cancel
Showing results for 
Search instead for 
Did you mean: 

Save-while-active request prevented by pending record changes

Former Member
0 Kudos

Hello together,

while doing our daily full backup with SAVE-While-Active function, we recognize the following error:

CPI8367    Save-while-active request prevented by pending record changes.

 Cause . . . . . :   The save-while-active request for job                    
   676599/ITUSER/DATASOLO has been ended because all commitment definitions  
   with pending changes for job 673801/QUSER/QXDARECVR were not committed or  
   rolled back before the save active wait time (SAVACTWAIT parameter) was    
   exceeded.

Then I am using the Command

DSPJOB JOB(673801/QUSER/QXDARECVR) OPTION(*CMTCTL)

to find the error details.

I can find the following help in CPF8367:

3 - If a commitment definition is preventing a save-while-active request 
 from continuing, informational message CPI8366 is shown on the status line 
 of the Commitment Definition Status display.

But in the display I can't find the informational message CPI8366 !!!??

Display Commitment Definition Status             SYSNO
                                                            21.09.09  14:15:42
Job:   QXDARECVR      User:   QUSER          Number:   673801                                                                                
Thread  . . . . . . . . . . . . . :   *NONE                                   
Logical Unit of Work ID . . . . . :   APPN.DEEBMHVA.X'6E10A9604809'.00917     
Lock Space ID . . . . . . . . . . :   UDB_580000000007797E                    
Commitment definition . . . . . . :   *JOB                                    
Activation group  . . . . . . . . :                                           
ASP group . . . . . . . . . . . . :   IASP144                                                                                
Job active  . . . . . . . . . . . :   YES                                     
Server job  . . . . . . . . . . . :                                           
Lock scope  . . . . . . . . . . . :   *JOB                                    
Transaction timeout . . . . . . . :                                           
Maximum lock wait . . . . . . . . :                                           
Number of commits . . . . . . . . :   915                                     
Number of rollbacks . . . . . . . :   1

Accepted Solutions (1)

Accepted Solutions (1)

0 Kudos

Hi Carsten,

this may be an IBM problem that needs investigation. In addition to the joblogs of the backup job and the job shown as having uncommitted changes, you could collect the journal entries for the last shown commit cycle using the command: DSPJRN JRN(R3SIDDATA/QSQJRN) RCVRNG(CURCHAIN) CCIDLRG(<commit cycle identifier>) OUTPUT(PRINT). You could also check if there are any commit blocks (internal objects for commitment control) left over in the database library. This is done with the command DMPSYSOBJ OBJ(QDBDB*) CONTEXT(R3SIDDATA).

I would suggest to open a PMR with that information.

Kind regards,

Christian Bartels.

Answers (2)

Answers (2)

Former Member
0 Kudos

Hi Volker,

now I found the reason. Not exactly the last reason, but a reason why three jobs with the name QXDARECVR a holding locks on some tables.

We do have two Windows Application Server in addition to the Central Instance. When I look to the display at sm50 for this two application Server, then I can see, that on one application server was one backroundjob with 20.000 seconds and on the other application Server there has been two backroundjobs with serveral tousands of seconds. These job are doing nothing and holds some commitments open. After canceling this jobs the sm50 backround jobs, the server Job QXDARECVR cancels itself too.

But the reason why this sm50 backround job stop working, I don't know. This jobs are running in period. For no reason this job stopped in the past same times. And because of that, the backup will not work proper and the save-while-active checkpoint will not be found.

In the joblog of the QXDARECVR Job, I can find a message with the message ID SAP4000. This indicates that the job is running for an application server.

best regards,

Carsten Schulz

0 Kudos

Hi Carsten,

if the problem can be seen again, then SAP Note 37104 ("Error analysis: Background processing system") may provide some insight about what is going on. In the "Solution" section, there is a chapter for symptom "Job remains in status "Active"", that may help. A double-click on one of the processes in SM50 and a DSPJOB OUTPUT(*PRINT) for the QXDARECVR job(s) may also be helpful to understand what status these processes are in.

Kind regards,

Christian Bartels.

Former Member
0 Kudos

But if I press F6 and then choose Record level:

Display Record Level Status                                                                                
System:   SYSNO 
 Job:   QXDARECVR      User:   QUSER          Number:   673801                                                                                
Commitment definition  . . . . . . . . :   *JOB                                                                                
-------------Changes--------------         
 File        Library     Member          Commit    Rollback     Pending         
 AFFW        R3SIDDATA   AFFW                13           0           0         
 AFKO        R3SIDDATA   AFKO                73           0           0         
 AFPO        R3SIDDATA   AFPO                80           0           0         
 AFRU        R3SIDDATA   AFRU               320           0           0         
 AFVC        R3SIDDATA   AFVC               324           0           0         
 AFVU        R3SIDDATA   AFVU               324           0           0         
 AFVV        R3SIDDATA   AFVV               324           0           0         
 AFWI        R3SIDDATA   AFWI               511           0           0         
 ARFCSDATA   R3SIDDATA   ARFCSDATA            4           0           0         
 ATAB        R3SIDDATA   ATAB                 0           0           0

I have no idea where to find the real reason why the backup end abnormal.

best regards,

Carsten Schulz

Former Member
0 Kudos

Hi Carsten,

you should be able to reproduce this over the day as well I would say.

I do see 2 reasons:

- this job is crazy and when you kill it, all is OK (the job gets restarted)

- there is a wrong counter in one of your tables, that is very ugly - this would happen even when SAP would be down.

I would try to reproduce and then to kill this one and then open a PMR with the results ...

Regards

Volker Gueldenpfennig, consolut international ag

http://www.consolut.net - http://www.4soi.de - http://www.easymarketplace.de

Former Member
0 Kudos

Hi together,

I have the same problem again.

I will now post here the result of DSPJOB JOB(673803/QUSER/QXDARECVR) OUTPUT(PRINT) OPTION(CMTCTL) in several posts.

Please can someboby say me, how I can detect, why this job should hold some open commits?

5722SS1 V5R4M0 060210                      Display Job                                          3.10.09  8:46:56         Page   1
 Job . . . :   QXDARECVR    User . . . :   QUSER         Number . . . . :   673741
                                    Commitment Control Status
 Thread . . . . . . . . . . . . . . . :  *NONE
 Logical unit of work identifier  . . :  APPN.SYSTEMNAME.X'6E10A0F047CD'.02273
 Lock Space ID. . . . . . . . . . . . :  UDB_580000000007794D
 Commitment definition  . . . . . . . :  *JOB
   Description  . . . . . . . . . . . :    Job-level
 Activation group . . . . . . . . . . :
 ASP Group . . . . . . . . . . . .  . :  IASP144
                                  Commitment Definition Status
 Resource location  . . . . . . . . . :  LOCAL
 Default lock level . . . . . . . . . :  *CHG
 Role . . . . . . . . . . . . . . . . :
 State  . . . . . . . . . . . . . . . :  RESET
 Date/time stamp  . . . . . . . . . . :
 Resync in progress . . . . . . . . . :  NO
 Heuristic operation  . . . . . . . . :
 Job active . . . . . . . . . . . . . :  YES
 Server job . . . . . . . . . . . . . :
 Lock scope . . . . . . . . . . . . . :  *JOB
 Transaction timeout  . . . . . . . . :
 Maximum lock wait  . . . . . . . . . :
 Number of commits  . . . . . . . . . :  2272
 Number of rollbacks  . . . . . . . . :  0
 Default journal  . . . . . . . . . . :
   Library  . . . . . . . . . . . . . :
 Notify object  . . . . . . . . . . . :  *NONE
   Library  . . . . . . . . . . . . . :
   Object type  . . . . . . . . . . . :
   Member . . . . . . . . . . . . . . :
 Commitment options:
   Wait for outcome . . . . . . . . . :  WAIT
   Action if problems . . . . . . . . :  ROLLBACK
   Vote read-only permitted . . . . . :  NO
   Action if ENDJOB . . . . . . . . . :  WAIT
   Last agent permitted . . . . . . . :  SYSTEM
   OK to leave out  . . . . . . . . . :  NO
   Accept vote reliable . . . . . . . :  YES

Former Member
0 Kudos
Record Level Status
                                 ------------ Changes ------------ Lock                                        Commit Cycle
 File       Library    Member         Commit   Rollback    Pending Level Status          Journal               Identifier
 AGR_DATEU  R3SIDDATA  AGR_DATEU           0          0          0 *CS   OPEN            R3SIDDATA/QSQJRN
 ARFCSDATA  R3SIDDATA  ARFCSDATA          27          0          0 *CHG  OPEN            R3SIDDATA/QSQJRN      1744249845
 ARFCSSTATE R3SIDDATA  ARFCSSTATE         27          0          0 *CHG  OPEN            R3SIDDATA/QSQJRN      1744249845
 BALDAT     R3SIDDATA  BALDAT             13          0          0 *CHG  OPEN            R3SIDDATA/QSQJRN      1744249845
                       BALDAT              0          0          0 *CHG  OPEN            R3SIDDATA/QSQJRN
 BALHDR     R3SIDDATA  BALHDR              4          0          0 *CHG  OPEN            R3SIDDATA/QSQJRN      1744249845
 CMFK       R3SIDDATA  CMFK                0          0          0 *ALL  OPEN            R3SIDDATA/QSQJRN
 5722SS1 V5R4M0 060210                      Display Job                                          3.10.09  8:46:56         Page   2
 Job . . . :   QXDARECVR    User . . . :   QUSER         Number . . . . :   673741
Former Member
0 Kudos
5722SS1 V5R4M0 060210                      Display Job                                          3.10.09  8:46:56         Page   5
 Job . . . :   QXDARECVR    User . . . :   QUSER         Number . . . . :   673741
                                    Commitment Control Status
 Thread . . . . . . . . . . . . . . . :  *NONE
 Logical unit of work identifier  . . :  APPN.SYSTEMNAME.X'6E10A0F047CD'.02273
 Lock Space ID. . . . . . . . . . . . :  UDB_580000000007794D
 Commitment definition  . . . . . . . :  *JOB
   Description  . . . . . . . . . . . :    Job-level
 Activation group . . . . . . . . . . :
 ASP Group . . . . . . . . . . . .  . :  IASP144
                                           RDB Status
 Remote            Relational          Pending
 Location          Database            Changes
 *LOCAL            IASP144             YES

This post shows that there are pending Changes.

But the post bevor show no pending changes The list is longer then shown here, but there are overall no pending changes.

I am not able to determine who is causing this problem.

kind regards,

Carsten Schulz