[KAFKA-5351] Broker clean bounce test puts the broker into a 'CONCURRENT_TRANSACTIONS' state permanently

• Type: Sub-task
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Fix Version/s: 0.11.0.0
• Component/s: clients, core, producer
• Labels: exactly-once
• Assignee: Apurva Mehta
• Reporter: Apurva Mehta
• Created: 31/May/17 05:19
• Updated: 01/Jun/17 05:49
• Resolved: 01/Jun/17 05:49

Description

In the broker clean bounce test, sometimes the consumer just hangs on a request to the transactional coordinator because it keeps getting a `CONCURRENT_TRANSACTIONS` error. This continues for 30 seconds, until the process is killed.

[2017-05-31 04:54:14,053] DEBUG TransactionalId my-second-transactional-id -- Received FindCoordinator response with error NONE (org.apache.kafka.clients.producer.internals.TransactionManager)
[2017-05-31 04:54:14,053] DEBUG TransactionalId: my-second-transactional-id -- Sending transactional request (transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,053] TRACE TransactionalId: my-second-transactional-id -- Waiting 100ms before resending a transactional request (transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,154] TRACE TransactionalId: my-second-transactional-id -- Sending transactional request (transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) to node 1 (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,191] TRACE Got transactional response for request:(transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) (org.apache.kafka.clients.producer.internals.TransactionManager)
[2017-05-31 04:54:14,191] DEBUG TransactionalId my-second-transactional-id -- Received EndTxn response with error COORDINATOR_NOT_AVAILABLE (org.apache.kafka.clients.producer.internals.TransactionManager)
[2017-05-31 04:54:14,192] DEBUG TransactionalId: my-second-transactional-id -- Sending transactional request (type=FindCoordinatorRequest, coordinatorKey=my-second-transactional-id, coordinatorType=TRANSACTION) (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,192] TRACE TransactionalId: my-second-transactional-id -- Sending transactional request (type=FindCoordinatorRequest, coordinatorKey=my-second-transactional-id, coordinatorType=TRANSACTION) to node 3 (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,193] TRACE Got transactional response for request:(type=FindCoordinatorRequest, coordinatorKey=my-second-transactional-id, coordinatorType=TRANSACTION) (org.apache.kafka.clients.producer.internals.TransactionManager)
[2017-05-31 04:54:14,193] DEBUG TransactionalId my-second-transactional-id -- Received FindCoordinator response with error NONE (org.apache.kafka.clients.producer.internals.TransactionManager)
[2017-05-31 04:54:14,193] DEBUG TransactionalId: my-second-transactional-id -- Sending transactional request (transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,193] TRACE TransactionalId: my-second-transactional-id -- Waiting 100ms before resending a transactional request (transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,294] TRACE TransactionalId: my-second-transactional-id -- Sending transactional request (transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) to node 1 (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,294] TRACE Got transactional response for request:(transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) (org.apache.kafka.clients.producer.internals.TransactionManager)
[2017-05-31 04:54:14,295] DEBUG TransactionalId my-second-transactional-id -- Received EndTxn response with error CONCURRENT_TRANSACTIONS (org.apache.kafka.clients.producer.internals.TransactionManager)
[2017-05-31 04:54:14,295] DEBUG TransactionalId: my-second-transactional-id -- Sending transactional request (transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,295] TRACE TransactionalId: my-second-transactional-id -- Waiting 100ms before resending a transactional request (transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,395] TRACE TransactionalId: my-second-transactional-id -- Sending transactional request (transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) to node 1 (org.apache.kafka.clients.producer.internals.Sender)
[2017-05-31 04:54:14,398] TRACE Got transactional response for request:(transactionalId=my-second-transactional-id, producerId=2000, producerEpoch=0, result=COMMIT) (org.apache.kafka.clients.producer.internals.TransactionManager)
[2017-05-31 04:54:14,398] DEBUG TransactionalId my-second-transactional-id -- Received EndTxn response with error CONCURRENT_TRANSACTIONS (org.apache.kafka.clients.producer.internals.TransactionManager)

Issue Links

• links to: GitHub Pull Request #3184