Phase MAIN_SHDRUN/ACT_UPG is running for a longer time

[Former Member]

Hi All,

Im doing EHP4 on our ECC 6.0 System and im stuck at SHDRUN/ACT_UPG for more than 14 hours now, so i logged into

the shadow system and checked for long running jobs and i see one background job RDDMASGL which has been running for

more than 36,000 seconds. I checked at the OS Level if there are any space issues and i see that we have space at the OS

level.I selected 3 R3trans process before starting . What can be done to optimize the perfomance

Our system is on SQL Server 2005 enterprise edition

Regards,

Ershad Ahmed.

[former_member57954]

1828667 - ABAP Dictionary mass activation and tp parameter <SID>/batch_proc

[Former Member]

Thanks Every one who helped me during EHP4 , I was able to complete the task on our DEV Box.

Regards,

Ershad Ahmed

[Former Member]

Hi All

The Job RDDMASGL completed after 20 hours and then i got an error message

Unable to execute the SDT request Abap operation pre-execute on hostname

Error from unit observor

Tool died unexpectedly with null

so i restarted from the point where it started and i see that the Job RDDMASGL restarted again

Regards,

Ershad Ahmed

[markus_doehr2]

> so i restarted from the point where it started and i see that the Job RDDMASGL restarted again

That doesn't matter - it will "see" that it has activated most of the objects.

Markus

[Former Member]

Hi,

My problem is different.

The phase is failing due to the program RDDMASGL failing for TSV_BTR_PAGE_ALLOC_FAILED.

We have now increased the abap/heap_area_nondia and total as well as roll_area.

But the job has restarted all over again and running since 6 hours again after failing first time around in around 7 hours.

WE have implemented the "Note 1631790 - Mass activation during transport is started repeatedly" but it still began all over again in our shadow instance.

Please suggest.

THanks and Regards,

Karan Shah

[Former Member]

Hi All,

I lost connectivity to the SDT tool during EHP4 and i had to restart STARTUP.BAT from the OS level to login and SDT asked me

whether it should start from the phase where it stopped and i clicked on yes and when i check in SDT for ABAP it says

ABAP : Starting Handshake and in Java it says ABAP:Starting Handshake , when i click on Monitor ABAP part shows as running

where as Java says it is waiting. I checked for the Background job and it is still running..

Iam concerned whether does SDT needs any manual intervention now or is it actually running.

Regards,

Ershad Ahmed

[Former Member]

Hi All

I checked in shadow system and i see that RDDIMPDP was last run at the time when the job RDDMASGL was started and

after that i do not see any runs of RDDIMPDP, not sure if it has got anything to do with this

Regards,

Ershad Ahmed.

[Former Member]

Hello Markus,

From ST06 it shows me that our system has got 4 CPUS

Regards,

Ershad Ahmed

[markus_doehr2]

You could use 4 R3trans and 4 batch to increase the speed during the next run. As of now there's nothing you can do but wait.

Markus

[Former Member]

Hi Ershad,

We are also in the same phase, RDDMASGL job is running from last 72000 sec., sys is with 2 CPU's but utilization is 30 % only.

Regards,

Hari.

[markus_doehr2]

This is nothing to worry about.

ACT_UPG calls subsequent dialog processes in the shadow instance to activate objects. Depending in which "level" and dependency it is called, there may only run one processes actually despite you gave more than one in the configuration.

The phase may run for quite a long time. On one of our system with 2 CPU (Windows 2003 + SQL 2008) that phase ran 26 hours with 2 parallel processes which is nothing uncommon.

It's a good idea to do several runs, use the UPGANA.XML in the htdocs directory to check which phases ran for a long time and use different parameters the next time.

Markus

[Former Member]

ACT_UPG calls subsequent dialog processes in the shadow instance to activate objects. Depending in which "level" and dependency it is called, there may only run one processes actually despite you gave more than one in the configuration.
The phase may run for quite a long time. On one of our system with 2 CPU (Windows 2003 + SQL 2008) that phase ran 26 hours with 2 parallel processes which is nothing uncommon.

We had a run time of 25.000 s on the RDDMASGL. Even if the phase is not in the downtime I am interested in speeding it up. Can there anything be done in the SHD (more DIA WP) to help the RDDMASGL to finish its jobs faster... ? More CPUs can be given to the OS.

Kind regards, Hinrich

[Former Member]

Hello Markus

Thank you for the reply , i selected 3 Batch process when starting EHP4.

Regards,

Ershad Ahmed

[markus_doehr2]

> Thank you for the reply , i selected 3 Batch process when starting EHP4.

How many CPUs has that machine?

I selected one R3trans and one batch per CPU, we used 8 and the phase took ~ 12 hours. Since this runs during normal production uptime the normal load on the machine may also influence the performance.

Markus

[markus_doehr2]

> level.I selected 3 R3trans process before starting . What can be done to optimize the perfomance

There's nothing you can do but wait.

How many batch processes did you select?

Markus