fjd 0.1.7

File-based job distribution for everyone

File-based job distribution. A straightforward pull-model for computational tasks, working with the assumption that all CPUs/cores can access a shared home directory.

Installation

$ pip install fjd

Usage

Start one or more workers, like this:

$ recruiter.py hire <number of workers>

These will sit in Unix screen sessions and wait for job assignments.

Then, start a dispatcher:

$ dispatcher.py

Now the dispatcher waits for jobs in the jobqueue directory. Workers announce themselves in the workerqueue directory, where the dispatcher will find them. (These working directories will be created if they do not yet exist.)

All you have to do now is to put jobs in the queue. You do this by putting a file per job in the jobqueue directory. The file should adhere to the general configuration file standard. Here is an example:

[control]
executable: python example/ajob.py
logfile: logfiles/job0.dat

[params]
param1: value0

Where you specify which command to execute and where results should go.

Your executable (the “job”) gets this configuration file passed as a command line argument. This way, it can see for itself in which logfile to write to.

Take care to get the relative paths correct (or simply make them absolute): If the paths are relative, the path to the executable should be relative to the workers working directory, whereas the path to the logfile should be relative to the jobs working directory.

In addition, you can put other job-specific configuration in there for the executable to see, as I did here in the [params]-section (in fact, only the [control]-section is fjd-specific).

An example

You can see how it all comes together by looking at the simple example in the example directory where there is one script that represents a job and one that creates ten jobs similar to the one we saw above and puts them in the queue.

To run this example, run a script that creates the jobqueue. Recruit some workers and start a dispatcher. Then, lean back and observe. We have a script that does all of this in run-example.sh:

#/bin/bash

python create_jobs.py
recruiter.py hire 4
dispatcher.py

And this is the output you should see:

$ cd example
$ ./run-example.sh
[FJD] Hired 4 workers on localhost.
[FJD] Dispatcher started.
[FJD] Found some jobs to dispatch
[FJD] Found some jobs to dispatch
[FJD] Found some jobs to dispatch
[FJD] No more jobs to dispatch.

It does not matter in which order you do these three things - create jobs, hire workers and dispatch. The workers patiently wait for jobs and the dispatcher waits for workers.

When all jobs are done (the dispatcher in the examnple quit because there were none left), you can “fire” the workers (i.e. kill the Unix screen sessions):

$ recruiter.py fire

And you’ll see the results, the log files written by our example jobs:

$ ls logfiles/
job0.dat    job2.dat        job4.dat        job6.dat        job8.dat
job1.dat    job3.dat        job5.dat        job7.dat        job9.dat

CHANGES

(Changes are not tracked yet, too early …)

TODO:

• test recruiter on remote machines

• specify a custom group identifier for worker and dispatcher

• unit tests