A colleague downloaded an image and made a lot of changes to it (yes, he probably should have adapted the dockerfile instead).
To avoid having to do the work myself, he did an export by running docker container export and provided me the file.
After having done the import, the container started and the changes my colleague made were available, but some commands were not working. After some research it seemed that ‘jruby’ and ‘gem’ were still there but the path was not adapted so the OS could not find them.
The original container was from an image with ‘FROM jruby:1.7.21-jdk’, and now it seems that the imported container does not execute the path changes linked to that base image (the original dockerfile does not include any other path changes, so I can only assume the path changes come from the base image).
I know now that I can commit the container to an image, or better yet with an updated dockerfile for production, but for my understanding: is it normal that it does not execute the base image path changes when I run this after the import?