| c9d2c90 | hequn8128 | 24 January 2020, 00:44:30 UTC | Commit for release 1.9.2 | 24 January 2020, 00:44:30 UTC |
| 0bdd21a | Zhu Zhu | 16 January 2020, 04:15:10 UTC | [FLINK-14701][runtime] Fix MultiTaskSlot to not remove slots which are not its children | 23 January 2020, 14:05:53 UTC |
| 772d83e | Benoît Paris | 21 January 2020, 10:59:49 UTC | [FLINK-15726] [Blink Planner] [hotfix] Fix error message in StreamExecTableSourceScan & BatchExecTableSourceScan This closes #10914 | 22 January 2020, 12:17:21 UTC |
| 6061372 | Kostas Kloudas | 22 January 2020, 07:59:12 UTC | [hotfix] Fix error from cherry-picking 0ea00eabaa0eca | 22 January 2020, 08:16:52 UTC |
| 0ea00ea | wangyang0918 | 15 January 2020, 10:36:38 UTC | [FLINK-13758][client] Support to register DFS files as distributed cache All the Flink Standalone, Yarn, Mesos, Kubernetes session clusters are using RestClusterClient#submitJob to submit a job to an existing session. Before this commit, the Flink client will hang when trying to register DFS artifacts as distributed cache for session cluster. | 21 January 2020, 15:53:52 UTC |
| 90ad2e6 | 曹建华 | 21 January 2020, 08:56:23 UTC | [FLINK-15549] Fix Integer overflow in ResettableIterator. (#10903) The ResettableIterator had a data overflow problem if the number of elements in a single input exceeded Integer.MAX_VALUE. | 21 January 2020, 09:21:21 UTC |
| d44848c | huangxingbo | 21 January 2020, 03:04:46 UTC | [FLINK-15700][python][doc] Improves the Python API Tutorial doc This closes #10911. | 21 January 2020, 07:17:08 UTC |
| c9f9c54 | Dian Fu | 29 October 2019, 03:18:45 UTC | [FLINK-14557][python] Clean up the py4j package by removing the unused directory __MACOSX. This closes #10047. | 21 January 2020, 04:49:45 UTC |
| 578a709 | Benoit Hanotte | 13 January 2020, 09:48:25 UTC | [FLINK-15577][table-planner] Add Window specs to WindowAggregate nodes' digests The RelNode's digest is used by the Calcite HepPlanner to avoid adding duplicate vertices to the graph. If an equivalent vertex was already present in the graph, then that vertex is used in place of the newly generated one. This means that the digest needs to contain all the information necessary to identifying a vertex and distinguishing it from similar - but not equivalent - vertices. In the case of the `WindowAggregation` nodes, the window specs are currently not in the digest, meaning that two aggregations with the same signatures and expressions but different windows are considered equivalent by the planner, which is not correct and will lead to an invalid Physical Plan. This commit fixes this issue and adds a test ensuring that the window specs are in the digest, as well as similar aggregations on two different windows will not be considered equivalent. This closes #10854 (cherry picked from commit 244718553742c086eefc95f927d7b26af597d40a) | 20 January 2020, 07:17:48 UTC |
| a2b211a | huangxingbo | 19 January 2020, 06:57:50 UTC | [FLINK-15657][python][doc] Fix the python table api doc link | 19 January 2020, 09:58:30 UTC |
| b95ba46 | Chesnay Schepler | 17 January 2020, 11:05:09 UTC | [hotfix][build] Remove unused depMgmt entry | 17 January 2020, 11:05:09 UTC |
| ddf1c21 | Roman | 17 January 2020, 10:09:19 UTC | [FLINK-15615][docs] Fix reported file sink consistency guarantees | 17 January 2020, 10:13:41 UTC |
| 6a124f8 | vthinkxie | 13 January 2020, 11:19:09 UTC | [FLINK-15489][web] Avoid caching of JM and TM logs in the web ui This adds the 'Cache-Control: no-cache' HTTP header when requesting TM or TM logs. | 16 January 2020, 11:41:51 UTC |
