cancel
Showing results for 
Search instead for 
Did you mean: 

R/3 -> XI : IDOC disappears!

Former Member
0 Kudos

Hello forum,

I have really a strange situation here. It's the following:

We created our own IDOC type and our own program to send that IDOC (its an R/3 46c). we20 & we21 are configured to send the IDOC to XI.

Sending the IDOC is just fine - in bd87 I see a green light. Status 03

On the R/3 side i see NO entry in the sm58 - so tRFC seems to have worked fine.

However: In XI there is NOTHING.

sxmb_moni, sm58, idx5, smq2 - I see no sign that anything happens. No message. Nada

Also no metadata is loaded in idx2. Of course also nothing in the CC monitoring (IDOC adapter does not run on Java, right?)

I checked like 10 times the way out of the R/3. Going bd87 and there to the port, we20, sm59, remote login and I am in my XI. So I am pretty convinced that I am not sending to just some other system.

The only strange thing is that if try to load the metadata manually, I get the message that "segment versions could not be read". However, after going out and back into idx2 and I see the metadata. Retrying the sending afterwards gives the same result.

I browsed the forum and added a TRFC port in XI that references back to the R/3 in we21 - w/o any impact on the above described.

I still think that he is somehow sending the IDOC somwhere else - even if I verified that he is not.

Am I missing something? Does sbd. have an idea about what's wrong here?

Any hint is appreciated!

Helge

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Helge,

Just a quick check.

Have you selected proper timings in SM58 and IDX5?

Have you selected the user as * in SM58?

Regards,

Jaishankar

Former Member
0 Kudos

Yes, I did. Also checked all other SAP systems we have around here. These IDOCs are just gone....

Answers (1)

Answers (1)

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

hi

try load the metadata manually,

also chck you have the runtime variable in sxmb_adm > integration engine configuration>

runtime logging 1

runtime logging_sync 1

runtime trace_lavel = 3

Thanks

hope it helps

Rgds

Rodrigo

Former Member
0 Kudos

Hi Rodrigo,

I double checked that one. The parameters were already set as you suggested before my first try.

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

try to send it again if you set the parametes in integration engine configuration,

are you sure you are sending idocs to system you say?? dont get angry, but check it again, is rare you send and IDocs, status says OK and you dont have nothing in IDX5

Thanks

Rodrigo

ps: Award points if useful

Edited by: Rodrigo Pertierra on Mar 5, 2008 3:59 PM

Former Member
0 Kudos

Hi Rodrigo,

you are right - i also never saw such a situation in 4 years full time with the XI...

I double checked it again and asked 2 colleagues. They see the same as I do.

We are currently checking the trace files from the message gateway. Whenever we send an IDOC we see that file growing. But there are no errors or warnings. We increased the log level - nothing...

Helge

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

only with Idocs yo have problem??

what about SXI_cache?? is it ok?? if you go to sxi_monitor and try to do a filter can u see al data about scanarios??

thanks

Former Member
0 Kudos

yes - I only have this issue with IDOCs sent to the XI system. Everything else is working fine.

This is actually on that XI the first scenario like that. So it could be a basic installation error - this is why I am currently focussing on the message gateway.

Cache is fine nd the monitor shows all messages but the incoming IDOCs...

Former Member
0 Kudos

Now I did the following: I changed the passwd of my user in the XI system and sent the IDOCs agein.

With the SAME RESULT - no error.

So I conclude he is sending the IDOCs somwhere else than to the XI system.

However, for the sending requests I see entries in the log file of the message gateway. Could it be that the message gateway is sending the IDOCs to another system?

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

thats whats said you before, check if the systems you are sending the IDOCs is this PI. in r3 go to we02 and look for the port the idoc was sent, then go to we21, search the port and confirm that this port link to PI.

Thanks

Rodrigo

ps: Reward point if useful

Former Member
0 Kudos

Alright- after spending the night with it here's what I discovered:

The actual configuration is accurate. Its a problem on R/3 side. The IDOCs are just not sent. Hasso knows why...

I changed the idoc processing from 'immediate' to 'packaging'. If I release the buffered IDOCs in bd87 manually they are actually sent (better: they were not but I got an error in sm58 that I corrected; so this scenario works now).

However, switching back to the immediate processing it does not work at all.

Being more an XI than an ALE guy - does anyone have an idea why he (the dear R/3) could only pretend to send the IDOCs but actually doing nothin'?

I will now start searching - but if anybody could tell me which job I have to schedule in order to automatically send buffered IDOCs in the background I would be grateful, too.

Helge