SAP BW Data packet processing issues – Second attempt of processing

This document explains one of the issues of data packet processing that we faced in our project. Data packet shows the message and it is locked saying second attempt of data processing is successful and it remains active with yellow status with warning as shown below for data packet 20.

Screen shots

As you can see in the first screen shot below WBS element data load got stuck due to second attempt error and job remains active indefinitely unless we process this data load.

Before starting the below job there are few steps those needs to be followed.

Go to the data load request and set it’s QM status to red as shown in the first screen shot above.

Go to Edit – Init Update — Setting for further update

Data packet processing should be set to be processed in the background.

Once this setting is done go to data packet and right click and select the manual update option. This option is only shown if data packet is in yellow status. Once you start manual further update following job will start with BI_BOOK* ID. You need to monitor this job in SM37 till it gets completed successfully.

Once this is completed successfully see to it that QM status for this request turns green overall. If not you can make it green manually.

Treatment of warning can also be handled as below in Infopacakge settings

There are multiple reasons for this to happen.

· This can occur when doing massive parallel updating of time-dependant master data into the target characteristic. Solution is to reduce the number of Data Packets updating in parallel; effectively serializing the update by processing 1 Data Packet at a time.

· This can occur also during the nightly load window when there is a lot of other cube based activity occurring in the system due to different process chains scheduled at the same time. Solution is to re-arrange the nightly load window process chain to do only master data activities and then do transaction data after, not both at the same time.

· You could also review the sequence of loading the master data attributes. The system will behave more efficiently when you execute the master data loads in the order of their relational dependencies with each other. As you know, a characteristic has attributes. Each attribute is a separate characteristic in its own right. Solution is to load the simpler, base characteristics first and then load the characteristics that have them as attributes. This will help to ensure the SIDs are all in place first and reduce the subsequent master data loading to only need to check if the values exists (which it now does).

· Treatment of such warning can be handled as this is occurring say in master data you can make the QM status green as shown above and further processing of process chain will not get affected due to this.