You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "PatrickRen (via GitHub)" <gi...@apache.org> on 2023/03/10 07:29:06 UTC
[GitHub] [flink] PatrickRen opened a new pull request, #22146: [FLINK-31157][release] Add release note for version 1.17
PatrickRen opened a new pull request, #22146:
URL: https://github.com/apache/flink/pull/22146
## What is the purpose of the change
This pull request adds release note for version 1.17.
## Brief change log
- Add release note for 1.17
## Verifying this change
This change is a doc-only without test coverage.
## Does this pull request potentially affect one of the following parts:
- Dependencies (does it add or upgrade a dependency): (yes / **no**)
- The public API, i.e., is any changed class annotated with `@Public(Evolving)`: (yes / **no**)
- The serializers: (yes / **no** / don't know)
- The runtime per-record code paths (performance sensitive): (yes / **no** / don't know)
- Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn, ZooKeeper: (yes / **no** / don't know)
- The S3 file system connector: (yes / **no** / don't know)
## Documentation
- Does this pull request introduce a new feature? (yes / **no**)
- If yes, how is the feature documented? (**not applicable** / docs / JavaDocs / not documented)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] leonardBang commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "leonardBang (via GitHub)" <gi...@apache.org>.
leonardBang commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1135723623
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
> It is mentioned in the "Calcite Upgrade" section of Flink 1.17 release announcement that "The downside of this upgrade is that the execution plan changes for certain SQL queries. Savepoints are not backwards compatible in these cases".
>
> Would this prevent users from upgrading from Flink 1.16 to Flink 1.17 without losing job state? Is there specific actions that users should take to get around this issue? Maybe we should mention this in the release note.
+1, @PatrickRen You can refer previous release which also upgrade the calcite version
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] PatrickRen commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "PatrickRen (via GitHub)" <gi...@apache.org>.
PatrickRen commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1138025355
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
Thanks @snuyanzin for the context. As the [document](https://nightlies.apache.org/flink/flink-docs-master/docs/dev/table/concepts/overview/#stateful-upgrades-and-evolution) I think it's not easy to give such a migration path described by @lindong28, because we couldn't tell exactly which kind of query or plan is changed by the Calcite upgrade. This incompatibility issue should have already existed for many versions as optimizing rules keep evolving.
+1 for removing the statement from the announcement.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] PatrickRen commented on pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "PatrickRen (via GitHub)" <gi...@apache.org>.
PatrickRen commented on PR #22146:
URL: https://github.com/apache/flink/pull/22146#issuecomment-1480642481
Big thanks for the review @echauchot! Fixed in the latest commit.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] echauchot commented on pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "echauchot (via GitHub)" <gi...@apache.org>.
echauchot commented on PR #22146:
URL: https://github.com/apache/flink/pull/22146#issuecomment-1480752646
> Big thanks for the review @echauchot! Fixed in the latest commit.
My pleasure
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] snuyanzin commented on pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "snuyanzin (via GitHub)" <gi...@apache.org>.
snuyanzin commented on PR #22146:
URL: https://github.com/apache/flink/pull/22146#issuecomment-1465950320
thanks for addressing comments, lgtm from my side
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] echauchot commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "echauchot (via GitHub)" <gi...@apache.org>.
echauchot commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1144415693
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
Review Comment:
I compared with the `docs/content `version. This file is out of sync lacking the last paragraph. I guess the change was done only on the `docs/content` version. I think we should overwrite the `docs/content.zh` version with the `docs/content` version as they are identical, this paragraph apart.
##########
docs/content/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,214 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
Review Comment:
```suggestion
input is an upsert source (use upsert-kafka connector), the query plan will remove the
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
Review Comment:
```suggestion
The slow task detecting is improved for speculative execution. Previously, it only considered the
```
##########
docs/content/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,214 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
Review Comment:
I think, it needs rewording
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
Review Comment:
```suggestion
to reduce the number of config options the user needs to tune when trying to increase the network
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
Review Comment:
```suggestion
a second step to Flink >= 1.17.x.
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
Review Comment:
```suggestion
indirectly interacts with the reporter via the singleton instance (e.g., a test trying to assert
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
Review Comment:
```suggestion
depends on `TypeSerializerConfigSnapshot`, please upgrade your serializer to `TypeSerializerSnapshot` while still using 1.8.x <= Flink <= 1.16.x, take a savepoint and restore from that
```
##########
docs/content/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,214 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
Review Comment:
```suggestion
temporal join. Note this may cause incompatible plan changes compared to 1.16.0, e.g., when left
```
##########
docs/content/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,214 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
Review Comment:
```suggestion
The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib, one in lib/opt
```
##########
docs/content/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,214 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
Review Comment:
```suggestion
a ChangelogNormalize node for it. This is an incompatible plan change compared to 1.16.0.
```
##########
docs/content/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,214 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
Review Comment:
```suggestion
After the patch is applied, the filter will no longer be pushed down into both inputs of the event time
```
##########
docs/content/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,214 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
Review Comment:
```suggestion
ChangelogNormalize node which appeared in 1.16.0.
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
Review Comment:
```suggestion
Kinesis connector now shades and relocates transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
Review Comment:
```suggestion
tasks into account. Tasks which have a longer execution time but consumes more data may not be
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
Review Comment:
```suggestion
Flink <= 1.16.x, and you are using watermark alignment, but at the same time you are not able to
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
Review Comment:
```suggestion
Adaptive batch scheduler is now used for batch jobs by default. It will automatically decide the
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
Review Comment:
```suggestion
parallelism of operators. The keys and values of related configuration items are improved for ease
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
Review Comment:
```suggestion
the previous default value of `1g`.
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
Review Comment:
```suggestion
size is fixed. To go back to the previous behavior, the user can explicitly configure this option to
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
Review Comment:
```suggestion
Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means that all savepoints
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
Review Comment:
```suggestion
Tuple, ...), and your savepoint is from Flink >= 1.8.0, you don't have to do anything.
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
Review Comment:
```suggestion
for reading data from upstream tasks. Reducing the number of required buffers helps reduce the
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
+##### [FLINK-20873](https://issues.apache.org/jira/browse/FLINK-20873)
+##### [FLINK-21239](https://issues.apache.org/jira/browse/FLINK-21239)
+##### [FLINK-29932](https://issues.apache.org/jira/browse/FLINK-29932)
+Calcite upgrade brings optimization changing logical plans for some queries involving `Sarg` and
Review Comment:
```suggestion
Calcite upgrade brings optimizations that change logical plans for some queries involving `Sarg` and
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] snuyanzin commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "snuyanzin (via GitHub)" <gi...@apache.org>.
snuyanzin commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1132040358
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,201 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior, or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, and heap and managed memory sizes when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.27
+##### [FLINK-20873](https://issues.apache.org/jira/browse/FLINK-20873)
Review Comment:
Since finally in Flink 1.17.0 it was upgraded to 1.29.0, I think it would make sense to use the latest Calcite version in 1.17 branch
```suggestion
#### Upgrade Calcite version to 1.29.0
##### [FLINK-20873](https://issues.apache.org/jira/browse/FLINK-20873)
##### [FLINK-21239](https://issues.apache.org/jira/browse/FLINK-21239)
##### [FLINK-29932](https://issues.apache.org/jira/browse/FLINK-29932)
Calcite upgrade brings optimization changing logical plans for some queries involving `Sarg` and
queries with `count` of non-distinct values.
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] snuyanzin commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "snuyanzin (via GitHub)" <gi...@apache.org>.
snuyanzin commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1132039260
##########
docs/content/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,201 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior, or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, and heap and managed memory sizes when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.27
+##### [FLINK-20873](https://issues.apache.org/jira/browse/FLINK-20873)
Review Comment:
Since finally in Flink 1.17.0 it was upgraded to 1.29.0, I think it would make sense to use the latest Calcite version in 1.17 branch
```suggestion
#### Upgrade Calcite version to 1.29.0
##### [FLINK-20873](https://issues.apache.org/jira/browse/FLINK-20873)
##### [FLINK-21239](https://issues.apache.org/jira/browse/FLINK-21239)
##### [FLINK-29932](https://issues.apache.org/jira/browse/FLINK-29932)
Calcite upgrade brings optimization changing logical plans for some queries involving `Sarg` and
queries with `count` of non-distinct values.
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] PatrickRen commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "PatrickRen (via GitHub)" <gi...@apache.org>.
PatrickRen commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1136479162
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
Thanks for reviewing this @lindong28 @leonardBang !
> Would this prevent users from upgrading from Flink 1.16 to Flink 1.17 without losing job state? Is there specific actions that users should take to get around this issue? Maybe we should mention this in the release note.
@lindong28 TBH I'm not quite familiar with this part. Maybe @snuyanzin could provide some inputs on the migration path. But I could add some warns in this section that highlights the incompatibility.