| 05eca9a | Chesnay Schepler | 15 January 2020, 10:58:50 UTC | [FLINK-15554][azure] Bump jetty-util to 3.1.2 | 15 January 2020, 11:00:07 UTC |
| b935c52 | Zhu Zhu | 10 January 2020, 08:33:02 UTC | [hotfix][runtime] ExecutionGraph failure cause cannot be null if job is FAILED | 13 January 2020, 08:34:42 UTC |
| c0974b8 | Zhu Zhu | 10 January 2020, 08:30:40 UTC | [FLINK-15522][runtime] JobResult takes ExecutionGraph failure cause as its failure cause only if job is FAILED | 13 January 2020, 08:34:32 UTC |
| c33f0d2 | Chesnay Schepler | 09 January 2020, 10:26:29 UTC | [FLINK-15523][conf] Japicmp checks ConfigConstants | 12 January 2020, 20:46:10 UTC |
| 6dfbf33 | Chesnay Schepler | 09 January 2020, 15:18:10 UTC | [FLINK-15543][legal] Remove Apache Camel NOTICE entries | 10 January 2020, 11:55:21 UTC |
| 24f5c5c | Seth Wiesman | 09 December 2019, 21:48:02 UTC | [FLINK-14825][state-processor-api][docs] Rework state processor api documentation | 08 January 2020, 17:24:36 UTC |
| 7a82b38 | Paul Lin | 08 January 2020, 02:22:40 UTC | [FLINK-11135][configuration] Reorder Hadoop config loading in HadoopUtils This closes #7314 . | 08 January 2020, 02:24:30 UTC |
| 632a6db | Jark Wu | 07 January 2020, 09:48:46 UTC | [FLINK-14200][table] Fix NPE for Temporal Table Function Join when left side is a query instead of a source (#10782) | 07 January 2020, 09:48:46 UTC |
| c016086 | Benchao Li | 07 January 2020, 07:13:35 UTC | [FLINK-15478][table-planner-blink] Fix FROM_BASE64 code generation exception because of the wrong result type (#10766) | 07 January 2020, 07:15:55 UTC |
| bcf5400 | Rui Li | 02 January 2020, 12:13:40 UTC | [FLINK-15259][hive] HiveInspector.toInspectors() should convert Flink constant to Hive constant this closes #10747. | 05 January 2020, 22:49:00 UTC |
| 55044d3 | Jingsong Lee | 02 January 2020, 05:35:46 UTC | [FLINK-15443][jdbc] Fix mismatch between Java float and JDBC float (#10731) | 02 January 2020, 08:03:00 UTC |
| ecd4e42 | Zhenghua Gao | 31 December 2019, 06:27:39 UTC | [FLINK-15421][table-planner-blink] Fix cast exception for timestamp MAX/MIN retract aggregate functions This closes #10723 | 31 December 2019, 06:27:39 UTC |
| fd76c99 | libenchao | 27 December 2019, 12:38:14 UTC | [FLINK-15418][table-planner-blink] Set FlinkRelDistribution in StreamExecMatchRule This closes #10706 | 30 December 2019, 14:32:19 UTC |
| 2079f8d | libenchao | 29 December 2019, 10:34:05 UTC | [FLINK-15435][python] Fix unstable test for ExecutionConfigTests.test_equals_and_hash The default parallelism used by local environments is calculated according to the available processors. We should specify the parallelism explicitly to avoid test unstable. This closes #10715. | 29 December 2019, 15:16:26 UTC |
| 025aca8 | Jark Wu | 28 December 2019, 03:50:12 UTC | [FLINK-15413][table-planner-blink] Fix ScalarOperatorsTest failed in travis (#10708) | 28 December 2019, 03:50:12 UTC |
| baeb9b3 | Rui Li | 27 December 2019, 07:48:24 UTC | [FLINK-15412][hive] LocalExecutorITCase#testParameterizedTypes failed in travis this closes #10702. | 27 December 2019, 18:03:44 UTC |
| 85e0f15 | JingsongLi | 23 December 2019, 07:30:47 UTC | [FLINK-15361][parquet] ParquetTableSource should pass predicate in projectFields fix the problem, when after projectFields, ParquetTableSource will loose predicates. this closes #10660. | 23 December 2019, 17:49:00 UTC |
| ea5f241 | Yun Tang | 13 December 2019, 08:38:45 UTC | [FLINK-15065][docs] Correct default value of RocksDB options in documentation This refer to https://github.com/facebook/rocksdb/pull/6123 which correctis RocksDB javadoc | 20 December 2019, 13:23:22 UTC |
| c221872 | Yun Tang | 24 November 2019, 18:00:40 UTC | [FLINK-14846][doc] Correct the default writerbuffer size documentation of RocksDB Correct the default write buffer size of RocksDB to '64MB' | 20 December 2019, 13:13:21 UTC |
| 73c82c3 | Benchao Li | 20 December 2019, 06:47:54 UTC | [FLINK-15266][table-planner-blink] Fix NPE for code generated CASE WHEN operator in blink planner (#10594) | 20 December 2019, 06:50:19 UTC |
| ff6b350 | Kostas Kloudas | 19 December 2019, 18:35:53 UTC | [FLINK-14683] Fix RemoteStreamEnvironment's constructor | 19 December 2019, 18:39:09 UTC |
| 4e07b8c | bowen.li | 17 December 2019, 23:31:07 UTC | [FLINK-15240][hive] is_generic key is missing for Flink table stored in HiveCatalog backport from 1.10 | 17 December 2019, 23:31:53 UTC |
| cca2387 | bowen | 17 December 2019, 23:10:46 UTC | [FLINK-15234][hive] hive table created from flink catalog table cannot have null properties in parameters backport from 1.10 | 17 December 2019, 23:31:53 UTC |
| 0f701a0 | Till Rohrmann | 13 December 2019, 17:43:06 UTC | [hotfix] Set initial bestCandidateScore in LocationPreferenceSlotSelectionStrategy to Double.NEGATIVE_INFINITY | 17 December 2019, 15:36:45 UTC |
| 850e807 | Till Rohrmann | 13 December 2019, 14:41:33 UTC | [hotfix] Remove explicit capacity from HashMap creation in SlotSharingManager | 17 December 2019, 15:36:45 UTC |
| 02755b9 | Till Rohrmann | 12 December 2019, 18:22:58 UTC | [FLINK-15013] Fix selection of NON_LOCAL slots in LocationPreferenceSlotSelectionStrategy The LocationPreferenceSlotSelectionStrategy ignored NON_LOCAL slots due to initializing the initial candidate score to a positive value. As NON_LOCAL candidates have a value of 0, the initial candidate score needs to be negative. This closes #10555. | 17 December 2019, 15:36:45 UTC |
| 28a4453 | Till Rohrmann | 13 December 2019, 14:56:40 UTC | [hotfix] Add locality assertions to LocationPreferenceSlotSelectionStrategyTest tests | 17 December 2019, 15:36:44 UTC |
| 10cfe8e | Till Rohrmann | 13 December 2019, 14:51:55 UTC | [hotfix] Refactor LocationPreferenceSlotSelectionStrategyTest to use Hamcrest matchers | 17 December 2019, 15:36:44 UTC |
| d3b33c0 | Till Rohrmann | 12 December 2019, 17:38:29 UTC | [hotfix] Introduce SlotSharingManager#createAndRegisterRootSlot | 17 December 2019, 15:36:44 UTC |
| 119b4e7 | Till Rohrmann | 12 December 2019, 11:09:32 UTC | [FLINK-15013] Mark root as resolved before completing MultiTaskSlot In order to prevent a race condition where Executions are scheduled because their inputs' locations have been assigned but the underlying root slot not being marked as resolved and, hence, not being available for location based scheduling, this commit enforces that we first resolve the root slot before completing the associated MultiTaskSlot. | 17 December 2019, 15:36:44 UTC |
| 1caada5 | Till Rohrmann | 11 December 2019, 15:59:41 UTC | [hotfix] Simplify SlotSharingManagerTest by introducing factory method | 17 December 2019, 15:36:44 UTC |
| 9fc5627 | Till Rohrmann | 03 May 2019, 16:07:48 UTC | [FLINK-14251] Add FutureUtils#forward utility The forward function completes the second future with the result of the first future. This closes #9786. | 17 December 2019, 15:36:44 UTC |
| 48371c1 | Till Rohrmann | 17 December 2019, 15:35:27 UTC | [hotfix] Only release parent slot if there are oversubscribed children to release | 17 December 2019, 15:36:43 UTC |
