build(ci): Update the Presto Java image to 0.297

[czentgr]

No description provided.

[netlify]

✅ Deploy Preview for meta-velox canceled.

Name	Link
🔨 Latest commit	cb019d4
🔍 Latest deploy log	https://app.netlify.com/projects/meta-velox/deploys/69e246bebca06a0008f70433

[github-actions]

Build Impact Analysis

No build targets affected by this change.

---

Fast path • Graph from main@3da0390118fa4b0addbfbdeb03f438ec1b1237a6

[github-actions]

CI Failure Analysis

All 7 failed fuzzer jobs fail with the same root cause: the Presto Java server fails to start inside the new container image, causing every fuzzer that depends on Presto as a source of truth to abort.

🔴 Join Fuzzer — INFRA Failure View logs

🔴 RowNumber Fuzzer — INFRA Failure View logs

🔴 Expression Fuzzer with Presto SOT — INFRA Failure View logs

🔴 Window Fuzzer with Presto as source of truth — INFRA Failure View logs

🔴 TopNRowNumber Fuzzer — INFRA Failure View logs

🔴 Aggregation Fuzzer with Presto as source of truth — INFRA Failure View logs

🔴 Writer Fuzzer with Presto as source of truth — INFRA Failure View logs

Error (identical in all 7 jobs):

Error running command: java.net.ConnectException: Failed to connect to /127.0.0.1:8080

The Presto CLI (/opt/presto-cli --server 127.0.0.1:8080 --execute 'CREATE SCHEMA IF NOT EXISTS hive.tpch;') fails to connect after the 60-second startup wait. The Presto server ($PRESTO_HOME/bin/launcher run) never becomes reachable on port 8080. No fuzzer code ran — the jobs fail during Presto server initialization.

---

Correlation with PR changes:

This failure is directly caused by the PR changes. The PR makes two related changes:

1. Container image switch (.github/workflows/scheduled.yml): All fuzzer jobs are changed from ghcr.io/facebookincubator/velox-dev:presto-java → ghcr.io/czentgr/czentgr-test:presto-java
2. Presto version bump (docker-compose.yml, scripts/docker/java.dockerfile): PRESTO_VERSION changed from 0.295 → 0.297

The new container image ghcr.io/czentgr/czentgr-test:presto-java contains a Presto 0.297 server that fails to start. The uploaded server.log artifacts are only 205 bytes each, meaning the Presto server likely crashes immediately or never writes meaningful output.

Known issues:

• No open issues track this specific failure.
• The most recent Fuzzer Jobs run on main (run 24550311472) passed successfully, confirming this is not a pre-existing/flaky issue. The failure is specific to this PR's container image change.

Recommended fix:

• Investigate why Presto 0.297 fails to start in the ghcr.io/czentgr/czentgr-test:presto-java container. Check the server logs for JVM/startup errors (the uploaded artifacts may contain clues).
• Consider testing the container image locally first: run the container, start Presto, and verify it listens on port 8080 before pushing the CI change.
• If Presto 0.297 requires additional JVM args, memory settings, or configuration changes, update the workflow or the container image accordingly.
• If the image is still being built/tested, consider keeping the original ghcr.io/facebookincubator/velox-dev:presto-java container reference until the new image is validated.