MartijnVisser opened a new pull request, #259: URL: https://github.com/apache/flink-connector-kafka/pull/259
Backport of [FLINK-39699] (PR #254) from `main` to `v4.0`. ## Why Both `KafkaSinkITCase` and `KafkaWriterFaultToleranceITCase` are recurring failures in the Weekly Flink Connector Kafka CI on `v4.0` (Flink 2.0.1). FLINK-39699 stabilized both on `main` but was never backported. ## What Two of the three FLINK-39699 commits, both of which cherry-pick cleanly onto `v4.0`: - **Wait for completed checkpoint stats in KafkaSinkITCase** (`2d441987`) — fixes the intermittent `containsExactlyInAnyOrderElementsOf` assertion failure (e.g. `testAbortTransactionsAfterScaleInBeforeFirstCheckpoint`) by waiting for a completed checkpoint before asserting on committed records. - **Stabilize KafkaWriterFaultToleranceITCase exception-on-unavailable tests** (`887d5941`) — drains a warm-up record before stopping the broker so the operation under test reliably surfaces the expected `NetworkException`/`TimeoutException` instead of an "actual not to be null" failure. The third FLINK-39699 commit (`443c0f32`, partition-assignment wait) is intentionally **omitted**: it depends on the `createNewTopicAndWaitForPartitionAssignment` helper introduced separately by FLINK-39234, which is not on `v4.0`. The two commits here already cover both observed weekly failures. ## Verification - `mvn clean test-compile -pl flink-connector-kafka` succeeds on `v4.0`. - `spotless:check` passes. - The integration tests were not run against the live flaky scenario (Docker-based, flaky by nature); efficacy is best confirmed across weekly CI cycles. Generated-by: Claude Code (Opus 4.8) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
