ Test Failure

Analysis: Real bug: Test_PartitionBy deterministically fails across all 4 test suites with expected: "num" vs actual: "(num)" (ClickHouse now returns partition keys in parentheses), and Test_NullEngine consistently times out with CDC data never reaching ClickHouse.
Confidence: 0.9

⚠️ This appears to be a real bug - manual intervention needed

View workflow run

[jgao54]

test failure is unrelated.