Thanks Gwen. I went with Confluent 2.0 as it has Kakfa 0.9 that matches
with that in HDP 2.4. I installed confluent-kafka-connect-hdfs and
confluent-common and softlinked a couple jar into kafka libs/.

I was able to start Kafka Connect but kafka.out was showing the following
error:

[2016-08-05 20:57:01,187] ERROR Exception emitting metrics
(org.apache.hadoop.metrics2.sink.kafka.KafkaTimelineMetricsReporter)
org.apache.hadoop.metrics2.sink.timeline.UnableToConnectException:
java.net.ConnectException: Connection refused
at
org.apache.hadoop.metrics2.sink.timeline.AbstractTimelineMetricsSink.emitMetrics(AbstractTimelineMetricsSink.java:87)
at
org.apache.hadoop.metrics2.sink.kafka.KafkaTimelineMetricsReporter.access$200(KafkaTimelineMetricsReporter.java:58)
at
org.apache.hadoop.metrics2.sink.kafka.KafkaTimelineMetricsReporter$TimelineScheduledReporter.report(KafkaTimelineMetricsReporter.java:253)
at
org.apache.hadoop.metrics2.sink.kafka.ScheduledReporter.report(ScheduledReporter.java:185)
at
org.apache.hadoop.metrics2.sink.kafka.ScheduledReporter$1.run(ScheduledReporter.java:137)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308)
at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180)
at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:745)

Is there any configuration I should also look into? I started with

./connect-standalone.sh ../config/connect-standalone.properties
/etc/kafka-connect-hdfs/quickstart-hdfs.properties

And here is my quickstart-hdfs.properties:

name=hdfs-sink
connector.class=io.confluent.connect.hdfs.HdfsSinkConnector
tasks.max=1
topics=hdfs
hdfs.url=hdfs://sandbox.hortonworks.com:8020
flush.size=3

Thanks,
-B

On Fri, Aug 5, 2016 at 3:31 PM, Gwen Shapira <g...@confluent.io> wrote:

> The installation instructions from Confluent will still work for you :)
>
> If you are using deb/rpm packages, basically add the repositories as
> explained here:
> http://docs.confluent.io/3.0.0/installation.html#rpm-packages-via-yum
>
> and then:
> sudo yum install confluent-kafka-connect-hdfs
> or
> sudo apt-get install confluent-kafka-connect-hdfs
>
> This will put the connector config in /etc/kafka-connect-hdfs and the
> connector jars in /usr/share/java/
>
> You may need to move the jar so it is on the classpath for connect
> (I'm not sure what's the default kafka classpath for HDP).
>
> BTW. We (Confluent) are testing the HDFS connector with HDP (we
> basically install Confluent platform on one machine and HDP on another
> and use Connect to move data) - so this setup should work :)
>
> Gwen
>
>
> On Fri, Aug 5, 2016 at 8:47 AM, Banias H <banias4sp...@gmail.com> wrote:
> > Hi,
> >
> > We are using Hortonworks HDP 2.4 with Apache Kafla 0.9 and we have an
> > in-house solution to pull messages from Kafka to HDFS. I would like to
> try
> > using kakfa-connector-hdfs to push messages to HDFS. As far as I concern,
> > Apache Kafka 0.9 doesn't come with kafka-connector-hdfs. What is a solid
> > way to run kafka-connector-hdfs in HDP? I won't be able to install
> > Confluent platform there though... I would appreciate any pointers.
> Thanks.
> >
> > -B
>
>
>
> --
> Gwen Shapira
> Product Manager | Confluent
> 650.450.2760 | @gwenshap
> Follow us: Twitter | blog
>

Reply via email to