> You can refer previous releases which also upgrade the calcite version
@leonardBang Unfortunately I can't find previous release notes mentioning upgrading Calcite.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] snuyanzin commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "snuyanzin (via GitHub)" <gi...@apache.org>.
snuyanzin commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1133775105
##########
docs/content/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, and heap and managed memory sizes when the total memory
Review Comment:
```suggestion
potential changes of network memory size, heap and managed memory sizes when the total memory
```
Probably one `and` could be removed
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] snuyanzin commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "snuyanzin (via GitHub)" <gi...@apache.org>.
snuyanzin commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1133756421
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior, or dependencies,
Review Comment:
```suggestion
These release notes discuss important aspects, such as configuration, behavior or dependencies,
```
not sure, however it looks like no need for comma here...
I try to follow this rule https://www.grammarly.com/blog/comma-before-or/
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] PatrickRen commented on pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "PatrickRen (via GitHub)" <gi...@apache.org>.
PatrickRen commented on PR #22146:
URL: https://github.com/apache/flink/pull/22146#issuecomment-1465732527
Thanks for the review @snuyanzin ! Fixed in the latest commit
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] lindong28 commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "lindong28 (via GitHub)" <gi...@apache.org>.
lindong28 commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1138082345
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
Given that Flink does not guarantee savepoint compatibility between major/minor releases, I think removing that sentence from the release announcement is reasonable.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] snuyanzin commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "snuyanzin (via GitHub)" <gi...@apache.org>.
snuyanzin commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1136752190
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
As far as I know there is currently no full support of backward compatibility for Table SQL/API while major version upgrade and this is a topic for FLIP-190 [1].
A good side here is that there are changes only for logical plans. Neither physical plans nor ser/de for json plans have been impacted. That means that ser/de from json plans while upgrade should work.
The only issue is that i fail to find a proper doc for that... Only a dedicated jira issue[2]
[1] https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=191336489
[2] https://issues.apache.org/jira/browse/FLINK-26250
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] leonardBang commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "leonardBang (via GitHub)" <gi...@apache.org>.
leonardBang commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1135723623
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
> It is mentioned in the "Calcite Upgrade" section of Flink 1.17 release announcement that "The downside of this upgrade is that the execution plan changes for certain SQL queries. Savepoints are not backwards compatible in these cases".
>
> Would this prevent users from upgrading from Flink 1.16 to Flink 1.17 without losing job state? Is there specific actions that users should take to get around this issue? Maybe we should mention this in the release note.
+1, @PatrickRen You can refer previous releases which also upgrade the calcite version
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] lindong28 commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "lindong28 (via GitHub)" <gi...@apache.org>.
lindong28 commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1135387560
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
It is mentioned in the "Calcite Upgrade" section of Flink 1.17 release announcement that "The downside of this upgrade is that the execution plan changes for certain SQL queries. Savepoints are not backwards compatible in these cases".
Would this prevent users from upgrading from Flink 1.16 to Flink 1.17 without losing job state? Is there specific actions that users should take to get around this issue? Maybe we should mention this in the release note.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] PatrickRen merged pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "PatrickRen (via GitHub)" <gi...@apache.org>.
PatrickRen merged PR #22146:
URL: https://github.com/apache/flink/pull/22146
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] echauchot commented on pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "echauchot (via GitHub)" <gi...@apache.org>.
echauchot commented on PR #22146:
URL: https://github.com/apache/flink/pull/22146#issuecomment-1479142587
@PatrickRen reviewing ...
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] lindong28 commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "lindong28 (via GitHub)" <gi...@apache.org>.
lindong28 commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1136845292
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
One of the most important purposes of release notes is to explain to users how to upgrade existing jobs from Flink 1.16 (and ideally earlier versions as well) to Flink 1.17. Without good upgrade instructions, the upgrade experience may be bad and many users may be scared away from using Flink 1.17.
In this specific case, I hope that users can find the following information from the release notes:
- Will I be able to restart a job successfully with Flink 1.17 using the savepoint taken when the job was executed with Flink 1.16?
- If not, do I have to discard state and start the job from a very early point to re-process data? What impact should I be prepared for when upgrading to Flink 1.17?
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
One of the most important purposes of release notes is to explain to users how to upgrade existing jobs from Flink 1.16 (and ideally earlier versions as well) to Flink 1.17. Without good upgrade instructions, the upgrade experience may be bad and many users may be scared away from using Flink 1.17.
