-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Report Docker Image in Login Page or JupyterLab UI #16
Comments
A similar or even bigger issue came up in one of our group meetings yesterday as well. Basically for people who have been working on the cloud, when they return to their projects they can easily lose the reproducibility of their own projects because the packages keep marching forward automatically. (e.g. a student (@shanicetbailey) returned to their project repo to do revisions for their paper last week, and are unable to get most of their notebooks to work anymore.) One suggestion from @rabernat was to give the user the choice of the image version when they start their hub. So someone can find the image that works best for their work, and then not have to worry again whether some PR would wipe things. Is something like this possible @yuvipanda? Also tagging @jbusecke here, since he might have some opinion here. |
cc @jmunroe who will be interested in these perspectives/requests from the communities. |
My suggestion would be to focus on implementing 2i2c-org/infrastructure#1253, which would allow the user to explicitly select their image at login time. Not only would that solve the need raised here (allow user to know what docker image they are using)--it would go further and allow them to actively choose the image they want. |
I agree with @rabernat. This would help both with reproducibility and debugging (by enabling a relatively quick high level test to confirm if an updated image contains the culprit for any given issue). |
It would also be useful to report the kinda instance you are on - someone just unfortunately wasted precious time because they were on a different kind of instance than they thought they were on, and there's no real way to tell this from the UI :( |
I have faced this too. |
Context
Users of Hubs launch a session with a Docker Image but do not know which image is being used without some digging (like
echo JUPYTER_IMAGE
from a terminal). For troubleshooting problems with libraries this is very important information to have. See this discussion for example pangeo-data/pangeo-docker-images#387 (comment)Proposal
A couple ideas:
In the profile list of a hub in addition to resources list the image (e.g.
pangeo/ml-notebook:2022.09.21
), this is maybe tricky because it can be overriden by the configuratorPrint it out somewhere in the JupyterLab Interface (maybe the lower toolbar for example like https://github.com/jupyter-server/jupyter-resource-usage)
Updates and actions
No response
The text was updated successfully, but these errors were encountered: