gtpd2m28 | Data Communications Services Reference |
TPF system operations are generally independent from the Message Routing Facility. This is to say that the MR Facility does not alter system operations procedures which are outside its own domain. For example, procedures relative to a TPF system remain unaltered whether the CPU is a stand-alone processor or part of a computer network system.
The information in TPF Operations about the message routing commands describes, in detail, all of the facilities and the control language (commands) necessary to manage them.
This section is intended to review some of the operational procedures directly related to the MR Facility. It is assumed that all the phases associated with system generation and loading have been executed resulting in all the MR programs and system records to be stored in the appropriate areas on the online files.
The Routing Control Application Table (RCAT) records are built as a result of the initial (first time) IPL of the system, from the RCAT Initialization Table (RCIT) records. Subsequent IPLs of the system will use the keypointed records from file.
Following the initial system IPL, the operator is notified of the completion of the system initialization by the advisory message hh.mm.ss. RESTART-1052 STATE. At this point the operator would initialize the Routing Control Blocks by entering the appropriate command. Note that RCB initialization is only required after the initial system IPL. The need for selective RCB initialization may arise in the course of normal system operations in the event the system areas of an RCB is corrupted or as an emergency procedure if all the terminal log out attempts have failed.
The operating procedures to cycle up the system and activate the various types of communication lines are independent of the message routing facility and are described in detail in TPF Operations. Procedures relative to the activation of SCP systems are to be found in the appropriate books. It is assumed, for the purpose of continuing the review of the procedures, that all the systems in the computer network have been activated and are operational.
An application whose name is in an RCAT entry can be started in any system
state through a command of the type:
ZROUT STRT xxxx |
Where:
Explanation: Obviously, before the application is started all the necessary operative procedures to make all the needed resources (for example, files) available to the application, must have been completed.
After the Start request has been processed, terminals on the network are capable of having log-on requests (to that application) honored by the TPF system.
System Action: None.
Error Response: None.
An active application can be stopped in any system state through the use of
the following message:
ZROUT STPI xxxx |
Where:
Explanation: If the TPF system is in 1052 state the application status in the RCAT is set to STOPPED. When terminals are logged in, and the system is above 1052 state, a broadcast message is sent to notify the operators that the application is being stopped. The operators then have 15 minutes to finish what they are doing. When finished, the operators should log off. At the end of 15 minutes, terminals that are still logged in are automatically logged off and the application status is set to STOPPED. Permanently logged terminals are not affected.
System Action: None.
Error Response: None.
All the MR system records that during normal system operation reside in storage are keypointed through the Keypointing mechanism with the exception of the RPST and RCLT records residing in a non-EP processor in an loosely coupled complex. Therefore, the MR restart programs in the restart schedule are capable of restoring the MR facility to the status in which it was prior to the restart occurring.
No special operational procedures are needed to restart the MR facility. Exceptional conditions may require intervention such as RCB initialization, stopping an application etc.
At any instant, the processing capabilities of a TPF CPU depend on the system state (as defined in TPF literature) in which it is operating. When MDBF is included, the term system state refers to the state of the subsystem in which the application resides. In order to render the total computer network system relatively independent from the combined state of each CPU, the action taken by the Message Routing Facility on the data received from either internal or external sources, depends exclusively on the processing capabilities of its own CPU.
While this does not directly affect system operations, it is nevertheless opportune to be aware of these actions for they are dependent on whatever steps the system operator takes in any CPU relative to the changing of system states.
The MR action (reject or route) is shown in Table 7. In general, the following rules are observed:
Message Destination | Message Origin Type | 1052/UTIL System State | CRAS System State | MESW/NORM System State |
---|---|---|---|---|
Terminal or Program in Another CPU | Not Applicable | Reject | Route (1) | Route (1) |
SMP-Like Applications in the Host CPU | Program | Route | Route | Route |
SMP-Like Applications in the Host CPU | Terminal (2) | Route | Route | Route |
Application Not SMP-Like | Program | Reject | Route | Route |
Application Not SMP-Like | Terminal (2) | Reject | Route (3) | Route |
Application not SMP-Like in the Host CPU | Terminal (2) | Reject | Reject (3) | Route |
Terminal Connected to the Host CPU | Program | Not Applicable | Route | Route |
Notes: