Cloudera Docs

Post-migration tasks

After migrating Cloudera Data Science Workbench (CDSW) 1.10.0 or later to Cloudera Machine Learning (CML), you need to perform several tasks before moving users to CML.

You begin this task by assigning the user, group and the roles in the PVC control plane.

The CDSW to CML migration in Cloudera Manager updates the docker registry, engines, and model builds. You follow steps to assign roles to users and groups, import Grafana dashboards you previously exported, configure endpoints and DNS resolution, and configure LDAP.
  1. In Cloudera Manager, click User Management, and select Users.
  2. Click Actions > Upload Users.
  3. Select Groups, click Create Groups, and create groups.
  4. Start the CML workspace on the CML cluster, and check workloads.
    Start new sessions, jobs, models, and applications. Try starting existing workloads that were migrated from CDSW.
  5. Import the Grafana dashboards you exported earlier.
  6. Configure cluster endpoint connectivity per the information about required connections you had in your CDSW cluster.
  7. If you customized your DNS configuration on CDSW, manually configure your DNS in your private cloud cluster.
    If you did not customize your DNS configuration, the migration tool sets up the default DNS configuration in your private cloud cluster.
  8. Configure LDAP on CML, and grant user access on CML.
    CML in CDP Private Cloud supports only LDAP.
  9. Disable the CDSW cluster, and give users access to CML.