GitHub user tillrohrmann opened a pull request: https://github.com/apache/flink/pull/5267
[FLINK-8393] [flip6] Reconnect to last known JobMaster when connection is lost ## What is the purpose of the change Reconnect to the last known location of a lost `JobMaster` connection. ## Brief change log - In case of a heartbeat timeout or a disconnect call, the `TaskExecutor` tries to reconnect to the last known `JobMaster` location ## Verifying this change - Added `RegisteredRpcConnection#testReconnect` ## Does this pull request potentially affect one of the following parts: - Dependencies (does it add or upgrade a dependency): (no) - The public API, i.e., is any changed class annotated with `@Public(Evolving)`: (no) - The serializers: (no) - The runtime per-record code paths (performance sensitive): (no) - Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Yarn/Mesos, ZooKeeper: (yes) - The S3 file system connector: (no) ## Documentation - Does this pull request introduce a new feature? (no) - If yes, how is the feature documented? (not applicable) You can merge this pull request into a Git repository by running: $ git pull https://github.com/tillrohrmann/flink resumeLostJobMasterConnection Alternatively you can review and apply these changes as the patch at: https://github.com/apache/flink/pull/5267.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #5267 ---- commit f9bccf20b046e4f73a52ca2a4b842ca985dfaaa8 Author: Till Rohrmann <trohrmann@...> Date: 2018-01-09T16:50:37Z [FLINK-8392] [rpc] Let termination future be completed by AkkaRpcActor#postStop Revert the changes introduced by FLINK-7754. An RpcEndpoint's termination future is now completed from the AkkaRpcActor#postStop method. commit 21524394bc37372dc13eb5c3938de051cbd6f03e Author: Till Rohrmann <trohrmann@...> Date: 2018-01-08T17:23:27Z [FLINK-7910] [tests] Generalize Test(Stream)Environment to use JobExecutor This commit introduces the JobExecutor interface which abstracts the actual mini cluster from the Test(Stream)Environment. By letting the Flip-6 MiniCluster as well as the FlinkMiniCluster implement this interface, we can run all test base jobs either on the Flip-6 mini cluster or on the current mini cluster. This closes #4897. commit 5f1bdc1a8546e24e079753f92f22a397ccba24de Author: Till Rohrmann <trohrmann@...> Date: 2017-12-01T14:02:09Z [FLINK-8389] [flip6] Release all slots upon closing of JobManager connection commit 141f21d85f0047e4e5c0776e70ac6d83a03e5943 Author: Till Rohrmann <trohrmann@...> Date: 2018-01-09T13:11:20Z [hotfix] Add retrieval of key sets to DualKeyMap commit d46ba9c5f6c1320d73b4d0e65462bcf2c45ff28f Author: Till Rohrmann <trohrmann@...> Date: 2017-12-01T14:10:46Z [hotfix] Enable checkpointing RPC calls commit 737676e5912a2b43dd195ab2a940bc15af6630fb Author: Till Rohrmann <trohrmann@...> Date: 2018-01-09T08:28:34Z [hotfix] Add JavaDocs to OnCompletionActions commit f456248a2e75da4947cea7f2d863db129f0efc5f Author: Till Rohrmann <trohrmann@...> Date: 2018-01-09T15:44:59Z [hotfix] Refactor JobMasterTest to avoid using Mockito commit fa7b667a196980e05194b88ba352a12f330b5ad0 Author: Till Rohrmann <trohrmann@...> Date: 2018-01-09T19:37:08Z [FLINK-8393] Reconnect to last known JobMaster when connection is lost ---- ---