Xiaojian Zhou created GEODE-9989: ------------------------------------ Summary: add a few info level logs in PersistenceAdvisorImpl to identify splitbrain issue Key: GEODE-9989 URL: https://issues.apache.org/jira/browse/GEODE-9989 Project: Geode Issue Type: Bug Reporter: Xiaojian Zhou
In scenario like: {code:java} 03:33:03.644 dataStoregemfire4_4494 recovered from disk 03:33:03.732 dataStoregemfire4_4494 closing 03:33:03.735 dataStoregemfire4_4494 Initialization of region replicate_5 completed, send newId(let’s name it 4494) to gemfire2 03:33:03.754 dataStoregemfire2_4493 recovered from disk 03:33:03.770 dataStoregemfire2_4493 closing 03:33:03.792 dataStoregemfire2_4493 Initialization of region replicate_5 completed. send newId(let’s name is 4493) to gemfire4, but gemfire4 is offline. So gemfire4 does not know gemfire2’s newId 4493. 03:34:11.247 gemfire4_9779 restarted, it does not know 4493 03:34:11.269 gemfire2_9856 restarted, it sends oldId=4493, newId=9856 to gemfire4, but gemfire4 does not know either of gemfire2’s oldId and newId When gemfire2_9856 asked gemfire4_9779 for its state, gemfire4_9779 replied "I don't know you", then gemfire2_9856's starting ends with ConflictingPersistentDataException. {code} We need more log to identify the issue. -- This message was sent by Atlassian Jira (v8.20.1#820001)