| c41c8e5 | Fabian Paul | 10 November 2022, 09:39:02 UTC | Commit for release 1.15.3 | 10 November 2022, 09:39:02 UTC |
| 7b50b3a | Ahmed Hamdy | 02 November 2022, 18:21:00 UTC | [FLINK-29827][Connectors/Base] Registering callback for asyncWriter on restore from state. | 10 November 2022, 08:35:22 UTC |
| a59ca7f | Chesnay Schepler | 10 November 2022, 08:32:43 UTC | [hotfix] Add docs/themes/connectors to .gitignore Ignore a directory that is generated in 1.16+. | 10 November 2022, 08:32:43 UTC |
| 6feaa44 | Chesnay Schepler | 08 November 2022, 10:19:06 UTC | [FLINK-29927][akka] Use singleton ExtensionId | 09 November 2022, 09:40:11 UTC |
| b379995 | Yufan Sheng | 08 November 2022, 06:55:50 UTC | [FLINK-29495][Connector/Pulsar] Bump the Pulsar to latest 2.10.2 (#21250) | 08 November 2022, 06:55:50 UTC |
| b68f37c | Ryan Skraba | 21 October 2022, 15:31:27 UTC | [FLINK-29198][test] Fail after maximum RetryOnException Co-authored-by: Sergey Nuyanzin <sergey.nuyanzin@aiven.io> | 04 November 2022, 13:22:46 UTC |
| 17bc1a0 | Elphas Toringepi | 18 October 2022, 15:33:05 UTC | [FLINK-26890][Connector/Kinesis] Handle invalid shards in DynamoDB streams consumer | 04 November 2022, 12:33:01 UTC |
| 2a99acd | Qingsheng Ren | 04 November 2022, 06:58:49 UTC | [FLINK-28975][connector/base] Add IdlenessManager for main and per-split output in source operator to mark idleness only if both outputs are idle (#20831) | 04 November 2022, 06:58:49 UTC |
| 221e9f7 | Qingsheng Ren | 14 October 2022, 08:32:52 UTC | [FLINK-29567][connector/common] Change numRecordsSend / numBytesSend / numRecordsSendError back to numRecordsOut / numBytesOut / numRecordsOutError in sink | 03 November 2022, 02:24:01 UTC |
| e45e5c3 | Chesnay Schepler | 31 October 2022, 08:18:15 UTC | [FLINK-29803][table] Populate Scala API source jars | 01 November 2022, 09:23:39 UTC |
| d9413d6 | Carlos Castro | 07 October 2022, 22:45:21 UTC | [FLINK-29539][k8s] Make dnsPolicy of Flink pod configurable via pod template This closes #21178. | 28 October 2022, 09:54:25 UTC |
| 1d29f54 | Jacky Lau | 28 October 2022, 01:49:52 UTC | [FLINK-29749][client] Make 'flink info' command could support dynamic properties This closes #21166. | 28 October 2022, 01:49:52 UTC |
| f55061a | jingge | 27 October 2022, 11:20:00 UTC | [hotfix][doc] `FlinkKafkaConsumer` is deprecated and will be removed with Flink 1.17 instead of 1.15 since Flink 1.16 has code freeze and will be released soon. | 27 October 2022, 13:02:49 UTC |
| fd9dbc2 | Jacky Lau | 27 October 2022, 09:35:17 UTC | [FLINK-27579][client] Make the client.timeout and parallelism.default could take effect in CLI frontend This closes #21168. | 27 October 2022, 09:35:17 UTC |
| f84d039 | huangxingbo | 26 October 2022, 06:23:28 UTC | [FLINK-29479][python][hotfix] Fix the testPythonSystemEnvEnabled in PythonOptionsTest This closes #21157. | 26 October 2022, 08:26:07 UTC |
| eeda260 | Yuxin Tan | 21 October 2022, 09:40:42 UTC | [FLINK-29134][metrics] Do not repeatedly add useless metric updating tasks to avoid wasting resources This closes #21132. | 26 October 2022, 02:46:24 UTC |
| 723f6a2 | Yuxin Tan | 23 October 2022, 09:25:30 UTC | [hotfix][tests] Migrate MetricFetcherTest to JUnit5 | 26 October 2022, 02:46:23 UTC |
| 2fa7114 | huangxingbo | 26 October 2022, 02:15:49 UTC | [hotfix][python] Fix the compile error in PythonOptionsTest | 26 October 2022, 02:15:49 UTC |
