Problems with RFC Resource Allocation

You have defined your parameters so that resource bottle necks will not arise. Your applications make RFC calls of the permitted types only. Yet you have realized that more work processes than allowed are occupied with RFC calls.

This can happen in the following circumstances:

An asynchronous RFC transmits a synchronous RFC

The asynchronous RFC transmits one or more synchronous RFC calls in the destination system. The profile parameters are not valid for these calls.

This problem is due to an error in the kernel and can be fixed by a kernel patch. The attribute with makes parameters configurable is passed on from the first RFC (this must be the right type) to other RFC calls.

Work process is changed by roll-out/roll-in

The asynchronous RFC occupies a work process in the destination system, but is rolled out because, for example, it contains the ABAP command COMMIT WORK or WAIT. If it is later rolled back in, the system does not check if the quotas have been exceeded.

This problem is due to an error in the kernel and can be fixed by a kernel patch. You can find out the patch number from the SAP Note 595032.


With the rdisp/rfc_check parameter in the RFC server you can change the settings in the check for how many RFC work processes are available.

No comments:

topics