In this specific case, I hope that users can find the following information from the release note:
- Will I be able to restart a job successfully with Flink 1.17 using the savepoint taken when the job was executed with Flink 1.16?
- If not, do I have to discard state and start the job from a very early point to re-process data? What impact should I be prepared for when upgrading to Flink 1.17?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] flinkbot commented on pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "flinkbot (via GitHub)" <gi...@apache.org>.
flinkbot commented on PR #22146:
URL: https://github.com/apache/flink/pull/22146#issuecomment-1463391299
<!--
Meta data
{
"version" : 1,
"metaDataEntries" : [ {
"hash" : "5eb387243eb4e0d2cbdf73556444f38c9dd365fe",
"status" : "UNKNOWN",
"url" : "TBD",
"triggerID" : "5eb387243eb4e0d2cbdf73556444f38c9dd365fe",
"triggerType" : "PUSH"
} ]
}-->
## CI report:
* 5eb387243eb4e0d2cbdf73556444f38c9dd365fe UNKNOWN
<details>
<summary>Bot commands</summary>
The @flinkbot bot supports the following commands:
- `@flinkbot run azure` re-run the last Azure build
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] leonardBang commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "leonardBang (via GitHub)" <gi...@apache.org>.
leonardBang commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1135722167
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
+1
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] snuyanzin commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "snuyanzin (via GitHub)" <gi...@apache.org>.
snuyanzin commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1136871850
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
+#### Upgrade Calcite version to 1.29.0
Review Comment:
Sorry for the confusion...
Based on the current doc [1]
```
Currently, the framework cannot guarantee that state can be mapped from a savepoint to a new table operator topology.
In other words: Savepoints are only supported if both the query and the Flink version remain constant.
```
Also we tried to do such test for simple SQL jobs to restore in 1.17 from the state made in 1.16 and failed [2], same issue between 1.15 and 1.16.
Since this is the current state of the upgrade between major version and not Calcite upgrade only related feature,
does it make sense to remove this sentence `"The downside of this upgrade is that the execution plan changes for certain SQL queries. Savepoints are not backwards compatible in these cases"`? Because anyway it's mentioned explicitly in doc [1]
[1] https://nightlies.apache.org/flink/flink-docs-master/docs/dev/table/concepts/overview/#stateful-upgrades-and-evolution
[2] https://issues.apache.org/jira/browse/FLINK-31066
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [flink] HuangXingBo commented on a diff in pull request #22146: [FLINK-31157][release] Add release note for version 1.17
Posted by "HuangXingBo (via GitHub)" <gi...@apache.org>.
HuangXingBo commented on code in PR #22146:
URL: https://github.com/apache/flink/pull/22146#discussion_r1143014759
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
+
+### Dependency upgrades
+
+#### Upgrade FRocksDB to 6.20.3-ververica-2.0
+##### [FLINK-30836](https://issues.apache.org/jira/browse/FLINK-30836)
+Upgrade FRocksDB to 6.20.3-ververica-2.0.
+
+#### Upgrade the minimal supported hadoop version to 2.10.2
+##### [FLINK-29710](https://issues.apache.org/jira/browse/FLINK-29710)
+The minimum Hadoop version supported by Apache Flink has been updated to version 2.10.2. For Hadoop
+3, the minimum Hadoop version that is now supported is version 3.2.3.
+
Review Comment:
```suggestion
#### Update dependency version for PyFlink
##### [FLINK-29421](https://issues.apache.org/jira/browse/FLINK-29421)
For support of Python3.10, PyFlink updates a series dependencies version: apache-beam==2.43.0 pemja==0.3.0 cloudpickle==2.2.0
#####
```
##########
docs/content.zh/release-notes/flink-1.17.md:
##########
@@ -0,0 +1,203 @@
+---
+title: "Release Notes - Flink 1.17"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+# Release notes - Flink 1.17
+
+These release notes discuss important aspects, such as configuration, behavior or dependencies,
+that changed between Flink 1.16 and Flink 1.17. Please read these notes carefully if you are
+planning to upgrade your Flink version to 1.17.
+
+
+### Clusters & Deployment
+
+#### Only one Zookeeper version is bundled in `flink-dist`
+##### [FLINK-30237](https://issues.apache.org/jira/browse/FLINK-30237)
+The Flink distribution no longer bundles 2 different Zookeeper client jars (one in lib/opt
+respectively). Instead, only 1 client will be bundled within the flink-dist jar. This has no
+effect on the supported Zookeeper server versions.