| 91ccde9 | liuyongvs | 21 October 2022, 10:07:52 UTC | [FLINK-29479][python] Fix system env cause conflict with users python depdendency This closes #21110. | 25 October 2022, 11:30:18 UTC |
| 47bcb5a | Krzysztof Chmielewski | 14 October 2022, 14:05:08 UTC | [FLINK-29627][streaming] - Fix duplicate key exception during recover more than 1 committable. Recovery more than one Committable causes `IllegalStateException` and prevents cluster to start. When we recover the `CheckpointCommittableManager` we deserialize SubtaskCommittableManager instances from recovery state, and we put them into `Map<Integer, SubtaskCommittableManager<CommT>>`. The key of this map is subtaskId of the recovered manager. However, this will fail if we have to recover more than one committable. What was implemented as a fix is to call `SubtaskCommittableManager::merge` if we already deserialize manager for this subtaskId. Signed-off-by: Krzysztof Chmielewski <krzysiek.chmielewski@gmail.com> | 20 October 2022, 09:03:12 UTC |
| c500e97 | Sergey Nuyanzin | 13 October 2022, 20:08:30 UTC | [FLINK-29638][connectors][filesystems][formats] Update Jackson-BOM to 2.13.4.2 | 19 October 2022, 19:27:54 UTC |
| 7712d4b | Sergey Nuyanzin | 07 October 2022, 13:38:22 UTC | [hotfix] Extract Jackson BOM version into a property | 19 October 2022, 19:27:54 UTC |
| 40b3aab | Sergey Nuyanzin | 29 September 2022, 09:46:43 UTC | [FLINK-29468][connectors][filesystems][formats] Update Jackson-BOM to 2.13.4. | 19 October 2022, 19:27:54 UTC |
| f44ffe5 | tison | 19 October 2022, 01:19:43 UTC | [BP-1.15][FLINK-29613][Connector/Pulsar] Fix wrong batch size assertion (#21100) This cherry-picks #21069. Co-authored-by: Yufan Sheng <yufan@streamnative.io> | 19 October 2022, 01:19:43 UTC |
| 962b6e0 | tison | 18 October 2022, 13:14:45 UTC | Add jaxb-api back to pulsar-client-all dependencies. (#21092) Co-authored-by: Yufan Sheng <yufan@streamnative.io> | 18 October 2022, 13:14:45 UTC |
| 8181abe | Matthias Pohl | 14 October 2022, 10:22:04 UTC | [FLINK-25554][test] Remove timeout to enable a thread dump in case the test times out again. | 18 October 2022, 06:57:16 UTC |
| 5e4da2c | Juntao Hu | 17 October 2022, 10:42:48 UTC | [FLINK-29658][python] Fix the missing LocalTime support when converting Table to DataStream This closes #21086. | 18 October 2022, 06:52:19 UTC |
| f19f032 | Dawid Wysakowicz | 17 October 2022, 07:49:57 UTC | [FLINK-29645] BatchExecutionKeyedStateBackend is using incorrect ExecutionConfig when creating serializer | 17 October 2022, 15:25:56 UTC |
| 507b93e | Juntao Hu | 30 September 2022, 05:01:30 UTC | [FLINK-29477][python] Fix ClassCastException when collect primitive array This closes #21016. | 16 October 2022, 04:31:27 UTC |
| 126ae4d | Fabian Paul | 11 October 2022, 12:52:44 UTC | [FLINK-29512][streaming] Align the checkpointId of recovered CommittableWithLinages with the CommittableSummary Before this change during recovery for the CommitableCollector we initialized the SubtaskCommittableManager always with the initial checkpoint id (1) but the holding CheckpointCommittable with the checkpoint id in state. This could lead to that the emitted CommittableWithLinages update SubtaskCommittableManagers that they do not belong to cause "Unknown subtask for <id>" failures. | 14 October 2022, 13:38:43 UTC |
| a826fe8 | zoucao | 14 October 2022, 02:27:01 UTC | [FLINK-26726][hive]Hive enumerators do not assign splits to unregistered (failed) readers This closes #21038 | 14 October 2022, 02:27:01 UTC |
