Releasing transport requests with job roles take forever after upgrading to Netweaver 7

[Former Member]

Dear experts,

Some time ago we've out SAP ECC 5 system to ECC 6. Since then, it takes a very long time to release transport requests containing job roles.

For example, a request containing 154 job roles (1 template and 153 derived roles) took 4 hours to release in our DEV system. Before the upgrade this only took a couple of minutes.

Afterwards, the import in our QAS and PRD system don't take longer than normal.

Apparently this time is taken by the Export phase:

Checks at Operating System Level	17.04.2014 09:50:03	(0) Successfully Completed
Pre-Export Methods	17.04.2014 09:50:37	(0) Successfully Completed
Export	17.04.2014 13:51:02	(0) Successfully Completed

Does anyone know what causes this and how to solve?

Kind regards,

Nicolas

[Former Member]

Hello Nicolas,

What I understood from your description is that ,it takes very long time to release the job. Once it is released, execution does not take much time.

To me it seems to be background job/work process issue. Maybe all the background work process are busy to execute current jobs and no free work process available for the jobs which are in the queue. In that case, if possible try to increase number of bg work processes.

Check for any performance issues during that period. Check if there is any locking issue and jobs are stuck because of that.

Hope this helps to investigate the issue

Regards,

Archana

[Former Member]

Hi Nicholas,

Have you checked to see if UST12 is buffered? Do any related selects show up in the DBACOCKPIT list of long running selects? Do any missing indexes appear for this table?

Just a few ideas...

Thanks and regards,

Joe Haynes

Nimbl, LLC

[Former Member]

In general, it would be most beneficial to always provided the complete production version you are on -- Enhplevel and SP level. Have you looked at OSSN 1673833 or 1816508 to see these are applicable for your scenario?

[Former Member]

are you on oracle? have you performed the table / system stats after the upgrade?

[Former Member]

We are on IBM iSeries.

After the upgrade we performed a health check which showed that table UST12 is one of the biggest in the memory.

I see that a selection on this table is also included in transport requests with job roles, but I don't know if we can make this table more efficient...

[Former Member]

Thanks, can you please check the note 1819126 - Long runtime for role transport release

[Former Member]

We've tried to implement this note but without any effect.