Apache Flink 1.12.5 Released

August 6, 2021 - Jingsong Lee

The Apache Flink community released the next bugfix version of the Apache Flink 1.12 series.

This release includes 76 fixes and minor improvements for Flink 1.12.4. The list below includes a detailed list of all fixes and improvements.

We highly recommend all users to upgrade to Flink 1.12.5.

Updated Maven dependencies:


You can find the binaries on the updated Downloads page.

Release Notes - Flink - Version 1.12.5


  • [FLINK-19925] - Errors$NativeIoException: readAddress(..) failed: Connection reset by peer
  • [FLINK-20321] - Get NPE when using AvroDeserializationSchema to deserialize null input
  • [FLINK-20888] - ContinuousFileReaderOperator should not close the output on close()
  • [FLINK-21329] - "Local recovery and sticky scheduling end-to-end test" does not finish within 600 seconds
  • [FLINK-21445] - Application mode does not set the configuration when building PackagedProgram
  • [FLINK-21469] - stop-with-savepoint --drain doesn't advance watermark for sources chained to MultipleInputStreamTask
  • [FLINK-21952] - Make all the "Connection reset by peer" exception wrapped as RemoteTransportException
  • [FLINK-22015] - SQL filter containing OR and IS NULL will produce an incorrect result.
  • [FLINK-22105] - SubtaskCheckpointCoordinatorTest.testForceAlignedCheckpointResultingInPriorityEvents unstable
  • [FLINK-22157] - Join & Select a part of composite primary key will cause ArrayIndexOutOfBoundsException
  • [FLINK-22312] - YARNSessionFIFOSecuredITCase>YARNSessionFIFOITCase.checkForProhibitedLogContents due to the heartbeat exception with Yarn RM
  • [FLINK-22408] - Flink Table Parsr Hive Drop Partitions Syntax unparse is Error
  • [FLINK-22419] - testScheduleRunAsync fail
  • [FLINK-22434] - Dispatcher does not store suspended jobs in execution graph store
  • [FLINK-22443] - can not be execute an extreme long sql under batch mode
  • [FLINK-22494] - Avoid discarding checkpoints in case of failure
  • [FLINK-22496] - ClusterEntrypointTest.testCloseAsyncShouldBeExecutedInShutdownHook failed
  • [FLINK-22502] - DefaultCompletedCheckpointStore drops unrecoverable checkpoints silently
  • [FLINK-22547] - OperatorCoordinatorHolderTest. verifyCheckpointEventOrderWhenCheckpointFutureCompletesLate fail
  • [FLINK-22564] - Kubernetes-related ITCases do not fail even in case of failure
  • [FLINK-22592] - numBuffersInLocal is always zero when using unaligned checkpoints
  • [FLINK-22613] - FlinkKinesisITCase.testStopWithSavepoint fails
  • [FLINK-22683] - The total Flink/process memory of memoryConfiguration in /taskmanagers can be null or incorrect value
  • [FLINK-22698] - RabbitMQ source does not stop unless message arrives in queue
  • [FLINK-22704] - ZooKeeperHaServicesTest.testCleanupJobData failed
  • [FLINK-22721] - Breaking HighAvailabilityServices interface by adding new method
  • [FLINK-22733] - Type mismatch thrown in DataStream.union if parameter is KeyedStream for Python DataStream API
  • [FLINK-22756] - DispatcherTest.testJobStatusIsShownDuringTermination fail
  • [FLINK-22788] - Code of equals method grows beyond 64 KB
  • [FLINK-22814] - New sources are not defining/exposing checkpointStartDelayNanos metric
  • [FLINK-22815] - Disable unaligned checkpoints for broadcast partitioning
  • [FLINK-22819] - YARNFileReplicationITCase fails with "The YARN application unexpectedly switched to state FAILED during deployment"
  • [FLINK-22820] - Stopping Yarn session cluster will cause fatal error
  • [FLINK-22833] - Source tasks (both old and new) are not reporting checkpointStartDelay via CheckpointMetrics
