If anyone can help me with this, I am still struggling with this. 
1 Like
Are you sure you are connected? What does the terminal say after the command
gcloud compute ssh --zone=$ZONE jupyter@$INSTANCE_NAME – -L 8080:localhost:8080 ?
I was alll up and running on GCP for more than a week now and suddenly today when I use .
gcloud compute ssh --zone=$ZONE jupyter@$INSTANCE_NAME – -L 8080:localhost:8080 I get jupyter@ip permission denied(public key).
I am still able to ssh into the machine without jupyter so its not my gcp key is not working but not sure how to proceed.Any suggestion?
1 Like
I have the same issue as Ritesh. I can successfully connect to my instance, but when I try to access localhost:8080 I get this error message on my terminal.
Any suggestions on what is going wrong?
I’m also having issues with zone resources at the moment, even with two instances in two different zones. I’ve seen references in this thread (and in GCP) to changing the zone, but this seems to be an option only when creating an instance. Is this correct?
Relatedly, I’d like to have a disk that I can attach to both/all my VMs. It’s my impression that it’s not possible to attach a disk to an instance in another zone, though. Is this correct? I’d love to be wrong here. 
Worked, thanks. I was too honest and stupid when choosing region.
Instance setup, resources checkout and installation following the guide went as smooth as possible, so I have everything up and running.
3 Likes
Just applied the coupon. Thanks a lot!
1 Like
Yes, VERY considerate of you to share this with other FASTAI students. Thank You!
Can you think of any way we can use some of the credit to archive and share (between students) FASTAI or GCP public data sets. I am thinking if the data sets are clean/compatible with a lesson number it would only take a change of a few words on a line or two of FASTAI code to run your model on an entirely different data set with new classifiers.
Thanks Again!
1 Like
I am having a problem opening jupyter notebook on GCP. The error reads the port 8080 is in use and suggests another port but I could not open the notebook from the browser. Does anyone have the same problem and suggested solution? Many thanks111
jupyter@my-fastai-instance:~$ jupyter notebook
[I 00:30:56.501 NotebookApp] [nb_conda_kernels] enabled, 3 kernels found
[I 00:30:56.508 NotebookApp] Writing notebook server cookie secret to /run/user/1000/jupyter/notebook_cookie_secret
[W 00:30:56.635 NotebookApp] WARNING: The notebook server is listening on all IP addresses and not using encryption. This is not recommended.
[W 00:30:56.635 NotebookApp] WARNING: The notebook server is listening on all IP addresses and not using authentication. This is highly insecure and not recommended.
[I 00:30:56.635 NotebookApp] The port 8080 is already in use, trying another port.
jupyter_http_over_ws extension initialized. Listening on /http_over_websocket
[I 00:30:56.667 NotebookApp] JupyterLab extension loaded from /opt/anaconda3/lib/python3.6/site-packages/jupyterlab
[I 00:30:56.667 NotebookApp] JupyterLab application directory is /opt/anaconda3/share/jupyter/lab
[I 00:30:56.808 NotebookApp] [nb_conda] enabled
[I 00:30:56.841 NotebookApp] ✓ nbpresent HTML export ENABLED
[W 00:30:56.841 NotebookApp] ✗ nbpresent PDF export DISABLED: No module named ‘nbbrowserpdf’
[I 00:30:56.841 NotebookApp] Serving notebooks from local directory: /home/jupyter
[I 00:30:56.841 NotebookApp] The Jupyter Notebook is running at:
[I 00:30:56.841 NotebookApp] http://(my-fastai-instance or 127.0.0.1):8081/
[I 00:30:56.841 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
1 Like
I was having similar problems when I first set up GCP. But I was running package called “deep learning package manager”. It was in my startup directory and whenever I was starting my laptop the package was starting and monopolizing the port 8080. If you have any startup packages that might be doing this “disable” them in your system applications settings (windows). Be sure and restart your system after doing this!
Next thing you can try is when you get your GCP console up and click on an instance and then hit start button. Go to the top of the screen and hit the open terminal button. This will open a terminal at the bottom of your GCP console. Now go back and “right click” the SSH button and hit either “run command in terminal” button or copy the command line paste it into the terminal and hit enter.
Now look at the little command bar that is on the upper right lip of the terminal. Right click on that little book icon and select open notebook in port 8080.
NOW a jupyter notebook should open in a browser window.
Let me know if this does not work for you.
2 Likes
What do you see when you run:
systemctl status jupyterIt seems like the geco wifi network won’t allow any outbound ssh. Tried ssh to a couple of other places and nothing worked. Is anyone else having this problem?
Thank you!
export IMAGE_FAMILY="pytorch-1-0-cu92-experimental" # or "pytorch-1-0-cpu-experimental" for non-GPU instances
export ZONE="us-west2-b" # budget: "us-west1-b"
export INSTANCE_NAME="my-fastai-instance"
export INSTANCE_TYPE="n1-highmem-8" # budget: "n1-highmem-4"
gcloud compute ssh --zone=$ZONE jupyter@$INSTANCE_NAME -- -L 8080:localhost:8080
1 Like
Haven’t tried this. Just so be clear, I should tried this after connecting it right? I will do it now and see what it says.
When I upgraded with conda as per instructions my jupyter would no longer start.
This is because it changed a lot of the permissions, so most libraries were only accessible as root!
I worked around this using the following sledgehammer:
sudo find /opt/anaconda3/ -exec chmod a+xr {} \;
It takes a minute or so to run, but now jupyter is working again for me
It says Failed to connect to bus: No such file or directory. What does this mean?
gcloud compute instances move my-fastai-instance --zone us-east1-c --destination-zone us-east1-b
The only problem is I get:
ERROR: (gcloud.compute.instances.move) Instance cannot be moved while in state: TERMINATED
For a bit more context: systemd is what starts up jupyter on boot.
systemctl status jupyter
will let you know if something went wrong starting it.
If there has been a problem the best way to see what’s going on is to run
jupyter notebook
and see what happens.
I suspect all those messages in the terminal are because your requests are being forwarded to the server, but jupyter isn’t responding to them.
Hmm… I’m not sure what’s going on. I’d try terminating the instance and following the instructions from scratch again.
1 Like