SYSTAT Error handling scenarios

, ,
1

Hi All,

I have some basic design questions about technical Ack being sent to SAP system from webMethods:

  1. Is there any relation between IDOC generation table and SYSTAT table?
  2. If SYSTAT is failed to reach SAP system what will be the best possible
    alternative to handle this.
  3. How to validate/check not to send duplicate SYSTATs to SAP system.
  4. Is there any chance that SAP team can alert if a particular SYSTAT is not
    received by them for a corresponding IDOC sent to wM.

Please help me out.
Thanks a lot in advance.

Regards,
Datta

2

A Gentle Reminder to All,

Please help me out.
Thanks in advance!

Regards,
Datta

3

Hi Datta,

  1. What do you mean by IDOC generation table?

  2. Ok, I checked this: if a SYSTAT01 IDoc fails (e.g. network error while sending), then the systat mechanism does not delete the corresponding TIDs from its list. This means when the next SYSTAT01 IDoc is due, they will be included again (and this time the IDoc will hopefully be successful…).
    However, there is one pitfall: if the sweepTRX job (you have one scheduled, don’t you?) deletes the corresponding status information from the message store before the next SYSTAT01 IDoc is due, then the systat mechanism won’t be able to determine the correct status for these TIDs. Therefore the answer to your question is: if a SYSTAT01 IDoc fails, make sure the sweepTRX does not interfere! This could for example be achieved by temporarily disabling sweepTRX, or by making sure the systat repeat time is at least twice as high as the sweepTRX repeat time.

  3. The systat mechanism should do this automatically. I know there have been one or two bugs, where the systat mechanism sent duplicate information to SAP in certain situations, but these have been fixed in 2004 or 2005. If you still see this happen (and you have the latest SAP Adapter patches), then open a support request with SAG.

  4. Tell the SAP team the repeat time you have defined for your systat job. (Depending on the overall load, I recommend times like 1-2 hours, or if the system is really heavily under water, do it once a day at night.) Let’s double that time, so that if one SYSTAT01 IDoc fails, the next one will get a chance. So the SAP team should check the status of an IDoc 2 x repeat time hours after it was sent. If the IDoc is still in status 03 by then, they should raise an alert.

Lanzelot