+
+
+### Table API & SQL
+
+#### Incompatible plan change of event time temporal join on an upsert source
+##### [FLINK-29849](https://issues.apache.org/jira/browse/FLINK-29849)
+A correctness issue when do event time temporal join with a versioned table backed by an upsert
+source was resolved. When the right input of the join is an upsert source, it no longer generates
+a ChangelogNormalize node for it. This is an incompatible plan change compare to 1.16.0.
+
+#### Incompatible plan change of filter after temporal join
+##### [FLINK-28988](https://issues.apache.org/jira/browse/FLINK-28988)
+After the patch applied, the filter will not be pushed down into both inputs of the event time
+temporal join. Note this may cause incompatible plan changes compare to 1.16.0, e.g., when left
+input is an upsert source(use upsert-kafka connector), the query plan will remove the
+ChangelogNormalize node from which appeared in 1.16.0.
+
+### Connectors & Libraries
+
+#### Remove cassandra connector from master branch
+##### [FLINK-30312](https://issues.apache.org/jira/browse/FLINK-30312)
+The Cassandra connector has been externalized and is no longer released as part of the main Flink
+release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-cassandra.
+
+#### Remove Pulsar connector from master branch
+##### [FLINK-30397](https://issues.apache.org/jira/browse/FLINK-30397)
+The Pulsar connector has been externalized and is no longer bundled and released as part of the
+main Flink release. Downloads can be found at https://flink.apache.org/downloads.html and the
+source code at https://github.com/apache/flink-connector-pulsar.
+
+#### Remove HCatalog
+##### [FLINK-29669](https://issues.apache.org/jira/browse/FLINK-29669)
+The HCatalog connector has been removed from Flink. You can use the Hive connector as a replacement.
+
+#### Remove Gelly
+##### [FLINK-29668](https://issues.apache.org/jira/browse/FLINK-29668)
+Gelly has been removed from Flink. Current users of Gelly should not upgrade to Flink 1.17 but
+stay on an older version. If you're looking for iterations support, you could investigate
+[Flink ML Iteration](https://nightlies.apache.org/flink/flink-ml-docs-stable/docs/development/iteration/)
+as a potential successor.
+
+#### Support watermark alignment of source splits
+##### [FLINK-28853](https://issues.apache.org/jira/browse/FLINK-28853)
+Since Flink 1.17, source connectors have to implement watermark alignment of source split in order
+to use the watermark alignment feature. The required methods to implement are:
+`SourceReader#pauseOrResumeSplits` and `SplitReader#pauseOrResumeSplits`.
+
+If you are migrating from
+Flink <= 1.16.x, and you were using watermark alignment, but at the same time you are not able to
+upgrade/modify your connector, you can disable per split alignment via setting
+`pipeline.watermark-alignment.allow-unaligned-source-splits` to true. Note that by doing so,
+watermark alignment will be working properly only when your number of splits equals to the
+parallelism of the source operator.
+
+#### Remove deprecated MiniClusterResource
+##### [FLINK-29548](https://issues.apache.org/jira/browse/FLINK-29548)
+The deprecated `MiniClusterResource` in `flink-test-utils` has been removed. The
+`MiniClusterWithClientResource` is a drop-in replacement.
+
+#### Kinesis connector doesn't shade jackson dependency
+##### [FLINK-14896](https://issues.apache.org/jira/browse/FLINK-14896)
+Shade and relocate transitive Jackson dependencies of `flink-connector-kinesis`. If your Flink job
+was transitively relying on the these, you may need to include additional Jackson dependencies into
+your project.
+
+### Runtime & Coordination
+
+#### Speculative execution take input data amount into account when detecting slow tasks
+##### [FLINK-30707](https://issues.apache.org/jira/browse/FLINK-30707)
+The slow task detecting is improved for speculative execution. Previously, it only considers the
+execution time of tasks when deciding which tasks are slow. It now takes the input data volume of
+tasks into account. Tasks which has a longer execution time but consumes more data may not be
+considered as slow. This improvement helps to eliminate the negative impacts of data skew on slow
+task detecting.
