Chia-Ping Tsai created KAFKA-21089:
--------------------------------------
Summary: Ensure the producer gets closed in KafkaProducerTest to
avoid cascading failures
Key: KAFKA-21089
URL: https://issues.apache.org/jira/browse/KAFKA-21089
Project: Kafka
Issue Type: Bug
Reporter: Chia-Ping Tsai
Assignee: Chia-Ping Tsai
from:
[https://github.com/apache/kafka/actions/runs/34763157891/job/103740045036?pr=23451]
{code:java}
Found 93 test failures:
FAILED ❌ KafkaProducerTest > testCloseIsForcedOnPendingFindCoordinator()
FAILED ❌ KafkaProducerTest >
testSubscribingCustomMetricsWithSameNameDoesntAffectProducerMetrics()
FAILED ❌ KafkaProducerTest > testRetriesAndIdempotenceForIdempotentProducers()
FAILED ❌ KafkaProducerTest > testCloseWhenWaitingForMetadataUpdate()
FAILED ❌ KafkaProducerTest > testTopicExpiryInMetadata()
FAILED ❌ KafkaProducerTest >
"testMetadataWithPartitionOutOfRange(boolean).isIdempotenceEnabled=true"
FAILED ❌ KafkaProducerTest >
"testMetadataWithPartitionOutOfRange(boolean).isIdempotenceEnabled=false"
FAILED ❌ KafkaProducerTest >
"testInitTransactionsWithKeepPreparedTxnAndTwoPhaseCommit(boolean,
boolean).keepPreparedTxn=true, enable2PC=false"
FAILED ❌ KafkaProducerTest >
"testInitTransactionsWithKeepPreparedTxnAndTwoPhaseCommit(boolean,
boolean).keepPreparedTxn=true, enable2PC=true"
FAILED ❌ KafkaProducerTest >
"testInitTransactionsWithKeepPreparedTxnAndTwoPhaseCommit(boolean,
boolean).keepPreparedTxn=false, enable2PC=true"
FAILED ❌ KafkaProducerTest > testDeliveryTimeoutAndLingerMsConfig()
FAILED ❌ KafkaProducerTest >
testSendOffsetsToTransactionTriggersMetadataRefreshThenNegotiatesV6()
FAILED ❌ KafkaProducerTest > testHeadersSuccess()
FAILED ❌ KafkaProducerTest > testTopicRefreshInMetadata()
FAILED ❌ KafkaProducerTest > testNullGroupMetadataInSendOffsets() {code}
Many test cases in KafkaProducerTest don't close the producer in a
try-with-resources or finally block. If an assertion fails in the middle of the
test, the sender thread is leaked, and every subsequent test in the class fails
in {{detectLeaks}}
Also, the root cause of that failure is a too-small timeout (2000ms) for
{{client.waitForRequests}}
{code:java}
client.waitForRequests(1, 2000); {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
