Failures when propagating SETROPTS commands

Before propagating a SETROPTS command, RACF® performs much the same validation that it has does if it is not propagating the command. The SETROPTS command must be valid on the RACF data sharing group member where it is issued for the command to be propagated. SETROPTS commands other than SETROPTS LIST fail if the coordinator is in read-only mode.

If the validation is successful, RACF attempts to process the command on the coordinating system. If the command fails on the coordinator, RACF does not attempt to propagate the command to the peer systems.

If the coordinator can process the command successfully, RACF requests that all peer members of the RACF data sharing group process the command. A propagated SETROPTS command can run on a peer system even if the peer system is in read-only mode. If a peer system encounters an error while running the command and it is running in non–data sharing mode, the command continues to run on all systems on which it can. However, if a peer system encounters an error while running the command and it is in data sharing mode, RACF terminates the command on all systems. RACF issues message IRRX006I on the console to identify the member of the group for which the command failed. Check the failing member's system log for additional RACF messages that further identify the problem.

If XCF fails trying to propagate a command, the command might run on some members of the data sharing group but not others. When you have resolved the XCF problem, reissue the SETROPTS command.

During the processing of a SETROPTS command on a sysplex, RACF records diagnostic information in symptom records written to SYS1.LOGREC if either of the following occurs:

Any member system in the sysplex sends or receives an emergency cancel during SETROPTS processing.
Each member system in the sysplex processes a SETROPTS command in two phases. If an error occurs after the completion of phase one but before the completion of phase two on a member system, that system sends a message called an emergency cancel, which it propagates to other sysplex member systems. When a member system receives an emergency cancel it terminates SETROPTS processing in progress.
The coordinator for a SETROPTS command leaves the sysplex before the command completes. The coordinator is the sysplex member system on which the SETROPTS command originated.

The IBM® Support Center can help with interpretation of the symptom records during problem determination.