Skip to main content
This page includes common troubleshooting topics for Anaconda Notebooks. If you encounter an issue that is not listed here, you can obtain support for Anaconda through the Anaconda community forums. For Anaconda installation or technical support options, visit our support offerings page. For questions about JupyterLab, refer to the official JupyterLab documentation.

​
Toolbox environment card says the environment is installed, but no environment in the Launcher or runtime selector

Cause
The environment might not have fully synchronized with the Launcher or runtime selector yet.
Solution
First, try refreshing the browser page. If the environment still doesn’t appear after refreshing, check the Environments page in Navigator to see if it’s available. If the environment is still unavailable after waiting two minutes, close all open Jupyter sessions and restart Jupyter from Navigator or the command line.

​
Quickstart environment installation errored

Cause
Installation errors can occur due to misconfigured channels in your .condarc file, organizational policies blocking installation, connectivity issues, or stale Jupyter kernel states.
Solution
First, check your .condarc file and verify that your channels are set to either defaults or conda-forge. If your channels are set correctly and you’re still experiencing issues, check with your administrator to see if they have defined any policies that could be blocking installation.

If neither of the above resolves the issue, you can try:
  • Refreshing the page
  • Checking your internet connection
  • Shutting down all running Jupyter kernels and restarting your Jupyter instance

​
CLI or Navigator gives “error starting kernel” message

Cause
Uninstalling a quick start environment using either the CLI or Navigator can result in an Error starting kernel: [Errno 2] No such file or directory message, because the Jupyter kernel references aren’t properly removed when the environment is deleted.
Solution
To resolve this issue:
  1. Open Anaconda Prompt (Terminal in macOS/Linux)
  2. Run: 
    jupyter kernelspec list
  3. Find the name of the quick start environment you deleted.
  4. Then run: 
    jupyter kernelspec uninstall <ENV_NAME>
    Replace <ENV_NAME> with the name of the deleted environment.

​
The GPU runtime does not start up

Cause
The GPU runtime can fail to start due to a temporary issue with the runtime initialization or connection.
Solution
Reload the Anaconda Notebooks page and then select “Launch a Remote Runtime” from the Launcher. If you’re still unable to run a GPU runtime, file a support ticket.

​
My code isn’t running in GPU runtime

Cause
Code fails to run due to missing import statements or variables.
Solution
Check for missing import statements and verify that variables were correctly transferred to the GPU runtime.

​
Published Panel app is blank

Cause
The app can appear blank if the notebook cannot be executed sequentially from top to bottom, which can happen when cells have dependencies that aren’t met or are out of order.
Solution
If you added content to your app but there’s still nothing showing up, ensure that your notebook can be run from top to bottom. The easiest way to test this is to select Run in the top menu, then select Run All Cells.

​
All CPU seconds have been used up in Notebooks

Cause
You’ve used up all your available CPU seconds.
Solution
CPU seconds are consumed by active runtimes, which can continue to run in the background even after you’ve logged out of Notebooks or switched to a different browser tab.

To prevent your CPU seconds from being used while you are not actively using Notebooks, make sure you shut down all runtimes in all open instances of Notebooks before logging off for the day. You can do this from the Running Terminals and Kernels tab:


If the problem persists, please raise the issue in the Anaconda community forums or file a support ticket.

You can upgrade your subscription tier to access additional CPU seconds. See our pricing page for further details.

​
Error when importing packages in Notebooks

Cause
The most common cause of errors is a lack of required package(s) installed in your environment. The default environment we provide, based on the Anaconda distribution, contains hundreds of the most common python packages for data science, but it doesn’t include everything. You may need to create a custom environment to install the package you need.
Solution

Make sure you have the right runtime selected

The default anaconda-<YEAR>.<MONTH>-py<PYTHON_VERSION> runtimes have a broad selection of packages, but you may have created a custom environment for your notebook. Separate environments are represented as “runtimes” in JupyterLab. You can view and switch between available runtimes by clicking the runtime name in the upper-right corner of the content pane.

List the packages available in an environment

If one of your imports is failing, start by verifying that the package is present in the runtime environment. You can view which packages are available in your current environment from the terminal by running conda list.To view packages in a specific environment, run conda list --name <ENV_NAME>.To see a list of available environments, run conda info --envs. An asterisk appears beside your current active environment.

Create a custom runtime

If none of the Anaconda provided environments contain the package(s) you need, create a custom runtime for your notebook project.

​
”Missing ipykernel dependency” error

Cause
A missing ipykernel dependency error is caused by an environment dependency list not including the required ipykernel package.
Solution
To resolve this issue, add the ipykernel package to your environment’s dependency list. For example:
    channels:
    - defaults
    dependencies:
    - python=3.11
    - numpy
    - pandas
    - ipykernel

​
Custom runtime not appearing in Launcher or Assign a Runtime modal

Cause
The most common reason a runtime or environment is not appearing in the Launcher or Assign a Runtime modal is that it doesn’t contain a .
Solution
For more information and instructions on installing a kernel package in your runtime/environment, see Creating custom runtimes.

​
”File load error,” “unhandled error,” or “unexpected error” message

Cause
If you receive a “file load error,” “unhandled error,” or “unexpected error,” like in the following figure, you have most likely exceeded the storage space for your current tier.

Solution
Follow the steps in the storage question above to remove items from your Notebooks instance, or upgrade to a higher subscription tier.

​
Notebooks won’t open from Anaconda Platform

Cause
Your browser’s pop-up blocker (automatically enabled on Firefox and Safari) can prevent Notebooks from opening.
Solution
Disable your pop-up blocker and try opening Notebooks again from Anaconda Platform.

​
Notebook cannot access external sites after subscription upgrade

Cause
Notebook kernels inherit network permissions when a session is started. Any notebook sessions that were started while your account was using the free tier are still operating under PythonAnywhere’s allowlist restrictions.
Solution
Stop and restart all Notebook sessions to establish a kernel with unrestricted internet access.
It can take up to an hour for the upgrade to take effect.