Marton Greber has uploaded this change for review. ( http://gerrit.cloudera.org:8080/23608
Change subject: KUDU-3662: Add stress test for checkpoint recovery ...................................................................... KUDU-3662: Add stress test for checkpoint recovery Add TestReplicationCheckpointStress to verify data durability under extreme failure conditions. The test aggressively triggers the checkpoint race condition by: - Creating table with multiple partitions (high split count) - Inserting data continuously across partitions - Repeatedly crashing and restoring from checkpoints mid-replication - Verifying no data loss after multiple crash/recovery cycles Test validates that the checkpoint race condition fix in MetricWrappedKuduEnumerator correctly handles: - Splits finishing between snapshot and checkpoint completion - Multiple concurrent splits in various states (active, buffered) - At-least-once semantics with idempotent UPSERT sink operations Test stability: Timeouts carefully tuned based on checkpoint and discovery intervals to avoid flakiness. Verified by running 1024 iterations on dist-test with debug build - no flakiness detected. Change-Id: Ic463c143a5330bf8db7c5b026a8b4dbe83b7a769 --- A java/kudu-replication/src/test/java/org/apache/kudu/replication/TestReplicationCheckpointStress.java 1 file changed, 275 insertions(+), 0 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/08/23608/1 -- To view, visit http://gerrit.cloudera.org:8080/23608 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: kudu Gerrit-Branch: master Gerrit-MessageType: newchange Gerrit-Change-Id: Ic463c143a5330bf8db7c5b026a8b4dbe83b7a769 Gerrit-Change-Number: 23608 Gerrit-PatchSet: 1 Gerrit-Owner: Marton Greber <[email protected]> Gerrit-Reviewer: Kudu Jenkins (120)
