Query on the pipeline flow of message in PI

[Former Member]

Hi Experts

                            At the time of sending an information from one system to another , the normal flow is : adapter Engine -> integration engine -> adapter engine. Incase there is an issue in reaching the target system , PI gives a retry  for 3 times normally, and the same gets reflected in the log of message monitoring of adapter engine.

  What I wish to know is at this point of RETRY, how, and from what point is this process carried out. And at the time of retry , does it start just after the integration engine  , like identifying the receiver agreement etc?

    Also , does it take the same node of the cluster , which it took at the first attempt  ?  I a raising the 2nd question because , i am facing a scenerio  where idoc information is sent to a webservice . The issue is , we are sending the same info again and again , and someties , its hitting the URL and sometimes its givibg a socket error. The strange fact is same message , same inteface, everything same , but sometimes the message is successful and sometimes it failes , and when it FAILS, it never PASSES even after multiple retries.

Guys , if you have any further queries, pls come back, but this issue is really taking huge escalation !!

Error Details :

*************************************************************************************

1. 15.11.2012 16:12:42.053        Information             MP: processing local module localejbs/sap.com/com.sap.aii.af.soapadapter/XISOAPAdapterBean
2. 15.11.2012 16:12:42.058        Information             SOAP: request message entering the adapter with user J2EE_GUEST
3. 15.11.2012 16:12:42.294        Error        SOAP: call failed: java.net.SocketException: Connection reset
4. 15.11.2012 16:12:42.294        Information             SOAP: sending a delivery error ack ...
5. 15.11.2012 16:12:42.294        Information             SOAP: sent a delivery error ack
6. 15.11.2012 16:12:42.295        Error        SOAP: error occured: com.sap.engine.interfaces.messaging.api.exception.MessagingException: java.net.SocketException: Connection reset
7. 15.11.2012 16:12:42.295        Error        MP: exception caught with cause com.sap.engine.interfaces.messaging.api.exception.MessagingException: java.net.SocketException: Connection reset
8. 15.11.2012 16:12:42.295        Error        Adapter Framework caught exception: java.net.SocketException: Connection reset
9. 15.11.2012 16:12:42.296        Error        Delivering the message to the application using connection SOAP_http://sap.com/xi/XI/System failed, due to: com.sap.engine.interfaces.messaging.api.exception.MessagingException: java.net.SocketException: Connection reset.
10. 15.11.2012 16:12:42.308        Error        The message status was set to NDLV.

*************************************************************************************

Regards,

Arnab.

[Harish]

Hi Arnab,

RETRY will start the message from the current state, so if message failed in Interface determination in RETRY it will start from Interface determination.

your question regarding the socket error some time is seems to be due to network connectivity.

do you have multiple server of PI and having different IP address.

If you PI have multiple server with different IP address then it might be issue with Firewall is not open for all the server at your end or parter'e end.

Regards,

Harish

[Former Member]

In my case , I do not have multiple servers of PI , but yes , I do have multiple cluster nodes. Having said that that, as we know , node sharng is a load sharing process, and why is that this problem NEVER aloows a RETRY message to be EVER succesful. However using same payload and all other features same , why is it that out of 10 times , at least 4-5 times , it is SUCCESSFUL?