| 6b48827 | Krzysztof Chmielewski | 10 October 2022, 14:34:36 UTC | [FLINK-29509][streaming] Set subtaskId and numberOfSubtasks on CheckpointCommittableManagerImpl during deserialization. When we recover the `CheckpointCommittableManager` we were ignoring the subtaskId it is recovered on. This becomes a problem when a sink uses a post-commit topology because multiple committer operators might forward committable summaries coming from the same subtaskId. This ticket implements a fix to use the subtaskId already present in the CommittableCollectorSerializer when recreating CheckpointCommittableManagerImpl during recovery. Signed-off-by: Krzysztof Chmielewski <krzysiek.chmielewski@gmail.com> | 13 October 2022, 06:57:04 UTC |
| bcff5af | empcl | 12 October 2022, 08:37:50 UTC | [FLINK-27384][hive] Fix the modified partitions are missed in temporal table with create-time mode This closes #20437. | 12 October 2022, 08:37:50 UTC |
| aa78e3f | Martijn Visser | 10 October 2022, 12:31:01 UTC | [FLINK-29495][Connector/Pulsar] Refactor Pulsar tests from JUnit4 annotation to JUnit5 annotation for disabling tests on Java 11 | 11 October 2022, 18:01:19 UTC |
| d5921dd | Dawid Wysakowicz | 11 October 2022, 12:48:11 UTC | [FLINK-22243] Remove adaptive scheduler Web UI limitation from docs | 11 October 2022, 17:18:29 UTC |
| 5ad5618 | Chesnay Schepler | 05 October 2022, 09:17:33 UTC | [FLINK-29503][rest] Add backpressureLevel field Enum fields have naming restrictions in some languages. | 11 October 2022, 11:07:47 UTC |
| 383d451 | Chesnay Schepler | 05 October 2022, 08:57:46 UTC | [FLINK-29504][rest] Add schema to jar upload content | 11 October 2022, 11:04:37 UTC |
| ae20e52 | Hong Liang Teoh | 04 October 2022, 07:44:29 UTC | [FLINK-29395][Connector/Kinesis] Handle empty deaggregated records in FanOutRecordPublisher | 10 October 2022, 13:42:11 UTC |
| 5d748f3 | huangxingbo | 10 October 2022, 12:16:56 UTC | [FLINK-29483][python] Fix vectorized python scalar function failed in thread mode | 10 October 2022, 12:16:56 UTC |
| ea4f9aa | Chesnay Schepler | 10 October 2022, 10:04:51 UTC | [hotfix] Fix broken link | 10 October 2022, 10:06:53 UTC |
| 74bc6d2 | Dawid Wysakowicz | 04 October 2022, 11:50:28 UTC | [FLINK-29500] InitializeOnMaster uses wrong parallelism with AdaptiveScheduler We change the signature of JobVertex#initializeOnMaster/finalizeOnMaster to pass a Context object. In this context we can pass the actual parallelism the vertex will be run with. | 06 October 2022, 15:07:05 UTC |
| b848cc2 | Danny Cranmer | 06 October 2022, 07:40:50 UTC | [FLINK-29476][connector][kinesis] Backport unit test to Flink 1.15 | 06 October 2022, 07:40:50 UTC |
| 9715b5e | Alexander Fedulov | 29 September 2022, 15:19:23 UTC | [FLINK-29476][connector][kinesis] Retry connection termination with EOFException | 06 October 2022, 07:33:39 UTC |
| d6e8e42 | Dawid Wysakowicz | 30 September 2022, 09:14:19 UTC | [FLINK-22243] Reactive Mode parallelism changes are not shown in the job graph visualization in the UI Update the JsonPlan of ExecutionGraph with the correct parallelism instead of using the original one from the JobGraph. | 06 October 2022, 06:31:20 UTC |
