Defining file dependencies in dynamic scheduling

You can manage file dependencies with dynamic agents, pools, and dynamic pools

File dependencies introduction

You use file dependencies in dynamic scheduling to control job and job stream processing that is based on the existence of one or more files or directories. When you specify a file dependency, HCL Workload Automation processes check if the specified file or directory exists before job and job stream processing starts.

You can select one or more of the following conditions, which are associated to the file, that must be true before the jobs or job stream processing starts:
  • The file exists.
  • The file exists and is a directory.
  • The file exists and is a regular file.
  • The file exists and is readable.
  • The file exists and its size is greater than zero.
  • The file exists and is writable.

File dependencies behavior

The file dependencies have a different behaviour for dynamic agents, pools, and dynamic pools.
Dynamic agents

HCL Workload Automation manages the file dependency resolution for dynamic agents, in the same way as for the workstation that is defined in the classic scheduling as fault-tolerant agent, master domain manager and its backup, domain manager and its backup, and so on.

If you define the JOB_A job on DYN_A dynamic agent, which depends on the FILE_A file, before JOB_A job runs, HCL Workload Automation processes perform a file existence check on the DYN_A workstation. When the FILE_A file is found, the dependency is resolved and the JOB_A job runs on the DYN_A dynamic agent.

Pools and dynamic pools
A pool contains several dynamic agent workstations with similar hardware or software characteristics. A dynamic pool contains several dynamic agent workstations that are dynamically defined based on the resource requirements you specify. If you define a file dependency for jobs or job streams that are defined in a pool or a dynamic pool, HCL Workload Automation processes perform a file existence check on a dynamic agent workstation of the pool or dynamic pool, randomly selected.
Note: To check the file dependencies, you must create the file in a file system that is accessible by all the workstations in the pool.

The file dependency is resolved when the first file is found on a dynamic agent workstation. The job does not necessarily run on the dynamic agent workstation where the file is located, but runs on one of the active workstations in the pool or dynamic pool when the file dependency is resolved.

IBM Workload Scheduler processes perform a file check, and the code selects the workstations that are available inside the pool. However, The file check does not perform for all the workstations.

If you have to make the file dependency work. Ensure the path on which the file has been created; and is accessible for all the workstations in the pool i.e. You have to share the file system on which the file is created, with all the workstations in the pool.

In your environment, you have POOL_A that contains DYN_A1, DYN_A2, DYN_A3 dynamic agents. If you define the JOB_A job in POOL_A pool, which depends on the FILE_A file, before JOB_A job runs, HCL Workload Automation processes perform a file existence check on the DYN_A1, DYN_A2, and DYN_A3 workstations. If the FILE_A file is found on the DYN_A2 workstation, the file dependency is resolved. The JOB_A job automatically runs on DYN_A3 which is the active dynamic agent workstation in the POOL_A pool.

How to define file dependencies

You can define file dependencies for jobs and job streams on dynamic agents, pools, or dynamic pool workstations, by setting the opens keyword in the job or job stream scheduling definition. For more information about the opens keyword syntax, see opens.

The following example shows that HCL Workload Automation processes check that the /opt/SF1-DIR/myfileSF1.txt file exists and is readable on the DYN-AGT-SF1 dynamic agent workstation before the SF1-JOB-HOSTNAME-0001 job runs on the DYN-AGT-SF1 workstation:
SCHEDULE MDMWKSNY1#NY1-JS1
VARTABLE VARTABLENY
OPENS DYN-AGT-SF1#"/opt/SF1-DIR/myfileSF1.txt" (-r %p)
:
DYN-AGT-SF1#SF1-JOB-HOSTNAME-0001
 TASK
    <?xml version="1.0" encoding="UTF-8"?>
<jsdl:jobDefinition xmlns:jsdl="http://www.ibm.com/xmlns/prod/
scheduling/1.0/jsdl" 
xmlns:jsdle="http://www.ibm.com/xmlns/prod/scheduling/1.0/jsdle" na
me="executable">
  <jsdl:application name="executable">
    <jsdle:executable interactive="false">
            <jsdle:script suffix="">hostname >>/opt/SF1-DIR/myfileSF1.txt 
            </jsdle:script>
        </jsdle:executable>
  </jsdl:application>
</jsdl:jobDefinition>
 RECOVERY STOP
END

See also

You can define the file dependencies for jobs and job streams also, by using the Dynamic Workload Console.

For more information about how to define file dependencies by using Dynamic Workload Console, see:

Designing your Workload.