| db531e7 | Arvid Heise | 13 December 2019, 14:26:38 UTC | [FLINK-13589] Fixing DelimitedInputFormat for whole file input splits. The DelimitedInputFormat drops bytes when using whole file input splits. This commit replicates the logic of regular input splits also for whole file input splits. | 17 December 2019, 02:15:21 UTC |
| 4d0633a | bowen.li | 15 December 2019, 06:53:11 UTC | [hotfix][doc][hive] rename page from 'Hive' to 'Hive Integration' | 15 December 2019, 06:55:14 UTC |
| 1325727 | kevin.cyj | 03 December 2019, 09:55:49 UTC | [FLINK-15030][runtime] Fix deadlock problem of blocking ResultPartition with minimum required buffers. Currently, BoundedBlockingSubpartition relies on the add of next BufferConsumer to flush and recycle the previous one, which means at least (numberOfSubpartitions + 1) buffers is needed to make the bounded blocking ResultPartition work. However, ResultPartitionFactory gives only (numberOfSubpartitions) required buffers to BufferPool of bounded blocking ResultPartition which may lead to deadlock problem. This commit fixes the problem by increasing the number of required buffers by 1. | 14 December 2019, 16:45:21 UTC |
| 9ea0493 | godfreyhe | 02 December 2019, 07:05:34 UTC | [FLINK-15001] [table-planner-blink] The digest of sub-plan reuse should contain retraction traits for stream physical nodes This closes #10561. | 14 December 2019, 03:26:25 UTC |
| f702cfb | Arvid Heise | 05 December 2019, 10:26:29 UTC | [FLINK-14574][core/e2e] Wrap FileSystem to use plugin context classloader. Filesystems loaded through PluginManager are wrapped in PluginFileSystem, which swaps the context classloader for all direct methods. Thus, lazily loaded classes (security, optimizations) are loaded with the expected classloader. In particular, java.util.ServiceLoader will use that context classloader to load new services. This change will fix the s3-hadoop plugin while writing. Also added minio-based s3 batch tests that write to s3. Modified existing s3 batch tests to also write to s3. Modified existing s3 streaming test to checkpoint onto s3. | 13 December 2019, 12:56:40 UTC |
| 9219b23 | Arvid Heise | 05 December 2019, 10:12:58 UTC | [FLINK-14574][e2e] Providing minio based s3 setup allowing users without AWS secrets to test plugins. | 13 December 2019, 12:56:40 UTC |
| d8fec10 | Arvid Heise | 05 December 2019, 10:11:41 UTC | [FLINK-14574][e2e] Replacing s3Util with dockerized aws cli enabling customized endpoints and other features that would needed to be manually implemented. | 13 December 2019, 12:56:40 UTC |
| 7c585f4 | Arvid Heise | 29 November 2019, 09:41:53 UTC | [hotfix][e2e] Removing unused s3 methods | 13 December 2019, 12:56:40 UTC |
| 5f0720e | Arvid Heise | 06 December 2019, 08:21:27 UTC | [hotfix][e2e] Cleaning logs before executing single tests. Logs of successful tests are cleaned up automatically. However, logs of failing tests remain to ease debugging. However, if the developer does not clean them before executing a new test with exception checks, the new test will fail independent of the actual result as the old logs will cause the checks to fail. With this fix, failing tests will not cause subsequent tests to fail because of parsing old logs. | 13 December 2019, 12:56:40 UTC |
| 8a73d68 | Yangze Guo | 10 December 2019, 02:03:49 UTC | [FLINK-14951][tests] Harden the thread safety of State TTL backend tests | 11 December 2019, 09:18:44 UTC |
| f3d892c | Piotr Nowojski | 10 December 2019, 18:13:39 UTC | fixup! [FLINK-15076][task] Fix SourceStreamTask cancellation | 10 December 2019, 18:25:39 UTC |
| 7f797e0 | Piotr Nowojski | 06 December 2019, 19:51:48 UTC | [FLINK-15076][task] Fix SourceStreamTask cancellation Source thread should be interrupted more or less the same way how task thread is being interrupted. This is important for example as in the scenario presented in the SourceStreamTaskTest#testCancellationWithSourceBlockedOnLock(). SourceFunction is blocked while holding checkpointLock, which might prevent task thread from cancelling properly if the SourceFunction is not interrupted. | 10 December 2019, 12:44:42 UTC |
| f65be9b | hpeter | 19 November 2019, 00:57:40 UTC | [FLINK-14953][formats] use table type to build parquet FilterPredicate This closes #10371 (cherry picked from commit b5d958d7bfabbc3ef7e75efdcf61be43ad4044be) | 10 December 2019, 06:27:03 UTC |
| 86d5183 | Rong Rong | 06 December 2019, 23:39:23 UTC | [FLINK-11120][table] Fix timestampadd on non-zero, negative interaval that cross over day boundary, refine documentation. This closes #8746. (cherry picked from commit 9320f344f37019b9b389ef06477c8049e9fa3218) | 07 December 2019, 16:51:11 UTC |
| 3b420f6 | XuQianJin-Stars | 15 June 2019, 08:10:42 UTC | [FLINK-11120][table] Fix timestampadd operation not returning correct result. (cherry picked from commit f5dafd3dda029e1529f8ba823777aec627707b97) | 07 December 2019, 16:51:11 UTC |
| bfb4e41 | Zhu Zhu | 15 November 2019, 06:58:31 UTC | [FLINK-14735][scheduler] Improve scheduling of all-to-all partitions with ALL input constraint for legacy scheduler This closes #10283 | 07 December 2019, 13:51:43 UTC |
| 1b49092 | Nico Kruber | 06 December 2019, 16:02:56 UTC | [FLINK-15113][config] add fs.azure.account.key to list of sensitive configs | 06 December 2019, 17:53:16 UTC |
| b9ce2a1 | Chesnay Schepler | 06 December 2019, 16:19:32 UTC | [FLINK-13862][docs] Update execution plans docs | 06 December 2019, 16:20:19 UTC |
| e4ad3b0 | Chesnay Schepler | 13 November 2019, 15:20:36 UTC | [FLINK-13945][build][docs] Fix instructions for building against custom hadoop versions | 06 December 2019, 15:16:50 UTC |
| 9a95489 | Hwanju Kim | 06 December 2019, 13:34:18 UTC | [FLINK-14949][Runtime/Task] Task cancellation can be stuck against out-of-thread error (#10387) This commit added an additional safety net against fatal error thrown from starting task cancellation by notifying the fatal error to taskmanager. Without it, if a fatal or out-of-memory error happens while initiating task cancellation, one or more critical threads that are responsible for cancellation either gracefully or forcefully may not be spawned, thereby job state machine being permanently stuck in a non-terminal state such as FAILING (by stuck task cancellation). The fatal error notification added by this patch can restart JVM cleaning up the state letting job state machine make progress. NOTE: Instead of catching fatal/OOM errors from each thread start, added a catch at a higher level by internalizing cancelOrFailAndCancelInvokable, as it would make sense to notify fatal error no matter where such error comes from during the function. In addition, doing so made unit testing easier by mocking. | 06 December 2019, 13:37:51 UTC |
| 1be72a8 | Lining Jing | 05 December 2019, 04:49:21 UTC | [FLINK-15063][metric]fix input group and output group of the task metric are reversed | 06 December 2019, 08:55:04 UTC |
| d9f8abb | Till Rohrmann | 04 December 2019, 14:23:08 UTC | [hotfix] Let YarnResourceManagerTest run asynchronous tasks in main thread | 04 December 2019, 21:16:24 UTC |
| b308de2 | Till Rohrmann | 04 December 2019, 11:11:37 UTC | [hotfix] Correct log statements in YarnResourceManager | 04 December 2019, 21:16:24 UTC |
