Skip to Content

SAP PI Adapter User defined Message Search

Dec 09, 2016 at 08:35 AM


avatar image


I have configured UDMS and it works fine as such. But so far I only found the button to manually indexing the messages that are already processed. What needs to be configured that every new message will be indexed as well ? Is there a background job that I need to setup or something else to be activated.

I have seen in several blocks that on the ABAP stack there is some config, but that looks like not applicable for the Java Stack ;-(



10 |10000 characters needed characters left characters exceeded
* Please Login or Register to Answer, Follow or Comment.

3 Answers

Best Answer
Vadim Klimov Dec 09, 2016 at 12:03 PM

Hello Helmut,

This is one of key differences between implementation of Lean Message Search / User-Defined Message Search functionality in Integration Engine (ABAP) and Adapter Engine (Java): in Integration Engine, there are various options for extractor execution - which are namely, execute it during message processing (synchronous extractor execution) or by scheduled background job for processed and persisted messages (asynchronous extractor execution), whereas in Adapter Engine, extractors of configured and active UDMS filters are only applied during message processing (and, as mentioned by Evgeniy, can be applied retrospectively for already processed and still persisted messages by means of indexing job).

There are pros and cons of having possibility of asynchronous extractor execution (in Integration Engine). Major advantage is that it doesn't impact message processing time (in contrast to synchronous extractor execution, which becomes one of message processing steps) - that one makes certain sense for message processing performance optimization especially in cases when many extractors are applied to the processed message, or when enabling UDMS for interfaces processing large messages, or combination of both. Payoff for performance gain resulting in decoupling message processing from UDMS indexing, is delayed search: newly processed message will not show up in UDMS search results until scheduled background indexing job is completed. In Adapter Engine, for some reasons, asynchronous extractor execution was not included in UDMS functionality, causing lack of corresponding job scheduling option that you pointed out.



10 |10000 characters needed characters left characters exceeded
Evgeniy Kolmakov Dec 09, 2016 at 09:42 AM

Hi Helmut!

As far as I know you don't need to configure any jobs for that. When you activate your UDMS, all newly processed messages are indexed automaticaly. The button, in turn, gives you the ability to index messages that have been processed before you activated UDMS.

Regards, Evgeniy.

10 |10000 characters needed characters left characters exceeded
Helmut Skolaut Jan 18, 2017 at 06:55 AM

Hi Vadim, Hi Evgeniy,

thanks for both of you about your input. Since the new SDN platform I have not been alerted about your comments and have not seen them earlier. Sorry for late feedback.

I can confirm that there is no job to be configured. I don't find the reason why in the beginning the new instances have not been UDMS indexed automatically, but I have not changed anything (except maybe a reboot?) but I now works suddenly.

Regards Helmut

10 |10000 characters needed characters left characters exceeded