EHP6 Upgrade shadow instance restart

[Former Member]

Hi Experts,

We are performing EHP6 upgrade of our ECC system. Currently, the upgrade is at Preprocessing phase at MAIN_SHDRUN/ACT_UPG, but it stopped, and when we checked in the shadow instance, the job RDDMASGL cancelled after 50000 sec run due to TSV_TNEW_PAGE_ALLOC_FAILED. Prior to starting the upgrade, we updated below parameters with the assumption that the values will be copied when SUM creates the shadow instance, so that we could avoid the stoppage within the MAIN_SHDRUN/ACT_UPG step.

ztta/roll_extension = 3000000000 (3GB)

abap/heap_area_dia = 4000000000 (4GB)

abap/heap_area_nondia = 4000000000 (4GB)

abap/heap_area_total = 4000000000 (4GB)

However, we still hit the error and stoppage, and still need to update the parameters in the shadow instance and restart the shadow instance. Somehow, the parameter value were not copied in the shadow instance parameters.

Can anyone advise how can we avoid this stoppage during the phase MAIN_SHDRUN/ACT_UPG? Is it possible to maintain and restart the shadow instance before the upgrade reaches the MAIN_SHDRUN/ACT_UPG during the pre-processing phase. Is there as note we can follow? Thanks.

We would appreciate your help.

Regards,

Philip

[Reagan]

Hello

Try setting these values and restart the instance.

ztta/roll_extension = 64000000000

abap/heap_area_total = 5000000000

abap/heap_area_nondia = 5000000000

ztta/roll_area = 10000000

abap/buffersize = 400000000

If possible restart the primary and the shadow instance after cleaning the shared memory segments.

Regards

RB

[Former Member]

Hi Reagan,

Thanks for the response. Currently, the preprocessing is running. We will consider the values you mentioned if the process stops again. Hopefully though not.

Our question actually is, is it possible to apply the parameters, and perform shadow instance at any point while the pre-processing phase is running?

Regards,

Philip

[Reagan]

Hello Philip

The upgrade tool will stop for the SPDD task. You can set the parameters in the shadow instance profile and restart the system before you continue from the SPDD phase. if you are using a transport request to perform the SPDD phase then you will can modify the shadow system profile from the OS level. The profiles will be created in the /<upgrade_directory>/abap/SID/SYS/profile directory. If the profiles are already created you can set the parameters in the profile and during the start of the shadow instance for the first time the system will use those parameters.

1387739 - Out of memory errors during shadow system operation

1674812 - Performance improvement for ACT_UPG

Regards

RB