SAPKB71007 gets stuck in XPRA_EXECUTION

[Former Member]

Dear all,

We have a Netweaver PI 7.10

We are applying a SP queue containing SAPKB71007 and SAPKB71008. SAPKB71007gets stuck in XPRA_EXECUTION. Looking into SAPRB71007.BXP at /usr/sap/trans/tmp/ we can see the following last line:

2 EPU122XPost-import method "SBTM_TT_DESIGN_AFTER_IMPORT" started for "BTTY" "L", date and time: "20110225163904"

And that's all, no changes in the file are made any more.

I have checked that there are no old entries in TRBAT nor TRJOB. The problem occurs when RDDEXECL starts to read table REPOSRC with operation 'Direct Read'. This operation gets stuck and, as far as I can see, there is no result. It just cancels and returns an error in SPAM.

Additional info: We are using the latest SPAM/SAINT level (041).

Any ideas will be very appreciated. Thank you.

Francisco Martínez

Edited by: Francisco José Martínez Carretero on Feb 28, 2011 12:13 PM

[Former Member · Accepted solution]

Hello

What is written on Queue Import logs?

[Former Member · Accepted solution]

Hi Kaan,

The only relevant line is that one I wrote above, executing one of the methods of XPRA_EXECUTION. Other logs do not contain any useful information. But I can post some of the last lines if you think you need to have a look.

Thanks for the response.

[sunny_pahuja2 · Accepted solution]

Hi,

Is there any information in Slog and Alog ?

Thanks

Sunny

[Former Member · Accepted solution]

Hi Sunny,

Nothing relevant:

These are the last lines of SLOG:

START EXECUTION OF REPORTS BXP R 20110228115420 DDIC lxdsap03

START tp_getprots BXP R 20110228115420 DDIC lxdsap03

ERROR SAPKB71007 BXP R 0012 20110228120527 SAP DDIC lxdsap03

STOP tp_getprots BXP R 20110228120537 DDIC lxdsap03

STOP EXECUTION OF REPORTS BXP R 20110228120537 DDIC lxdsap03

And these ones belong to ALOG:

ALL BXP.ALL t 0000 20110228115420 DDIC lxdsap03

SAPKB71007 BXP.ALL R 0012 20110228120527 SAP DDIC lxdsap03

Thanks

More info:

The following error is stored in the developer trace of the process running RDDEXECL:

B Mon Feb 28 15:08:34 2011

B ***LOG BY4=> sql error -913 performing SEL on table REPOSRC

B ***LOG BY0=> [CLI Driver][DB2] DSNT408I SQLCODE = -913, ERROR: UNSUCCESSFUL EXECUTION CAUSED BY DEADLOCK OR TIMEOUT. REASON CODE 00C9008E, TYPE OF

RESOURCE , AND RESOURCE NAME

DSNT418I SQLSTATE = 57033 SQLSTATE RETURN CODE

DSNT415I SQLERRP = DSNXRRC SQL PROCEDURE DETECTING ERROR

DSNT4

A TH VERBOSE LEVEL FULL

However, there is no other process nor user running anything in the system

Thanks and regards

Francisco

Edited by: Francisco José Martínez Carretero on Feb 28, 2011 3:10 PM

[Former Member · Accepted solution]

Hi Francisco,

just a hint. this could indicate a lack of memory. had a similar issue while applying FINBASIS (EXECUTION OF REPORTS AFTER PUT return code 12). solved by increasing profile parameters as follows:

abap/shared_objects_size_MB 20 increased to 200

enque/table_size 16000 increased to 25000

Edited by: Gerd Schuster on Feb 28, 2011 3:55 PM

[Former Member · Accepted solution]

Hi Gerd,

I tried changing the parameter values according to your hints but unfortunately the problem still remains.

Thank you for your answer.

