cancel
Showing results for 
Search instead for 
Did you mean: 

Data Load Problem in CFIN system - IDOC / SM58 logon issue

rameez_khan
Active Participant
0 Kudos

Dear All,

We have done IDOC setup between ECC and CFIN system. ECC is on prem and CFIN system is on Azure Cloud.

We have setup IDOC between 2 system which go via RFC, no PI involved.

Now we want to do initial data load and we send around 20K IDOC's for customer from ECC to CFIN.

However IDOC's get stuck in SM58 with error of "Logon Failed".

Note: parameter is set: rdisp/tm_max_no to 3000

Has anybody faced similar issue in past ? How do we resolve this ?

Thanks for help

Rameez Khan

Sandra_Rossi
Active Contributor
0 Kudos

Yes. Possibilities in SM59: wrong user, wrong password, wrong host address, remote user locked, etc.

rameez_khan
Active Participant
0 Kudos

No Sandra, Because some idocs are sent to CFIN system. 1000 Idocs get processed but rest 19K are stuck in SM58

Nikhil_Rao_ABAP
Participant
0 Kudos

Is it possible to entirely serialize the sending or send in smaller packets? one possibility could also be that many idocs/idoc packets make simultaneous calls to your CFIN system and hog up all the dialog sessions leaving no session open for remaining packets, which get the logon failed. The one's that do(did) get a session end(ed) up getting processed.

Regards,

Nikhil Rao

rameez_khan
Active Participant
0 Kudos

Hello Nikhil,

Thank you for your suggestion. Idocs are getting processed in background and as correctly suggested all background processes are getting used up. Currently we have around 40 work process. And also I have kept batch job for idocs sending 3K an hour. Even when sending 3K hourly I am getting this error.

But we cannot keep 1K or 2K in an hour. Initial data load volume is huge.

Thanks

Rameez

Nikhil_Rao_ABAP
Participant
0 Kudos

Hello Rameez, yes agreed that you would have background jobs set up in both systems( even multiple ones since you noticed that even the BTC processes were used up).... but my guess is this: the BG jobs in ECC will spawn "logged on dialog processes" in CFIN to send the idocs/idoc packets( so you need to make sure that this is a situation that is well controlled).... in addition, I now noticed that the idoc processing is also happening in the CFIN system via BG jobs set up there(i.e. the idocs arrive in the cfin system and are queued up there to be picked up for processing later).... note that we could have a scenario that the parent job/s is/are running in BG in CFIN, but they too are spawning parallel dialog sessions in the CFIN system for parallel processing( not only will these be contending with each other, but also with the sending processes that the ECC job is spawning in your CFIN system). So run a few combinations to ensure enough availability.

Accepted Solutions (0)

Answers (0)