This is a basic question for which I didn’t see anything on in the docs or forum searches.

I have a development machine which has all kinds of stuff installed to do machine learning, like anaconda, python, etc. I want to containerize all of my development using Docker. Would it be a good practice to uninstall everything of importance on the host machine, since everything is going to be running on containers? I guess the main question I have is whether my docker build can/will use any of these ML from my local host machine when building the container.

1 Answer 1


Ultimately, as long as you have the disk space and RAM, you can run both in parallel with very few issues.

The biggest problem you may run into is that the default ports of the applications you run in a container may already be in use. For example SQL server install runs on port 1433 by default. If you try to run a SQL server container and do not map a different port on the host machine to 1433 on the container you will run into errors. This is easy enough do though:

docker run -e "ACCEPT_EULA=Y" -e "SA_PASSWORD=" -p 1431:1433 -d mcr.microsoft.com/mssql/server:2019-CU3-ubuntu-18.04

In the above example, MSSQL server will run on port 1431 and map to port 1433 in the container. On your local machine you can access either your SQL server on port 1433 or the container SQL server on 1431.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.