Upgrade

ENTERPRISE

Upgrading DC/OS Data Science Engine

DC/OS Data Science Engine does not support in-place upgrades. In order to transition to a new version you must manually back up your data located in your jupyter_data folder and upload it to a new instance of DC/OS Data Science Engine, running on the latest version using the following guide.

  1. Navigate to the Notebook terminal from Jupyter Notebook UI and run the following command to compress the contents of your jupyter_data folder:

    nobody@95602332-7dd1-4a24-b2a5-b0fa028c0998:/mnt/mesos/sandbox$ tar -cvf jupyter_data.tar.gz jupyter_data
    Copy

    NOTE: If you have files you would like to keep from outside of the jupyter_data folder, you should also back up these files, but keep in mind that only the data in the jupyter_data folder persists between service restarts. If you copy files from outside of the jupyter_data directory, please ensure you do not copy any hidden files that could interfere with the operation of DC/OS Data Science Engine after being uploaded to the new instance.

  2. Use the Download button to download jupyter_data.tar.gz.

    DC/OS Data Science Engine Notebook Download

  3. Deploy a new instance of DC/OS Data Science Engine, specifying the upgrade version. You may need to tear down your existing DC/OS Data Science Engine instance depending on the available resources in the cluster.

  4. From the new instance’s UI, select the Upload button to transfer jupyter_data.tar.gz to the sandbox of the new DC/OS Data Science Engine instance.

    DC/OS Data Science Engine Notebook Upload

  5. Then run the following command to uncompress the contents of the archive to the jupyter_data folder:

    nobody@e08af121-e9e7-4c0f-8188-23bfead5080a:/mnt/mesos/sandbox$ tar -xvf jupyter_data.tar.gz
    Copy
    DC/OS Data Science Engine Notebook Unzip