[IPython-User] notebook server dies when logging out of terminal

Moritz Emanuel Beber moritz.beber@gmail....
Tue Feb 26 05:56:20 CST 2013


On 02/26/2013 08:41 AM, K.-Michael Aye wrote:
> I am using a notebook server profile on our fast linux server so that I
> can log onto the notebook server with the browser on my Macbook.
> This works pretty well as long as I stay connected to the terminal
> where I am logged into the server.
> I am lauching my notebookserver on a remote server like this:
> ipython notebook --profile=nbserver &
> to send the nbserver into the background, so that when I would
> disconnect because I am sending the macbook to sleep for a while, while
> changing the room and, most likely, the kind of network connection.
> However, for some reason the nbserver always becomes unresponsive in a
> way. I still can log out and into the dashboard via the password
> protection, but cells from a newly openened notebook are not being
> executed, the '*' stays forever in front of each cell that was
> Shift-Returned.
> Further information:
> I have configured the nbserver to use https:.
> Anybody has any idea what might be going wrong or how I could set up a
> notebook server that stays alive independent from a ssh login?
> Best regards,
> Michael
> _______________________________________________
> IPython-User mailing list
> IPython-User@scipy.org
> http://mail.scipy.org/mailman/listinfo/ipython-user

For most processes you would have to start them with 'nohup' and 
redirect their stdout and stderr streams to be able to just log out 
without them being affected. I highly recommend 'screen' as Matthias 
already mentioned. You can even start the notebook inside a screen 
session with a single command:

$ screen -dmS your_screen_session_name_here ipython notebook 

'screen' is a great utility to learn anyway since you can have "multiple 
terminals" through a single ssh session. When you later want to interact 
with programs running inside of screen, simply type:

$ screen -r
$ screen -r your_screen_session_name_here

if you have multiple screen instances.


More information about the IPython-User mailing list