Page History

...

Accessing parameters in shell scripts

• Environment variables are used to make job (i.e. task) Job and order parameters accessible are exposed as environment variables to shell scripts.
• Environment variables are named using a predefined prefix and the name of the original parameter: 
  Environment variable names are written with uppercase letters.
  • The default value for the environment variable prefix is is SCHEDULER_PARAM_ . 
  • Environment variable names are provided with uppercase letters.
  • For example, a task parameter param1 can be accessed by the SCHEDULER_PARAM_PARAM1 environment variable name.

  • If the SCHEDULER_PARAM_ prefix is not to be used for the environment variables and instead the parameter names themselves then the following setting can be added to the ./config/scheduler.xml file:

    Code Block
    language xml
    <params> <param name="scheduler.variable_name_prefix" value="*NONE"/> ... </param>

    Note that: there is a risk of JobScheduler overwriting already existing environment variables with new ones of the same name that have been derived from job or order parameters if a prefix such as SCHEDULER_PARAM_ is not used.For details see Which environment variables are provided by JobScheduler?

Examples

The following two examples show a task parameter defined in a job <param> tag can be called in a shell script.

 <job>
    <params>
        <param name="param1"  value="Test"/>
    </params>
    <script language="shell">
        <![CDATA[
 rem This is an example shell script to show the use of parameters
 echo Param1 has the value %SCHEDULER_PARAM_PARAM1%
        ]]>
    </script>
    <run_time/>
 </job>

 <job>
    <params>
        <param name="param1" value="Test"/>
    </params>
    <script language="shell">
        <![CDATA[
 # This is an example shell script to show the use of parameters
 echo "Param1Parameter param1 has the value $SCHEDULER_PARAM_PARAM1"
        ]]>
    </script>
    <run_time/>
 </job

To use parameters as command line arguments for a script add environment variables as arguments in the command line as follows:

  anyshellscript.sh $SCHEDULER_PARAM_PARAM1 $SCHEDULER_PARAM_PARM2 ...

...

Passing parameters to subsequent shell jobs in a job chain

• JobScheduler allows the creation or overwriting of order parameters for subsequent jobs in a job chain.
• JobScheduler parses a temporary file for name/value pairs each time the status of after termination of each job in a job chain changes. 
  • The name of this temporary file is held in the SCHEDULER_RETURN_VALUES environment variable.
  • The name/value pairs are automatically added as order parameters.
  • A parameter set in one job shell script can be written to this temporary file and made available to subsequent jobs using the following (for Windows):
    
    • myParam=myValue >> %SCHEDULER_RETURN_VALUES%
  • A task parameter set in one job can be written to this temporary file and thereby made available to subsequent jobs in a job chain using the following (for Windows):
    • myParam=%SCHEDULER_PARAM_MYPARAM% >> %SCHEDULER_RETURN_VALUES%
  • A parameter set in one job is called in the shell script of a Such parameters are accessible in subsequent job using the following (for Windows):
    • echo myParam = %SCHEDULER_PARAM_MYPARAM%

The following code blocks show two jobs that demonstrate the setting and retrieval of shell script parameters in a job chain:

<job order="yes" stop_on_error="no" name="job1_shell_with_parameter_set">
    <params>
        <param name="param1" value="Value1-Job1" />
        <param name="param2" value="Value2-Job1" />
    </params>
    <script language="shell">
        <![CDATA[
rem This is a sample shell script to demonstrate the setting of parameters 
echo param1 = %SCHEDULER_PARAM_PARAM1%
echo param2 = %SCHEDULER_PARAM_PARAM2%PARAM2% 
 
rem creating order parameters for subsequent jobs
echo param2 = some-other-value >> %SCHEDULER_RETURN_VALUES%
echo param3 = Value3 >> %SCHEDULER_RETURN_VALUES%
        ]]>
    </script>
    <run_time/>
</job>

<job order="yes" stop_on_error="no" name="job2_shell_with_parameter_get">
    <params>
        <param name="param1" value="Value1-Job2" />
        <param name="param2" value="Value2-Job2" />
    </params>
    <script language="shell">
        <![CDATA[
echo param1 has the value: %SCHEDULER_PARAM_PARAM1%
echo param2 has the value: %SCHEDULER_PARAM_PARAM2%
echo param3 has the value: %SCHEDULER_PARAM_PARAM3%
echo SCHEDULER_RETURN_VALUES = %SCHEDULER_RETURN_VALUES%
        ]]>
    </script>
    <run_time/>
 </job>

...

 <job_chain>
    <job_chain_node state="100"
                    job="job1_shell_with_parameter_set"
                    next_state="200"
                    error_state="error"/>
    <job_chain_node state="200"
                    job="job2_shell_with_parameter_get"
                    next_state="success"
                    error_state="error"/>
    <job_chain_node state="success"/>
    <job_chain_node state="error"/>
 </job_chain>

Example Behavior

The screen shot below shows the log file for the example order and the values for the environment variables.

...

• The param1 task parameters set in both the first and second jobs are only valid within the respective jobs.
• The param2 task parameter set in the first job and which is forwarded as an environment variable to the second job overwrites the param2 task parameter set in the second job.
• The param3 shell script parameter set in the shell script in the first job is passed to the script in the second job.

Example Download

The example described above can be downloaded from the following link:

...

The example behavior described above will be visible in the order log file.

Further References

Job and Order Parameters

• How to set and read job and order parameters
• How to access order parameters and job parameters using the JobScheduler API
• How to specify parameters using the "on_exit_code" command
• How to clone parameters from a persistent order