Skip to content

Installation

Install instructions will be here.

Docker Quick-start

Do you have a link to a sky docker image file? If not, these are not the install instructions you are looking for. If you do have a link, go ahead and download the file. If you are installing on Windows or an Intel-based Mac, you want the amd64 image. For installing on more recent Apple Silicon-based Macs, use an arm64 image.

In addition to the image file, you'll want to download and install Docker Desktop for your computer. Once it's installed, make sure it's running (you should be able to open and see Docker's dashboard).

Then, open a console terminal and change to the directory where the docker image sky.tar file has been saved. It might be named something slightly different from sky.tar, in which case change the filename in the commands you type to match the filename of the image file you have. The first step is to run

docker load --input sky.tar

This command may take a few moments to complete, as the content of the downloaded image file will be imported into docker. You'll need several free gigabytes of hard disk space to complete this step, so if you encounter unexplained EOF errors here that may be the problem.

After the docker image is loaded, you should see a sky image in the docker desktop dashboard. If you can see it, you should then be able to start a container with this image using the following command:

docker run -p 8899:8899 --rm --volume "$(pwd)":/tmp/workplace/work sky:latest

docker run -p 8899:8899 --rm --volume "%cd%":/tmp/workplace/work sky:latest

Within this command, we have:

  • -p 8899:8899 tells docker to expose the container's port 8899 (which has been configured to be the port served by Jupyter Lab) to localhost.
  • --rm means to remove the container when it exits, so there isn't an extraneous container image file left on your file system.
  • --volume "...":/tmp/workplace/work makes the current working directory available inside Jupyter Lab in a directory named work (the container has been configured with /tmp/workplace as the base location for Jupyter).
  • sky:latest tells Docker to use the latest version of the user image that was installed in the docker load step above. If your installed docker image (as shown in the Docker Dashboard) has a different name than sky then change the command to give the corrent image name.