5de3094 | Aljoscha Krettek | 23 November 2017, 17:11:01 UTC | Commit for release 1.4.0 | 23 November 2017, 17:11:01 UTC |
3b58038 | Aljoscha Krettek | 20 November 2017, 17:06:11 UTC | [hotfix] Always explicitly set hadoop.version in create_binary_release Before, the "hadoop2" profile would create a binary release for whatever happens to be the default hadoop.version. | 23 November 2017, 17:09:26 UTC |
13631b9 | twalthr | 23 November 2017, 13:30:15 UTC | [FLINK-8118] [table] Improve KafkaTableSource documentation | 23 November 2017, 14:34:22 UTC |
2fb2458 | Xingcan Cui | 22 November 2017, 16:00:39 UTC | [FLINK-8118] [table] Allow to specify the offsets of KafkaTableSources This closes #5056. | 23 November 2017, 14:34:22 UTC |
828ef09 | Stefan Richter | 22 November 2017, 16:52:35 UTC | [FLINK-5465] [streaming] Wait for pending timer threads to finish or to exceed a time limit in exceptional stream task shutdown. This closes #5058. (cherry picked from commit d86c6b6) | 23 November 2017, 14:25:01 UTC |
e100861 | Piotr Nowojski | 23 November 2017, 12:08:43 UTC | [hotfix][docs] Improve Kafka exactly-once docs | 23 November 2017, 14:02:49 UTC |
62bf001 | Piotr Nowojski | 22 November 2017, 14:55:20 UTC | [hotfix][kafka] Remove unused method in kafka tests | 23 November 2017, 13:45:16 UTC |
27564c3 | Piotr Nowojski | 22 November 2017, 14:53:08 UTC | [FLINK-8132][kafka] Re-initialize transactional KafkaProducer on each checkpoint Previously faulty scenario with producer pool of 2. 1. started transaction 1 with producerA, written record 42 2. checkpoint 1 triggered, pre committing txn1, started txn2 with producerB, written record 43 3. checkpoint 1 completed, committing txn1, returning producerA to the pool 4. checkpoint 2 triggered , committing txn2, started txn3 with producerA, written record 44 5. crash.... 6. recover to checkpoint 1, txn1 from producerA found to "pendingCommitTransactions", attempting to recoverAndCommit(txn1) 7. unfortunately txn1 and txn3 from the same producers are identical from KafkaBroker perspective and thus txn3 is being committed result is that both records 42 and 44 are committed. With this fix, after re-initialization txn3 will have different producerId/epoch counters compared to txn1. | 23 November 2017, 13:45:07 UTC |
736b908 | Piotr Nowojski | 22 November 2017, 10:37:48 UTC | [hotfix][kafka] Throw FlinkKafkaProducer011Exception with error codes instead of generic Exception | 23 November 2017, 13:45:00 UTC |
8a052bf | Bowen Li | 10 October 2017, 05:31:17 UTC | [FLINK-6505] Proactively cleanup local FS for RocksDBKeyedStateBackend on startup | 23 November 2017, 12:51:41 UTC |
35517f1 | Fabian Hueske | 13 November 2017, 13:54:54 UTC | [FLINK-2170] [connectors] Add OrcRowInputFormat and OrcTableSource. This closes #5043. | 22 November 2017, 23:35:58 UTC |
7868ea4 | uybhatti | 03 March 2017, 22:55:22 UTC | [FLINK-2170] [connectors] Add OrcRowInputFormat and OrcTableSource. This closes #4670. | 22 November 2017, 22:12:40 UTC |
2815796 | twalthr | 22 November 2017, 16:43:08 UTC | [FLINK-8136] [table] Fix code generation with JodaTime shading This closes #5054. | 22 November 2017, 17:20:19 UTC |
6cb8b5b | Aljoscha Krettek | 22 November 2017, 15:22:28 UTC | [hotfix] Make aws docs version agnostic | 22 November 2017, 15:23:07 UTC |
a964ef6 | Cristian | 22 November 2017, 12:17:39 UTC | [hotfix][docs] Fix typo in Trigger doc This closes #5051. | 22 November 2017, 14:47:37 UTC |
7ed2cef | Aljoscha Krettek | 22 November 2017, 11:00:09 UTC | [FLINK-8131] Update to Kafka 0.11.0.2 | 22 November 2017, 13:32:08 UTC |
56c78cb | zentol | 14 November 2017, 10:55:33 UTC | [FLINK-8070][yarn][tests] Print errors found in log files This closes #5012. | 22 November 2017, 10:56:41 UTC |
