for clientAck, that is needed, have a look at: org.apache.activemq.ActiveMQMessageConsumer#dispatch
The automatic dlq processing occurs if onMessage throws an exception, but it is bypassed for transactions and clientack mode as the acking is out of control of the consumer in that case. If you use autoack or activemq's individual ack mode, you will get auto dlq processing (the PoisonAck) if onMessage throws a runtime exception. On 31 May 2011 17:02, rasitha1 <rasitha.wijesin...@gmail.com> wrote: > Hi, > > Can you please expand a bit on this? > > "Typically you NAK a message by sending it to some bad message > destination; i.e. you still consume it." > > I'm looking to use "individualDeadLetterStrategy" to separate failed > messages into separate DLQs. But my client code (when using > ClientAcknowledge mode) just creates a message manually to the default DLQ. > (which I'm not sure is the right way) > > Looking at the source I see you need to send a special message > (AckType=PoisonAck) to tell the broker that the message failed. How should I > go about doing that? > > At the end I need to get failed messages into separate DLQs while using > ClientAcknowledge mode. > > Thanks, > Ras > > > > -- > View this message in context: > http://activemq.2283324.n4.nabble.com/Dead-Letter-Queue-tp2356685p3563553.html > Sent from the ActiveMQ - User mailing list archive at Nabble.com. > -- http://fusesource.com http://blog.garytully.com