Description
Preliminary Checks
- This issue is not a question, feature request, RFC, or anything other than a bug report. Please post those things in GitHub Discussions: https://github.com/nebari-dev/nebari/discussions
Summary
Every time conda-store rebuilds an environment, its actually storing it as a hashed version of the displayed env name. Because of the way jupyter works, if you update an env you're using in a notebook, the notebook will still be using the old environment.
This has to do with the juypter server caching the absolute path to the current kernel.
How conda store works:
- symlink global/dashboard -> /path/to/environment/dashboard-1
When you build a new environment conda-store changes the symlink - /path/to/environment/dashboard-1 (exists but not symlinked)
- symlink global/dashboard -> /path/to/environment/dashboard-2
Jupyter server or something along those lines while we give it the symlink it resolves the path an "remembers" /path/to/environment/dashboard-1
We need a note in our environment building docs and maybe the FAQ that mentions at a high level that when you rebuild your environment, you'll need to
- switch from you environment to a different kernel
- switch back to your desired environment
Restart will not be effective here.
Steps to Resolve this Issue
Proper resolution lies with the conda-store project. We just need to document existing functionality.
Metadata
Metadata
Assignees
Labels
Type
Projects
Status
Todo 📬
Status
New 🚦