python - IPython Notebook with remote ipyparallel Controller -

i'm trying setup remote cluster on group of servers own using ipyparallel library. figured if share $ipythondir between ipcontrollers, ipengines , notebook connect , work, not case current setup.

what i'm attempting accomplish such ipcontroller , ipengines sitting on cluster waiting jupyter notebook connect controller , use it's cluster computing resources.

currently cannot notebook connect controller though ports open, servers directly accessible, , ipythondir shared.

when open notebook , go clusters tab see parallel profile, it's not started. odd because ipcontroller , ipengines started , waiting connection notebook.

this boils down to:

• is possible run notebook on different server ipcontroller?
• if above possible, why can not notebook connect cluster, , instead when click start on profile makes local cluster.

thanks!

yes possible if notebook kernel running on same server ipcontroller. notebook can displayed browser. use functionality regularly.

the way have done have ipython profile available on server. in case it's windows server , profiles set under c:\users\<user>\.ipython\. in case profile folder called profile_my32bitcluster , when creating client, specify profile use:

from ipyparallel import client  rc = client(profile='my32bitcluster') dview = rc[:]  # test pushing out dataframe across engines, modifying # , returning modified dataframes... df = pd.dataframe(data={'x':[1,2,3,4,5], 'y':[1,4,9,16,25]})  dview.push({'df':df})  def myfunc(x):     import sys     import os     import pandas pd     global df     df['z'] = df['x'] * x     return df  results = dview.map_sync(myfunc, [2,3,4]) 

i hope helps.