Skip to Content
Oct 17, 2021 at 01:50 PM

Work processes hang after remote server failure


We have system A communicating with system B via RFC.

If there is a hardware failure or network issue in one of the application servers in system B, all work processes in system A hang on connection to that application server for ten minutes. Eventually all connections fail.

When the connection fail the error and return code in work process developer traces in system A are

Error RFCIO_ERROR_SYSERROR in abrfcpic.c : 3259
CPIC-CALL: 'ThSAPOCMINIT' : cmRc=17 thRc=239

I'd like to know where the ten minute timeout comes from. I've checked all gw and rdisp parameters in system A and the only one that is set to ten minutes (600 seconds) is gw/deallocate_timeout

During the hang work processes have status "On Hold" and reason "RFC".

We know there is a hardware issue behind this so the reason is clear, but we don't know where the ten minute wait is coming from. We'd like to reduce the wait time to avoid impact on the business.