9d28619 | Nico Kruber | 17 November 2017, 13:13:30 UTC | [FLINK-7841] [docs] Update AWS docs with respect to S3 file system changes This closes #5029 | 22 November 2017, 10:41:16 UTC |
f26edb8 | Gabor Gevay | 20 November 2017, 14:12:17 UTC | [FLINK-8117] [runtime] Eliminate modulo operation from round-robin partitioners This closes #5041 | 22 November 2017, 09:09:36 UTC |
5f523e6 | Gabor Gevay | 20 November 2017, 15:51:43 UTC | [hotfix][docs] Fix some typos in the documentation. This closes #5039. | 21 November 2017, 13:43:54 UTC |
ddbc015 | Bowen Li | 04 November 2017, 04:08:24 UTC | [FLINK-7977][build] Bump version of compatibility check for Flink 1.4 This closes #4945. | 20 November 2017, 16:29:30 UTC |
314087e | Aljoscha Krettek | 20 November 2017, 15:32:06 UTC | [hotfix] Fix create_release_branch.sh to use correct branch name | 20 November 2017, 15:32:06 UTC |
3d4146c | Aljoscha Krettek | 20 November 2017, 15:09:10 UTC | [FLINK-8115] Fix Kafka download link in end-to-end test | 20 November 2017, 15:10:09 UTC |
28b3115 | Joerg Schad | 18 November 2017, 06:23:02 UTC | [FLINK-8102][docs] Fixed formatting issues in Mesos documentation. | 20 November 2017, 14:57:23 UTC |
f001264 | yew1eb | 10 October 2017, 18:52:35 UTC | [hotfix][license] Add missing licenses This close #4794. | 20 November 2017, 14:51:04 UTC |
2b4f1b1 | Till Rohrmann | 30 October 2017, 13:15:20 UTC | [FLINK-7943] Make ParameterTool thread safe This commit changes the serialization of the ParameterTool such that only the data map is contained. The defaultData and the unrequestedParameters maps are not serialized because they are only used on the client side. Additionally, the defaultData and unrequestedParameters map are being made thread safe by using ConcurrentHashMaps. This closes #4921. | 20 November 2017, 14:51:03 UTC |
2570328 | zentol | 20 November 2017, 12:58:21 UTC | [FLINK-8110][dist] Relocate jackson services in flink-dist | 20 November 2017, 14:47:51 UTC |
ee4420f | zentol | 20 November 2017, 14:07:32 UTC | [FLINK-8114][py] Fix forwarding of arguments | 20 November 2017, 14:44:12 UTC |
497c36f | zentol | 20 November 2017, 11:58:27 UTC | [FLINK-8109][py] Check for existence of plan/additional files | 20 November 2017, 14:44:08 UTC |
8eadda3 | zentol | 20 November 2017, 11:57:43 UTC | [FLINK-8108][py] Fix bounds check | 20 November 2017, 14:44:03 UTC |
d0ea9d9 | Aljoscha Krettek | 17 November 2017, 16:19:51 UTC | [FLINK-8099] Reduce default restart delay to 1 second | 20 November 2017, 13:12:44 UTC |
2f3f8c7 | Piotr Nowojski | 17 November 2017, 13:40:30 UTC | [FLINK-8086][kafka] Ignore ProducerFencedException during recovery ProducerFencedException can happen if we restore twice from the same checkpoint or if we restore from an old savepoint. In both cases transactional.ids that we want to recoverAndCommit have been already committed and reused. Reusing mean that they will be known by Kafka's brokers under newer producerId/epochId, which will result in ProducerFencedException if we try to commit again some old (and already committed) transaction. Ignoring this exception might hide some bugs/issues, because instead of failing we might have a semi silent (with a warning) data loss. | 20 November 2017, 12:26:15 UTC |
1a6121a | Piotr Nowojski | 17 November 2017, 13:31:07 UTC | [hotfix][kafka] Improve logging in FlinkKafkaProducer011 | 20 November 2017, 12:26:08 UTC |
3fe70d7 | Xpray | 17 November 2017, 03:01:27 UTC | [FLINK-8095] [table] Introduce ProjectSetOpTransposeRule This closes #5026. | 20 November 2017, 11:50:40 UTC |
