This topic describes:

• The flow of a directed command in a RACF® remote sharing facility (RRSF) environment
• The types of errors experienced in an RRSF environment
• Handshaking between RRSF nodes
• The connection states of RRSF nodes and the transitions between them
• Actions to recover from an RRSF failure
• The recording of RRSF errors

The INMSG data set is used to temporarily hold requests that are being sent to the local node from itself or a remote node, such as commands directed to the local node or output from RACF commands that were directed to a remote node. The OUTMSG data set is used to temporarily hold requests that are being sent to a target node, such as commands directed from the local node. If RRSF experiences a failure, it:

• Detects the error when it first occurs. This:
  • Prevents additional damage to the RACF subsystem address space and the RACF database.
  • Allows either the system programmer or the IBM® support center to diagnose and fix the problem.
• Protects the rest of the RRSF network from damage by isolating the error to the failing node.
• Saves the work request for retry when the error has been corrected.

Figure 1 shows an illustration of the flow of a directed command in an RRSF environment. The steps shown in this figure are:

1. A user issues a RACF command. This command enters the local node's OUTMSG data set while waiting to be sent to the remote node.
2. When this command is sent to the remote node to be processed, it enters the remote node's INMSG data set.
3. The command is processed in the remote node's RACF subsystem address space and the results are placed in the remote node's OUTMSG data set while waiting to be sent to the local node.
4. The results are sent to the local node and placed in the local node's INMSG data set until the output is returned to the user.