Dump during processing on multiple work process.

Question

Hi ,
I have designed a prohgram "report1" which creates variants for another program "report2" and schedules those variants in back ground on multiple work processess.
When run the "report1", it creates 19 variants for "report2" and those variants will be scheduled for background process on multiple workprocess.
My problem is, when i execute "report1", out of 19 jobs 7 jobs have been cancelled due to dump.
The dump analysis is as follows*
Runtime errors         TSV_TNEW_PAGE_ALLOC_FAILED
Short dump has not been completely stored. It is too big.
h1 No storage space available for extending the internal table.
What happened?*
You attempted to extend an internal table, but the required space was
not available.
Error analysis*
The internal table (with the internal identifier "IT_77") could not be
enlarged any further. To enable error handling, the internal table had
to be deleted before this error log was formatted. Consequently, if you
navigate back from this error log to the ABAP Debugger, the table will
be displayed there with 0 lines.
When the program was terminated, the internal table concerned returned
the following information:
Line width: 1240
Number of lines: 340000
Allocated lines: 340000
New no. of requested lines: 10000 (in 1250 blocks).
      Last error logged in SAP kernel*
Component............ "EM"
Place................ "SAP-Server SR-3110_ISI_30 on host SR-3110 (wp 26)"
Version.............. 37
Error code........... 7
Error text........... "Warning: EM-Memory exhausted: Workprocess gets PRIV "
Description.......... " "
System call.......... " "
Module............... "emxx.c"
Line................. 1886.
Source code extract*
001640   FORM get_new_info .
001650
001660     DATA: wa_res LIKE zrisu_openitems_view,
001670           it_temp_final TYPE TABLE OF zrisu_openitems_view.
001680
001690     REFRESH: it_temp_final, it_final.
001700   * collect the free memory
001710     CALL METHOD cl_abap_memory_utilities=>do_garbage_collection.
001720
001730   * First find all the open FI document items on the system satisfying
001740   *  the selection criteria.
001750     SELECT opbel opupw opupk opupz gpart vtref vkont abwbl abwtp stakz
001760            waers studt betrw mahnv blart xblnr
001770            INTO CORRESPONDING FIELDS OF TABLE it_temp_final
001780            FROM dfkkop
001790            PACKAGE SIZE c_blksiz
001800            FOR ALL ENTRIES IN s_opbel
001810            WHERE opbel = s_opbel-low
001820              AND augst = space
001830              AND gpart IN s_part
001840              AND abwtp = space
001850              AND abwbl = space
001860              AND bldat IN s_bldat
001870              AND blart IN s_blart.
            APPEND LINES OF it_temp_final TO it_final.  ---> dump occurs at this statement
001890       REFRESH it_temp_final.
001900     ENDSELECT.
001910     REFRESH s_opbel.
001920
01930     SORT it_final BY opbel opupw opupk opupz.
01940
01950     LOOP AT it_final INTO wa_final.
01960   * Check the documentdate if relevant
01970       IF wa_res-opbel EQ wa_final-opbel.
01980   * Use information from last record  if same document  to make the
01990   *    program faster.
02000         IF wa_res-txt EQ 'DELETE NEXT RECORD'.
02010   * The record should be deleted for the same reason as the last item
02020           DELETE it_final.
02030           CONTINUE.
02040         ELSE.
02050           wa_final-bldat = wa_res-bldat.
02060         ENDIF.
02070       ENDIF.
If i run the jobs for which i got dump individualyy one after other, then the jobs executed successfully.
Pls help me in this regrad.
Can i use filed-symbols in thsi case to avoid dump?
Thanks in advance .
Taj

Answer 1

HI
It seems memory available to a particular process in your case was not enough to continue.
Could you try to implement the same by convertin Report2 as an FM and calling it as:
CALL FUNCTION func STARTING NEW TASK task
So now instead of making different variants you could just loop over the FM with different values.
If this does not seem feasible restrict your data anyhow.
Revet with findings!