4d55987 | Dian Fu | 17 November 2017, 02:53:31 UTC | [FLINK-8096] [table] Fix time attribute materialization when writing to TableSink This closes #5025. | 20 November 2017, 11:18:26 UTC |
abdc7d1 | Nico Kruber | 02 November 2017, 18:38:48 UTC | [FLINK-7988][s3] fix HadoopS3FileSystemITCase leaving test directories behind in S3 This closes #4950. | 18 November 2017, 09:58:53 UTC |
36b8075 | Nico Kruber | 09 November 2016, 20:04:50 UTC | [FLINK-4228][yarn/s3] fix for yarn staging with s3a defaultFs + includes a new unit tests for recursive uploads to hfds:// targets + add a unit test for recursive file uploads to s3:// via s3a [FLINK-4228][yarn/s3] turn the dependencies around Instead of having flink-s3-fs-hadoop depend on flink-yarn_<scala_version>, let flink-yarn depend on the s3 filesystem and implement the test there. This is safer with regards to the scala-independent flink-s3-fs-hadoop project. [FLINK-4228][yarn] change the S3 upload tests to use Hadoop's S3 implementations This is how YARN would use it and what should really be tested. [FLINK-4228][yarn] enable S3 tests for newer Hadoop versions - requires the 'include_hadoop_aws' build profile (or property) to be set - requires a newer aws-sdk version (than Hadoop pulls in) to work with our httpcomponents version - we also add a check that at least one S3 implementation is tested to not silently ignore all tests because of such a missing dependency This closes #4939. | 18 November 2017, 09:58:53 UTC |
666b1b2 | Stephan Ewen | 25 July 2017, 15:26:38 UTC | [FLINK-7266] [core] Prevent attempt for parent directory deletion for object stores This closes #4397 | 17 November 2017, 16:22:24 UTC |
a0dbe18 | Stephan Ewen | 25 July 2017, 15:19:25 UTC | [FLINK-7265] [core] Introduce FileSystemKind to differentiate between FileSystem and ObjectStore | 17 November 2017, 16:22:24 UTC |
e784f3a | Vetriselvan1187 | 13 November 2017, 16:54:43 UTC | [FLINK-8061][QS] Remove trailing * in QSClient javadocs. | 17 November 2017, 14:31:15 UTC |
3753ae2 | kkloudas | 17 November 2017, 08:26:10 UTC | [FLINK-8057][QS] Change error message in KvStateRegistry.registerKvState(). | 17 November 2017, 10:21:18 UTC |
1a68d75 | kkloudas | 16 November 2017, 16:45:49 UTC | [FLINK-8059][QS] QS client throws FlinkJobNotFoundException for queries with unknown jobIds. | 17 November 2017, 10:20:55 UTC |
12b0c58 | kkloudas | 16 November 2017, 16:02:16 UTC | [FLINK-8055][QS] Deduplicate logging messages about QS start. | 17 November 2017, 10:19:35 UTC |
6314e48 | kkloudas | 15 November 2017, 14:38:36 UTC | [FLINK-8065][QS] Improve error message when client already shut down. | 17 November 2017, 10:19:08 UTC |
96b350a | kkloudas | 15 November 2017, 14:32:42 UTC | [FLINK-8062][QS] Make getKvState() with namespace private. | 17 November 2017, 10:18:47 UTC |
d0324e3 | kkloudas | 14 November 2017, 14:05:45 UTC | [FLINK-8063][QS] QS client does not retry when an UnknownKvStateLocation is thrown. | 17 November 2017, 09:37:18 UTC |
42e2441 | Xpray | 06 November 2017, 15:47:33 UTC | [FLINK-7986] [table] Introduce FilterSetOpTransposeRule This closes #4956. | 16 November 2017, 15:16:22 UTC |
e7f7d0c | Xingcan Cui | 15 November 2017, 03:01:13 UTC | [FLINK-8069] [table] Add preserving WatermarkStrategy. This closes #5016. | 16 November 2017, 10:47:16 UTC |
f14fcef | Fabian Hueske | 09 November 2017, 14:07:17 UTC | [FLINK-8016] [docs] Add documentation for KafkaJsonTableSinks. This closes #4990. | 16 November 2017, 10:46:04 UTC |
2fd5311 | Fabian Hueske | 07 November 2017, 16:59:43 UTC | [FLINK-8014] [table] Add Kafka010JsonTableSink. - Refactor KafkaTableSink tests. | 16 November 2017, 10:45:54 UTC |
