[IPython-User] No response to client from ipcontroller

Adam Davis addavis@fosterandpartners....
Mon Jan 30 12:17:08 CST 2012


I have succesfully run ipcluster for a function defined with the
@lview.parallel decorator on a single host. I am trying to run the
ipcontroller / ipengine set up on this same host as a first step in
testing on multiple hosts. However, the controller doesn't seem to
respond to the Client instance in my python code.

I get the following (excerpted) messages in the ipcontroller log:

	2012-01-30 11:35:57.752 [IPControllerApp] client::client
'6142679f-36ad-4876-a096-ecbe675a9a3f' requested 'registration_request'
	2012-01-30 11:35:57.752 [IPControllerApp]
registration::register_engine(0, '6142679f-36ad-4876-a096-ecbe675a9a3f',
['6142679f-36ad-4876-a096-ecbe675a9a3f'],
'6142679f-36ad-4876-a096-ecbe675a9a3f')
	2012-01-30 11:35:58.752 [IPControllerApp] heartbeat::sending
72.0300002098
	2012-01-30 11:35:59.752 [IPControllerApp] heartbeat::sending
73.0300002098
	2012-01-30 11:35:59.752 [IPControllerApp]
heartbeat::handle_new_heart('6142679f-36ad-4876-a096-ecbe675a9a3f')
	2012-01-30 11:35:59.752 [IPControllerApp] registration::finished
registering engine 0:'6142679f-36ad-4876-a096-ecbe675a9a3f'
	2012-01-30 11:35:59.752 [IPControllerApp] engine::Engine
Connected: 0

	2012-01-30 11:36:07.548 [IPControllerApp] client::client
'56566026-4447-442d-8607-f1b00f9740a6' requested 'registration_request'
	2012-01-30 11:36:07.548 [IPControllerApp]
registration::register_engine(1, '56566026-4447-442d-8607-f1b00f9740a6',
['56566026-4447-442d-8607-f1b00f9740a6'],
'56566026-4447-442d-8607-f1b00f9740a6')
	2012-01-30 11:36:07.766 [IPControllerApp] heartbeat::sending
81.0450000763
	2012-01-30 11:36:08.766 [IPControllerApp] heartbeat::sending
82.0450000763
	2012-01-30 11:36:08.766 [IPControllerApp]
heartbeat::handle_new_heart('56566026-4447-442d-8607-f1b00f9740a6')
	2012-01-30 11:36:08.766 [IPControllerApp] registration::finished
registering engine 1:'56566026-4447-442d-8607-f1b00f9740a6'
	2012-01-30 11:36:08.766 [IPControllerApp] engine::Engine
Connected: 1

... that lead me to believe all is ok with the Engines establishing
connection to the controller.

Subsequently, I get the following when my client code connects:

	2012-01-30 11:36:59.094 [IPControllerApp] client::client
'186a760b-a105-4283-a5e3-a5f57feda744' requested 'connection_request'
	2012-01-30 11:36:59.094 [IPControllerApp] client::client
['186a760b-a105-4283-a5e3-a5f57feda744'] connected

However, where the ipcluster then returns the results and outputs log
messages like:

	2012-01-30 16:34:43.357 [IPControllerApp] task::task
'cf5dd2c6-94f9-4d7e-8df6-aadc9d3f5498' arrived on 1

	2012-01-30 16:34:44.795 [IPControllerApp] task::task
'cf5dd2c6-94f9-4d7e-8df6-aadc9d3f5498' finished on 1

... the ipcontroller/engine setup doesn't ever return results. I have
searched the mailing list archives and following this post:
http://mail.scipy.org/pipermail/ipython-user/2011-October/008530.html, I
tried invoking the ipcontroller as:

ipcontroller --ip=0.0.0.0

I have also tried editing ipcontroller_config.py to listen on 0.0.0.0 as
well, but the effect is the same in each case. Are there any obvious
things I may be doing incorrectly (or failing to do)? Otherwise, are
there further suggestions for debugging beyond what is available through
the controller and engine logs? Many thanks.

-Adam


More information about the IPython-User mailing list