| 2b8b670 | Xintong Song | 20 January 2021, 03:45:46 UTC | Commit for release 1.10.3 | 20 January 2021, 03:45:46 UTC |
| 6e38231 | Xintong Song | 15 January 2021, 08:05:36 UTC | [FLINK-20906][legal] Update copyright year to 2021 for NOTICE files. This closes #14661 | 15 January 2021, 08:39:16 UTC |
| e6ebcd1 | yushujun | 19 June 2020, 08:33:33 UTC | [hotfix][table][docs] Fix typo in System (Built-in) Functions page This closes #12716. | 26 November 2020, 14:03:06 UTC |
| 46680ce | leiqiang | 18 November 2020, 09:27:59 UTC | [hotfix][table][docs] Fix table systemFunction lag description. This closes #14115. | 26 November 2020, 13:45:17 UTC |
| d2409c1 | Till Rohrmann | 13 November 2020, 11:35:51 UTC | [FLINK-15170][test] Replace System.currentTimeMillis with System.nanoTime in WebFrontendITCase Replacing System.currentTimeMillis with System.nanoTime ensures that the test case won't suffer from jumping clocks as they can occur with System.currentTimeMillis. This should fix the observed test instability. This closes #14065. | 19 November 2020, 13:33:40 UTC |
| 2fbef47 | HuangXingBo | 19 November 2020, 03:39:06 UTC | [FLINK-20218][python] Import urllib.parse before importing apache-beam (#14125) | 19 November 2020, 03:39:06 UTC |
| 2eaba74 | Dian Fu | 18 November 2020, 08:33:14 UTC | [FLINK-20183][python] Fix the default PYTHONPATH is overwritten in client side This closes #14108. | 18 November 2020, 08:33:14 UTC |
| e876073 | Matthias Pohl | 15 November 2020, 21:20:24 UTC | [hotfix][runtime] A customized filename can be specified through Content-Disposition that also allows passing of path information which was not properly handled. This is fixed now. We just use the filename instead of interpreting any path information that was passed through a custom filename. Two tests were added to verify the proper behavior: 1. a custom filename without path information was used 2. a custom filename with path information was used The change required adapting the MultipartUploadResource in a way that it is used not as a @ClassRule but as a @Rule instead. This enables us to initialize it differently on a per-test level. The change makes the verification of the uploaded files configurable. | 18 November 2020, 06:21:27 UTC |
| a14665d | Jiayi Liao | 16 November 2020, 09:30:17 UTC | [FLINK-16753] Use CheckpointException to wrap exceptions thrown from StreamTask (#14071) | 16 November 2020, 09:30:17 UTC |
| 10f155e | SteNicholas | 12 November 2020, 10:03:07 UTC | [BP-1.10][FLINK-20013][network] BoundedBlockingSubpartition may leak network buffer if task is failed or canceled | 12 November 2020, 20:10:14 UTC |
| fef13fd | Till Rohrmann | 11 November 2020, 17:21:07 UTC | Revert "[FLINK-20033] Make JobManagerJobStatusListener call back run directly" This reverts commit 9840907ec8c16d4d50d92388b883b959cc10f793. This closes #14039. | 12 November 2020, 13:55:41 UTC |
| d8fa676 | Till Rohrmann | 11 November 2020, 17:21:00 UTC | Revert "[FLINK-20033] Ensure that stopping a JobMaster will suspend the running job" This reverts commit 10bfd2971dd960ac799ba02156683fc7a740383d. | 11 November 2020, 17:21:00 UTC |
| fd624ce | Matthias Pohl | 10 November 2020, 12:56:01 UTC | [FLINK-20076][runtime][test] Fixed DispatcherTest.testOnRemovedJobGraphDoesNotCleanUpHAFiles. We missed starting the JobGraphStore which caused an exception and prevented us testing the actual behavior even though the test succeeded. This closes #14019. | 11 November 2020, 09:44:32 UTC |
| e3beef3 | Dawid Wysakowicz | 06 November 2020, 10:07:31 UTC | [FLINK-20018] Allow escaping in 'pipeline.cached-files' and 'pipeline.default-kryo-serializers' This commit enables escaping in options that expect a map of string-string entries. It lets users pass options such as e.g. pipeline.cached-files=name:file1,path:'oss://bucket/file1' | 11 November 2020, 08:51:49 UTC |
