Checks performed before migration

Consider the list of issues and workarounds prior to migration.

Configuration readiness

In CDSW, Kube-DNS is used for handling the DNS, while in Cloudera AI, core DNS is used. If you add any custom kube-DNS configuration to the CDSW, those configuration details will not be copied to the Cloudera AI automatically. Checking this flags and verifies if any custom configuration is added to CDSW.

Troubleshooting failures

Check if any custom configurations are added to the kube-DNS and if the custom configurations are necessary.

Registry readiness

Verify the registry readiness in Cloudera AI, by checking if the created Cloudera AI cluster has proper registry permissions.

Troubleshooting failures

Check if the cdp-private-installer-embedded-registry-docker-pull-secret secret, in the control plane namespace is present in the Cloudera AI.

If the above configuration is present, check if the docker configuration in this secret is correct.

Host Mount readiness

CDSW supports the Host Mount feature, however, this is not supported in Cloudera AI. Migrate the workloads containing the host mounts is not possible.

Troubleshooting failures

The workloads containing the host mounts must be modified in CDSW before the migration. Convert the workloads using the engines to ML Runtimes.

Engine type readiness

The legacy engine is deprecated in Cloudera AI. If the CDSW contains any workloads using the engine or custom engines, those workloads must be converted to ML Runtimes before the migration.

Troubleshooting failures

Convert the engine-based or custom engine-based workloads to runtime-based workloads.

NFS filesystem readiness

The CDSW to Cloudera AI migration tool supports migrating project files from the CDSW internal NFS server to the Cloudera AI project storage. Checking readiness verifies the NFS filesystem size for the migration.

Troubleshooting failures

Check the storage configuration of the NFS storage in the Cloudera AI. Ensure enough storage is in the NFS storage to perform the migration.

Runtime Addons' readiness

Cloudera AI supports numerous ML Runtime addons such as Spark and Hadoop CLI. This check verifies that all the ML Runtime addons are installed in the Cloudera AI properly.

Troubleshooting failures

In the Cloudera AI UI, go to Site Administrator- > Runtime and check the status of the ML Runtime addons. If any of the ML Runtime addons is in the wrong state, select the Reload action from the Actions drop-down list.

Service readiness

Check and verify that all the services active in CDSW are enabled and started in Cloudera AI.

Troubleshooting failures

Ensure that the necessary services are started in the Cloudera AI during the workbench provisioning.

Versions readiness

CDSW to Cloudera AI migration is supported only from CDSW version 1.10.0. This check verifies the source CDSW version.

Troubleshooting failures

Update the CDSW to version 1.10.0.