+
+#### Use adaptive batch scheduler as default scheduler for batch jobs
+##### [FLINK-30682](https://issues.apache.org/jira/browse/FLINK-30682)
+Adaptive batch scheduler are now used for batch jobs by default. It will automatically decide the
+parallelism of operators. The keys and values of related configuration items are improved for easy
+of use. More details can be found in the
+[document](https://nightlies.apache.org/flink/flink-docs-release-1.17/docs/deployment/elastic_scaling/#adaptive-batch-scheduler).
+
+#### Simplify network memory configurations for TaskManager
+##### [FLINK-30469](https://issues.apache.org/jira/browse/FLINK-30469)
+The default value of `taskmanager.memory.network.max` has changed from `1g` to `Long#MAX_VALUE`,
+to reduce the number of config options user needs to tune when trying to increase the network
+memory size. This may affect the performance when this option is not explicitly configured, due to
+potential changes of network memory size, heap and managed memory size when the total memory
+size is fixed. To go back to the previous behavior, user can explicitly configure this option to
+the previous default value `1g`.
+
+A threshold is introduced for controlling the number of required buffers among all buffers needed
+for reading data from upstream tasks. Reducing the number of required buffers helps reduces the
+chance of failures due to insufficient network buffers, at the price of potential performance
+impact. By default, the number of required buffers is only reduced for batch workloads, while stay
+unchanged for streaming workloads. This can be tuned via
+`taskmanager.network.memory.read-buffer.required-per-gate.max`. See the description of the config
+option for more details.
+
+### Metric Reporters
+
+#### Only support reporter factories for instantiation
+##### [FLINK-24235](https://issues.apache.org/jira/browse/FLINK-24235)
+Configuring reporters by their class is no longer supported. Reporter implementations must provide
+a MetricReporterFactory, and all configurations must be migrated to such a factory.
+
+#### `UseLogicalIdentifier` makes datadog consider metric as custom
+##### [FLINK-30383](https://issues.apache.org/jira/browse/FLINK-30383)
+The Datadog reporter now adds a "flink." prefix to metric identifiers if "useLogicalIdentifier" is
+enabled. This is required for these metrics to be recognized as Flink metrics, not custom ones.
+
+#### Use separate Prometheus CollectorRegistries
+##### [FLINK-30020](https://issues.apache.org/jira/browse/FLINK-30020)
+The PrometheusReporters now use a separate CollectorRegistry for each reporter instance instead of
+the singleton default registry. This generally shouldn't impact setups, but it may break code that
+indirectly interacted with the reporter via the singleton instance (e.g., a test trying to assert
+what metrics are reported).
+
+### Checkpoints
+
+#### Drop TypeSerializerConfigSnapshot and savepoint support from Flink versions < 1.8.0
+##### [FLINK-29807](https://issues.apache.org/jira/browse/FLINK-29807)
+Savepoints using `TypeSerializerConfigSnapshot` are no longer supported. That means, all savepoints
+from Flink < 1.8.0 are no longer supported. Furthermore, savepoints from Flink < 1.17.0 created
+with custom serializer using deprecated since Flink 1.8.0 class `TypeSerializerConfigSnapshot` are
+also no longer supported.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro,
+Tuple, ...), and your savepoint is from 1.8.0 <= Flink, you don't have to do anything.
+
+If you are only using built-in serializers (Pojo, Kryo, Avro, Tuple, ...), and your savepoint is
+from Flink < 1.8.0, please first upgrade your job to 1.8.x <= Flink <= 1.16.x, before upgrading in
+the second step to Flink >= 1.17.x.
+
+If previously you were using a custom serialiser that
+depends on `TypeSerializerConfigSnapshot`, please first while still using 1.8.x <= Flink <= 1.16.x
+upgrade your serializer to `TypeSerializerSnapshot`, take a savepoint and restore from that
+savepoint in Flink >= 1.7.0.
+
+### Python
+
+#### Remove duplicated operators in the StreamGraph for Python DataStream API jobs
+##### [FLINK-31272](https://issues.apache.org/jira/browse/FLINK-31272)
+This duplicated operator issue has been addressed since 1.15.4, 1.16.2 and 1.17.0. For jobs which
+are not affected by this issue, there are no backward compatibility issues. However, for jobs which
+are affected, it may not be possible to restore from savepoints generated from versions
+1.15.0 ~ 1.15.3 and 1.16.0 ~ 1.16.1.
+
Review Comment:
```suggestion
#### Adds support of Python 3.10 and removes support of Python 3.6
##### [FLINK-29421](https://issues.apache.org/jira/browse/FLINK-29421)
PyFlink 1.17 will support Python 3.10 and remove the support of Python 3.6
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org