How to download a specific git-lfs file into a Docker container

Question

I have a git repo with several large files. Git-LFS is enabled. I want to bring in one of the files into a Docker container. I have installed git-lfs in the container. So far I have:

RUN git clone --no-checkout --depth 1 https://github.com/my-org/my-data-repo.git
RUN cd my-data-repo
RUN git lfs pull -I data/my-large-file.csv

The file actually get's downloaded but the Docker build process fails because I get the following error:

Error updating the git index: (1/1), 90 MB | 4.8 MB/s                                                                                                                                                                                       
error: data/my-large-file.csv: cannot add to the index - missing --add option?
fatal: Unable to process path data/my-large-file.csv


Errors logged to .git/lfs/logs/20200709T142011.864584.log
Use `git lfs logs last` to view the log.

How can do this without an exception being thrown which kills the Docker build process?

Answer 1

One of your issues is :

RUN cd my-data-repo
RUN git lfs pull -I data/my-large-file.csv

doesn't work as you expect :

• a first process is started, in which cd my-data-reop sets the current directory to my-data-repo,
• this process is then ended (along with its working directory)
• a second process is then started, which runs git lfs pull ... from the intital directory

---

You can either group your commands :

RUN cd my-data-repo && git lfs pull -I data/my-large-file.csv

or use the WORKDIR command (as suggested here) :

WORKDIR my-data-repo
RUN git lfs pull -I data/my-large-file.csv