Useful resources
- Our website nero-docs.stanford.edu.
- Our Nero Youtube channel
- Our Stanford Slack channels #nero-announce, #nero-users.
Q: I am trying to connect to JupyterLab, but I am getting a “504 Gateway Time-out” error.
If you are unable to access JupyterLab, you can restart your JupyterLab server by going to https://nero.compute.stanford.edu/hub/home and clicking on “Stop my server”, followed by “Start my server”.
Q: I am trying to connect to JupyterLab. However, I am getting the error “Spawn Failed> Server at https:///user/ didn’t respond in 30 seconds ”.
There could be a variety of reasons. Here are some common ones:
⚠You created a custom conda environment and there is likely something wrong with it.
You can try to ssh into Nero by opening a terminal and typing:
$ ssh <your-sunet>@nero.compute.stanford.edu
From the Nero login node, you can revert back to the base environment by typing:
$ mv ~/.srcc/jupyter/override ~/.srcc/jupyter/override.bak
More information on how to create custom conda environments on JupyterLab is available here.
⚠The maximum storage quota has been reached on one of the directories you have access to.
This means that this directory has reached its storage quota. As a reminder, home directories are allocated 10GB of storage, while PI directories are allocated at least 1TB of data. You can try to ssh into Nero by opening a terminal and typing:
$ ssh <your-sunet>@nero.compute.stanford.edu
From the Nero login node, you can move or delete some files in the directory that is full. You can check the storage quota of a directory by typing:
$ cd /path/to/directory
$ df . -h
Q: I restarted my server, but I am now getting an error that says “Please use a different workspace. This workspace is already in use in another JupyterLab window. Please enter another workspace name.”
This simply means that you have an existing tab with another Nero session open in your browser. If you are unable to access JupyterLab, you can restart your JupyterLab server by going to https://nero.compute.stanford.edu/hub/home and clicking on “Stop my server”, followed by “Start my server”.
💡Tip:We are aware that this issue is recurrent with Safari, even when all other Nero tabs have been closed. You may want to use Nero with Chrome or Firefox instead.
Q: I see an error “Error Invalid response: 403” when I try to launch a Jupyter Notebook.
When you launch JupyterLab for the first time, make sure that you navigate to a directory in which you have read/write privileges before you try creating a Notebook. These directories are:
- /home/
- /share/pi/
Q: I use RStudio and I am getting a “500 error” when I try to connect to my session.
There could be a variety of things going on but here are a few commands to try.
On a Nero terminal, you can delete old R sessions by typing:
$ rm -rf ~/.rstudio/sessions/*
Restart your JupyerHub session by clicking on “Hub” > “Control Panel” > “Stop my server”, then “Start my server”.
💡Tip:In general, it is good practice to quit your RStudio session when it is not in use. You can do so from your Rstudio session, by clicking on File > Quit Session.
Q: My JupyterLab session freezes. My browser eventually produces an error.
You probably have too many tabs in your Nero JupyterLab session opened. This also happens if you try to preview a csv file within the Nero JupyterLab environment. We recommend that you avoid opening too many tabs on Nero JupyterLab, and that you use a programming language to import your CSV file, and preview the data (e.g: using the df.head() command in Python).
Q: How do I install a Python module on Nero?
The short answer is:
Open a Terminal on Nero and type:
$ pip install --user <module-name>
For more instructions on how to install a Python package in your own conda environment, please visit our documentation website.
Q: How do I install an R package on Nero?
Whether you are using RStudio, the R kernel on a Jupyter notebook, the short answer is:
Open a Terminal on Nero
Type the following commands:
$ module load r
$ R
$ install.packages(“package-name”)
# Accept the suggested settings (type “Yes)
# When prompted to select a CRAN mirror, choose “USA (OR)”. To choose this you would type the number to the left of “USA (OR)”
For github R packages you will want to clone the repo locally, then install in R. More information is available here.
Q: How can I use a GPU with a Nero Jupyter Notebook?
Please carefully read our documentation on this topic here.
💡Tip: As mentioned on our website, make sure to use the right Nero IP address when setting up port-forwarding connection.
💡Tip: If you are using port 8888 as shown in the documentation, but are getting an error, this means that another user is already connected to that port. Use another port number (anything above 1024 works).
Q: I am trying to use an API to download data onto Nero. However, I am getting a time-out error.
Nero is an environment for high-risk data. As such, it only enables API calls to websites that have been whitelisted by the Nero team. The error you are seeing is actually by design.
💡Tip: You can nevertheless download the data to your local machine and import it onto Nero using a variety of methods. Make sure to check our documentation and Youtube channel for instructions on how to upload data onto Nero.
Q: When I try to access storage and/or BigQuery on my Nero Google Cloud Project, I get one of the following errors “Sorry, the server was not able to fulfill your request”, “You do not have permissions to see this information” and/or “AccessDeniedException: 403 Request is prohibited by organization’s policy. vpcServiceControlsUniqueIdentifier:”
There could be a variety of reasons for these errors. Here are some of the most common ones:
⚠You may not be connected to the full-tunnel VPN, which is a necessary requirement for accessing Nero GCP. The first two minutes of this video show how to connect to the full-tunnel VPN.
⚠You are using a new Macbook, which uses an IP format not currently supported by the Stanford VPN client. If you are still having trouble seeing data from the console, and think you have your networking and VPN configured correctly, open Chrome and search in Google for “what is my ip”. You should see an address in the form 171.66.XX.XX.
If you see a different range, you are probably not on the VPN, or you did not select the Full Traffic option.
If you see an address formatted like “2001:0db8:85a3:0000:0000:8a2e:0370:7334” you need to disable IPv6.
The current Cisco VPN does not support tunneling of IPv6 traffic. It only supports IPv4. If you are using a modern from Apple, you probably have IPv6 enabled, which will cause problems. To disable IPv6 you must go into System Preferences –> Network, and click the “Advanced…” button for your current network:
On the “TCP/IP” tab you should select “Link-local only” for the IPv6 configuration:
