alpinegizmo commented on a change in pull request #12257: URL: https://github.com/apache/flink/pull/12257#discussion_r429283693
########## File path: docs/dev/connectors/kafka.md ########## @@ -23,125 +23,59 @@ specific language governing permissions and limitations under the License. --> +Flink provides an [Apache Kafka](https://kafka.apache.org) connector for reading data from and writing data to Kafka topics with exactly-once guaruntees. + * This will be replaced by the TOC {:toc} -This connector provides access to event streams served by [Apache Kafka](https://kafka.apache.org/). - -Flink provides special Kafka Connectors for reading and writing data from/to Kafka topics. -The Flink Kafka Consumer integrates with Flink's checkpointing mechanism to provide -exactly-once processing semantics. To achieve that, Flink does not purely rely on Kafka's consumer group -offset tracking, but tracks and checkpoints these offsets internally as well. - -Please pick a package (maven artifact id) and class name for your use-case and environment. -For most users, the `FlinkKafkaConsumer010` (part of `flink-connector-kafka`) is appropriate. - -<table class="table table-bordered"> - <thead> - <tr> - <th class="text-left">Maven Dependency</th> - <th class="text-left">Supported since</th> - <th class="text-left">Consumer and <br> - Producer Class name</th> - <th class="text-left">Kafka version</th> - <th class="text-left">Notes</th> - </tr> - </thead> - <tbody> - <tr> - <td>flink-connector-kafka-0.10{{ site.scala_version_suffix }}</td> - <td>1.2.0</td> - <td>FlinkKafkaConsumer010<br> - FlinkKafkaProducer010</td> - <td>0.10.x</td> - <td>This connector supports <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-32+-+Add+timestamps+to+Kafka+message";>Kafka messages with timestamps</a> both for producing and consuming.</td> - </tr> - <tr> - <td>flink-connector-kafka-0.11{{ site.scala_version_suffix }}</td> - <td>1.4.0</td> - <td>FlinkKafkaConsumer011<br> - FlinkKafkaProducer011</td> - <td>0.11.x</td> - <td>Since 0.11.x Kafka does not support scala 2.10. This connector supports <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-98+-+Exactly+Once+Delivery+and+Transactional+Messaging";>Kafka transactional messaging</a> to provide exactly once semantic for the producer.</td> - </tr> - <tr> - <td>flink-connector-kafka{{ site.scala_version_suffix }}</td> - <td>1.7.0</td> - <td>FlinkKafkaConsumer<br> - FlinkKafkaProducer</td> - <td>>= 1.0.0</td> - <td> - This universal Kafka connector attempts to track the latest version of the Kafka client. - The version of the client it uses may change between Flink releases. Starting with Flink 1.9 release, it uses the Kafka 2.2.0 client. - Modern Kafka clients are backwards compatible with broker versions 0.10.0 or later. - However for Kafka 0.11.x and 0.10.x versions, we recommend using dedicated - flink-connector-kafka-0.11{{ site.scala_version_suffix }} and flink-connector-kafka-0.10{{ site.scala_version_suffix }} respectively. - </td> - </tr> - </tbody> -</table> - -Then, import the connector in your maven project: +## Dependency + +Apache Flink ships with multiple Kafka connectors; universal, 0.10, and 0.11. +This universal Kafka connector attempts to track the latest version of the Kafka client. +The version of the client it uses may change between Flink releases. +Modern Kafka clients are backwards compatible with broker versions 0.10.0 or later. +For most users the universal Kafka connector is the most appropriate. +However, for Kafka 0.11.x and 0.10.x versions, we recommend using dedicated ``0.11`` and ``0.10`` respectively. +For details on Kafka compatibility, please refer to the official [Kafka documentation](https://kafka.apache.org/protocol.html#protocol_compatibility). +<div class="codetabs" markdown="1"> +<div data-lang="universal" markdown="1"> {% highlight xml %} <dependency> - <groupId>org.apache.flink</groupId> - <artifactId>flink-connector-kafka{{ site.scala_version_suffix }}</artifactId> - <version>{{ site.version }}</version> + <groupId>org.apache.flink</groupId> + <artifactId>flink-connector-kafka{{ site.scala_version_suffix }}</artifactId> + <version>{{ site.version }}</version> +</dependency> +{% endhighlight %} +</div> +<div data-lang="011" markdown="1"> +{% highlight xml %} +<dependency> + <groupId>org.apache.flink</groupId> + <artifactId>flink-connector-kafka-011{{ site.scala_version_suffix }}</artifactId> + <version>{{ site.version }}</version> </dependency> {% endhighlight %} - -Note that the streaming connectors are currently not part of the binary distribution. -See how to link with them for cluster execution [here]({{ site.baseurl}}/dev/projectsetup/dependencies.html). - -## Installing Apache Kafka - -* Follow the instructions from [Kafka's quickstart](https://kafka.apache.org/documentation.html#quickstart) to download the code and launch a server (launching a Zookeeper and a Kafka server is required every time before starting the application). -* If the Kafka and Zookeeper servers are running on a remote machine, then the `advertised.host.name` setting in the `config/server.properties` file must be set to the machine's IP address. - -## Kafka 1.0.0+ Connector - -Starting with Flink 1.7, there is a new universal Kafka connector that does not track a specific Kafka major version. -Rather, it tracks the latest version of Kafka at the time of the Flink release. - -If your Kafka broker version is 1.0.0 or newer, you should use this Kafka connector. -If you use an older version of Kafka (0.11 or 0.10), you should use the connector corresponding to the broker version. - -### Compatibility - -The universal Kafka connector is compatible with older and newer Kafka brokers through the compatibility guarantees of the Kafka client API and broker. -It is compatible with broker versions 0.11.0 or newer, depending on the features used. -For details on Kafka compatibility, please refer to the [Kafka documentation](https://kafka.apache.org/protocol.html#protocol_compatibility). - -### Migrating Kafka Connector from 0.11 to universal - -In order to perform the migration, see the [upgrading jobs and Flink versions guide]({{ site.baseurl }}/ops/upgrading.html) -and: -* Use Flink 1.9 or newer for the whole process. -* Do not upgrade the Flink and operators at the same time. -* Make sure that Kafka Consumer and/or Kafka Producer used in your job have assigned unique identifiers (`uid`): -* Use stop with savepoint feature to take the savepoint (for example by using `stop --withSavepoint`)[CLI command]({{ site.baseurl }}/ops/cli.html). - -### Usage - -To use the universal Kafka connector add a dependency to it: - +</div> +<div data-lang="010" markdown="1"> {% highlight xml %} <dependency> - <groupId>org.apache.flink</groupId> - <artifactId>flink-connector-kafka{{ site.scala_version_suffix }}</artifactId> - <version>{{ site.version }}</version> + <groupId>org.apache.flink</groupId> + <artifactId>flink-connector-kafka-010{{ site.scala_version_suffix }}</artifactId> + <version>{{ site.version }}</version> </dependency> {% endhighlight %} +<span class="label label-danger">Attention</span> The ``0.10`` sink does not support exactly-once writes to Kafka. +</div> +</div> -Then instantiate the new source (`FlinkKafkaConsumer`) and sink (`FlinkKafkaProducer`). -The API is backward compatible with the Kafka 0.11 connector, -except of dropping specific Kafka version from the module and class names. +Flink's streaming connectors are not currently part of the binary distribution. +See how to link with them for cluster execution [here]({{ site.baseurl}}/dev/projectsetup/dependencies.html). ## Kafka Consumer -Flink's Kafka consumer is called `FlinkKafkaConsumer010` (or 011 for Kafka 0.11.0.x versions, etc. -or just `FlinkKafkaConsumer` for Kafka >= 1.0.0 versions). It provides access to one or more Kafka topics. +Flink's Kafka consumer - `FlinkKafkaConsumer` (or `FlinkKafkaConsumer011` for Kafka 0.11.x version, +or just `FlinkKafkaConsumer010` for Kafka 0.10.x versions) - provides access to read from one or more Kafka topics. Review comment: ```suggestion Flink's Kafka consumer - `FlinkKafkaConsumer` (or `FlinkKafkaConsumer011` for Kafka 0.11.x, or `FlinkKafkaConsumer010` for Kafka 0.10.x) - provides access to read from one or more Kafka topics. ``` ########## File path: docs/dev/connectors/kafka.md ########## @@ -23,125 +23,59 @@ specific language governing permissions and limitations under the License. --> +Flink provides an [Apache Kafka](https://kafka.apache.org) connector for reading data from and writing data to Kafka topics with exactly-once guaruntees. + * This will be replaced by the TOC {:toc} -This connector provides access to event streams served by [Apache Kafka](https://kafka.apache.org/). - -Flink provides special Kafka Connectors for reading and writing data from/to Kafka topics. -The Flink Kafka Consumer integrates with Flink's checkpointing mechanism to provide -exactly-once processing semantics. To achieve that, Flink does not purely rely on Kafka's consumer group -offset tracking, but tracks and checkpoints these offsets internally as well. - -Please pick a package (maven artifact id) and class name for your use-case and environment. -For most users, the `FlinkKafkaConsumer010` (part of `flink-connector-kafka`) is appropriate. - -<table class="table table-bordered"> - <thead> - <tr> - <th class="text-left">Maven Dependency</th> - <th class="text-left">Supported since</th> - <th class="text-left">Consumer and <br> - Producer Class name</th> - <th class="text-left">Kafka version</th> - <th class="text-left">Notes</th> - </tr> - </thead> - <tbody> - <tr> - <td>flink-connector-kafka-0.10{{ site.scala_version_suffix }}</td> - <td>1.2.0</td> - <td>FlinkKafkaConsumer010<br> - FlinkKafkaProducer010</td> - <td>0.10.x</td> - <td>This connector supports <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-32+-+Add+timestamps+to+Kafka+message";>Kafka messages with timestamps</a> both for producing and consuming.</td> - </tr> - <tr> - <td>flink-connector-kafka-0.11{{ site.scala_version_suffix }}</td> - <td>1.4.0</td> - <td>FlinkKafkaConsumer011<br> - FlinkKafkaProducer011</td> - <td>0.11.x</td> - <td>Since 0.11.x Kafka does not support scala 2.10. This connector supports <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-98+-+Exactly+Once+Delivery+and+Transactional+Messaging";>Kafka transactional messaging</a> to provide exactly once semantic for the producer.</td> - </tr> - <tr> - <td>flink-connector-kafka{{ site.scala_version_suffix }}</td> - <td>1.7.0</td> - <td>FlinkKafkaConsumer<br> - FlinkKafkaProducer</td> - <td>>= 1.0.0</td> - <td> - This universal Kafka connector attempts to track the latest version of the Kafka client. - The version of the client it uses may change between Flink releases. Starting with Flink 1.9 release, it uses the Kafka 2.2.0 client. - Modern Kafka clients are backwards compatible with broker versions 0.10.0 or later. - However for Kafka 0.11.x and 0.10.x versions, we recommend using dedicated - flink-connector-kafka-0.11{{ site.scala_version_suffix }} and flink-connector-kafka-0.10{{ site.scala_version_suffix }} respectively. - </td> - </tr> - </tbody> -</table> - -Then, import the connector in your maven project: +## Dependency + +Apache Flink ships with multiple Kafka connectors; universal, 0.10, and 0.11. Review comment: ```suggestion Apache Flink ships with multiple Kafka connectors: universal, 0.10, and 0.11. ``` ########## File path: docs/dev/connectors/kafka.md ########## @@ -23,125 +23,59 @@ specific language governing permissions and limitations under the License. --> +Flink provides an [Apache Kafka](https://kafka.apache.org) connector for reading data from and writing data to Kafka topics with exactly-once guaruntees. + * This will be replaced by the TOC {:toc} -This connector provides access to event streams served by [Apache Kafka](https://kafka.apache.org/). - -Flink provides special Kafka Connectors for reading and writing data from/to Kafka topics. -The Flink Kafka Consumer integrates with Flink's checkpointing mechanism to provide -exactly-once processing semantics. To achieve that, Flink does not purely rely on Kafka's consumer group -offset tracking, but tracks and checkpoints these offsets internally as well. - -Please pick a package (maven artifact id) and class name for your use-case and environment. -For most users, the `FlinkKafkaConsumer010` (part of `flink-connector-kafka`) is appropriate. - -<table class="table table-bordered"> - <thead> - <tr> - <th class="text-left">Maven Dependency</th> - <th class="text-left">Supported since</th> - <th class="text-left">Consumer and <br> - Producer Class name</th> - <th class="text-left">Kafka version</th> - <th class="text-left">Notes</th> - </tr> - </thead> - <tbody> - <tr> - <td>flink-connector-kafka-0.10{{ site.scala_version_suffix }}</td> - <td>1.2.0</td> - <td>FlinkKafkaConsumer010<br> - FlinkKafkaProducer010</td> - <td>0.10.x</td> - <td>This connector supports <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-32+-+Add+timestamps+to+Kafka+message";>Kafka messages with timestamps</a> both for producing and consuming.</td> - </tr> - <tr> - <td>flink-connector-kafka-0.11{{ site.scala_version_suffix }}</td> - <td>1.4.0</td> - <td>FlinkKafkaConsumer011<br> - FlinkKafkaProducer011</td> - <td>0.11.x</td> - <td>Since 0.11.x Kafka does not support scala 2.10. This connector supports <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-98+-+Exactly+Once+Delivery+and+Transactional+Messaging";>Kafka transactional messaging</a> to provide exactly once semantic for the producer.