Is there a way to reuse the rcc dependencies?

I’m automating a test and running that automation in a pipeline in gitlab. I’m using rcc to run this automation but as the container is ephemeral, everything downloaded from the “conda.yaml” file has to be downloaded again at each test run. What I want to know is if there is a way to store these dependencies in the docker image and use them when I run the test, and not download it again every time.

Hi @gomesamantha2303,

Yes there is.

If you know the conda.yaml (or yamls) that you are using then you can simply rebuild the environment when creating the container image. Just run rcc ht vars conda.yaml commands when creating the container and the RCC cache is stored into your container.
This will make your container a bit bigger put caching is way faster.

BR, Kari


Hi Kari,

hm, did you probably rather mean the image creation? Having the dependencies within the image would skip the whole env recreation if conda.yaml is still the same.

Good spot Simon, edited to avoid confusion.

I keep saying development is 95% naming, so these are crucial distinctions :wink:

1 Like