| 10bfd29 | Till Rohrmann | 07 November 2020, 16:22:52 UTC | [FLINK-20033] Ensure that stopping a JobMaster will suspend the running job This commit adds a test which ensures that stopping a JobMaster will suspend the running job. This closes #13980. | 09 November 2020, 09:10:17 UTC |
| 9840907 | Till Rohrmann | 07 November 2020, 16:20:25 UTC | [FLINK-20033] Make JobManagerJobStatusListener call back run directly Since the ExecutionGraph can only be modified using the JobMaster's main thread, we no longer need to trigger a jobStatusChanged async message when the JobManagerJobStatusListener.jobStatusChanges callback is triggered. This ensures that job status changes will directly be reported. | 07 November 2020, 16:45:19 UTC |
| 260cf3b | Robert Metzger | 14 October 2020, 06:17:10 UTC | [FLINK-19237] Fix rejected slot offer bug in JobMaster The JobMaster was rejecting slot offers from TaskExecutors erroneously in cases where the TaskExecutor ID is reused. This fixes the JobMaster.suspend() behavior to do proper bookkeeping of TaskExecutors. | 07 November 2020, 16:45:19 UTC |
| 85fc5c1 | Chesnay Schepler | 03 November 2020, 17:56:20 UTC | [FLINK-19901][metrics] Fix caching offset for variables | 03 November 2020, 18:01:56 UTC |
| bbeecc3 | Piotr Nowojski | 28 October 2020, 11:17:47 UTC | [FLINK-19671][codestyle] Revert .editorconfig change violating our coding style > Rules about breaking the long lines: > > Break the argument list or chain of calls if the line exceeds limit or earlier if you believe that the breaking would improve the code readability > If you break the line then each argument/call should have a separate line, including the first one > Each new line should have one extra indentation (or two for a function declaration) relative to the line of the parent function name or the called entity | 29 October 2020, 08:19:47 UTC |
| 64f4f98 | Yangze Guo | 27 October 2020, 12:49:01 UTC | [FLINK-19252][security] Proactively create the working dir for Jaas security module | 27 October 2020, 12:50:58 UTC |
| 933041b | Till Rohrmann | 21 October 2020, 12:18:59 UTC | [FLINK-19557] Trigger LeaderRetrievalListener notification upon ZooKeeper reconnection in ZooKeeperLeaderRetrievalService We have to trigger the LeaderRetrievalListener notification upon reconnecting to ZooKeeper because the NodeCache might not trigger the nodeChanged call if the server state is the same as the state cached in the NodeCache. Therefore, we would miss to tell the listener that the old leader (before the connection loss) is still the valid leader. This closes #13727. | 23 October 2020, 12:51:26 UTC |
| d0611c5 | Till Rohrmann | 21 October 2020, 10:44:28 UTC | [hotfix] Close TestingServer in ZooKeeperLeaderElectionConnectionHandlingTest | 23 October 2020, 12:51:25 UTC |
| 85bc836 | Dian Fu | 23 October 2020, 06:44:06 UTC | [FLINK-19675][python] Fix PythonCalcExpandProjectRule to handle cases when the calc node contains WHERE clause, composite fields access and Python UDF at the same time (#13746) | 23 October 2020, 06:44:06 UTC |
| 88c06e3 | Roman Khachatryan | 20 October 2020, 15:02:41 UTC | [FLINK-19401][checkpointing] Download checkpoints only if needed If checkpoints in ZK are the same as already downloaded then skip download. This prevents excessive S3 reads and eventually failure to start the job, particularly in multi-region recovery. ZooKeeperHighAvailabilityITCase is dropped is it now fails false-positively. Instead, unit testRetrievalFailsRecovery was added in the previous commit. | 22 October 2020, 08:10:52 UTC |
| 6790e79 | Roman Khachatryan | 21 October 2020, 09:29:33 UTC | [tests][checkpointing] Test retrieval error fails ZKStore.recover The added test, coupled with the next change (download checkpoints only if necessary) allows to drop ZK-HA-ITCase. Otherwise, ZK-HA-ITCase would fail false-positively. | 22 October 2020, 08:10:52 UTC |
| 9fe6c78 | Till Rohrmann | 20 October 2020, 11:47:38 UTC | [FLINK-19022] Fail TaskManagerRunner fatally if TaskExecutor terminates unexpectedly An unexpected termination of the TaskExecutor indicates that the system has reached an invalid state. Therefore, the runner will fail fatally if this happens. This closes #13705. | 20 October 2020, 21:02:51 UTC |
| 4ffc99c | Till Rohrmann | 20 October 2020, 08:45:32 UTC | [FLINK-19022] Fail DispatcherResourceManagerComponent fatally if ResourceManager terminates unexpectedly An unexpected termination of the ResourceManager indicates an invalid state. Hence, the DispatcherResourceManagerComponent should react to it by calling the FatalErrorHandler. | 20 October 2020, 12:53:29 UTC |
| 2c6bef2 | Till Rohrmann | 20 October 2020, 09:36:02 UTC | [FLINK-19022] Fail fatally if DispatcherGatewayService terminates unexpectedly in AbstractDispatcherLeaderProcess This commit ensures that we fail fatally if the DispatcherGatewayService termiantes unexpectedly in the AbstractDispatcherLeaderProcess. This ensures that the system won't stay in an invalid state. | 20 October 2020, 12:13:29 UTC |
| e0b4de2 | Till Rohrmann | 19 October 2020, 13:33:27 UTC | [FLINK-19022] Let TaskExecutor, Dispatcher, RM fail fatally if onStart throws a Throwable If TaskExecutor., Dispatcher. or ResourceManager.onStart throws a Throwable, the system is no longer in a valid state. Hence, this commit ensures that the components will properly call FatalErrorHandler.onFatalError in such a case. | 20 October 2020, 12:04:54 UTC |
| 26344b2 | Till Rohrmann | 19 October 2020, 13:29:20 UTC | [FLINK-19022][rpc] Log if RpcEndpoint failed its termination In order to not swallow the failure reason of failed RpcEndpoints, we are now logging the cause of a RpcEndpoint failure. | 20 October 2020, 12:04:54 UTC |
| b11918e | Till Rohrmann | 20 October 2020, 09:34:52 UTC | [hotfix] Correct TestingDispatcherGatewayService closeAsync behavior Let TestingDispatcherGatewayService complete the termination future when calling closeAsync. | 20 October 2020, 12:04:53 UTC |
| 1843cbe | Till Rohrmann | 19 October 2020, 14:55:48 UTC | [hotfix] Introduce ResourceManagerService to make DispatcherResourceManagerComponent easier to test By passing an interface instead of ResourceManager, it is easier to test the DispatcherResourceManagerComponent. | 20 October 2020, 12:04:53 UTC |
| 91829d3 | Till Rohrmann | 19 October 2020, 14:50:14 UTC | [hotfix] Let DispatcherResourceManagerComponent accept AutoCloseableAsync instead of WebMonitorEndpoint This commit improves the testability of DispatcherResourceManagerComponent by allowing to pass in a AutoCloseableAsync implementation. | 20 October 2020, 12:04:53 UTC |
| 58643ea | Till Rohrmann | 20 October 2020, 09:06:43 UTC | [hotfix] Add CoreMatchers.willNotComplete to assert that a future times out | 20 October 2020, 12:04:40 UTC |
| 072c78e | Till Rohrmann | 16 October 2020, 10:34:46 UTC | [FLINK-13553] Ignore KvStateServerHandlerTest because it is unstable | 19 October 2020, 11:00:36 UTC |
| c09e959 | Aljoscha Krettek | 08 June 2020, 16:56:11 UTC | [FLINK-19671] Update .editorconfig to conform to checkstyle and import ordering This puts in place an opinionated code style that matches our checkstyle/import ordering requirements. | 16 October 2020, 08:39:18 UTC |
| e262e7d | Aljoscha Krettek | 08 June 2020, 15:41:00 UTC | [FLINK-19671] Raze .editorconfig by resetting to IntelliJ default This creates an .editorconfig that has all the additional configuration options that IntelliJ provides but leaves them at the default. I'm doing this so that the next commit can change those defaults and we can thus track where we diverge from the defaults. | 16 October 2020, 08:39:12 UTC |
| ffd352e | wineandcheeze | 14 October 2020, 06:36:54 UTC | [FLINK-19523][conf] Hide sensitive command-line configurations | 14 October 2020, 06:39:20 UTC |
| fb34ce7 | Till Rohrmann | 09 October 2020, 11:01:48 UTC | [FLINK-17458] Stop TaskSubmissionTestEnvironment.testingRpcService when closing the environment This closes #13569. | 13 October 2020, 14:07:56 UTC |
| 0d78e80 | Matthias Pohl | 08 October 2020, 14:27:24 UTC | [FLINK-17341][runtime] Add TaskSlotTable.getActiveTaskAllocationIdsPerJob Introduced new method for retrieving AllocationIDs of active TaskSlots based on the JobID in a Set. This newly introduced method replaces getActiveSlots(JobID) returning an Iterator to avoid ConcurrentModificationExceptions. This closes #13564. | 09 October 2020, 16:06:26 UTC |
| b042a48 | Dawid Wysakowicz | 21 September 2020, 19:26:30 UTC | [FLINK-19339] Support unions with logical types in Avro >= 1.9.x Avro 1.9.x introduced yet another mechanism for registering/looking up conversions for logical types. If a logical type is part of a union a static field MODEL$ of type SpecificData is added to the generated Avro class with registered conversions for logical types. In this commit we try to use that SpecificData in AvroSerializer and AvroDeserializationSchema whenever available instead of instantiating a new unrelated one. The change is backwards compatible and if the field is not available we fallback to the old ways. | 24 September 2020, 06:05:09 UTC |
| 6c49a8e | Zhu Zhu | 18 September 2020, 07:01:33 UTC | [FLINK-FLINK][docs] Update the version in docs/_config.yml to be the latest released minor version | 19 September 2020, 07:30:53 UTC |
| a20fb4a | Roman Khachatryan | 27 July 2020, 21:07:44 UTC | [FLINK-15467][task] Wait for sourceTaskThread to finish before exiting from invoke. This closes #13383. | 16 September 2020, 19:20:40 UTC |
| 05ff71c | Arvid Heise | 11 September 2020, 17:19:30 UTC | [FLINK-19109][streaming] Disallow chaining of ContinuousFileReaderOperator. When CFRO is not actively reading from a split, its inside some poll loop where no mail is being processed. Hence, downstream operators cannot rely on processing timers among other things being reliably triggered. | 16 September 2020, 13:50:42 UTC |
| 4749e4c | Aljoscha Krettek | 03 September 2020, 16:52:11 UTC | [FLINK-19135] Strip ExecutionException in (Stream)ExecutionEnvironment.execute() In FLINK-14850 we changed the execute() method to be basically final JobClient jobClient = executeAsync(...); return jobClient.getJobExecutionResult(userClassloader).get(); Unfortunately, this means that execute() now throws an ExecutionException instead of a ProgramInvocationException or JobExecutionException as before. The ExecutionException is wrapping the other exceptions that we were throwing before. We didn't notice this in tests because most tests use Test(Stream)Environment which overrides the execute() method and so doesn't go through the PipelineExecutor logic or the normal code path of delegating to executeAsync(). This change brings us back to the previous behaviour. | 10 September 2020, 11:59:11 UTC |
| 85c1817 | Zhijiang | 07 September 2020, 03:55:48 UTC | [FLINK-18832][datastream] Add compatible check for blocking partition with buffer timeout From the requirement it is no need to enable buffer timeout for batch jobs since the downstream can only consume data when the upstream finishes. Furthermore the current implementation of BoundedBlockingSubpartition does not consider the concurrent issue from the flusher thread by enabling buffer timeout. So it is nice to check this compatibility during job graph generation in advance and give a friendly message hint for users. This closes #13209. | 08 September 2020, 08:07:06 UTC |
| dfdf4a1 | Zhijiang | 20 August 2020, 07:28:52 UTC | [hotfix][datastream] Fix the formatting of StreamEdge class | 08 September 2020, 08:07:06 UTC |
| 40f81b7 | liujiangang | 24 August 2020, 09:41:15 UTC | [FLINK-18959][Runtime] Try to revert MiniDispatcher for archiveExecutionGraph and shutdown cluster upon cancel. This closes #13227. | 04 September 2020, 07:39:55 UTC |
| 56bb246 | Yun Gao | 03 September 2020, 14:30:39 UTC | [FLINK-14087][datastream] Clone the StreamPartitioner to avoid being shared at runtime. | 03 September 2020, 14:38:56 UTC |
| 41fa48b | Jimmy.Zhou | 01 September 2020, 09:23:02 UTC | [FLINK-19105][docs] Fix documentation errors about FileSystem usage This closes #13297. | 03 September 2020, 09:17:05 UTC |
| 9794ba6 | Andrey Zagrebin | 27 August 2020, 15:54:28 UTC | [FLINK-19055] Wait less time for all memory GC in tests (MemoryManager#verifyEmpty) The waiting happens in a separate thread in production and does not disrupt anything. The problem is that tests use the same method which blocks the testing thread for too long. We do not need such big waiting time in tests. The PR refactors code to get back to the smaller waiting time in tests. This closes #13265. | 31 August 2020, 08:02:58 UTC |
| feaa832 | Andrey Zagrebin | 27 August 2020, 15:59:45 UTC | [hotfix] Add a production factory method for MemoryManager, use builder in tests | 31 August 2020, 08:02:58 UTC |
| ffcef94 | finchen | 26 August 2020, 16:03:03 UTC | [FLINK-18287][docs][python] Correct the documentation of Python Table API in SQL pages This closes #13253. | 27 August 2020, 01:46:15 UTC |
| 830438f | Zhu Zhu | 25 August 2020, 02:27:31 UTC | Update japicmp configuration for 1.10.2 | 25 August 2020, 02:27:31 UTC |
| 13c4b42 | Jark Wu | 19 August 2020, 02:07:22 UTC | [FLINK-18212][table-planner-blink] Fix Lookup Join failed when there is a UDF equal condition on the column of temporal table This closes #13183 | 19 August 2020, 02:07:22 UTC |
| 7a08234 | Matthias Pohl | 18 August 2020, 06:30:28 UTC | [hotfix][docs] Replaced outdated 'RocksDBStateBackend.setOptions(..)' by 'RocksDBStateBackend.setRocksDBOptions(..)' and fixed typo in parameter list (PptionsFactory -> RocksDBOptionsFactory). This closes #13185. | 18 August 2020, 07:11:16 UTC |
| a568d93 | Yangze Guo | 17 August 2020, 03:04:46 UTC | [FLINK-18081][doc] Fix broken links in Kerberos Authentication Setup and Configuration | 17 August 2020, 06:01:41 UTC |
| 69c5553 | Kezhu Wang | 12 August 2020, 03:35:05 UTC | [FLINK-18815] Close safety net guarded closeable iff it is still registered This closes #13124. | 13 August 2020, 08:59:59 UTC |
| dd81861 | Kezhu Wang | 12 August 2020, 03:26:16 UTC | [FLINK-18815] Change Thread.sleep(2) to Thread.sleep(0) to fail SafetyNetCloseableRegistryTest#testClose more often | 13 August 2020, 08:59:59 UTC |
| 3e17833 | Andrey Zagrebin | 21 July 2020, 16:17:40 UTC | [FLINK-18646] Verify memory manager empty in a separate thread with larger timeout UnsafeMemoryBudget#verifyEmpty, called on slot freeing, needs time to wait on GC of all allocated/released managed memory. If there are a lot of segments to GC then it can take time to finish the check. If slot freeing happens in RPC thread, the GC waiting can block it and TM risks to miss its heartbeat. Another problem is that after UnsafeMemoryBudget#RETRIGGER_GC_AFTER_SLEEPS, System.gc() is called for each attempt to run a cleaner even if there are already detected cleaners to run. This leads to triggering a lot of unnecessary GCs in background. The PR offloads the verification into a separate thread and calls System.gc() only if memory cannot be reserved and there are still no cleaners to run after long waiting. The timeout for normal memory reservation is increased to 2 second. The full reservation, used for verification, gets 2 minute timeout. This closes #12980. | 13 August 2020, 08:18:41 UTC |
| bced52d | Till Rohrmann | 12 August 2020, 14:55:43 UTC | [hotfix] Remove mocking from RestServerEndpointITCase | 13 August 2020, 08:17:43 UTC |
| fd7b371 | Till Rohrmann | 12 August 2020, 14:50:52 UTC | [FLINK-18902][rest] Allow request serving while the REST handlers are shutting down In order to serve the job execution result when using the per job mode, the REST handlers must be able to serve requests while the RestServerEndpoint is being shut down. Otherwise, it is not possible to serve the asynchronous operation results such as the job execution result. This commit solves this problem by checking whether the InFlightRequestTracker allows to enqueue a new request or not in the AbstractHandler. If it rejects a new request, this means that the handler has been shut completely shut down and that we should close the connection to the client. If the InFlightRequestTracker still accepts requests, then this means that the RestHandler might still want to serve an asynchronous result. This closes #13133. | 13 August 2020, 08:17:36 UTC |
| e581d7f | Roman Khachatryan | 10 August 2020, 13:44:22 UTC | [FLINK-18821][network] Fix indefinite wait in PartitionRequestClientFactory.createPartitionRequestClient | 13 August 2020, 03:33:35 UTC |
| 6690317 | Roman Khachatryan | 10 August 2020, 13:41:23 UTC | [hotfix][tests] Unignore PartitionRequestClientFactoryTest class | 13 August 2020, 03:33:35 UTC |
| 3317d40 | Zhijiang | 19 June 2020, 03:01:38 UTC | [hotfix][build] Adjust the file permission for update_japicmp_configuration.sh | 11 August 2020, 06:45:51 UTC |
| fafee5e | Zhu Zhu | 10 August 2020, 08:56:13 UTC | [FLINK-18859][tests] Increase timeout of ExecutionGraphNotEnoughResourceTest#testRestartWithSlotSharingAndNotEnoughResources to make it more stable | 10 August 2020, 14:54:43 UTC |
| 5696cf2 | Matthias Pohl | 07 August 2020, 14:01:36 UTC | [FLINK-17503][runtime] [logs] Refactored log output. The log output was cleaned up: - Only the key of the option is logged out instead of the whole instance's toString() method. - A new utility method was introduce that adapts the log output in a way that an adapted extension is used for maximum values instead of logging the actual max value. This closes #13086. | 08 August 2020, 10:12:24 UTC |
| 8c13486 | Maximilian Michels | 31 July 2020, 15:46:04 UTC | [FLINK-16510] Allow configuring shutdown behavior to avoid JVM freeze This adds the configuration option `cluster.processes.halt-on-fatal-error` which defaults to `false`. If set to `true`, a custom SecurityManager will be installed on top of the existing SecurityManager to exit forcefully via `Runtime#halt`. Exiting that way may be necessary due to the Java 8 JVM freezing during a graceful shutdown when using the G1 garbage collector. This closes #13079. | 07 August 2020, 10:29:26 UTC |
| b717877 | Max Kuklinski | 23 January 2020, 15:53:39 UTC | [FLINK-13689] [Connectors/ElasticSearch] Fix thread leak in Elasticsearch connector when cluster is down | 07 August 2020, 06:47:13 UTC |
| d954278 | xcomp | 30 July 2020, 19:04:01 UTC | [FLINK-18677][fix] Added handling of suspended or lost connections within the ZooKeeperLeaderRetrievalService. The listener needs to be notified in case of a connection loss so that it is able to initiate necessary actions on its side. [FLINK-18677][runtime] [style] Replaced spaces by TABs to follow the Apache Flink code styles. [FLINK-18677][runtime] [fix] Synchronize notifyLeaderLoss through lock and removed redundant code. The redundant code was moved into notifyIfNewLeaderAddress(String, UUID) which is then used by notifyLeaderLoss() and within nodeChanged(). Additionally, the method call of notifyLeaderLoss() is guarded now by a lock to synchronize the state change (i.e. lastLeaderAddress and lastLeaderSessionID). [FLINK-18677][runtime] The exception was added to make it more explicit that the method is not expected to be called. [FLINK-18677][runtime] Decreased wait time the queue to be filled since we're not expecting any objects. The test does not expect any calls happening. Hence, no CompletableFuture instance will be queued. The longer wait time would just result in a longer running test. [FLINK-18677][runtime] Added infinite wait time to happy test. The previous implementation had a fixed timeout. Slower machines might need longer to process the test which might result in test failures. The new implementation removes the timeout so that the test wouldn't fail just because of a poor performance of the machine the test is running on. [FLINK-18677][runtime] Moved log messages out of synchronization blocks. This closes #13055. | 05 August 2020, 09:38:28 UTC |
| 52a77e6 | Tartarus0zm | 29 July 2020, 12:08:14 UTC | [FLINK-18663][rest] Exit early if shutdown has started | 03 August 2020, 20:21:30 UTC |
| bda1383 | Tartarus0zm | 29 July 2020, 10:56:28 UTC | [FLINK-18663][rest] Improve exception handling - ensure that request finalization runs even if handleException throws an exception - catch NPE in handleException, which occurs if the client closes the connection | 03 August 2020, 20:21:25 UTC |
| 7627f54 | Chesnay Schepler | 03 August 2020, 08:49:46 UTC | [FLINK-16619][coordination] Log reception of slot reports only once | 03 August 2020, 20:15:34 UTC |
| eb0a56b | Till Rohrmann | 30 July 2020, 12:14:24 UTC | [FLINK-18772] Disable web submission for per-job/application mode deployments When running Flink in per-job/application mode, it will instantiate a MiniDispatcherRestEndpoint. This endpoint does not instantiate the web submission REST handlers. However, it still displayed the submit job link in the web ui. This commit changes the behaviour so that we no longer display this link when running Flink in per-job/application mode. This closes #13030. | 03 August 2020, 13:50:41 UTC |
| 47ea234 | wangyang0918 | 27 July 2020, 09:48:09 UTC | [hotfix][k8s] Correct dependencies in Kubernetes pom and notice file | 29 July 2020, 12:06:04 UTC |
| 183dfc3 | wangyang0918 | 27 July 2020, 09:04:12 UTC | [FLINK-18517][k8s] Bump fabric8 version from 4.5.2 to 4.9.2 This closes #12995. | 29 July 2020, 12:05:58 UTC |
| e268680 | Zhijiang | 16 July 2020, 08:22:25 UTC | [FLINK-18595][network] Fix the deadlock issue by task thread and canceler thread in RemoteInputChannel Assuming two remote channels as listeners in LocalBufferPool, the deadlock happens as follows 1. While the Canceler thread calling ch1#releaseAllResources, it will occupy the bufferQueue lock and try to call ch2#notifyBufferAvailable. 2. While task thread exiting to call CachedBufferStorage#close, it might release exclusive buffers for ch2. Then ch2 will occupy the bufferQueue lock and try to call ch1#notifyBufferAvailable. 3. ch1 and ch2 will both occupy self bufferQueue lock and wait for other side's bufferQueue lock to cause deadlock. Regarding the solution, we can check the released state outside of bufferQueue lock in RemoteInputChannel#notifyBufferAvailable to return immediately. | 28 July 2020, 06:19:51 UTC |
| bfe6c2e | Andrey Zagrebin | 18 June 2020, 09:04:06 UTC | [FLINK-16225] Implement user class loading exception handler This closes #12563. | 22 July 2020, 12:06:28 UTC |
| 1442300 | Gao Yun | 17 March 2020, 07:20:39 UTC | [FLINK-15962][network] Reduce the default chunk size to 4M in netty stack | 22 July 2020, 11:58:56 UTC |
| c1f8763 | Rui Li | 21 July 2020, 06:49:28 UTC | [FLINK-18644][doc][hive] Remove obsolete hive connector docs This closes #12939 | 21 July 2020, 06:49:28 UTC |