</td> - </tr> - <tr> - <td>flink-connector-kafka{{ site.scala_version_suffix }}</td> - <td>1.7.0</td> - <td>FlinkKafkaConsumer<br> - FlinkKafkaProducer</td> - <td>>= 1.0.0</td> - <td> - This universal Kafka connector attempts to track the latest version of the Kafka client. - The version of the client it uses may change between Flink releases. Starting with Flink 1.9 release, it uses the Kafka 2.2.0 client. - Modern Kafka clients are backwards compatible with broker versions 0.10.0 or later. - However for Kafka 0.11.x and 0.10.x versions, we recommend using dedicated - flink-connector-kafka-0.11{{ site.scala_version_suffix }} and flink-connector-kafka-0.10{{ site.scala_version_suffix }} respectively. - </td> - </tr> - </tbody> -</table> - -Then, import the connector in your maven project: +## Dependency + +Apache Flink ships with multiple Kafka connectors; universal, 0.10, and 0.11. +This universal Kafka connector attempts to track the latest version of the Kafka client. +The version of the client it uses may change between Flink releases. +Modern Kafka clients are backwards compatible with broker versions 0.10.0 or later. +For most users the universal Kafka connector is the most appropriate. +However, for Kafka 0.11.x and 0.10.x versions, we recommend using dedicated ``0.11`` and ``0.10`` respectively. Review comment: ```suggestion However, for Kafka versions 0.11.x and 0.10.x, we recommend using the dedicated ``0.11`` and ``0.10`` connectors, respectively. ``` ########## File path: docs/dev/connectors/kafka.md ########## @@ -216,46 +139,37 @@ For convenience, Flink provides the following schemas: <div data-lang="AvroDeserializationSchema" markdown="1"> {% highlight xml %} <dependency> - <groupId>org.apache.flink</groupId> - <artifactId>flink-avro</artifactId> - <version>{{site.version }}</version> + <groupId>org.apache.flink</groupId> + <artifactId>flink-avro</artifactId> + <version>{{site.version }}</version> </dependency> {% endhighlight %} </div> <div data-lang="ConfluentRegistryAvroDeserializationSchema" markdown="1"> {% highlight xml %} <dependency> - <groupId>org.apache.flink</groupId> - <artifactId>flink-avro-confluent-registry</artifactId> - <version>{{site.version }}</version> + <groupId>org.apache.flink</groupId> + <artifactId>flink-avro-confluent-registry</artifactId> + <version>{{site.version }}</version> </dependency> {% endhighlight %} </div> </div> -When encountering a corrupted message that cannot be deserialized for any reason, there -are two options - either throwing an exception from the `deserialize(...)` method -which will cause the job to fail and be restarted, or returning `null` to allow -the Flink Kafka consumer to silently skip the corrupted message. Note that -due to the consumer's fault tolerance (see below sections for more details), -failing the job on the corrupted message will let the consumer attempt -to deserialize the message again. Therefore, if deserialization still fails, the -consumer will fall into a non-stop restart and fail loop on that corrupted -message. +When encountering a corrupted message that cannot be deserialized for any reason the deserialization schema should return null which will result in the record being skipped. +Due to the consumer's fault tolerance (see below sections for more details), failing the job on the corrupted message will let the consumer attempt to deserialize the message again. +Therefore, if deserialization still fails, the consumer will fall into a non-stop restart and fail loop on that corrupted message. ### Kafka Consumers Start Position Configuration -The Flink Kafka Consumer allows configuring how the start position for Kafka -partitions are determined. - -Example: +The Flink Kafka Consumer allows configuring how the start position for Kafka partitions are determined. Review comment: ```suggestion The Flink Kafka Consumer allows configuring how the start positions for Kafka partitions are determined. ``` ########## File path: docs/dev/connectors/kafka.md ########## @@ -152,16 +86,14 @@ The constructor accepts the following arguments: - "bootstrap.servers" (comma separated list of Kafka brokers) - "group.id" the id of the consumer group Review comment: I believe the consumer group id is optional. Do we want to get into that topic here? ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
