cancel
Showing results for 
Search instead for 
Did you mean: 

Benchmark - rdisp/max_wprun_time parameter

nilocraig
Explorer
0 Kudos

Lately we are experiencing several time-out dumps in our system caused by a few programs/routines. We have adjusted and fine tuned the code, but with the growing database, the long processing time is justified. These routines cannot be executed in background as per our business requirement.

We still use the default 600 seconds (10 minutes) setting for our rdisp/max_wprun_time parameter and we would like to increase it, but our admin guys say this will cause an impact on system performance.

That is why we decided to conduct a small survey to verify how this parameter is set in other businesses, if it has been changed from default and what is the current value. And finally if you experienced any issues when changing this value.

Based on our preliminary calculations, 1800 seconds would be enough, and this affects only about 3% of our programs.

Any help or input would be appreciated.

Best regards,

Nilo Craig

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Hi Nilo,

Yes, 1800 seconds is quite normal for an ERP system. Where as values of as high as high 3600 are quite common in BW and CRM applications.

The impacts which I can think of is that a long running dialog job might hold up resources like extended memory( shared memory which other processes can also use) and heap memory(private memory)  for a longer duration and only when they are processing a big dataset.

A number of such jobs running together can impact the system but it again depends upon what kind of dialog load you are having on your system and how your system is tuned.

Few of these dialog jobs running for long should not impact your system much if you have enough dialog resources and application servers in your production environment.

Thanks

Amit

Answers (3)

Answers (3)

Former Member
0 Kudos

Hello Nilo.

There is a point to consider.

The number of btc work process:

If the number of btc wp is not enough, increasing the max_wprun_time parameter can delay other jobs.

You can analysis the utilization of the btc wp in sm50, the number of free background wp, the average of utilization.

Sriram2009
Active Contributor
0 Kudos

Hi Nilo

1. In your case you can enable the ST12 trace check the ABAP & DB usage for single user with help of ABAPer and then take required action

Kindly refer the transaction ST12 trace usage steps

ST12 Trace - Step by step instruction on how to use it for analysis

2. During the time out is this any bottle nick of hardware resources(like memory or network or process) ?

     During the peak time could you check the transaction code ST02 over all memory & buffer values status?

3. You can refer the SCN link

http://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/a4dcf62d-0a01-0010-afa5-e8edcecbc...

Normally we can define the parameter rdisp/max_wprun_time  in ECC - 10 Min & BI - based on your reports developed

Regards

Sriram

alwina_enns
Employee
Employee
0 Kudos

Hello Nilo,

probably you could proceed as described in the note "25528 - Parameter rdisp/max_wprun_time":

  • "Every program that calls up a "COMMIT WORK." at regular intervals resets its
    clock to 0 and can have longer run times than defined by the parameter
    rdisp/max_wprun_time."


For a BW system recommendations on this parameter are in the note "192658 - Setting parameters for BW systems".

Regards,
Alwina