SUM upgrade phase Execution at step: ABAP: Phase MAIN_NEWBAS/XPRAS_AIMMRG is running too long

[Former Member]

hi Experts,

We are upgrading our solution manager 7.1 SP11 to SP13 through Software Update Manager [SUM] SP13 on windows environment.

but from almost 30 hours i was stuck at this very step: " ABAP: Phase MAIN_NEWBAS/XPRAS_AIMMRG is running ... "

Java installation is completed successful. I dont see any dumps in shadow systems. in SM66 i see one background job running from 30 hours.and in SM37 i see the job RDDEXECL active from the same time stamp.

please advice we are in serious concern about this.

thanks,

Raja

the following 2 images are SM37 job and its job log

[Former Member · Accepted solution]

thanks everyone for their help, it really did help me to solve it.

This is how i finally solved it...

My Issue:

"

2EETW000 sql error:  1205 Transaction (Process ID 117) was deadlocked on lock resources with another process and has been chosen as the deadlock victim. Rerun the transaction.

2EETW125 SQL error "1205" during "1205" access: "Transaction (Process ID 117) was deadlocked on lock resources with another process and has been chosen as the deadlock victim. Rerun the transaction."

1 ETP111 exit code           : "12"

Found this in XPRASUPG.ELG log.

in SM37 the job that is keep on running for last 2 days  in active state "RDDEXECL "

so even though log says there is an error due to dead lock, it never showed up..as the job RDDEXECL keep on running.

we have 3 approaches here

1. restart SUM

2. Restart shadow instance

3. restart the job " RDDEXECL"

i have choosed option 3 as im little worried to try other 2 options as im in middle of execution phase.if it hadnt worked i would have to go through the restart of SUM/Shadow instance.

thanks,

Raja

[former_member74206 · Accepted solution]

Hi Raja,

                I have exactly the same issue.. when you said you chose option 3 restart of the job "RDDEXCEL" what exactly did you do? 

Did you go into sm36 and recreate a job with program RDDEXCEL ?

My Job is still running so do I cancel it or wait for it to crash?

Regards,

                Ryan

[former_member74206 · Accepted solution]

For those that want to know I cancelled the running job RDDEXCEL.. Then made a copy of the job and ran it. The upgrade execution started again (visisble in updated log files).

[srini_sandaka2 · Accepted solution]

Hi All,

We have also facing the same issue in execution phase during Solution manager 7.1 SP12 to SP14 stack upgrade.

In the phase we are not able to login into system.Please advise how to fix the issue.

Thanks

Vijay.

[Former Member · Accepted solution]

Hello Srini,

What is the error you are getting when trying to login to SAP.  Try to login with DDIC

Regards,

Mahendra

[j_bayrhammer · Accepted solution]

Hello,

I repeat my post from above:

we are at same point.

I read note 1947874 and checked job status in SM37.

After that all RDDEXECL jobs were shown as finished.

I can't login into shadow instance. So I had a look into via dpmon pf=Instance.

There I can see, that all dialog workprocesses are in use. Thats the cause why I can't login.

But tables are not changing, instance seems blocked.

Is it an option to restart shadow instance now? Or where can I see, what's the problem?

Regards,

Julia

[Former Member]

Hi Raja,

As you are performing Solman 7.1 SP11 to SP13 upgrade and step "MAIN_NEWBAS/XPRAS_AIMMRG" should not take more than 2 hrs to complete; 

So there might be chances of deadlock occurrence. Please check the same in DB01. also please check upgrade logs in <drive>:\usr\sap\SUM\abap\log provide the screenshot of logs.

Thanks.

Regards,

Jaskirat Singh

[Former Member]

Hi Singh,

I have attached the logs which shows that there is a deadlock error.

But i dont see them in the shawdow system DB02.

as far as i understand i think the step " MAIN_NEWBAS/XPRAS_AIMMRG " found a deadlock, and waited so long that he got terminated.

can we re-start at this point or what should be my approach here..?

XPRASUPG.txt.zip

SAPI-70220INCTSPLGN.txt.txt.zip

[Former Member]

HI Raja,

Did you check the deadlock in Tcode DB01.

Also you should not restart upgrade at this point. If you encountered an error in SUM GUI then you continue to resume upgrade by clicking on continue.

Also have a look at below Snote which says repetition of task solves the deadlock issue; But don't apply Snote correction in your system.

1155807 - Database deadlock when executing DD_INT_UPDATE_DDFTX

Regards,

Jaskirat Singh

[j_bayrhammer]

Hello,

we are at same point.

I read note 1947874 and checked job status in SM37.

After that all RDDEXECL jobs were shown as finished.

I can't login into shadow instance. So I had a look into via dpmon pf=Instance.

There I can see, that all dialog workprocesses are in use. Thats the cause why I can't login.

But tables are not changing, instance seems blocked.

Is it an option to restart shadow instance now? Or where can I see, what's the problem?

Regards,

Julia

[former_member185239]

Hi Raja,

- Are you getting any dump related to this in ST22?

-  Check the status of the job RDDEXECL

    run the transaction sm37 and job->check.

- Attached the logs for the phase and also the job log.

With Regards

Ashutosh Chaturvedi

[Former Member]

No dumps in ST22

Job status in SM37 is fine and active

attached the logs for reply to post of " Jaskirat Singh"

[manumohandas82]

Hi Raja ,

If the status of the process does not change for a long time in SM50 then more probable that the work process is stalled .

You may also take a trace using ST12 , ST05  . Check for any DB deadlocks or long held locks .

Is there any CPU / Memory contention ?

Believe you could wait for some more time , in the meantime  take a trace or check in STAD

Thanks ,

Manu

[manumohandas82]

Hi Raja ,

This phase generally takes a long time to complete . If the Job is active  [ Check in sm66 , sm50 , etc ] and updating there is nothing to worry .

The following Note should help you

1947874 - Performance analysis for XPRAS_AIMMRG phase

Troubleshoot as recommended in the Note

Thanks ,

Manu

[Former Member]

thank u Manu for the quick response.

in sm 50 i see the job is on hold but in sm 37 it is active..

and how to check if it is updating anything in the system? i dont see any logs in the backend systems after the job started in May 17, 5:24 AM MST.