Apache Flink 1.6.3 Released

December 22, 2018 -

The Apache Flink community released the third bugfix version of the Apache Flink 1.6 series.

This release includes more than 80 fixes and minor improvements for Flink 1.6.2. The list below includes a detailed list of all fixes.

We highly recommend all users to upgrade to Flink 1.6.3.

Updated Maven dependencies:


You can find the binaries on the updated Downloads page.

List of resolved issues:


  • [FLINK-10097] - More tests to increase StreamingFileSink test coverage
  • [FLINK-10252] - Handle oversized metric messages
  • [FLINK-10367] - Avoid recursion stack overflow during releasing SingleInputGate
  • [FLINK-10863] - Assign uids to all operators in general purpose testing job


  • [FLINK-8336] - YarnFileStageTestS3ITCase.testRecursiveUploadForYarnS3 test instability
  • [FLINK-9635] - Local recovery scheduling can cause spread out of tasks
  • [FLINK-9646] - ExecutionGraphCoLocationRestartTest.testConstraintsAfterRestart failed on Travis
  • [FLINK-9878] - IO worker threads BLOCKED on SSL Session Cache while CMS full gc
  • [FLINK-10149] - Fink Mesos allocates extra port when not configured to do so.
  • [FLINK-10166] - Dependency problems when executing SQL query in sql-client
  • [FLINK-10309] - Cancel with savepoint fails with java.net.ConnectException when using the per job-mode
  • [FLINK-10357] - Streaming File Sink end-to-end test failed with mismatch
  • [FLINK-10359] - Scala example in DataSet docs is broken
  • [FLINK-10364] - Test instability in NonHAQueryableStateFsBackendITCase#testMapState
  • [FLINK-10419] - ClassNotFoundException while deserializing user exceptions from checkpointing
  • [FLINK-10425] - taskmanager.host is not respected
  • [FLINK-10455] - Potential Kafka producer leak in case of failures
  • [FLINK-10463] - Null literal cannot be properly parsed in Java Table API function call
  • [FLINK-10481] - Wordcount end-to-end test in docker env unstable
  • [FLINK-10482] - java.lang.IllegalArgumentException: Negative number of in progress checkpoints
  • [FLINK-10491] - Deadlock during spilling data in SpillableSubpartition
  • [FLINK-10566] - Flink Planning is exponential in the number of stages
  • [FLINK-10567] - Lost serialize fields when ttl state store with the mutable serializer
  • [FLINK-10570] - State grows unbounded when "within" constraint not applied
  • [FLINK-10581] - YarnConfigurationITCase.testFlinkContainerMemory test instability
  • [FLINK-10642] - CodeGen split fields errors when maxGeneratedCodeLength equals 1
  • [FLINK-10655] - RemoteRpcInvocation not overwriting ObjectInputStream's ClassNotFoundException
  • [FLINK-10663] - Closing StreamingFileSink can cause NPE
  • [FLINK-10669] - Exceptions & errors are not properly checked in logs in e2e tests
  • [FLINK-10670] - Fix Correlate codegen error
  • [FLINK-10674] - Fix handling of retractions after clean up
  • [FLINK-10681] - elasticsearch6.ElasticsearchSinkITCase fails if wrong JNA library installed
  • [FLINK-10690] - Tests leak resources via Files.list
  • [FLINK-10693] - Fix Scala EitherSerializer duplication
  • [FLINK-10715] - E2e tests fail with ConcurrentModificationException in MetricRegistryImpl
  • [FLINK-10750] - SocketClientSinkTest.testRetry fails on Travis
  • [FLINK-10752] - Result of AbstractYarnClusterDescriptor#validateClusterResources is ignored
  • [FLINK-10753] - Propagate and log snapshotting exceptions
  • [FLINK-10763] - Interval join produces wrong result type in Scala API
  • [FLINK-10770] - Some generated functions are not opened properly.
  • [FLINK-10773] - Resume externalized checkpoint end-to-end test fails
  • [FLINK-10809] - Using DataStreamUtils.reinterpretAsKeyedStream produces corrupted keyed state after restore
  • [FLINK-10816] - Fix LockableTypeSerializer.duplicate()
  • [FLINK-10821] - Resuming Externalized Checkpoint E2E test does not resume from Externalized Checkpoint
  • [FLINK-10839] - Fix implementation of PojoSerializer.duplicate() w.r.t. subclass serializer
  • [FLINK-10842] - Waiting loops are broken in e2e/common.sh
  • [FLINK-10856] - Harden resume from externalized checkpoint E2E test
  • [FLINK-10857] - Conflict between JMX and Prometheus Metrics reporter
  • [FLINK-10880] - Failover strategies should not be applied to Batch Execution
  • [FLINK-10913] - ExecutionGraphRestartTest.testRestartAutomatically unstable on Travis
  • [FLINK-10925] - NPE in PythonPlanStreamer
  • [FLINK-10946] - Resuming Externalized Checkpoint (rocks, incremental, scale up) end-to-end test failed on Travis
  • [FLINK-10990] - Enforce minimum timespan in MeterView
  • [FLINK-10992] - Jepsen: Do not use /tmp as HDFS Data Directory
  • [FLINK-10997] - Avro-confluent-registry does not bundle any dependency
  • [FLINK-10998] - flink-metrics-ganglia has LGPL dependency
  • [FLINK-11011] - Elasticsearch 6 sink end-to-end test unstable
  • [FLINK-11017] - Time interval for window aggregations in SQL is wrongly translated if specified with YEAR_MONTH resolution
  • [FLINK-11029] - Incorrect parameter in Working with state doc
  • [FLINK-11041] - ReinterpretDataStreamAsKeyedStreamITCase.testReinterpretAsKeyedStream failed on Travis
  • [FLINK-11045] - UserCodeClassLoader has not been set correctly for RuntimeUDFContext in CollectionExecutor
  • [FLINK-11083] - CRowSerializerConfigSnapshot is not instantiable
  • [FLINK-11087] - Broadcast state migration Incompatibility from 1.5.3 to 1.7.0
  • [FLINK-11123] - Missing import in ML quickstart docs
  • [FLINK-11136] - Fix the logical of merge for DISTINCT aggregates