| 0071401 | wangyang0918 | 03 December 2019, 14:09:03 UTC | [FLINK-15036][yarn] Container startup error should be handled inside of the YarnResourceManager's main thread | 04 December 2019, 21:16:24 UTC |
| c388c26 | Jark Wu | 04 December 2019, 07:23:18 UTC | [FLINK-14960][e2e] Fix failed dependency shading test of table modules | 04 December 2019, 18:08:54 UTC |
| cfef075 | Dawid Wysakowicz | 04 December 2019, 09:15:18 UTC | [FLINK-14910][datastream] Checking for auto generated uids only for PhysicalStreamTransformations This closes #10417 | 04 December 2019, 18:04:43 UTC |
| 399ec11 | Chesnay Schepler | 29 November 2019, 15:50:05 UTC | [FLINK-14995][kinesis][legal] Update NOTICE | 01 December 2019, 16:47:37 UTC |
| 6039e11 | Mads Chr. Olesen | 27 November 2019, 13:56:16 UTC | [FLINK-14976][cassandra] Release semaphore on all Throwable's in send() | 29 November 2019, 18:53:02 UTC |
| cf82a14 | Stefan Richter | 28 November 2019, 16:21:52 UTC | [FLINK-10377] Support checkpoint overtaking a savepoint in TwoPhaseCommitSink The precondition checkState(pendingTransactionIterator.hasNext(), "checkpoint completed, but no transaction pending"); in TwoPhaseCommitSinkFunction.notifyCheckpointComplete() seems too strict, because checkpoints can overtake checkpoints and will fail the precondition. In this case the commit was already performed by the first notification and subsumes the late checkpoint. I think the check can be removed. This can happen in the following scenario: # savepoint is triggered # checkpoint is triggered # checkpoint completes (but it doesn't subsume the savepoint, because checkpoints subsume only other checkpoints). # savepoint completes | 28 November 2019, 16:23:06 UTC |
| e7f540c | chunpinghe | 27 November 2019, 11:01:38 UTC | [FLINK-11835][tests] Wait until job was recovered before unblocking task | 27 November 2019, 11:03:39 UTC |
| 1663c21 | Konstantin Knauf | 22 November 2019, 21:00:32 UTC | [FLINK-14930][fs][oss] Document credential providers | 27 November 2019, 10:45:48 UTC |
| 49cc9fe | Konstantin Knauf | 22 November 2019, 20:12:20 UTC | [FLINK-14930][fs][oss] Fix shading prefix The OSS filesystem only supports OSS credential providers, which use the "com.aliyun" relocation and not the hadoop one. | 27 November 2019, 10:45:47 UTC |
| c1f2294 | Victor Wong | 26 November 2019, 15:04:13 UTC | [FLINK-14817][doc] Fix misleading documentation using method chaining of Configuration (#10323) | 26 November 2019, 15:38:54 UTC |
| 7a15e8b | Chesnay Schepler | 24 November 2019, 17:21:12 UTC | [FLINK-14104][build] Use flink-shaded-jackson 9.0 | 26 November 2019, 14:06:51 UTC |
| dbf2171 | Chesnay Schepler | 24 November 2019, 21:14:40 UTC | [hotfix][build] Remove redundant shaded-jackson version | 26 November 2019, 14:06:51 UTC |
| 53c1c18 | Chesnay Schepler | 14 November 2019, 13:51:19 UTC | [FLINK-13995][legal] Properly exclude netty license directory | 26 November 2019, 10:49:15 UTC |
| 2dd3ba8 | godfrey he | 25 November 2019, 02:43:25 UTC | [FLINK-13708] [table-planner-blink] transformations should be cleared after execution in blink planner (#10290) | 25 November 2019, 02:43:25 UTC |
| 43ac587 | Chesnay Schepler | 13 November 2019, 10:40:04 UTC | [FLINK-14104][build] Add opt-in jackson 2.10.1 profile | 21 November 2019, 15:34:32 UTC |
| 4ed1965 | Chesnay Schepler | 14 November 2019, 11:06:39 UTC | [FLINK-14104][build] Bump shade-plugin version to 3.1.1 Earlier versions fail with jackson 2.10.1, likely due to java9+ features present in the jar. | 21 November 2019, 15:34:32 UTC |
