-
Notifications
You must be signed in to change notification settings - Fork 54
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
Move VICTOR's hub image to use a Dockerfile #1934
Comments
3.7 is *quite* old, and folks get caught up in it because it also sometimes forces much older versions of packages - and it can be quite confusing to debug (see 2i2c-org/infrastructure#1934) for example.
Helps a *lot* with figuring out *why* certain packages were installed with the versions they were installed with. Makes it easier to debug why a older version of gdal is provided, for example - like in 2i2c-org/infrastructure#1934
Replicating my message in freshdesk: Hi Samuel, In general, it is always good to get packages from conda - even if you got gdal from apt, packages that are coming from conda that depend on gdal will still use the gdal from conda! So this usually causes confusion with respect to versioning. So I dug into why the conda version was only pulling in an older version of gdal. The error message Sarah pulled out indicated that when gdal newer version is specified, conda just tries to find compatible versions and fails. I dug a little deeper, and discovered this is because the default version of python used is 3.7, and that is the last version of gdal that can be resolved with that! So volcanocyber/VICTOR-notebook#1 bumps up the version of python to 3.10, and that brings us a newer gdal :) I've also opened 2i2c-org/hub-user-image-template#20 to set 3.10 as the new default in our template so others don't run into this problem. I'll also be opening an issue in repo2docker soon to bump the default there as well. Try merging that PR and testing it out? Hope that helps! Thank you :)
|
I have now opened jupyterhub/repo2docker#1219 as well upstream to bump up the default python to 3.10. In addition, there is jupyterhub/repo2docker#909 that would allow us to move away from 18.04 - but I think that should be orthogonal to the issue here. |
We have solved this without them needing to use a dockerfile. |
Helps a *lot* with figuring out *why* certain packages were installed with the versions they were installed with. Makes it easier to debug why a older version of gdal is provided, for example - like in 2i2c-org/infrastructure#1934
Helps a *lot* with figuring out *why* certain packages were installed with the versions they were installed with. Makes it easier to debug why a older version of gdal is provided, for example - like in 2i2c-org/infrastructure#1934
Context
ref: https://2i2c.freshdesk.com/a/tickets/288
In this ticket, a specific version of
gdal
is required in the image but wouldn't correctly install. We eventually tracked this down to the fact that the base image used by repo2docker is Ubuntu 18.04 (bionic). Information was requested on how an install of Ubuntu 22.04 or a newer version of apt could be achieved. I don't think this is possible without reconfiguring the repo to use a Dockerfile - which is perhaps work we should do ourselves, rather than leave to the community.Image repo: https://github.com/volcanocyber/VICTOR-notebook
Proposal
No response
Updates and actions
No response
The text was updated successfully, but these errors were encountered: