WS inevoked, but intermediate message event still waiting

[former_member207065]

Hi experts,

In my process, an intermediate message event is used to wait for a non-SAP system's notification. Occasionally the intermediate message event seemed not receiving the message and was still waiting after its WS invoked by the non-SAP system. This caused inconsistency. I want to know why, and how to avoid this problem?

Regards,

Abe

[former_member207065 · Accepted solution]

Hi all,

Sorry for the late reply. I will try Manish's solution to guarantee the messaging. In addition, I found the latest BPM API had provided capability of sending messages to IME, see MessageManager. This is also a good workaround.

Regards,

Abe

[former_member207065]

I have read thisUnderstanding Intermediate Events, Asynchronous... | SCN

and found WS-RM can make messaging reliable. But this doc also says "Using WS-RM has a performance impact" and "WS-RM cannot be used for local transport in one physical system".

So is there any other solution to achieve reliable messaging?

[Former Member]

That may be one of the reason. BPM engine by default is not a guranteed delivery server like PI where all asynchronous msgs are guranteed. In BPM even if your bpm webserver accepts the msg(will be give a 200 ok status to caller) it all depends on process instance available. If no process instance is availble or for some other reason cannot instantiate one the msg will be lost, the only place where you will see a trace might be in soa logs or developer traces. Now you have two options

1)If you are on 731/74 version route yr msg through PI and use XI3.0 protocol which makes it guaranteed.

2)use ws-rm protocol but then the caller should have capability to send ws-rm.

[junwu]

if the bpm server is down, the ws call should be failed. do they have proper error handling at the calling side?

[former_member207065]

Thank you Jun. The server has not gone down when this happened, just before down. So the WS call succeeded and did not throw any exception.

[junwu]

can u help to understand what is before down??????? to me "before down" means "online".....

or you mean, a few minutes before down?

[junwu]

and how do u know it happened just before down?

[ROHITGERA1]

Hi Abe,

I think for this you have to give the Corelation Condition to the Intermediate Message Event and check weather the Intermediate message event is receiving that notification or not.

you can give the Corelation Condition in the Property of the Intermediate Message Event.

Thanks,

Rohit

[former_member207065]

Hi Rohit,

Thank you. But I have given the Correlation Condition to the Intermediate Message Event.

Thanks,

Abe

[former_member207065]

And this problem only occurred before our production system went down. Is there any way to monitor the messages of the intermediate message event?

[ROHITGERA1]

Hi Abe,

You can check the web service manually for the Intermediate message event through WS navigator.

you have to confirm weather web service is invoking properly or not
or there is some error when invoking the web service ,

and check the BPM History also in the Process Instance in NWA.

Thanks,

Rohit

[former_member207065]

Thank you. I checked all. No error or exception. Is there any workaround to guarantee the receiving of IME's message. Because IME's WS is asynchronous, other system can not get a response to confirm IME is triggerred successfully.