Runtime Error in POSTING_INTERFACE_CLEARING

[sid_07_]

Hello Experts,

I am using POSTING_INTERFACE_CLEARING FM for doing mass clearing of account receivables and account payable. I have used this FM inside an Inbound Idoc FM. The Idoc will be triggered immediately.

The POSTING_INTERFACE_CLEARING is called inside a loop on Idoc data records. The clearing FM will be called twice for each iteration.

When I tried to post 5 thousand records( clearing FM will be called 10K times), It took around 2 hours for this processing and last 200 records were not processed due to below runtime errors.

1. Runtime error SYSTEM_NO_ROLL has occurred
2. Runtime error TSV_TNEW_OCCURS_NO_ROLL_MEMORY has occurred
3. Runtime error LOAD_NO_SPACE_FOR_TABLE has occurred
4. Runtime error SYSTEM_NO_MEMORY has occurred

These errors occurred in POSTING_INTERFACE_CLEARING and not in the Inbound IDOC FM.

I checked with the Basis, they are asking to split the records across multiple Idocs and the solution should work.

But my question is 10K records is not a big number and it is not a time out issue as well. Also I am clearing all the variables in my inbound FM at appropriate place. Is there anything else which I am missing.

[FredericGirod]

When did you do the commit in the Idoc customizing ?

[sid_07_]

frdric.girod the commit is being done in the POSTING_INTERFACE_CLEARING FM itself. I get the new document no. in the exporting parameter of this FM.

[RaymondGiuseppi]

'These errors occurred in POSTING_INTERFACE_CLEARING and not in the Inbound IDOC FM.' yes but AFAIK this FM clear its memory at start of each call, so the increasing memory consumption should come from the caller (so your IDOC processing)

Did you try to analyze memory with tools such as Abap memory inspector or in the dump data?

[sid_07_]

raymond.giuseppi I checked the dump data, but could not find any table with large no. of records. The memory analysis I could do only on a test data with very few records.

Regarding the increasing memory consumption on IDOC processing, yes there is an internal table to which records are appended at the end of each iteration but still the Idoc processing FM is executed for all the 5K records as I can see the logs are updated with runtime error message.

So I believe the runtime error is occurring for that clearing FM execution only.