TypeError in dask-jobqueue/core.py when starting up workers (v0.4.0)

Error Seems pretty straightforward. Adding a quick int() fixed things on my end. LMK if you want a PR and or CI coverage… (I didn’t see a contributing guide anywhere in the docs) For context grid is an SGECluster object. Thanks for the great tool.

In [6]: grid.scale(2)

DEBUG:Scaling up to 2 workers.
DEBUG:Found 0 active/pending workers.
DEBUG:starting 2 workers
In [7]: tornado.application - ERROR - Exception in callback <functools.partial object at 0x7fba4979c208>
Traceback (most recent call last):
  File ".../miniconda/envs/dask/lib/python2.7/site-packages/tornado/ioloop.py", line 758, in _run_callback
    ret = callback()
  File ".../miniconda/envs/dask/lib/python2.7/site-packages/tornado/stack_context.py", line 300, in null_wrapper
    return fn(*args, **kwargs)
  File ".../miniconda/envs/dask/lib/python2.7/site-packages/dask_jobqueue/core.py", line 375, in scale_up
    self.start_workers(n - active_and_pending)
  File ".../miniconda/envs/dask/lib/python2.7/site-packages/dask_jobqueue/core.py", line 299, in start_workers
    for _ in range(num_jobs):
TypeError: range() integer end argument expected, got float.

[mrocklin]

Thanks for the bug report @kaelancotter . A pull request would be welcome. Contributor guidelines are available at documentation for the parent project: http://dask.pydata.org/en/latest/develop.html

[kaelancotter]

I took a look through the travis ci logs for python 2.7 sge. Right now it looks like the sge_master image in the compose stack that executes pytest uses a default 3.x python from miniconda3 instead of the travis_python_version. I think fixing the container to use the travis python version will greatly improve 2.7 test coverage, including coverage of the original type error. I hipe Im not overstepping. 😃