Regards

Francisco

[former_member128291 · Accepted solution]

Dear Francisco,

Kindly note that SQL0913 is an object lock situation. A user or process already accesses

the object REPOSRC is locking the object. Please take care that no user is working on your system and no batch job is running during the import of support packages. To release the lock situation you can stop and start the SAP system. Please ensure you have a valid backup of your system made from a piont in time before you began to apply the support packages as, should the manual import fail, it may be required to restore the system from backup.

You can further drill down on the exact cause of this error. What is reported in work process trace files and also dev_disp, dev_rd, dev_ms etc. log files. This can be most probably due to some memory issues or maximum number of connections reached error.

Check and let me know the results.

Regards,

Abhishek

[Former Member · Accepted solution]

Dear Abhishek,

Thanks for your response.

However, as I wrote in my third comment, there is no other user in the system nor process running. In fact I can see very clearly in SM50 that the only WP that is trying to access table REPOSRC has been launched by RDDEXECL. No other process is accessing the table. I checked this with my DB2 colleagues.

Additionally, no useful information can be found in the dev files.

Thanks and Regards

Francisco

[former_member128291 · Accepted solution]

Dear Francisco,

Strange, i suspected something out of the trace files.

Did you checked the work process trace file too? Is there no error mentioned?

Also check the SM21 logs.

If there are no clues and you cannot afford to restart the system then i would suggest you to raise an OSS message with SAP.

Regards,

Abhishek

[Former Member · Accepted solution]

Hi,

Did you take a look in SM12 ?

Maybe there is a lock error.

Best regards,

Alin

[Former Member · Accepted solution]

Hi, do you have in the meantime any solution.. I have exactly the same problem.

Maybe you can help.

Regards

Robert

[Former Member · Accepted solution]

Hi,

I opened an OSS message, which has reached the SAP CORE Support team, they are working on a solution for that. As soon as I have a solution I'll post it here.

Regards

Francisco

[Former Member · Accepted solution]

Hi Fransisco,

Let us know what SAP says on this issue.

Meanwhile you can just have a look on SAP Note: 954792.

It says something on the error in XPRA_EXECUTION phase.

Regards,

Sujit,

[Former Member · Accepted solution]

The sql0913 has a reason code listed to know if it is a lock-wait timeout (usually 3600 seconds) or a deadlock. See http://publib.boulder.ibm.com/infocenter/db2luw/v9/index.jsp?topic=/com.ibm.db2.udb.msg.doc/doc/sql0... for details.

In dba cockpit you should be able to look at the deadlock logs and see process information of the other process holding a lock. You may have to open this log twice as there are 2 files and when you open it it switches files so you are liikely looking ath the other file that does not have your error in it. So reopen and it switches back to the file your message is in.

you may need to do a db2 snapshot to cross reference the sap deadlock information showing app handles to process information shown in db2 snapshot.

db2 get snapshot for applications on sid > db2snap1

more db2snap1

Could there be another os level process with the lock. perhaps, R3trans, R3load etc.

Edited by: Fett Patrick on Mar 15, 2011 12:14 PM

[Former Member · Accepted solution]

Thanks Sujit,

I already read that OSS note. Unfortunately it does not apply to our case.

As I detected some problems with certain ABAP methods (they couldn't be generated using SGEN or the "activate" icon from SE38) I reported that to the OSS in the message. It was clearly a problem related with that methods (a generation just restarted the WP with no further message), so they started to investigate that line. Now they are working on it. I'll keep you informed.

Regards

Francisco

[Former Member · Accepted solution]

Hello Francisco

any news or sulution from SAP concerning your problem?

Regards

Robert

[Former Member · Accepted solution]

Hi,

No news yet. They are still working on it.

Regards

Francisco

[Former Member · Accepted solution]

Hi all,

SAP Support has released a new OSS note for our issue: 1551400.

It is related to the activation or compiling of certain ABAP programs with deeply nested constants, for example CL_SBTM_TYP_UTIL==============CP.

Next Friday 25th March they will release a new kernel patch in a few days to fix the problem.

Thanks and Regards

Francisco