  • [FLINK-22856] - Move our Azure pipelines away from Ubuntu 16.04 by September
  • [FLINK-22886] - Thread leak in RocksDBStateUploader
  • [FLINK-22898] - HiveParallelismInference limit return wrong parallelism
  • [FLINK-22908] - FileExecutionGraphInfoStoreTest.testPutSuspendedJobOnClusterShutdown should wait until job is running
  • [FLINK-22927] - Exception on JobClient.get_job_status().result()
  • [FLINK-22946] - Network buffer deadlock introduced by unaligned checkpoint
  • [FLINK-22952] - docs_404_check fail on azure due to ruby version not available
  • [FLINK-22963] - The description of taskmanager.memory.task.heap.size in the official document is incorrect
  • [FLINK-22964] - Connector-base exposes dependency to flink-core.
  • [FLINK-22987] - Scala suffix check isn't working
  • [FLINK-23010] - HivePartitionFetcherContextBase::getComparablePartitionValueList can return partitions that don't exist
  • [FLINK-23030] - PartitionRequestClientFactory#createPartitionRequestClient should throw when network failure
  • [FLINK-23045] - RunnablesTest.testExecutorService_uncaughtExceptionHandler fails on azure
  • [FLINK-23074] - There is a class conflict between flink-connector-hive and flink-parquet
  • [FLINK-23076] - DispatcherTest.testWaitingForJobMasterLeadership fails on azure
  • [FLINK-23119] - Fix the issue that the exception that General Python UDAF is unsupported is not thrown in Compile Stage.
  • [FLINK-23120] - ByteArrayWrapperSerializer.serialize should use writeInt to serialize the length
  • [FLINK-23133] - The dependencies are not handled properly when mixing use of Python Table API and Python DataStream API
  • [FLINK-23135] - Flink SQL Error while applying rule AggregateReduceGroupingRule
  • [FLINK-23164] - JobMasterTest.testMultipleStartsWork unstable on azure
  • [FLINK-23166] - ZipUtils doesn't handle properly for softlinks inside the zip file
  • [FLINK-23182] - Connection leak in RMQSource
  • [FLINK-23184] - CompileException Assignment conversion not possible from type "int" to type "short"
  • [FLINK-23201] - The check on alignmentDurationNanos seems to be too strict
  • [FLINK-23223] - When flushAlways is enabled the subpartition may lose notification of data availability
  • [FLINK-23233] - OperatorEventSendingCheckpointITCase.testOperatorEventLostWithReaderFailure fails on azure
  • [FLINK-23248] - SinkWriter is not closed when failing
  • [FLINK-23417] - MiniClusterITCase.testHandleBatchJobsWhenNotEnoughSlot fails on Azure
  • [FLINK-23429] - State Processor API failed with FileNotFoundException when working with state files on Cloud Storage


  • [FLINK-17857] - Kubernetes and docker e2e tests could not run on Mac OS after migration
  • [FLINK-18182] - Upgrade AWS SDK in flink-connector-kinesis to include new region af-south-1
  • [FLINK-20695] - Zookeeper node under leader and leaderlatch is not deleted after job finished
  • [FLINK-21229] - Support ssl connection with schema registry format
  • [FLINK-21411] - The components on which Flink depends may contain vulnerabilities. If yes, fix them.
  • [FLINK-22708] - Propagate savepoint settings from StreamExecutionEnvironment to StreamGraph
  • [FLINK-22747] - Update commons-io to 2.8
  • [FLINK-22757] - Update GCS documentation
  • [FLINK-22774] - Update Kinesis SQL connector's Guava to 27.0-jre
  • [FLINK-22939] - Generalize JDK switch in azure setup
  • [FLINK-23009] - Bump up Guava in Kinesis Connector
  • [FLINK-23052] - cron_snapshot_deployment_maven unstable on maven
  • [FLINK-23312] - Use -Dfast for building e2e tests on AZP