Bad performance in sending IDOCs from WAS to R3

[Former Member]

Hi all,

We have the following scenario:

WAS 7.0 – XI – R3

with protocol: SOAP – XI – IDOC

When IDOC comes from WAS to R3, it appers always at first time a 405 error (we see it in ‘sxmb_moni’ transaction of WAS). After 3 or 4 minutes it is reprocessed automatically and it arrives OK without doing anything.

(It doesn’t happen from R3 to WAS, we mean that it works OK in this direction)

Does anybody know why it happens? How can we perform it faster?

Thank you in advance,

Kind regards,

Xavier

[udo_martens · Accepted solution]

Hi Xavier,

it is reprocessed because of a job, which is trying to restart failed asynchronous messages periodicly. Plz look to SM58 and SMQ1 if there are failed messages and try to find out the reason. Look as well to SM21 for informations, may be there is a network or destination problem. And of course it is easier ti find out your problem if you post the hole error message from SXMB_MONI. Put the trace there to 3 (SXMB_ADM / Configuration Integration Engine / Specific Configuratin / RUNTIME / TRACE_LEVEL), to have maximum of details.

Regards,

Udo

[Former Member · Accepted solution]

Hi Udo,

We have tried to see any information in the transactions that you have indicated us.

In SM58 and in SMQ1 there were not any entries.

In SM21 appears a message with “warning symbol” that says: “Perform rollback”. It is also indicated:

Process Dialog work process No001

Program: SAPMSSY1

Package: STSK

Does it say you anything?

Do you know another point to investigate where is the source of this problem? Do you think is a standard error and there’s any note that we haven’t found?

Thanks in advance

Xavier

Message was edited by: Xavier Ester Duxans

Message was edited by: Xavier Ester Duxans

[Former Member · Accepted solution]

Hi Xavier,

Assuming htat you are talking about HTTP 405 error,

please can you have a look at note 906435

Regards

Vijaya

[Former Member · Accepted solution]

Hi Vijaya,

First of all, we are sorry for the delay. We have had some problems with the J2EE Engine, and it was not possible to realise tests to try to solve the problem of this message.

We have read the note you indicated us (906435) which says that 3 reasons can cause our problem and propose a solution for each one.

1) We have added and modified some parameters as it is indicated in the XI Tuning Guide, but the problem persists. So, could you possibly recommend us which parameters may help us?

We are working with SAP AII 4.0, XI and R3. There is a CPU for each system. The problem is always in the direction from AII to R3. There’s always only one user connected to SAP, who sends only one IDOC at the same time, and he doesn’t do any action in the system until the IDOC arrives to R3.

Do you think it could be a problem of memory? Do we need any specific conditions of this in our case?

2) We have modified the parameter “icm/server_port_0” as it is indicated and nothing has changed.

3) The third point is OK, we have ICM 7.00. So, we don’t need to check our kernel patch level.

What other information we can provide you to try to help us?

Thanks in advance.

Best regards,

Xavier