Cloudera Docs

Restoring a Cloudera Data Visualization version in Cloudera Machine Learning/Cloudera Data Science Workbench

Restoring a previous version means reverting your Cloudera Data Visualization application to an earlier version. You can restore your system to a known stable state, especially when issues arise during an upgrade or if the new version introduces major problems that impact functionality, and you need to return to the previous Cloudera Data Visualization version to maintain system functionality. If you decide to downgrade, follow the steps outlined below.

To restore a previous version, you should have either connected your system to an external database (for example: MySQL, MariaDB, Postgres, Oracle) or created a backup of the SQLite metadata database.
  1. Stop the application.
    On the Applications page in Cloudera Machine Learning/Cloudera Data Science Workbench, click Actions > Stop to stop the selected Cloudera Data Visualization application.
  2. Back up the SQLite metadata database.
    1. Navigate to your project's Overview page and click New Session.
    2. When your session is running, click >_ Terminal Access above the session log pane to open a terminal window.
      The default working directory is /home/cdsw, where all your project files reside.
    3. Navigate to /home/cdsw/.arc.
      This directory contains the Cloudera Data Visualization logs, files, settings, and the arcviz.db database.
    4. Run the following command to create a copy of your current Cloudera Data Visualization metadata.
      sqlite3 arcviz.db “.backup arcviz.db-BACKUP-[***version***]”

      You can use this backup when reverting to a previous Cloudera Data Visualization version.

  3. Create a new Cloudera Data Visualization application in your Cloudera Machine Learning/Cloudera Data Science Workbench project.
  4. Copy the backed-up SQLite database to /home/cdsw/.arc/ in the new project.
  5. Once the backup is restored, start the new Cloudera Data Visualization application.