Fault-tolerant workstations and replanning

About this task

To replan, you can configure IBM® Z Workload Scheduler to automatically send the Symphony file to IBM Workload Scheduler after you modify the current plan. To send the new Symphony file automatically, set the CPUAUTOLNK parameter of the CPUREC initialization statement to ON.

Otherwise, you can control manually when the Symphony file is delivered to the distributed agent and set the CPUAUTOLNK parameter to OFF. You can improve scheduling performance in these ways:
  • Reduce network congestion by controlling when the file is delivered
  • Keep jobs from being scheduled on a workstation
For the syntax and description of CPUAUTOLNK, refer to Scheduling End-to-end with Fault Tolerance Capabilities.
When replanning, the distributed agent can receive and run the Symphony file and report job status back to IBM Z Workload Scheduler only if the workstation is active and linked. From the MODIFYING WORK STATIONS IN THE CURRENT PLAN panel, you can view and modify the status of the fault-tolerant agent. Select option 5.5 from the main menu to display the panel:
Figure 1. EQQMWSLL - Modifying work stations in the current plan
EQQMWSLL -------- MODIFYING WORK STATIONS IN THE CURRENT PLAN  Row 1 to 8 of 8 
Command ===>                                                  Scroll ===> CSR  
                                                                               
Enter the row command S to select a work station for modification, or          
I to browse system information for the destination. If the work station        
is virtual enter the row command  V to vary the global status.                 
                                                                               
Row Work station               F V L S T R Completed     Active Remaining      
cmd name text                  A           oper  dur.    oper   oper  dur.     
'   CPUA Workstation 1           N   A C A     0    0.00      0     1    0.08  
'   CPUV Workstation 2           N   A C A     0    0.00      0     0    0.00  
S   CPU2 Workstation 3           N L A C A     0    0.00      0     0    0.00  
'   VWS1 Virtual workstation   Y Y   A C A     0    0.00      0     0    0.00  
'   PRT1 Printer pool            N     P A     0    0.00      0     0    0.00  
'   SETP Used to prepare JCL     N     G S     0    0.00      0     0    0.00  
'   STC1 Started task            N   O C A     1    0.00      0     0    0.00  
'   WTO1 Messages                N   A G C     0    0.00      0     0    0.00  
******************************* Bottom of data ********************************
In the S column, the possible values for status are the following:
A
Active. The workstation is running and scheduling its own plan.
O
Offline. The workstation is not running and scheduling its own plan.
The Fully Active (FA) column shows if all the destinations of the virtual workstation are active:
Y
All the destinations are active.
N
Some or all the destinations are not active.
In the L column, the possible values are the following:
F
Fully linked. The workstation is fully linked through its primary connection (either the domain manager or the OPCMASTER, if the fault-tolerant agent is a domain manager) and all secondary connections (the full-status fault-tolerant agents that are in the same domain). You can define a fully linked workstation only if ENABLESWITCHFT is set to YES in the TOPOLOGY statement. Fully linked workstations can be only fault-tolerant agents, domain managers, and standard agents.
L
Linked. Communication is established between IBM® Z Workload Scheduler and the workstation.
U
Unlinked. There is not an established connection to send and exchange events.
Job status includes information such as job started, job run, and job completed. When the new plan is produced, the distributed agents are stopped, and the scheduling on the agents is interrupted. The scheduling is resumed only when the new Symphony file is delivered.
When you send the Symphony file (CPUAUTOLNK=OFF) manually, you must manually link the fault-tolerant agent. Perform the following steps:
  1. In the MODIFYING WORK STATIONS IN THE CURRENT PLAN panel (EQQMWSLL - Modifying work stations in the current plan), use the S row command to select the fault-tolerant agent.
  2. In the MODIFYING A WORK STATION IN THE CURRENT PLAN panel, enter one of the following commands:
    Figure 2. EQQMWSTP - Modifying the status of a fault-tolerant workstation in the CP
    EQQMWSTP ------- MODIFYING A WORK STATION IN THE CURRENT PLAN -----------------
    Command ===>                                                                   
                                                                                   
    Select one of the following command:                                           
     L LINK             To send command link work station                          
     U UNLINK           To send command unlink work station                        
     S START            To send command start work station                         
     P STOP             To send command stop work station                          
                                                                                   
    Work station          : FMAX        COMPUTER AUTOMATIC FTW                     
    Type                  : Computer    FT Work Station                            
    Status                : Active                                                 
    Status link           : Linked
    L
    To link the workstation
    U
    To unlink the workstation
    S
    To start workstation
    P
    To stop workstation
  3. The MODIFYING WORK STATION STATUS IN THE CURRENT PLAN panel displays the change.

In some situations the Symphony file in IBM Workload Scheduler is not created. For example, there might be a recovery from a disaster situation. In this example, the new current plan is created, but the Symphony file is not. In this situation, you can keep the current plan up-to-date by selecting the 5 (SYMPHONY RENEW) option from the PRODUCING OPC DAILY PLANS panel.

The following are some other examples about the need of renewing the Symphony file:
  • You change the TCP/IP port number on a workstation.
  • You modify the CPUREC or DOMREC initialization statements.
  • You make changes to the job in the script library EQQSCLIB.
  • End-to-end scheduling with fault tolerance capabilities has configuration errors, for example the specified directories do not exist, the file system is full, or the workstation is not correctly defined.