cancel
Showing results for 
Search instead for 
Did you mean: 

Long runtime of loading stage on a process chain

Former Member
0 Kudos

Hi experts,

Just want to ask what is the possible cause of a long runtime of a specific loading step under the process chain

usually it runs 1.5hrs-3 hours with only 12,000-15,000 records being process.

Many thanks in advance!

Accepted Solutions (1)

Accepted Solutions (1)

RamanKorrapati
Active Contributor
0 Kudos

Hi,

its depend on the load on your servers and  application servers(free).

At which step its taking more time? dtp or info pack?

thanks

Former Member
0 Kudos

Hi thanks for the response,

Im running 3.x so I don't have DTP I believe its on Info pack

thanks!

RamanKorrapati
Active Contributor
0 Kudos

Hi,

During your info pack load, have you seen how many application servers are free at SM50.

Even by seeing your info pack back end job at ECC SM37, we will know where its taking more time to load data.

From info pack monitor menu Environment--> job overview--> job in the source system, enter ecc user id and pass word, you will ger your job, select job and click on log details. observe the log details step by step.

Mean while check ECC Tx - SM58, some queue may be stcuked here. need to process them by manual(F6).

Thanks

former_member186082
Active Contributor
0 Kudos

Hi Christian,

Is it a full or delta load?

Check the number of records in source and if this more, definitely load time will be high.

Check the job in source, no issues until it is progressing. If it is not progressing or cancelled, you have to step in and take some actions.

Regards,

Chandra.

Former Member
0 Kudos

it is a full load

Answers (5)

Answers (5)

Former Member
0 Kudos

Hi Christian,

The load my be extend may be the below mentioned reason

1) Backend locks

2) check the processes in sm50

3) Server slow

4) check how many background jobs running in sm37.

5) SM51,SM66 in source system - What is running for this extractor,which program is being read,which table etc...

6) Check also the table space available  in St04 .

Regards,

Baskaran

former_member202718
Active Contributor
0 Kudos

Hi Christian,

Where are you getting stuck

  •      ECC-PSA ? if its ECC to PSA then check the CMOD or FM.
  •      PSA to DSO ?
  •      DSO to Cube ?

If its to DSO or Cube then check for the Routines - (Start or End Routine).

And check for the code in the routines.

Example 1:-Also if you are doing a Look Up to another DSO to populate a New Filed and then using a Non Key to do extract the data then it would take time as its Non Key in that case build Secondary indexes in DSO.

Example 2:-Also If you have multiple loops and Selects statements inside Loop.

rgds

SVU

Former Member
0 Kudos

Hi Christian,

In general, the following will affect loading time:

1) Number of records/fields to process

2) Number of times require data lookup (database or itab)

3) Size of the data to lookup (related to point 2)

And you may also look for loop in the logic of any routine (start/end/field/expert).

As loop increase number of operatiosn, and it is kind of increasing the number of point 1 and 2, specially nested loop.

I will suggestion you look at the process monitor first to indicate which part cost most of the times.

(Start routine/Transformation/End routine/Update?)

And then try to check what is happening.

In general, if cost most time in:

1) Start routine/End routine

Looking for loop, select into itab and try to time it

2) Transformation

As transformation is really a loop, try to spot if any field routine have a database look up or complex logic, try to comment them one by one and see if it make any major difference

(In the order by your understanding to data)

3) Update

Too many records? (From source or generated by routines)

Regards

Bill

Former Member
0 Kudos

It has start routine from source to target,,,

how does a star routine affects the loading step? thank you so much!

MGrob
Active Contributor
0 Kudos

Hi

A startroutine can contain code that is inefficient in updating/deleting data or do a lookup from another target which can impact the runtime. You should analyze the infopackage in which step it burns the most time. Inserting the data by handling the updaterule or selecting the data from the source etc.

hope that helps

Martin

Former Member
0 Kudos

Hi,

It might the cause of the datasource in the source system. Can you check the datasource in RSA3 in Debug Mode with the help of Abaper.

The problem persists, then you need to split the loads by using the infopackage selections.

Thanks,

Purushotham.