Job Execution

Job Execution Environment Variables

Scale provides several environment variables to the job execution that are available for developer consumption.

ALLOCATED_CPUS: JSON decimal
The number of allocated CPUs as defined in the seed manifest interface.resources / job configuration

ALLOCATED_DISK: JSON decimal
The number of allocated disks as defined in the seed manifest interface.resources / job configuration

ALLOCATED_GPUS: JSON decimal
The number of allocated GPUs as defined in the seed manifest interface.resources / job configuration

ALLOCATED_MEM: JSON decimal
The amount of allocated memory as defined in the seed manifest interface.resources / job configuration

ALLOCATED_SHAREDMEM: JSON decimal
The amount of allocated shared memory as defined in the seed manifest interface.resources / job configuration

SCALE_JOB_ID: JSON string
The Scale Job Id

SCALE_EXE_NUM: JSON string
The Scale job execution number

SCALE_RECIPE_ID: JSON string
The Scale recipe ID the running job is located in

SCALE_BATCH_ID: JSON string
The batch ID the running job is in

INPUT_METADATA: JSON object
A JSON object of the inputs name and public URL provided to the job, as well as the name(s) and public URL(s) of the encompassing recipe.

This JSON object is in the following form:

{
    "JOB": {
        {"input_a": ["http://public/facing/url/1", "http://public/facing/url/2"]},
        {"input_b": ["http://public/facing/url/3"]}
    },
    "RECIPE": {
        {"input_a": ["http://public/facing/url/a"]},
        {"input_b": "json_value"}
    }
}