Skip to content
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

install additional conda packages may fail #96

Closed
cehbrecht opened this issue Jun 24, 2020 · 3 comments · Fixed by #97
Closed

install additional conda packages may fail #96

cehbrecht opened this issue Jun 24, 2020 · 3 comments · Fixed by #97
Assignees
Labels
bug Something isn't working
Milestone

Comments

@cehbrecht
Copy link
Member

cehbrecht commented Jun 24, 2020

This playbook installs additional conda packages (gunicorn, ...) needed for deployment. Installation may fail due to package dependencies (exsiting env with "conda-forge" .... update with "default" channel):

- name: Install additional Conda packages.

Is there a conda option to prevent this? Use pip instead?

@cehbrecht cehbrecht added the bug Something isn't working label Jun 24, 2020
@cehbrecht cehbrecht added this to the 0.4.0 milestone Jun 24, 2020
@cehbrecht cehbrecht self-assigned this Jun 24, 2020
@tlvu
Copy link

tlvu commented Jun 24, 2020

Is there a conda option to prevent this? Use pip instead?

Have to conda install with the same list of channels, keeping also the same precedence ordering as the original environment.yml file, ex: https://github.com/Ouranosinc/PAVICS-e2e-workflow-tests/blob/441edde3b381eff7ce82e5a171323b31196553be/docker/Dockerfile#L30

@cehbrecht
Copy link
Member Author

@tlvu I think the easiest for now would be just to use pip ... I only need to install gunicorn and optionally psycopg2. Otherwise I would need to figure out the channel list used in the environment.

@tlvu
Copy link

tlvu commented Jun 25, 2020

@tlvu I think the easiest for now would be just to use pip ... I only need to install gunicorn and optionally psycopg2. Otherwise I would need to figure out the channel list used in the environment.

Mostly the channel list is only conda-forgeand defaults. If there are extra channels, they should "theoretically" not overshadow the releases in conda-forge and defaults so we could probably have skipped them safely.

Pip is fine when the new packages are pure python packages, not containing C-extensions that requires other system packages.

So bottom line is pip might not work in all cases, but if it works for this specific case, fine.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants