MAIN_SHDRUN/ACT_UPG phase running very slow EHP7 upgrade

[Former Member]

Hi Gurus,

We are performing EHP7 upgrade of  existing Development system (ERP ECC 6.0) system from ECC 6.0 EHP3

System OS : WIndows Server 2012

System DB : MS SQL Server 2012

Source Kernel Version : 720 Support Pack 525

Phase MAIN_SHDRUN/ACT_UPG phase of Pre-processing step is running very slow.

It is running for more than 48 hrs and log update frequency is very slow.

There was issue in shadow instance and job RDDMASGL (job was failing with TSV_PAGE_NEW_ALLOC_FAIL error) in shadow was failing yesterday.

I have changed memory parameters of shadow instance as per note 1387739 - Out of memory errors during shadow system operation

These changes in shadow instance were done on 6^th Jan 1:00 PM and shadow instance was restarted after updating parameter values in shadow instance profie and upgrade was continued from point it stopped.

abap/heap_area_total = 10000000000   (Old value 8588886016)    
abap/heap_area_nondia = 10000000000    (Old value 0)
ztta/roll_area = 20000000               (Old NA)

However there the upgrade is continued and there is no error since yesterday but logs are updated very slowly and it seems upgrade is running slow.

There is not update of log in /abap/log directory since 24 hours

Logs are updated in /abap/tmp but only one file is getting updated for last 24 hrs and its size currently is 108 MB (filling very slow). This file for which status and size is updated for last 24 hrs is SAPA742EPU.CP1

Below is screen for same

Any other file status is not updated since 6^th Jan 1:26 PM in tmp folder

Below is the screen for /abap/log directory and no logs updated since 24 hrs

Can anybody suggest if Shadow instance memory parameter change would be an issue.

Stopping shadow during ACT_UPG and retry would be a right option.

Please suggest.

[former_member594566]

Hi,

I was also facing the same issue during SUM phase MAIN_SHDRUN/ACT_UPG during up-gradation of Solution Manager 7.1 to 7.2. The current phase was running from more than 48 hours and no logs were updated neither it threw any error.

I followed the SAP Note 1387739 - Out of memory errors during shadow system operation

and made below changes into Shadow instacne profile.

abap/heap_area_total = 10000000000

abap/heap_area_nondia = 10000000000

ztta/roll_area = 20000000

Then I restarted the Shadow System and started the phase again. Now it completed in just only six hours.

Regards,

Ravinder Singh

[Former Member]

Hi Anand,

As long as you are not seeing any error in the log it is running fine i think.

regarding of slowness is depends on many things. for more information check below notes.

1918279 - performance analysis for ACT_UPG phase

1674812 - Performance improvement for ACT_UPG

https://wiki.scn.sap.com/wiki/display/SL/Analysing++activation+errors+of+single+objects

Thanks,

Venkat

[Former Member]

Also please have a look on it. This is very useful.

2235057 - Performance issues when running a SAP Upgrade with MS SQL Server