EHP7 upgrade pre processingphase MAIN_SHDRUN/ACT_UPG running from long time

[Former Member]

Hi experts,

I'm doing an EHP7 upgrade on our windows cluster system.

Currently on Phase MAIN_SHDRUN/ACT_UPG Pre Processing and this phase

is taking too much time.

I have checked the logs its not updating, I tried few oracle & memory related sap notes still same .

any suggestions

[Reagan]

Check whether you are hitting the issue mentioned in the SAP note 1635605 - CLIENT HANGS ON INSERT INTO TABLE WITH SECUREFILE LOB and if yes then set the parameter _use_zero_copy_io to false and restart the Oracle database. This has been discussed before.

[Former Member]

Hi Reagan,

As you mentioned I have implemented this parameter before it self .

any other suggestions? even I tried  whole restart and resume the upgrade .

implemented the below parameters

SAP recommendation :

abap/heap_area_total = 10000000000

abap/heap_area_nondia = 10000000000

ztta/roll_area = 20000000

still the same it's running

[Reagan]

There is no information about the parameter _use_zero_copy_io in your initial post.

If that is already set then check the logs in the <DIR_PUT>/abap/tmp directory

This is one of the longest phases in the SUM so you need to monitor the logs in the above mentioned tmp directory and monitor the progress. Once a specific activation is done the tool moves the log file to the <DIR_PUT>/abap/log directory.

If the log files are not getting <DIR_PUT>/abap/tmp directory then connect to the shadow system and check for any dumps appearing in the system (Tx ST22).

If possible restart the shadow system and retry the phase.

Cheers

Reagan

[Former Member]

Hi Reagan,

I have checked the logs  in <DIR_PUT>/abap/tmp directory  it's not updating anything  from last few hours ,

As mentioned I tried many times the restarting the shadow instance and re-run the phase  and no dumps as well .

Just want to check did I miss anything related to windows cluster system ? I just brought the DB & SAP into one node and running the upgrade by stopping the 2nd node .

and changed the autostart  to NO .

[Reagan]

Connect to the shadow system and check the status of the Activation job and also check the executing server. Make sure that job is executed in the right AS.

[Former Member]

Hi Krishna,

Pls read the below notes.It may helpful.

1674812 - Performance improvement for ACT_UPG

1919276: Performance issues when running a SAP Upgrade

Regards

Maruthi V