Cluster Jobs

Run a script on whichever node fits best

Cluster Settings Cluster Info

Submit a job

Name

Script (a .agi file that defines run(job))

Arguments (JSON, optional)

Input paths (comma separated, used to run near the data)

Map/reduce dataset (optional glob; turns this into a map/reduce job)

With a dataset the script must define mapper(files, emit) and reducer(key, values); the files are processed on the nodes that hold them.

Required features

Timeout (s)

The script runs with your permissions on the chosen node. cluster:/ paths work as usual.

Job detail

Select a job to see its log and output

Log

Output

Jobs

Job	Status	Node	Progress	When