"Buffered" Multiple Queues (Content-Specific) ccBPM

[Former Member]

Hello all,

If I try to configure the inbound processing (SWF_INB_CONF) for an integration process as "Buffered possible" and Multiple Queues (Content-Specific) I get the following message.

"Combination of delivery mode Buffering Possible and queue handling Multiple Queues (Content-

Specific) is not permitted"

Could anyone explain me, why this is not permitted? Does anyone know a workaround?

BR,

Rene.

[samiullah_qureshi]

Hi,

Incase if there are two messages in two different queues they both would be waiting for the same buffer. Instead the messages should be directly delivered to the process instance.

I dont think XI has a provision for the messages to wait in queue for a buffer.

Also you can refer following blog to better understand buffering in XI:-

[https://www.sdn.sap.com/irj/sdn/weblogs?blog=/pub/wlg/10195] [original link is broken] [original link is broken] [original link is broken];

Regards,

Sami.

[bhavesh_kantilal]

hello Rene

Any Special reason you need to have Buffering possible?

I will try to give my understanding of why what you are trying does not work -

When you have a Single Queue - Message would go to the single Queue and then are then sent across to Process Specific Buffer.

In your case you want to combine multiple Queues ( Content Specific and hence I assume you have a Correlation ). In this case, multiple queues would be created in the runtime depending on the message / correlation and hence having a buffer after this doesn't make much sense to me.

Maybe there is a better answer, or rather a better way to explain this, and I would love to hear it as well but just that to me this combination doesn't seem to make sense.

Regards

Bhavesh

[Former Member]

Hi Bhavesh,

yes i use a correlation and for performance reasons i have to use multiple queues (content specific). The messages for a specific process instance will not arrive in a specific order, so I have to buffer the messages after correlation for each process instance (each process instance should has a seperate queue).

How can i solve this problem?

Thx for help!

BR,

Rene.

[bhavesh_kantilal]

Guess one of the two,

1. Move the correlating Receive Steps to multiple branches of a Fork Step.

2., If 1 is not possible; then you will have to use a Single Queue!

Can you provide the process of your BPM so we can possibly look at way to see if Buffering is not needed!

Regards

Bhavesh