cancel
Showing results for 
Search instead for 
Did you mean: 

update processes are occupied fully by one user ID

Former Member
0 Kudos

Dears,

The update processes are occupied fully in two SAP Instances by one user ID, this happened for the second time due to the same user and same transaction,

it caused some IDOC generated very very slowly, 

actually there are still some free update processes in the other two instances, but why IDOC processing update cannot assign to the free update processes?

Thanks,

Michael

Accepted Solutions (0)

Answers (5)

Answers (5)

former_member182657
Active Contributor
0 Kudos

Any recent changes under kernel release ? Try to patch the current kernel with the latest one available at SMP and re analysis your system.

Regards,

Gaurav

former_member182657
Active Contributor
0 Kudos

Hi Hong,

Are you facing occupation of update WP while if you execute/work for any other Z or standard reports.Please confirm.

Additionally share some check results during the execution of same transaction.

1.Find out are there any records being locked under SM12 for the same transaction.

2.Locks held till the commit statement happens in the code ,check these exclusive locks with DB01 or v$lock,and find out the SQL causing the issue.

Hope this will help you.

Regards,

Gaurav

Former Member
0 Kudos

Hi Gaurav,

User is executing Z reports, and it generate many update tasks, there are about 4 update wp run more than 30 minutes, after I canceled them by manual then it recovered normal.

1. some VBAK and VBRK tables are locked in SM12 for the same transaction.

2.there is no any locks in DB01.

system always report the alert 'High update response time', we increased two dialog instances last month.

SAP recommand us to adjust some oracle parameters, but I want to get the root cause first, because it never happened before.

Thanks,

Michael

Sriram2009
Active Contributor
0 Kudos

Hi Hong

You can enable the ST12 trace, check the ABAP & DB usage with help of ABAPer

Regards

Sriram

divyanshu_srivastava3
Active Contributor
0 Kudos

Hi Michael,

Check for deadlocks in this system.

Also, what about the logs ?

Check with the user what he is doing. Ask him to recheck what he is doing ?

Regards,

Divyanshu

Former Member
0 Kudos

Hi Divyanshu,

There is no any deadlocks in this system when issue happened.

only two dialog instance's update were all occupied by the same user and same transaction.

Users replied he just do the same operation as normal.

Abaper also analized the trace log and think it's normal.

But now almost everyday some user's transaction will occupied many update work processes,

then it will cause other user's update request always waitting in the system, it impacted seriously for other update business, the system's running speed is also normal when one user's transaction occupied many update work processes and running about 1 hours.

So what any other logs can we checked for this issue ?

Thanks,

Michael

alwina_enns
Employee
Employee
0 Kudos

Hello Michael,

please check the traces of the affected update work processes, if any errors/deadlocks are reported.

Regards,
Alwina

former_member182657
Active Contributor
0 Kudos

Hi Hong,

As per attached screenshot there is one entry with customized table ZCNC_NEW_OP_VAL for which you're facing the issue.First i would suggest you to consult with your development team.

In addition share SM21 system log details for analysis.

Regards,

Gaurav