cancel
Showing results for 
Search instead for 
Did you mean: 

Random number of iDocs going missing

Former Member
0 Kudos

Hi,

A client is experiencing some issues on their QAS systems with a random number of iDocs going missing during the transfer from ECC to PI.

I have read through countless discussions already and I'm drawing a blank.

The setup is as follows:

PI QAS System:  Windows/Oracle (11.2.0).  Netweaver  731.

ECC QAS System:  Windows/Oracle (11.2.0).  ECC 6.0 EPH2.  Netweaver 7.0 SP29

I have a job I can run with a variant which attempts to transfer 500 iDocs from ECC to PI.

On the last attempt 162 were found on PI after the job ran.  The time previous to that, 197 were found on PI.

Status of all iDocs in ECC = 3

SMQ1/SMQ2/SM58 have been checked - all clear.

SM21 is swapped with warnings though.  I've attached an example. 

I'm assuming the RFC is dropping out part way through the process maybe??  SM59 tests all are fine.

Any suggestions on where to look next would be greatly appreciated.

Thanks

Accepted Solutions (1)

Accepted Solutions (1)

iaki_vila
Active Contributor
0 Kudos

Hi Paul,

What PI IDOC adapter are you using?, if you are using the ABAP one you can try to trace the IDOCs with the IDX5 transaction:

Comparison of IDoc Adapter (IE) and IDoc Adapter (AAE) - SAP NetWeaver PI Reference Documentation - ...

Regards.

Former Member
0 Kudos

Thanks for the reply.

There is something I totally missed.

ECC:  SM58 is not all clear.  Dozens of:

IDOC_INBOUND_ASYNCHRONOUS:  PI:  Logon Failed

Why would only SOME fail and others are successful??

Thanks

former_member184720
Active Contributor
Former Member
0 Kudos

Thanks for the suggestion.  I have reviewed the Note.  These parameter values are exactly the same in the Production environment.  And this does not happen in Production.  Only in QAS.

So in this case, I don't feel that increasing the parameter values will help.

Any further suggestions would be welcomed.............

Answers (2)

Answers (2)

Former Member
0 Kudos

For anyone with a similar issue in the future, two changes were made here which ultimately fixed the issue.

Firstly we increased the value of rdisp/max_arq from 500 to 2000 - this increased the number of iDocs moving across but still didn't fix the issue.

And also, we noticed the RFC was missing a language setting.  We maintained that and bingo!  All started working!

So a combination of the two has fixed it.

Former Member
0 Kudos

Hello Paul,

your SM21 log also points to logon issues with your NetWeaver component.

Are you running a clustered PI environment?

What happens if you re-execute the messages in SM58 (check the user you're doing this with)?

Regards,

Jörg