| 1c97e1d | Zhu Zhu | 20 November 2019, 14:11:36 UTC | [FLINK-14641][docs] Fix description of metric `fullRestarts` (#10128) Pointing out that this metric also includes fine grained restarts in 1.9.2 and later versions. | 20 November 2019, 14:12:52 UTC |
| 7d04551 | danrtsey.wy | 15 November 2019, 07:01:34 UTC | [FLINK-14382][yarn] Fixes the Flink plugin mechanism on yarn cluster by shipping only the plugin directory instead of adding it to the classpath. And add class isolation check in dummy FS. | 20 November 2019, 14:05:32 UTC |
| ce10d75 | danrtsey.wy | 08 November 2019, 09:35:51 UTC | [FLINK-14382][e2e] Update e2e tests to support two inputs from two dummy FS. | 20 November 2019, 14:05:32 UTC |
| 7785460 | danrtsey.wy | 08 November 2019, 09:30:45 UTC | [hotfix] Refactor test_yarn_kerberos_docker.sh by moving common functions to common_yarn_docker.sh. | 20 November 2019, 14:05:32 UTC |
| 673a23c | danrtsey.wy | 11 November 2019, 14:42:51 UTC | [FLINK-14382][e2e] Add another dummy fs for e2e test. It is same as dummy fs with different schema. | 20 November 2019, 14:05:32 UTC |
| 908175d | danrtsey.wy | 11 November 2019, 14:40:30 UTC | [FLINK-14382][examples] Adding multiple --input support to WordCount | 20 November 2019, 14:05:32 UTC |
| 2cecd2b | danrtsey.wy | 11 November 2019, 14:31:24 UTC | [FLINK-14382][java] Add MultipleParameterTool to support parsing multiple parameters of args. For example, --multi multiValue1 --multi multiValue2. | 20 November 2019, 14:05:32 UTC |
| 3c08499 | danrtsey.wy | 10 November 2019, 07:30:45 UTC | [FLINK-13184][yarn] Use dynamic properties instead of uploading taskmanager-conf.yaml to hdfs. This will reduce the time cost of launching a TaskExecutor so that YarnResourceManager could start a large number of TaskExecutors timely. | 20 November 2019, 05:25:04 UTC |
| 1f1f9a3 | danrtsey.wy | 10 November 2019, 07:19:32 UTC | [FLINK-13184][yarn] Use NMClientAsync instead of NMClient to avoid starting TaskExecutor blocking call. The start container requests will be executed in a thread pool of NMClientAsync. | 20 November 2019, 03:22:40 UTC |
| 8fb54e5 | Dian Fu | 14 November 2019, 01:35:58 UTC | [FLINK-14066][python] Support to install pyflink in windows This closes #10188. | 20 November 2019, 01:24:35 UTC |
| 9fe6259 | 林小铂 | 18 November 2019, 09:16:43 UTC | [FLINK-13747][client] Add integration tests for classloading in program execution | 19 November 2019, 09:09:44 UTC |
| 0f30c26 | 林小铂 | 18 November 2019, 09:15:15 UTC | [FLINK-13747][client] Make client respect classloading policy | 19 November 2019, 09:09:44 UTC |
| 3b97ec9 | Leonard Xu | 14 November 2019, 15:19:30 UTC | [FLINK-14784][table] CsvTableSink miss delimiter when row start with null member. This closes #10199 (cherry picked from commit 65ecfab8d5a3f9ef7877ea35ff22fb72ec2db0a5) | 15 November 2019, 07:14:42 UTC |
| 4e86b3e | Nico Kruber | 19 August 2019, 21:48:57 UTC | [FLINK-13791][docs] Speed up sidenav by using group_by _includes/sidenav.html parses through pages_by_language over and over again trying to find children when building the (recursive) side navigation. By doing this once with a group_by, we can gain considerable savings in building the docs via `./build_docs.sh` without any change to the generated HTML pages: This closes #9487 | 14 November 2019, 13:22:58 UTC |
| 345abdf | Nico Kruber | 14 August 2019, 21:05:00 UTC | [FLINK-13726][docs] Build docs with jekyll 4.0.0.pre.beta1 This significantly reduces the build times, on my machine from 140s to 47s! This closes #9444 | 14 November 2019, 13:22:57 UTC |
