Hello, I think you can achieve what you want with the JMSSourceConverter. Looking at the Flume documentation it says that "The JMS source allows pluggable converters, though it’s likely the default converter will work for most purposes. The default converter is able to convert Bytes, Text, and Object messages to FlumeEvents. In all cases, the properties in the message are added as headers to the FlumeEvent".
Perhaps you can implement the JSMMessageConverter interface ( https://github.com/apache/flume/blob/dfa0627573b9a75a25dc7149a7d63c9bac953ff4/flume-ng-sources/flume-jms-source/src/main/java/org/apache/flume/source/jms/JMSMessageConverter.java). The default converted also implements this interface ( https://github.com/apache/flume/blob/dfa0627573b9a75a25dc7149a7d63c9bac953ff4/flume-ng-sources/flume-jms-source/src/main/java/org/apache/flume/source/jms/DefaultJMSMessageConverter.java) and also provides a good reference point for implementing you own. You custom code would go into plugins.d directory in flume. More details here: https://flume.apache.org/FlumeUserGuide.html#jms-source On Thu, Aug 31, 2017 at 12:05 PM, Kajal K Das <kajal....@in.ibm.com> wrote: > Hi All, > > We are designing a near-real time pipeline in which we will receive > messages from IBM MQ. On the Hadoop side, we are planning to use Flume to > ingest the messages using JMS source. > The messages will include some custom header information, which we need to > extract as well. We discovered that interceptors cannot be configured with > a JMS source. > In that case: what are the alternatives in order to extract the message > headers using Flume? Or should we try something else (e.g. Kafka)? > > > > Regards > Kajal Kr. Das > >
