HRMD_A ALE Problems

Question

A while back I set up HR Master Data distribution. Initially, the client was running HR on the same server as FI, MM, SD etc. They wanted to move to a new HR server for security reasons.
I set up the ALE in order to connect the HR system to the R/3 system. Among others, i used message type HRMD_A. A problem I have come across is that sometimes the ALE tries to send objects across in the incorrect order, and the IDocs fail.
For example, if I create a new position, and then make some changes to that position, or assign relationships etc - The ALE sometimes tries to send the changes or relationships before it sends the actual position - This of course causes an error on the receiving system because the object does not exist when the changes are being attempted...
What are the steps that I should make sure have been taken to ensure no errors in the ALE?
What are the background jobs that I should have scheduled etc?

Answer 1

Dear David/Manas,
Manas : Please refer to note 752194 and also the documentation in SE38 of report RHALEINI. Serialization can assure that IDOCs are processed with the correct order even if customer only use message type HRMD_A.
David :
2.1 Re-distribute all the data just to make sure that all is in sync. Is that correct?
- Yes you need to make sure your initial distribution has been done correctly.
2.2 Before I run transaction PFAL, I assume that I need to de-activate change pointers u2013 or not?
- No you do not need to deactivate the change pointers.
2.3 I then follow your steps 1-4
3.1 After that, I should set up the serialization for sending and receiving on the box:
3.2 Which system must I do this on u2013 R/3 or HR?
- You need to install the serialization in your source system (where PFAL will be running). please refer to the documentation of report RHALEINI for further info.
3.3 I assume that I will need to "Define New Serialization Groups" for HRMD_A?
- Refer to note 752194 and the documentation of report RHALEINI. It will guide you in this process.
3.4 The distribution model has already been maintained as mentioned in point 1.1. Do I need to add info-type 1028 to the filter?
-  Unless you need this infotype
3.5 I will then need to u201CDefine Inbound Processingu201D.
For inbound:
-> go to transaction WE20 -> Select Partner
Select Inbound Message Type and change the processing method from
"Trigger Immediately" to "Trigger by background program".
3.6 Lastly I will need to create a variant for program RBDSER01 and schedule a background job u2013 How often should this job run? Which server should this job be run on? R/3 or HR?
This report is used when you are using different application servers and different message types. So in your scenario, you can use report RBDMIDOC to create your idoc from change pointers and then run report RSEOUT00 to send these idocs to the target system.
Hopefully I clarify your request,
Best Regards
Christine