Workprocess getting restarted after DB Roll back

Question

SAP Version : 4.7 EE
Database : DB2 UDB
OS : AIX
I have an error reported by end user saying Workprocess are getting restarted when he is running a Transaction when i checked the system log and therafter the Workprocess Log I have observed the following , I have also found the heap memory is getting full and at the same time the DB is doing a rollback, can somebody explain if this is only cause of the heap memory or its Db problem over here , attached below is the Log
Thu Feb  4 03:30:55 2010
  WP has reached abap/heaplimit = 40894464 bytes
Thu Feb  4 03:33:20 2010
  SelMemClass: heap quota (DIA) exceeded 2000000000 2000002464
Thu Feb  4 03:33:23 2010
  TH VERBOSE LEVEL FULL
RABAX: level LEV_RX_PXA_RELEASE_MTX entered.
RABAX: level LEV_RX_PXA_RELEASE_MTX completed.
RABAX: level LEV_RX_VMIT_CLEANUP entered.
RABAX: level LEV_RX_VMIT_CLEANUP completed.
RABAX: level LEV_RX_ROLLBACK entered.
ERROR => Call to DbSlRollbackDB6 while DBIF already active in DbSlExeReadDB6*[dbsldb6.c    514]
  ***LOG BZY=> unexpected return code 1          calling DbSlRollba [dbcon#9 @ 2932] [dbcon   2932 ]
  ***LOG BYJ=> database function ROLLBACK   for connection R/3        failed [dbcon#9 @ 2935] [dbcon   2935 ]
ERROR => DISTRIBUTED_TA_FAILED:
[dbcon.c      1649]
  0: name = R/3, con_id = 000000000 state = ACTIVE      , perm = YES, reco = NO , timeout = 000, con_max = 255, con_opt = 255, occ
  ***LOG R39=> ThIRollBack, db_rollback ( 016384) [thxxhead.c   12637]
  in_ThErrHandle: 1
ERROR => ThIRollBack: db_rollback (step 4, th_errno 18, action 2, level 1) [thxxhead.c   9723]

Answer 1

Karthik,
We hit the exact same problem when activating an APO integration model in ECC last week. It looks like an SQL statement that returns a very large result set is executed. While the result set is fetched the ABAP heap is exhausted. Since no more memory is available the ABAP engine abruptly cancels the execution and begins error handling. The database interface rejects to execute a ROLLBACK statement while the SQL statement is still active. As the rollback fails the task handler restarts the SAP work process and no ABAP short dump or job log entries can be written. We first attempting to resolve this problem by changing
+abap/heap_area_dia + to 4GB->8GB->12GB. However, finally, the solution that worked for us was to cleanup the change pointers and reactivating the integration model with 4GB heap.
Regards,
Sameer