Skip to content

example1.md

Latest commit

 

History

History
53 lines (42 loc) · 1.76 KB

example1.md

File metadata and controls

53 lines (42 loc) · 1.76 KB
Raw

Super Power 1: Running a Fixed Set of Jobs

The core command for scheduling a set of Cloud jobs is super run. Using the -p option to super run, Super can run a fixed number of UNIX command lines, in the Cloud. The output of the N jobs will be joined and flowed to stdout.



Example

super run -p3 -- printenv JOB_INDEX
[Job 1] 1
[Job 2] 2
[Job 3] 3

Here we specified -p3, and printed out a convenience environment variable JOB_INDEX inside of each Cloud job. The output flows to our terminal. To help distinguish the output flowing from many concurrent jobs, the log lines emitted by a job with job index k are prefixed by [Job k].

Options for super run

  • -p: Use a fixed number of jobs.
  • -q: Do not show the [Job k] prefix. If you pipe the output of super run to other commands on your laptop, Super will automatically operate in -q mode.
  • -g: Add extra debugging information to the output of super run.
  • -f: You may provide your pipeline as a file, and via -f - you may do so via stdin.
  • -i: Specify a custom base Docker image. It is recommended that you extend starpit/sh:0.0.5.
  • -c, -m: Specify a combination of CPU and memory allocation requests. Only certain combinations are supported.

Other Super Powers