  • [FLINK-4173] - Replace maven-assembly-plugin by maven-shade-plugin in flink-metrics
  • [FLINK-10353] - Restoring a KafkaProducer with Semantic.EXACTLY_ONCE from a savepoint written with Semantic.AT_LEAST_ONCE fails with NPE
  • [FLINK-10608] - Add avro files generated by datastream-allround-test to RAT exclusions
  • [FLINK-10613] - Remove logger casts in HBaseConnectorITCase
  • [FLINK-10614] - Update test_batch_allround.sh e2e to new testing infrastructure
  • [FLINK-10637] - Start MiniCluster with random REST port
  • [FLINK-10678] - Add a switch to run_test to configure if logs should be checked for errors/excepions
  • [FLINK-10692] - Harden Confluent schema E2E test
  • [FLINK-10883] - Submitting a jobs without enough slots times out due to a unspecified timeout
  • [FLINK-10906] - docker-entrypoint.sh logs credentails during startup
  • [FLINK-10916] - Include duplicated user-specified uid into error message
  • [FLINK-10951] - Disable enforcing of YARN container virtual memory limits in tests
  • [FLINK-11005] - Define flink-sql-client uber-jar dependencies via artifactSet


  • [FLINK-10606] - Construct NetworkEnvironment simple for tests
  • [FLINK-10607] - Unify to remove duplicated NoOpResultPartitionConsumableNotifier
  • [FLINK-10827] - Add test for duplicate() to SerializerTestBase


  • [FLINK-10793] - Change visibility of TtlValue and TtlSerializer to public for external tools