Fixing the Flaky Test for org.apache.fluss.client.table.scanner.batch.KvSnapshotBatchScannerITCase.testKvSnapshotLease

[hemanthsavasere]

Purpose

• Fix a race condition in FlussClusterExtension.triggerSnapshot() that caused testKvSnapshotLease to intermittently fail with expected: [1L, 1L, 1L] but was: [1L, 0L, 1L].

Linked issue: close #2807

Brief change log

File: fluss-server/src/test/java/org/apache/fluss/server/testutils/FlussClusterExtension.java

• Before: Read currentSnapshotId() synchronously right after triggerSnapshot(); if unchanged, returned null and skipped the bucket
• After: Poll currentSnapshotId() in a loop (10ms sleep, 30s timeout) until the ID increments, confirming the async trigger was actually processed; fail with a clear message on timeout

Tests

• Test: fluss-client/src/test/java/org/apache/fluss/client/table/scanner/batch/KvSnapshotBatchScannerITCase.java
• Fix target: fluss-server/src/test/java/org/apache/fluss/server/testutils/FlussClusterExtension.java:747-775

API and Format

• No API changes. Changes only related to tests.

Documentation

• No new features are introduced