hachikuji commented on code in PR #12149: URL: https://github.com/apache/kafka/pull/12149#discussion_r874042264
########## clients/src/main/java/org/apache/kafka/clients/producer/internals/TransactionManager.java: ########## @@ -1292,8 +1292,10 @@ public void handleResponse(AbstractResponse response) { reenqueue(); } else if (error == Errors.COORDINATOR_LOAD_IN_PROGRESS || error == Errors.CONCURRENT_TRANSACTIONS) { reenqueue(); - } else if (error == Errors.TRANSACTIONAL_ID_AUTHORIZATION_FAILED || - error == Errors.CLUSTER_AUTHORIZATION_FAILED) { + } else if (error == Errors.CLUSTER_AUTHORIZATION_FAILED) { + log.warn("Retrying upon missing initProducerId due to cluster authorization error"); + reenqueue(); Review Comment: Hmm, I think we do want the error to get propagated back to the application. Most users would expect the application to fail so that they have a chance to see the issue and fix it. High level, what I was thinking we could do is fail the request by calling `result.fail`, then transition back to `UNINITIALIZED`. Then if the user wants to, they can call `initTransactions()` to retry. -- 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: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org