[
https://issues.apache.org/jira/browse/CAMEL-9467?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15171575#comment-15171575
]
Claus Ibsen commented on CAMEL-9467:
------------------------------------
Anbu
We are getting closer. I wonder if you can make scala only for testing, eg
{code}
<groupId>org.apache.kafka</groupId>
<artifactId>kafka_2.11</artifactId>
<version>${kafka-version}</version>
{code}
should have
{code}
<scope>test</scope>
{code}
The idea is that camel-kafka is only using the pure Java client of Kafka - no
Scala needed.
If that is possible then we are getting much closer.
In the consumer I think some logic is needed to keep the consumer pooling. You
only do 1 poll and then process the returned records. But if there is new
records in the future, then they are not polled?
Take a look at the class javadoc of the KafkaConsumer how it uses a Runnable to
do the polling. We would need someting similar, to setup a thread pool for
those worker threads that does the polling.
The old code had this in the BatchingConsumerTask, though make it simpler, and
just do like the class javadoc from kafka client.
> Camel-Kafka: Upgrade Producer/Consumer and their configuration properties
> -------------------------------------------------------------------------
>
> Key: CAMEL-9467
> URL: https://issues.apache.org/jira/browse/CAMEL-9467
> Project: Camel
> Issue Type: Task
> Components: camel-kafka
> Reporter: Andrea Cosentino
> Assignee: Andrea Cosentino
> Fix For: 2.17.0
>
> Attachments: camel-kafka.zip
>
>
> From release 0.9.0.0 the Kafka Consumer/Producer configuration properties
> seems to be changed a little.
> Also we have to use the new Java implementation of Kafka Producer/Consumer,
> actually we are using the Scala version
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