8883fa2 | twalthr | 16 November 2017, 09:20:16 UTC | [FLINK-7389] [table] Remove Calcite PushProjector This closes #5022. | 16 November 2017, 10:45:46 UTC |
b59fae3 | Shuyi Chen | 09 November 2017, 08:05:20 UTC | [FLINK-7003] [table] Fix 'SELECT *' for tables with nested schema. This closes #4989. | 16 November 2017, 10:45:38 UTC |
397f0d1 | twalthr | 16 November 2017, 10:00:30 UTC | [FLINK-7698] [table] Tests joins with null literals | 16 November 2017, 10:06:14 UTC |
13962e1 | twalthr | 15 November 2017, 11:07:16 UTC | [FLINK-7942] [table] Reduce aliasing in RexNodes This closes #5019. | 15 November 2017, 17:29:25 UTC |
084ff68 | Fabian Hueske | 15 November 2017, 09:12:55 UTC | [FLINK-7490] [table] Use correct classloader to compile generated code that calls UDAGGs. This closes #5018. | 15 November 2017, 16:55:25 UTC |
06a922e | twalthr | 26 September 2017, 10:10:33 UTC | [FLINK-7678] [table] Support composite inputs for user-defined functions This closes #4726. | 15 November 2017, 16:34:36 UTC |
32bfc38 | twalthr | 14 November 2017, 10:06:54 UTC | [FLINK-8013] [table] Support aggregate functions with generic arrays This closes #5011. | 15 November 2017, 15:19:17 UTC |
49aeb8f | twalthr | 15 November 2017, 13:39:21 UTC | [FLINK-7451] [table] Disable testing of the charset in TableTestBase | 15 November 2017, 13:50:01 UTC |
81e3a88 | Jark Wu | 15 August 2017, 11:24:57 UTC | [FLINK-7451] [table] Support non-ascii character literals in Table API and SQL This closes #4544. | 15 November 2017, 13:49:52 UTC |
d6d35fa | Nico Kruber | 14 November 2017, 13:36:22 UTC | [FLINK-7973] disable JNI bridge for relocated hadoop classes in s3-fs-* | 15 November 2017, 09:45:31 UTC |
5c6eaab | zentol | 08 November 2017, 10:58:38 UTC | [FLINK-7419][build][avro] Shade jackson dependency in flink-dist This closes #4981. | 14 November 2017, 20:05:13 UTC |
195e3da | zentol | 14 November 2017, 13:47:00 UTC | [FLINK-8071][build] Bump shade-plugin asm version to 5.1 This closes #5014. | 14 November 2017, 20:05:06 UTC |
2774335 | Vetriselvan1187 | 06 November 2017, 17:11:41 UTC | [FLINK-7998][examples] Fix TPCHQuery3 examples This closes #4959. | 14 November 2017, 14:57:25 UTC |
7c7f24e | zentol | 14 November 2017, 09:58:43 UTC | [FLINK-8056][dist] Use 'web.port' instead of 'jobmanager.web.port' This closes #5010. | 14 November 2017, 14:44:05 UTC |
fcc79c0 | Alejandro Alcalde | 07 November 2017, 12:09:46 UTC | [FLINK-8006] [Startup Shell Scripts] - Fixing $pid This closes #4968. | 14 November 2017, 14:15:15 UTC |
8d8c52f | zentol | 07 November 2017, 16:13:57 UTC | [FLINK-8011][dist] Set flink-python to provided This closes #4973. | 14 November 2017, 14:15:15 UTC |
a4b9996 | gyao | 10 November 2017, 13:54:32 UTC | [hotfix][docs] Fix broken link to FLINK-7811 This closes #4995. | 14 November 2017, 14:15:15 UTC |
24970a9 | Michael Fong | 13 November 2017, 02:17:02 UTC | [FLINK-4500][docs] Update cassandra documentation regarding data loss As of FLINK-4500 the Cassandra connector will wait for pending updates to finish upon checkpoint. This closes #5002. | 14 November 2017, 14:15:14 UTC |
c2d3d6a | yew1eb | 11 November 2017, 13:08:21 UTC | [hotfix][docs] Fix typos in deployment AWS documentation This closes #5000. | 14 November 2017, 14:15:14 UTC |
cf099f1 | yew1eb | 11 November 2017, 12:45:42 UTC | [hotfix][docs][javadocs] Remove double "of" This closes #4999. | 14 November 2017, 14:15:14 UTC |
6f9ab72 | Piotr Nowojski | 13 November 2017, 15:51:18 UTC | [FLINK-7845][runtime] Make NettyMessage public This a walkaround strange javaassist bug. The issue should go away once we upgrade netty dependency. Please check the ticket for more information. This closes #5007. | 14 November 2017, 14:15:14 UTC |
