Job getting failed - CD Module

[Former Member]

Hi Guys,

I'm finding an issue in Prod systems where one of the jobs is getting failed frequently.

There is a program which selects data from the table DFKKKO. I'm pasting the portion of the code, where the job is actually geting cancelled.

*******************************************************************************

005600 DATA: BEGIN OF t_dfkkko occurs 0,

005610 opbel like dfkkko-opbel,

005620 storb like dfkkko-storb,

005630 cputm like dfkkko-cputm,

005640 END OF t_dfkkko.

005650

005660 DATA: t_dfkkko2 like t_dfkkko occurs 0 with header line.

005670

005680 select opbel

005690 storb

005700 from dfkkko

005710 into table t_dfkkko

005720 for all entries in t_itab1

005730 where opbel = t_itab1-doc1r+0(12).

005740

-

---

> select

005760 opbel

005770 cputm

005780 from dfkkko

005790 into corresponding fields of table t_dfkkko2

005800 for all entries in t_dfkkko

005810 where opbel = t_dfkkko-storb.

In ST22, I'm getting the following error message:

Runtime errors <b>TSV_TNEW_BLOCKS_NO_ROLL_MEMORY</b>

I believe it has something to do with the low memory issue. But I dont know how to tackle it. Will refreshing all the internal tables do?

Does anybody has a suitable answer for this?

Let me know if you want me to send the whole code as a word doc file.

Thanks in advance.

Pritam

[Former Member]

Hi Pritam,

The error might be at the first SELECT, SAP often shows the next statement as the one it stops at.

Look at DELETEing ADJACENT DUPLICATES on table T_ITAB1- you may be reading the same data more than once.

[Former Member]

Before both SELECTs, you have to make sure that the tables used in FOR ALL ENTRIES has values. For the first SELECT:

CHECK NOT t_itab1[] IS INITIAL.

and for the other one:

CHECK NOT t_dfkkko[] IS INITIAL.

Rob

[Former Member]

Hi Pritam,

Did you checked the buffer size, is it enough big.

And also try clearing buffer memory before running ur program.

Hope this helps.

- Hitesh

[Former Member]

In addition to the other hints already given:

- Trace the program with ST05 to see the SQL code generated (select line and press "Explain").

- if the query is fine but the amount of data is simply too big, use the PACKAGE SIZE statement and process your data in chunks. 50.000 is usually a good package size.

Hope that helps

Frank

[Former Member]

It was a memory issue and the basis guys were able to reolve it. Thanks a lot for your inputs.