| 5079130 | Dawid Wysakowicz | 04 October 2022, 06:29:31 UTC | [FLINK-26469] Adaptive job shows error in WebUI when not enough resource are available When the JobDetailsHandler returned an empty list of nodes, the Web UI incorrectly presented it as an exception while initializing the job. It is a correct situation when the job is INITIALIZING, but not yet running. This is especially visible in combination with AdaptiveScheduler where such a state can be present for longer periods when the scheduler waits for required resources. Instead of presenting a message "Job failed during initialization of JobManager" we present details page with a note that the Job is not yet running and not tasks/nodes details. | 05 October 2022, 13:20:33 UTC |
| 33b290e | Yufan Sheng | 29 September 2022, 17:40:14 UTC | [FLINK-29381][Connector/Pulsar] Fixes the split assignment for Key Shared subscription. | 30 September 2022, 03:46:28 UTC |
| 6075746 | Yufan Sheng | 28 September 2022, 07:28:14 UTC | [FLINK-26182][Connector/pulsar] Create e2e tests for the Pulsar source and sink based on the connector testing framework. | 30 September 2022, 03:46:28 UTC |
| 022df77 | Yufan Sheng | 27 September 2022, 10:04:24 UTC | [FLINK-26182][Connector/pulsar] Extract common logic from Pulsar source testing tools. | 30 September 2022, 03:46:28 UTC |
| 2e4eb09 | Yufan Sheng | 27 September 2022, 10:04:24 UTC | [FLINK-26182][Connector/Pulsar] Enable the transaction in the Pulsar TestContainers in a better way. | 30 September 2022, 03:46:28 UTC |
| 1ccf832 | Yufan Sheng | 27 September 2022, 10:04:23 UTC | [FLINK-26182][Connector/Pulsar] Drop the embedded Pulsar runtime for in favor of mock Pulsar runtime with local metastore support. | 30 September 2022, 03:46:28 UTC |
| dbfdeba | pvary | 29 September 2022, 13:00:50 UTC | [FLINK-28733][scripts] jobmanager.sh supports dynamic parameters | 29 September 2022, 13:00:50 UTC |
| 4760601 | Roman Khachatryan | 31 August 2022, 11:24:56 UTC | [FLINK-29158] Fix logging in DefaultCompletedCheckpointStore This closes #20723. | 28 September 2022, 03:17:00 UTC |
| a5ae2fa | Chesnay Schepler | 21 September 2022, 16:10:08 UTC | [FLINK-29397][runtime] Check if changelog provider is null | 22 September 2022, 23:17:01 UTC |
| 9ee1589 | Jun He | 17 September 2022, 02:56:07 UTC | [FLINK-29325][docs] Fix documentation bug on how to enable batch mode for streaming examples This closes #20849. | 21 September 2022, 02:50:11 UTC |
| eb65655 | harker2015 | 20 September 2022, 15:31:44 UTC | [FLINK-29324][Connectors/Kinesis] Fix NPE for Kinesis connector when closing (#20853) * [FLINK-29324] Fix NPE for Kinesis connector when closing * [FLINK-29324] Add unit test case | 20 September 2022, 15:48:10 UTC |
| 7784ea7 | huangxingbo | 20 September 2022, 04:31:33 UTC | [hotfix] Add 1.15 release note in Chinese doc | 20 September 2022, 04:34:58 UTC |
| 5a207e5 | Yufan Sheng | 05 September 2022, 16:02:04 UTC | [FLINK-28934][Connector/pulsar] Support connector testing tools for Pulsar unordered source. | 16 September 2022, 01:34:50 UTC |
| 252b81c | Yufan Sheng | 05 September 2022, 16:01:17 UTC | [FLINK-27400][Connector/pulsar] Filter system topics for Pulsar connector. | 16 September 2022, 01:34:50 UTC |
| c9bfb8d | Yufan Sheng | 05 September 2022, 15:59:52 UTC | [FLINK-27611][Connector/pulsar] Fix ConcurrentModificationException during checkpoint on Pulsar unordered reader. | 16 September 2022, 01:34:50 UTC |
| e79d3b4 | Yufan Sheng | 05 September 2022, 15:57:53 UTC | [FLINK-28084][Connector/pulsar] Disable retry and delete reconsume logic on PulsarUnorderedPartitionSplitReader. | 16 September 2022, 01:34:50 UTC |
| 96d4b67 | Yufan Sheng | 05 September 2022, 15:51:10 UTC | [FLINK-27388][Connector/pulsar] Change the topic setup logic in Pulsar runtime operator. | 16 September 2022, 01:34:50 UTC |