8b7698d | Stefan Richter | 13 November 2017, 13:31:45 UTC | [FLINK-8053] [checkpoints] Default to asynchronous snapshots for FsStateBackend and MemoryStateBackend. (cherry picked from commit 2906698) | 14 November 2017, 10:54:10 UTC |
8f2d0fa | Nico Kruber | 10 November 2017, 15:42:10 UTC | [hotfix] let end-to-end tests check for empty .out files again | 13 November 2017, 16:41:15 UTC |
97a3491 | Nico Kruber | 10 November 2017, 13:04:29 UTC | [hotfix] ignore a warning from the error check of the S3 e2e tests | 13 November 2017, 16:41:15 UTC |
3574f8b | Nico Kruber | 10 November 2017, 12:16:40 UTC | [hotfix] fix presto end-to-end test not cleaning up | 13 November 2017, 16:41:15 UTC |
e666e62 | Aljoscha Krettek | 08 November 2017, 17:57:22 UTC | [hotfix] Make end-to-end test scripts more robust This uses traps to ensure that we properly do cleanups, remove config values and shutdown things. | 13 November 2017, 16:41:15 UTC |
9f68212 | Aljoscha Krettek | 08 November 2017, 16:49:17 UTC | [FLINK-7973] Add shaded S3 FileSystem end-to-end tests | 13 November 2017, 16:41:15 UTC |
25a28ab | Nico Kruber | 06 November 2017, 18:53:37 UTC | [FLINK-7973] Fix shading and relocating Hadoop for the S3 filesystems - do not shade everything, especially not JDK classes! -> instead define include patterns explicitly - do not shade core Flink classes (only those imported from flink-hadoop-fs) - hack around Hadoop loading (unshaded/non-relocated) classes based on names in the core-default.xml by overwriting the Configuration class (we may need to extend this for the mapred-default.xml and hdfs-defaults.xml): -> provide a core-default-shaded.xml file with shaded class names and copy and adapt the Configuration class of the respective Hadoop version to load this file instead of core-default.xml. Add checkstyle suppression pattern for the Hadoop Configuration classes Also fix the (integration) tests not working because they tried to load the relocated classes which are apparently not available there Remove minimizeJar from shading of flink-s3-fs-presto because this was causing "java.lang.ClassNotFoundException: org.apache.flink.fs.s3presto.shaded.org.apache.commons.logging.impl.LogFactoryImpl" since these classes are not statically imported and thus removed when minimizing. Fix s3-fs-presto not shading org.HdrHistogram Fix log4j being relocated in the S3 fs implementations Add shading checks to travis | 13 November 2017, 16:41:15 UTC |
ce1cb8f | twalthr | 13 November 2017, 13:09:45 UTC | [FLINK-7657] [table] Add all basic types to RexProgramExtractor | 13 November 2017, 13:22:27 UTC |
bb04187 | Kent Murra | 27 September 2017, 20:48:55 UTC | [FLINK-7657] [table] Add time types FilterableTableSource push down This closes #4746. | 13 November 2017, 13:22:16 UTC |
e2b92f2 | Stefan Richter | 13 November 2017, 10:50:07 UTC | [FLINK-8040] [tests] Fix test instability in ResourceGuardTest (cherry picked from commit ad8ef6d) | 13 November 2017, 12:50:46 UTC |
431ae36 | Aljoscha Krettek | 10 November 2017, 17:13:26 UTC | [FLINK-7702] Add maven-bundle-plugin to root pom Before, we had it in places that require it. This doesn't work when running mvn javadoc:aggregate because this will only run for the root pom and can then not find the "bundle" dependencies. | 13 November 2017, 11:05:03 UTC |
6aeac3f | Aljoscha Krettek | 10 November 2017, 15:28:46 UTC | [FLINK-7702] Remove Javadoc aggregation for Scala code genjavadoc generated some Java code that was making Javadoc fail. | 13 November 2017, 11:04:54 UTC |
b3df579 | Piotr Nowojski | 10 November 2017, 14:15:11 UTC | [hotfix][docs] Change mailing list link in quickstart to flink-user Previously it was pointing to flink-dev | 10 November 2017, 16:43:47 UTC |
da435f1 | Aljoscha Krettek | 10 November 2017, 09:54:16 UTC | [FLINK-6163] Document per-window state in ProcessWindowFunction | 10 November 2017, 14:56:13 UTC |
5f992e8 | Piotr Nowojski | 10 November 2017, 12:57:51 UTC | [hotfix][build] Disable dependency convergence in flink-dist Previously mvn javadoc:aggregate goal was failing | 10 November 2017, 13:26:46 UTC |
896f13d | Michael Fong | 14 August 2017, 12:57:06 UTC | [FLINK-4500] CassandraSinkBase implements CheckpointedFunction This closes #4605. | 10 November 2017, 10:10:13 UTC |
2117eb7 | gyao | 08 November 2017, 10:46:45 UTC | [FLINK-8005] Set user-code class loader as context loader before snapshot During checkpointing, user code may dynamically load classes from the user code jar. This is a problem if the thread invoking the snapshot callbacks does not have the user code class loader set as its context class loader. This commit makes sure that the correct class loader is set. | 10 November 2017, 08:26:37 UTC |
005a871 | Piotr Nowojski | 07 November 2017, 11:13:59 UTC | [FLINK-7765][build] Enable dependency convergence by default Disable it in most modules. | 09 November 2017, 14:45:09 UTC |
7df7fc4 | Piotr Nowojski | 05 October 2017, 13:17:13 UTC | [hotfix][build] Deduplicate maven-enforcer version | 09 November 2017, 14:44:57 UTC |
d302c65 | Aljoscha Krettek | 09 November 2017, 14:34:44 UTC | [hotfix] Fix formatting in windowing documentation | 09 November 2017, 14:35:13 UTC |
02a19a1 | Dan Kelley | 08 November 2017, 01:27:44 UTC | [FLINK-8017] Fix High availability cluster-id key in documentation | 09 November 2017, 13:08:41 UTC |
c794329 | Fabian Hueske | 06 November 2017, 20:22:35 UTC | [FLINK-8002] [table] Fix join window boundary for LESS_THAN and GREATER_THAN predicates. This closes #4962. | 08 November 2017, 17:44:06 UTC |
51657fc | Fabian Hueske | 07 November 2017, 16:57:39 UTC | [FLINK-8012] [table] Fix TableSink config for tables with time attributes. This closes #4974. | 08 November 2017, 17:43:57 UTC |
1b20f70 | Xingcan Cui | 07 November 2017, 17:17:57 UTC | [FLINK-7996] [table] Add support for (left.time = right.time) predicates to window join. This closes #4977. | 08 November 2017, 17:43:51 UTC |
8c60f97 | Rong Rong | 31 October 2017, 18:05:38 UTC | [FLINK-7922] [table] Fix FlinkTypeFactory.leastRestrictive for composite types. This closes #4929. | 08 November 2017, 17:43:39 UTC |
a126bd3 | Xpray | 03 November 2017, 07:19:42 UTC | [FLINK-7971] [table] Fix potential NPE in non-windowed aggregation. This closes #4941. | 08 November 2017, 17:43:31 UTC |
f5a0b4b | zentol | 07 November 2017, 15:58:53 UTC | [FLINK-8009][build][runtime] Remove transitive dependency promotion This closes #4972. | 08 November 2017, 10:22:15 UTC |
17aae5a | Tzu-Li (Gordon) Tai | 07 November 2017, 11:35:33 UTC | [FLINK-8001] [kafka] Prevent PeriodicWatermarkEmitter from violating IDLE status Prior to this commit, a bug exists such that if a Kafka consumer subtask initially marks itself as idle because it didn't have any partitions to subscribe to, that idleness status will be violated when the PeriodicWatermarkEmitter is fired. The problem is that the PeriodicWatermarkEmitter incorrecty yields a Long.MAX_VALUE watermark even when there are no partitions to subscribe to. This commit fixes this by additionally ensuring that the aggregated watermark in the PeriodicWatermarkEmitterr is an effective one (i.e., is really aggregated from some partition). | 08 November 2017, 09:23:27 UTC |
49dc380 | zentol | 07 November 2017, 14:36:49 UTC | [FLINK-8010][build] Bump remaining flink-shaded versions | 08 November 2017, 09:22:22 UTC |