I want to get this working on port 80. I tried downloading the files and editing the port and building the image, but the result I got when building is not the as the image.
Can anyone point me in the right direction here?
Thanks in advance!
Just to mention it here: if you want to add a new user make sure you use "useradd -m someName". Otherwise the home directory might be missing and you run into the error "spawner failed to start ..." when you log in
Look at https://github.com/jupyterhub/jupyterhub#docker
The following worked:
- docker run -p 8000:8000 jupyter/jupyterhub jupyterhub --no-ssl
- docker exec -it <containerID> bash # containerID is what you get as the output of
- useradd <username> # some new user
- pip install jupyter
- Visit http://localhost:8000 and log in with the username and password in step 3
@hugokoopmans, did you have any luck?
I'm trying to do the same and not making any progress.
As stated previously, would be nice if the documentation here would be towards running a docker image instead of the a base install.
I intent to set up an immage to use multiuser and also running an R kernel besides the python kernel. Anything I need to know before I give it a spinn?
I have found that ipywidgets do not work when running jupyterhub through an extension of this docker image. I can import the ipywidgets module and create widgets, but they do not display (the widgets DO display properly when running jupyterhub outside of docker). In the js console I can see that it 404s when trying to load 8000/user/test/nbextensions/widgets/notebook/js/extension.js. Also the logs say: "ipywidgets package not installed. Widgets are unavailable." (however, as mentioned above, I can in fact import the module).
Has anyone else had this problem or know of a solution?
As stated by @mtbdeano Jupyter honors the local system settings which are configured for PAM authentication. Your final solution will want to be designed to meet your needs through that. Either by using the useradd/passwd commands or by configuration the local PAM service to connect with an external user registry like LDAP.
This container works great. The trick to adding users is as follows
docker exec <containerid> bash to access the root shell on the image then use regular unix
passwd to create users. The container uses PAM authentication, so a unix user is a valid jupyter user!
@dtenenba - The ipython/scipyserver image is quite similiar.
Sorry @simonbiggs , that results in a stack trace.
It seems that the documentation here is documentation for installing jupyterhub, not for using the jupyterhub docker image. That's unfortunate.