command and job executions using templates

* commands, jobs *

(job_submit, shell_command, system_command, tso_command)

Actually a developer can add any kind of commands/job to run which could be a risk (even if pull requests are reviewed). The pipeline user has higher authorites and a wrong command can damage the application or the system.

Instead of adding commands/jobs directly. There should be a pre-defined set of commands/jobs which can be run by a developer during deployments. So, they can only execute a command/job which is needed and approved and they cannot be altered.

Each command/job is a template and can be stored in git or in a folder. The command parameters are artifacts, as I suggested above.

For example:

shell_command for copy (cp)

the template looks like:

cp_templete.j2:

cp {{ source_path }}/{{ source_file }} {{ target_path }}/{{ target_file }}

the cp_command artifact:

mycpcomm.cmd

____commad: cp

________template: cp_templete.j2

________command_parameters:

____________source_path:

____________source_file:

____________target_path:

____________target_file:

(sorry the editor doesn't support indentations)

The module should capture the command output and display in the log. How to process the command output and use it in further step or ans a condition for fail os success? The command execution can be successful but the response may indicate a failure eg.: ls doc* may be empty which means files are missing.