| c67a31b | Yufan Sheng | 16 August 2022, 23:02:53 UTC | [FLINK-28934][Connector/pulsar] Fix split assignment in different Pulsar subscriptions. | 16 September 2022, 01:34:50 UTC |
| 1afcba0 | Chesnay Schepler | 13 September 2022, 18:51:24 UTC | [FLINK-29262][docs] Document API compatibility guarantees | 13 September 2022, 18:53:06 UTC |
| 4e8247c | Chesnay Schepler | 13 September 2022, 12:23:25 UTC | [hotfix][build] Wipe legacy japicmp exclusions | 13 September 2022, 12:41:08 UTC |
| 771000c | Chesnay Schepler | 13 September 2022, 12:19:50 UTC | [FLINK-29260][release] Wipe exclusion list when updating reference version | 13 September 2022, 12:41:08 UTC |
| 776ccfa | Matthias Pohl | 12 September 2022, 09:35:44 UTC | [FLINK-29253][runtime] Removes synchronous close call from DefaultJobManagerRunnerRegistry#localCleanupAsync localCleanupAsync is meant to be called from the Dispatcher's main thread. Any blocking calls should be avoided here. Instead, we could use closeAsync which is implemented by JobManagerRunner. Two additional tests needed to be touched because the previously used AutoCloseableAsync#close method added another FlinKException to the Stacktrace. | 13 September 2022, 07:11:10 UTC |
| cef796f | Chesnay Schepler | 01 February 2022, 14:10:37 UTC | [FLINK-25868][build] Enable japicmp for all modules | 12 September 2022, 19:04:35 UTC |
| 7e961c3 | Chesnay Schepler | 31 August 2022, 14:27:17 UTC | [FLINK-25868][build] Use custom japicmp plugin version | 12 September 2022, 19:04:35 UTC |
| c8a6233 | Chesnay Schepler | 01 February 2022, 14:09:19 UTC | [FLINK-25868][build] Disable japicmp in packaging/e2e modules Optimize build times and reduce noise (warnings would be printed as these modules are not being deployed, so the previous versions can not be found). | 12 September 2022, 19:04:35 UTC |
| 5448f56 | Chesnay Schepler | 01 February 2022, 14:08:05 UTC | [FLINK-25868][build] Limit japicmp to jar modules Skips quickstart and pom modules. | 12 September 2022, 19:04:35 UTC |
| b552505 | snuyanzin | 08 September 2022, 07:24:39 UTC | [FLINK-29223][coordination] Add missing output info for jobs already reached terminal state | 12 September 2022, 12:16:30 UTC |
| a3adb13 | wenbingshen | 09 September 2022, 09:03:40 UTC | [FLINK-29207][pulsar][BP-1.15] Fix Pulsar message eventTime may be incorrectly set to a negative number (#20799) | 09 September 2022, 09:03:40 UTC |
| d0ed7c1 | huangxingbo | 08 September 2022, 07:14:21 UTC | [hotfix][Azure][python] Upgrade to macOS-latest | 08 September 2022, 07:17:11 UTC |
| e1cbf9e | Danny Cranmer | 06 September 2022, 12:14:39 UTC | [FLINK-29205][connectors/kinesis] Passthrough use config to HTTP client when constructing Async Client for Kinesis EFO | 06 September 2022, 15:58:28 UTC |
| e7c7df4 | lincoln lee | 05 September 2022, 02:10:10 UTC | [FLINK-29138][table-planner] fix project can not be pushed into lookup source This closes #20727 | 05 September 2022, 02:10:10 UTC |
| ed8700d | Matthias Pohl | 29 August 2022, 12:48:38 UTC | [FLINK-28078][tests] Mitigate likelihood to run into test stability issues caused by CURATOR-645 CURATOR-645 covers a bug in the LeaderLatch implementation that causes a race condition if a child node, participating in the leader election, is removed too fast. This results in a different code branch being executed which triggers a reset of the LeaderLatch instead of re-collecting the children to determine the next leader. The issue occurs because LeaderLatch#checkLeadership is not executed transactionally, i.e. retrieving the children and setting up the watcher for the predecessor is not done atomically. This leads to the race condition where a children (the previous leader's node) is removed before setting up the watcher which results in an invalid handling of the situation using reset. Adding some sleep here (simulating the leader actually doing something) will reduce the risk of falling into the race condition because it will give the concurrently running LeaderLatch instances more time to set up the watchers properly. This is only meant as a temporary solution until CURATOR-645 is resolved and the curator dependency on the Flink side is upgraded. | 01 September 2022, 12:39:27 UTC |
| 6d73aab | Paul Lin | 20 August 2022, 07:35:25 UTC | [hotfix][build] Fix streaming-java oldVersion artifactId Remove temporary scala suffix settings for 1.15 in japicmp | 01 September 2022, 09:19:55 UTC |
| 3dad3a6 | huangxingbo | 01 September 2022, 02:51:17 UTC | [FLINK-29161][tests] Fix the built docker image name | 01 September 2022, 08:38:22 UTC |
| ad33fe8 | Yun Tang | 29 August 2022, 12:49:33 UTC | [FLINK-29130][state] Correct the doc description of local-recovery | 31 August 2022, 08:57:24 UTC |
| cbd6c8d | Yufan Sheng | 30 August 2022, 02:32:56 UTC | [FLINK-28609][Connector/Pulsar] PulsarSchema didn't get properly serialized. (#20699) | 30 August 2022, 02:32:56 UTC |
| 291d224 | Chesnay Schepler | 29 August 2022, 16:40:18 UTC | Update version to 1.15.2 | 29 August 2022, 16:40:18 UTC |
| fcff490 | wangyang0918 | 15 August 2022, 15:03:19 UTC | [FLINK-28265][k8s] Make KubernetesStateHandleStore#addEntry idempotent This closes #20673. | 27 August 2022, 06:55:08 UTC |
| 493a1aa | Roman Khachatryan | 15 August 2022, 17:30:36 UTC | [FLINK-28976][state] Don't add extra delay to the 1st materialization | 26 August 2022, 11:57:19 UTC |
| 605a711 | Roman Khachatryan | 15 August 2022, 17:29:58 UTC | [hotfix][state] Fix logging in Materializer and make FLINK-28976 more explicit (cherry picked from commit 0e083c119b24a9b5fa7af1ad274c9ee7111c07cc) | 26 August 2022, 11:57:19 UTC |
| 258c3e3 | Roman Khachatryan | 24 August 2022, 16:20:02 UTC | [hotfix][tests] Use negative changelog materialization interval in tests ...instead of a big value so that the subsequent change won't break them. | 26 August 2022, 11:57:19 UTC |
| 62d7cc0 | Danny Cranmer | 24 August 2022, 21:00:45 UTC | Update japicmp configuration for 1.15.2 | 24 August 2022, 21:01:16 UTC |
| a526391 | Yufan Sheng | 23 August 2022, 01:55:26 UTC | [FLINK-27917][Connector/Pulsar] Drop Consumer.seek() in the testing method for fixing the race condition. (#20578) | 23 August 2022, 01:55:26 UTC |
| 29ff78a | Sergey Nuyanzin | 14 August 2022, 08:28:09 UTC | [FLINK-28951][table-planner] Make header with one line comments This closes #20568. | 18 August 2022, 09:58:09 UTC |
| 51b3752 | Chesnay Schepler | 17 August 2022, 16:01:34 UTC | [FLINK-28488][kafka] Only forward measurable Kafka metrics and ignore others | 17 August 2022, 21:58:52 UTC |
| 9791850 | Danny Cranmer | 16 August 2022, 12:12:35 UTC | [FLINK-28861][table][docs] Update "upgrading" documentation to include notice regarding the Table API operator issue | 16 August 2022, 18:15:36 UTC |
| 44c6496 | Ingo Bürk | 16 August 2022, 08:27:11 UTC | [FLINK-28994][runtime-web] Enable withCredentials for outgoing HTTP requests Some environments require cookies to authenticate the Flink UI. By enabling the withCredentials flag, Angular will send cookies along with the request. Signed-off-by: Ingo Bürk <airblader@apache.org> This closes #20593. | 16 August 2022, 12:54:09 UTC |
| c90d0d5 | Hong Teoh | 15 August 2022, 23:11:56 UTC | [FLINK-28978][kinesis] Update AWS Regions validation to allow for future AWS regions | 16 August 2022, 10:40:20 UTC |
| 105d7c9 | Timo Walther | 10 August 2022, 13:30:50 UTC | [FLINK-28861][table] Make UID generation behavior configurable and plan-only by default Before this commit, due to changes for FLIP-190, every operator generated by the planner got a UID assigned. However, the UID is based on a static counter that might return different results depending on the environment. Thus, UIDs are not deterministic and make stateful restores impossible e.g. when going from 1.15.0 -> 1.15.1. This PR restores the old pre-1.15 behavior for regular Table API. It only adds UIDs if the operator has been created from a compiled plan. A compiled plan makes the UIDs static and thus deterministic. table.exec.uid.generation=ALWAYS exists for backwards compatibility and could make stateful upgrades possible even with invalid UIDs on best effort basis. This closes #20586. | 16 August 2022, 07:24:50 UTC |
| 65f7913 | Yufan Sheng | 16 August 2022, 06:09:08 UTC | [BK-1.15][FLINK-27199][Connector/Pulsar] Bump pulsar to 2.10.0 (#20582) * [FLINK-27199][Connector/pulsar] Use mocked pulsar runtime instead of embedded runtime and enable tests. * [FLINK-27199][Connector/Pulsar] Bump the pulsar-client-all to latest 2.10.0. * [FLINK-27199][Connector/Pulsar] Pulsar 2.10.0 deprecated the queue length, add new memory-based options. * [FLINK-27199][Connector/Pulsar] Move mocked Pulsar runtime to new local memory based metastore. Drop embedded Pulsar runtime. Add a bootstrap script for Pulsar docker runtime. | 16 August 2022, 06:09:08 UTC |
| 63c1c1b | Juntao Hu | 11 August 2022, 02:19:38 UTC | [FLINK-28908][python] Fix LIST type in Python DataStream API This closes #20539. | 15 August 2022, 04:36:17 UTC |
| ded03b7 | Zhu Zhu | 12 August 2022, 04:12:33 UTC | [FLINK-28878][tests] Increase slot request timeout of PipelinedRegionSchedulingITCase This helps to make the tests more stable. This closes #20559. | 15 August 2022, 03:43:51 UTC |
| 5f842cb | Yufan Sheng | 14 August 2022, 14:49:32 UTC | [FLINK-27399][Connector/Pulsar] Change initial consuming position setting logic for better handle the checkpoint. (#19972) (#20564) * Change the initial start cursor and stop cursor to better handle the consuming behaviors. * Create the initial subscription instead seek every time. This should fix the wrong position setting. * Fix the wrong stop cursor, make sure it stops at the correct space * Drop Consumer.seek() for https://github.com/apache/pulsar/pull/16171 | 14 August 2022, 14:49:32 UTC |
| e5570e3 | Qishang Zhong | 13 August 2022, 03:21:58 UTC | [FLINK-28817][connector/common] NullPointerException in HybridSource when restoring from checkpoint | 14 August 2022, 00:07:26 UTC |
| 3d910dc | PengYuan | 12 August 2022, 08:30:51 UTC | [FLINK-27856][k8s] Fix the NPE when no spec is configured in pod template This closes #20558. | 12 August 2022, 08:30:51 UTC |
| b24c516 | 鲍健昕 | 20 July 2022, 02:35:40 UTC | [FLINK-27570][runtime] Count checkpoint finalization failures in CheckpointFailureManager | 12 August 2022, 07:20:19 UTC |
