Apache Flink 1.6.2 Released

October 29, 2018 -

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

This release includes more than 30 fixes and minor improvements for Flink 1.6.1. The list below includes a detailed list of all fixes.

We highly recommend all users to upgrade to Flink 1.6.2.

Updated Maven dependencies:


You can find the binaries on the updated Downloads page.

List of resolved issues:


  • [FLINK-10242] - Latency marker interval should be configurable
  • [FLINK-10243] - Add option to reduce latency metrics granularity
  • [FLINK-10331] - Reduce number of flush requests to the network stack
  • [FLINK-10332] - Move data available notification in PipelinedSubpartition out of the synchronized block


  • [FLINK-5542] - YARN client incorrectly uses local YARN config to check vcore capacity
  • [FLINK-9567] - Flink does not release resource in Yarn Cluster mode
  • [FLINK-9788] - ExecutionGraph Inconsistency prevents Job from recovering
  • [FLINK-9884] - Slot request may not be removed when it has already be assigned in slot manager
  • [FLINK-9891] - Flink cluster is not shutdown in YARN mode when Flink client is stopped
  • [FLINK-9932] - Timed-out TaskExecutor slot-offers to JobMaster leak the slot
  • [FLINK-10135] - Certain cluster-level metrics are no longer exposed
  • [FLINK-10157] - Allow `null` user values in map state with TTL
  • [FLINK-10222] - Table scalar function expression parses error when function name equals the exists keyword suffix
  • [FLINK-10259] - Key validation for GroupWindowAggregate is broken
  • [FLINK-10263] - User-defined function with LITERAL paramters yields CompileException
  • [FLINK-10316] - Add check to KinesisProducer that aws.region is set
  • [FLINK-10354] - Savepoints should be counted as retained checkpoints
  • [FLINK-10363] - S3 FileSystem factory prints secrets into logs
  • [FLINK-10379] - Can not use Table Functions in Java Table API
  • [FLINK-10383] - Hadoop configurations on the classpath seep into the S3 file system configs
  • [FLINK-10390] - DataDog MetricReporter leaks connections
  • [FLINK-10400] - Return failed JobResult if job terminates in state FAILED or CANCELED
  • [FLINK-10415] - RestClient does not react to lost connection
  • [FLINK-10444] - Make S3 entropy injection work with FileSystem safety net
  • [FLINK-10451] - TableFunctionCollector should handle the life cycle of ScalarFunction
  • [FLINK-10465] - Jepsen: runit supervised sshd is stopped on tear down
  • [FLINK-10469] - FileChannel may not write the whole buffer in a single call to FileChannel.write(Buffer buffer)
  • [FLINK-10487] - fix invalid Flink SQL example
  • [FLINK-10516] - YarnApplicationMasterRunner does not initialize FileSystem with correct Flink Configuration during setup
  • [FLINK-10524] - MemoryManagerConcurrentModReleaseTest.testConcurrentModificationWhileReleasing failed on travis
  • [FLINK-10532] - Broken links in documentation
  • [FLINK-10544] - Remove custom settings.xml for snapshot deployments


  • [FLINK-9061] - Add entropy to s3 path for better scalability
  • [FLINK-10075] - HTTP connections to a secured REST endpoint flood the log
  • [FLINK-10260] - Confusing log messages during TaskManager registration
  • [FLINK-10282] - Provide separate thread-pool for REST endpoint
  • [FLINK-10291] - Generate JobGraph with fixed/configurable JobID in StandaloneJobClusterEntrypoint
  • [FLINK-10311] - HA end-to-end/Jepsen tests for standby Dispatchers
  • [FLINK-10312] - Wrong / missing exception when submitting job
  • [FLINK-10371] - Allow to enable SSL mutual authentication on REST endpoints by configuration
  • [FLINK-10375] - ExceptionInChainedStubException hides wrapped exception in cause
  • [FLINK-10582] - Make REST executor thread priority configurable