[jira] [Commented] (CAMEL-12031) KafkaConsumer stops consuming messages when exception occurs during offset commit

["Claus Ibsen (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CAMEL-12031?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16265771#comment-16265771 ] 

Claus Ibsen commented on CAMEL-12031:
-------------------------------------

Thanks for reporting yeah we can do a re-connect if kafka throws one of its exception. You are welcome to work on a patch as github PR

> KafkaConsumer stops consuming messages when exception occurs during offset commit
> ---------------------------------------------------------------------------------
>
>                 Key: CAMEL-12031
>                 URL: https://issues.apache.org/jira/browse/CAMEL-12031
>             Project: Camel
>          Issue Type: Bug
>          Components: camel-kafka
>    Affects Versions: 2.20.0
>            Reporter: Rafał Gała
>             Fix For: 2.20.2, 2.21.0
>
>
> When processing of messages takes longer than max session timeout, the consumer thread will end after receiving the *org.apache.kafka.clients.consumer.CommitFailedException*.
> {code:java}
>        @Override
>         public void run() {
>             boolean first = true;
>             boolean reConnect = true;
>             while (reConnect) {
>                 // create consumer
>                 ClassLoader threadClassLoader = Thread.currentThread().getContextClassLoader();
>                 try {
>                     // Kafka uses reflection for loading authentication settings, use its classloader
>                     Thread.currentThread().setContextClassLoader(org.apache.kafka.clients.consumer.KafkaConsumer.class.getClassLoader());
>                     this.consumer = new org.apache.kafka.clients.consumer.KafkaConsumer(kafkaProps);
>                 } finally {
>                     Thread.currentThread().setContextClassLoader(threadClassLoader);
>                 }
>                 if (!first) {
>                     // skip one poll timeout before trying again
>                     long delay = endpoint.getConfiguration().getPollTimeoutMs();
>                     log.info("Reconnecting {} to topic {} after {} ms", threadId, topicName, delay);
>                     try {
>                         Thread.sleep(delay);
>                     } catch (InterruptedException e) {
>                         Thread.currentThread().interrupt();
>                     }
>                 }
>                 first = false;
>                 // doRun keeps running until we either shutdown or is told to re-connect
>                 reConnect = doRun();
>             }
>         }
> {code}
> The *doRun()* method returns false and the loop ends. It should be possible to let the proces continue after failed offset commit.
> I think the catch block inside *doRun* method should look like this:
> {code:java}
>            ...
>             } catch (InterruptException e) {
>                 getExceptionHandler().handleException("Interrupted while consuming " + threadId + " from kafka topic", e);
>                 log.info("Unsubscribing {} from topic {}", threadId, topicName);
>                 consumer.unsubscribe();
>                 Thread.currentThread().interrupt();
>             } catch (org.apache.kafka.clients.consumer.CommitFailedException e) { //or even org.apache.kafka.common.KafkaException
>                 getExceptionHandler().handleException("Error consuming " + threadId + " from kafka topic", e);
>                 reConnect = true;
>             } catch (Exception e) {
>                 getExceptionHandler().handleException("Error consuming " + threadId + " from kafka topic", e);
>             } finally {
>                 log.debug("Closing {} ", threadId);
>                 IOHelper.close(consumer);
>             }
>             ...
> {code}



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)