| 2a3b642 | Jark Wu | 14 July 2020, 01:58:18 UTC | [FLINK-18539][datastream] Fix StreamExecutionEnvironment#addSource(SourceFunction, TypeInformation) doesn't use the user defined type information This closes #12878 | 14 July 2020, 01:58:18 UTC |
| cf047a4 | Aljoscha Krettek | 09 July 2020, 08:02:45 UTC | [FLINK-18478] Use AvroFactory.extractAvroSpecificSchema in AvroDeserializationSchema This makes AvroFactory and the used method public where they were package private before. This fixes the problem that AvroDeserializationSchema was not working with types generated from avrohugger. It could also just be seen as refactoring/code cleanup. | 09 July 2020, 14:48:07 UTC |
| d64a5a0 | Milan Nikl | 09 June 2020, 15:09:32 UTC | [FLINK-18097][history] Delete all job-related files on expiration | 08 July 2020, 19:10:10 UTC |
| 494e8f8 | Fabian Paul | 06 July 2020, 14:16:25 UTC | [FLINK-18501] Use inner class name for logging scheme mapping when filesystems are instantiated | 08 July 2020, 09:51:53 UTC |
| b9f6ab5 | Piotr Nowojski | 07 July 2020, 12:30:20 UTC | [hotfix][docs] Mark 1.10 as outdated | 07 July 2020, 12:59:51 UTC |
| a519b04 | Leonard Xu | 02 July 2020, 03:44:26 UTC | [FLINK-16309][docs] ElasticSearch 7 connector is missing in SQL connector list This closes #12801 | 02 July 2020, 03:44:26 UTC |
| 294d07a | Yubin Li | 29 June 2020, 08:46:54 UTC | [hotfix][docs] Fix Python example in the documentation about event_time This closes #12789. | 30 June 2020, 06:55:31 UTC |
| d40bd0b | Shuiqiang Chen | 29 June 2020, 11:57:53 UTC | [FLINK-17920][python][docs] Add the Python example of the Interval Join (#12779) | 29 June 2020, 12:10:08 UTC |
| 9b45486 | Yu Li | 16 June 2020, 06:01:29 UTC | [FLINK-17800][rocksdb] Support customized RocksDB write/read options and use RocksDBResourceContainer to get them | 26 June 2020, 16:15:35 UTC |
| 3f8649e | Yun Tang | 22 June 2020, 08:27:02 UTC | [hot-fix][rocksdb] Ensure RocksDBKeyedStateBackend disposed at RocksDBStateMisuseOptionTest This closes #12736. | 26 June 2020, 16:15:35 UTC |
| de6f3aa | Yun Tang | 04 June 2020, 17:16:27 UTC | [FLINK-17800][rocksdb] Ensure total order seek to avoid user misuse | 26 June 2020, 16:15:22 UTC |
| 23ad7e3 | zoudan | 09 June 2020, 08:12:11 UTC | [FLINK-18168][table-runtime-blink] Fix array reuse for BinaryArrayData in converters This closes #12542. | 25 June 2020, 07:08:27 UTC |
| b50704b | GuoWei Ma | 22 June 2020, 09:10:52 UTC | [FLINK-17639] Document FileSystems supported by the StreamingFileSink | 24 June 2020, 12:47:39 UTC |
| dcd7574 | Kostas Kloudas | 22 June 2020, 14:05:01 UTC | [FLINK-18352] Make DefaultClusterClientServiceLoader/DefaultExecutorServiceLoader thread-safe | 22 June 2020, 14:05:01 UTC |
| fd91aff | Yu Li | 11 June 2020, 11:32:39 UTC | [FLINK-18242][state-backend-rocksdb] Separate RocksDBOptionsFactory from OptionsFactory This closes #12673. | 20 June 2020, 09:00:04 UTC |
| 831daaa | Yu Li | 20 June 2020, 08:59:06 UTC | Revert "[FLINK-17800][roksdb] Ensure total order seek to avoid user misuse" This reverts commit b8ddbef9a5cc5dc769ba61bd5019dd96843c932f. | 20 June 2020, 08:59:06 UTC |
| f87937b | Yu Li | 20 June 2020, 08:58:59 UTC | Revert "[FLINK-17800][roksdb] Support customized RocksDB write/read options and use RocksDBResourceContainer to get them" This reverts commit 88c22864504d772764c5838afe0b944f1da50a3a. | 20 June 2020, 08:58:59 UTC |
| b1fdf6d | Chesnay Schepler | 16 June 2020, 13:13:58 UTC | [FLINK-18329][legal] Fix typo | 17 June 2020, 09:04:27 UTC |
| f7a4128 | vanderliang | 23 March 2020, 09:36:30 UTC | [FLINK-16717][k8s] Use headless service for rpc and blob port This closes #11488. | 16 June 2020, 13:04:24 UTC |
| 88c2286 | Yu Li | 16 June 2020, 06:01:29 UTC | [FLINK-17800][roksdb] Support customized RocksDB write/read options and use RocksDBResourceContainer to get them | 16 June 2020, 11:30:04 UTC |
