chia7712 commented on a change in pull request #8981:
URL: https://github.com/apache/kafka/pull/8981#discussion_r451219201



##########
File path: tests/kafkatest/tests/core/transactions_test.py
##########
@@ -47,7 +47,10 @@ def __init__(self, test_context):
         self.num_output_partitions = 3
         self.num_seed_messages = 100000
         self.transaction_size = 750
-        self.transaction_timeout = 40000
+        # This is reducing the transaction cleanup interval.
+        # IN hard_bounce mode, transaction is broke ungracefully. Hence, it 
produces unstable
+        # offsets which obstructs TransactionalMessageCopier from receiving 
position of group.
+        self.transaction_timeout = 5000

Review comment:
       > the broker failure will be detected after ZK session expiration
   
   It is unrelate to broker. The root cause is client 
(TransactionMessageCopier) is ungracefully closed in hard_bounce mode. The 
transactional message sent by client is not correctly completed so there are 
pending transaction (unstable offset) stored by broker. When the client is 
restarting, it fails to get offsets of partition due to unstable offset 
(TransactionMessageCopier needs to get position of partition to calculate 
remaining messages after restarting).
   
   The timeout of transaction is higher than the timeout of verification so 
this test fails if any transaction is not correctly abort.




----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


Reply via email to