SOA Composte retries intermittently after hitting error

Question

Hi,
The business flow looks like this
AQ adapter -> SOA Compsoite -> target system (DB calls)
Messages comes to AQ and SOA composite picks the message from AQ and makes a call to the target system .All data source participate in global transaction, if any error happens the message would get rolled backed to Error Queue
Issue
Whenever the target system is down, the corresponding event adapter that picks messages from AQ intermittently retries the message n number of times.
The retry count is random in some cases it kept going even after 40 retries.
Can you please suugest, is there any option to stop the retry count

Answer 1

Hi Alfredo/Dhaval,
Thanks for the update.
I have followed all the steps that you have mentioned , but still face the issue .
The reason for multiple retry is , In AQ the
state 0 means ready to be picked up,
state 2 means successfully completed , so none of the threads will pick this message
state 3 means faulted and none of threads will pick the message .
The time taken by AQ to change from state 0 to 2/3 is dependent on BPEL transaction .
In this scenario, bpel takes 5 mins to commit the transaction (as JTA is set to 5 mins),
which means the message state in AQ would be 0 for 5 mins(0 means ready to Pick) , in mean time other thread would pick the same message
(mindelaybetweenmessage kept as 30 secs, i cant keep it as 5 mins because of performance issue) and this is happening in a loop.
My doubt is , when a target system is down , ideally bpel should throw remote fault exception , but here the flow does not go to catch or catch all block
but instead the flow is stopped in Invoke where it makes a call to target system and after 5 mins JTA transaction time out errored message is occuring .