Unable to start jupyter notebook on the Deep Learning Virtual Machine (DLVM) in MS Azure

Question

I'm trying to train keras models on the Deep Learning Virtual Machine (DLVM)from MS Azure. I've setup a Ubuntu VM and connected to it via ssh from my mac but when I try to run jupyter notebook, I get this error : No such notebook dir: ''/dsvm/Notebooks'', for which I tried the solutions specified in Deep Learning Virtual Machine can't run jupyter "No such notebook dir: ''/dsvm/Notebooks''", but that gives me The Jupyter Notebook is running at:https://[all ip addresses on your system]:9999/ But chrome says my vm's ip : 9999 is not working.

According to the docs jupyter hub should be available at 8000 as soon as the vm starts but I can't access http://my-vm-ip:8000 either. when I tried systemctl status jupyterhub as recommended here Can't reach Jupyter Notebooks on Azure Deep Learning Virtual Machine, I can see that jupyter hub is indeed listening at 8000 and there is a firewall rule defined at the vm allowing tcp connections to 8000 over the internet.

How do I access jupyter from the DLVM?

Paul Shealy Paul Shealy · Accepted Answer · 2018-05-14T18:05:13

JupyterHub should be working and accessible on your VM. Are you connecting via https? Your post says you're using http, but we only support https.

Also, have you have confirmed that port 8000 is open on the Azure portal? Some VMs are created without this port open for various reasons. You might also try restarting the JupyterHub service to confirm it's working properly.

If you want to run Jupyter instead, you will need to fix an issue with the latest Ubuntu DSVM with the Jupyter config file. At a terminal:

sudo -s
source /anaconda/bin/activate py35
jupyter notebook --generate-config --config=/usr/local/etc/jupyter/jupyter_notebook_config.py

You should then open port 8888 in the firewall. Jupyter will then be accessible.

Unable to start jupyter notebook on the Deep Learning Virtual Machine (DLVM) in MS Azure

2 Answers