This is the first node of the workflow. As such, the platform takes cares of providing the inputs to the streaming executable: the list of Landsat catalogue entries you created earlier.
Tip
ciop-copy can handle catalogue entries and download the datasets using the online resources in the metadata
A node requires a job template including:
<jobTemplates>
<jobTemplate id="py-ndvi">
<streamingExecutable>/application/py-ndvi/run.py</streamingExecutable>
<defaultParameters>
</defaultParameters>
<defaultJobconf>
<property id="mapred.task.timeout">9900000000000</property>
</defaultJobconf>
</jobTemplate>
</jobTemplates>
Info
The code for the whole application is available here: https://github.com/Terradue/dcs-python-ndvi
While this job template doesn’t define any parameter, it defines the mapred.task.timeout property which is the wall-time in miliseconds between two logging entries.
Note
Log entries using ciop-log function in bash (or rciop.log in R and cioopy.log in Python) tell the platform the process is alive. If the wall-time is reached the execution is terminated with an error