[lake/tiering] Continue tiering after empty tables

[beryllw]

Purpose

Linked issue: close #4372

Brief change log

Tests

API and Format

Documentation

[beryllw]

@luoyuxia cc

[Copilot]

🟡 Changes recommended

One or more issues must be addressed before approval.

Get a fresh assessment by requesting another Copilot review.

Pull request overview

This PR reduces lake-tiering delays by continuing table claims after empty tables, with controlled scheduling and focused tests.

Changes:

• Adds injectable scheduling and delayed empty-table continuation.
• Refactors startup dependencies for testability.
• Adds tests for continuation, failover, failures, and shutdown.

File summaries

File	Description
fluss-flink/fluss-flink-common/src/test/java/org/apache/fluss/flink/tiering/source/enumerator/TieringSourceEnumeratorSchedulingTest.java	Updated as part of this pull request.
fluss-flink/fluss-flink-common/src/main/java/org/apache/fluss/flink/tiering/source/enumerator/TieringSourceEnumerator.java	Updated as part of this pull request.

Review details

Suppressed comments (1)

fluss-flink/fluss-flink-common/src/main/java/org/apache/fluss/flink/tiering/source/enumerator/TieringSourceEnumerator.java:534

• This guard does not check whether any reader still demands a split, so an empty-table round can schedule a continuation after the last reader has been satisfied or unregistered. When it fires, requestTieringTableSplitsViaHeartBeat still performs a heartbeat (it only suppresses requestTable), adding an unnecessary coordinator RPC instead of falling back directly to the periodic poll. Include the demand check in this guard.

        if (closed || emptyTableContinuationScheduled) {
            return;

• Files reviewed: 2/2 changed files
• Comments generated: 1
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.