4.5. Variable Expansion

Pegasus Planner supports notion of variable expansions in the DAX and the catalog files along the same lines as bash variable expansion works. This is often useful, when you want paths in your catalogs or profile values in the DAX to be picked up from the environment. An error is thrown if a variable cannot be expanded.

To specify a variable that needs to be expanded, the syntax is ${VARIABLE_NAME} , similar to BASH variable expansion. An important thing to note is that the variable names need to be enclosed in curly braces. For example

 ${FOO}  - will be expanded by Pegasus
 $FOO    - will NOT be expanded by Pegasus.

Also variable names are case sensitive.

Some examples of variable expansion are illustrated below:

  • DAX

    A job in the DAX file needs to have a globus profile key project associated and the value has to be picked up (per user) from user environment.

    <profile namespace="globus" key="project">${PROJECT}</profile> 
    
  • Site Catalog

    In the site catalog, the site catalog entries are templated, where paths are resolved on the basis of values of environment variables. For example, below is a templated entry for a local site where $PWD is the working directory from where pegasus-plan is invoked.

    <site  handle="local" arch="x86_64" os="LINUX" osrelease="" osversion="" glibc="">
            <directory  path="${PWD}/LOCAL/shared-scratch" type="shared-scratch" free-size="" total-size="">
                    <file-server  operation="all" url="file:///${PWD}/LOCAL/shared-scratch">
                    </file-server>
            </directory>    
            <directory  path="${PWD}/LOCAL/shared-storage" type="shared-storage" free-size="" total-size="">
                    <file-server  operation="all" url="file:///${PWD}/LOCAL/shared-storage">
                    </file-server>
            </directory>
            <profile namespace="env" key="PEGASUS_HOME">/usr</profile>
            <profile namespace="pegasus" key="clusters.num" >1</profile>
    </site>
    
  • Replica Catalog

    The input file locations in the Replica Catalog can be resolved based on values of environment variables.

    # File Based Replica Catalog
    production_200.conf file://$PWD/production_200.conf site="local" 

    Note

    Variable expansion is only supported for File based Replica Catalog, not Regex or other file based formats.

  • Transformation Catalog

    Similarly paths in the transformation catalog or profile values can be picked up from the environment i.e environment variables OS , ARCH and PROJECT are defined in user environment when launching pegasus-plan.

    # Snippet from a Text Based Transformation Catalog
    tr pegasus::keg{
        site obelix {
            profile globus "project" "${PROJECT}"
            pfn "/usr/bin/pegasus-keg"
            arch "${ARCH}"
            os "${OS}"
            type "INSTALLED"
        }
    }