Issues with respect to Database node failure or Production support system related issues we need to carry out further checks as follows

  1. Check jobs those are running in SAP BW system. We need to check in detail
  2. Check job status in SAP ECC system. You can see in Tcode SM37 running job log.
  3. Check job status in the third party scheduling tool like Autosys or Control M.
  4. Check job status which is loading the data from BW to other data warehouse via open hub.
  5. Check SM58 for any TRFCs which are stuck
  6. Check for BD87 for the IDOC stuck with issues.

To discuss above points I have used example of CCA load which failed, While rest of the running loads were failed lets say BPC data load and Planning loads, Z_0CO_OM_CCA_9_XXX process chain was running and adding the records.

So we should look into detail tab as shown in the screen shot below

DB Failover1.png

Check Transfer (IDOC and TRFC) for errors or messages.

DB Failover2.png

As you can see the first data packet itself failed to process completely at 19:18:26.

Packet 67 as shown below fail to get the IDOC data.

DB Failover3.png

Around same time in BD87 you can see there is one IDOC which is in yellow and stuck status

DB Failover4.png

After that data packets were successful. But you can see the processing timing are different.

Data Packet 68 processed at 04:56:54 on 8th March

DB Failover5.png

Data packet 69 processed at 19:42:47 on 7th March 2012 as shown below.

DB Failover6.png

So normally processing of data packets can be random. The last packet can be processed first and get updated in data target and the first arrived packet can be processed at the end.

DB Failover7.png

So since yellow packets can be processed later we never raise alarm till the maximum processing time gets completed for that particular load or we can check behavior of the load with the help of basis team.

So in the past also we faced such situation but definitely not for so many data packets but it used to stuck few packets like 4-5 data packets in say WBS load.

We used to manually update those packets in the data target.

DB Failover8.png

Also in case if there is any stuck IDOC in the system we can manually execute them in SM58 or BD87 with the help of basis team.(As BR team may not have authorization)

DB Failover9.png

DB Failover10.png

DB Failover11.png

Conclusion

  1. If there are many data packets have failed status or yellow status you can ask basis to kill the job immediately, delete the old request from data targets and restart the job again.
  2. If there are few data packets have issues wrt processing you can carry out manual update.
  3. If there are few IDOCs stuck we can request basis to correct/Process same through BD87 or SM58.
To report this post you need to login first.