Troubleshooting Cloudera AI command-line utility tool

Learn about the troubleshooting issues with the Cloudera AI command-line utility tool.

Retrying migration

The command line utility tool has been designed to resume the export or import operation from the exact point where it was left off in the event of failures.

Project files - The command-line utility tool employs rsync to facilitate the migration of project files. When the export or import command is rerun, the tool synchronizes the project files from the source to the target.

Project settings or artifacts, that is Models, Jobs or Applications - During a rerun, the project settings or artifacts, that is Models, Jobs or Applications, that have already been migrated are not to be updated. Instead, only the missing artifacts in the target workbench are migrated. This behavior aligns with the resume operation support in case of failures. The command-line utility tool is not designed to support a synchronization operation between the source and target projects. If you want to update the project that has already been migrated, delete the project in the target workbench and then rerun the migration.

Running the tool as a background process

There might be cases when the tool is required to run as a separate process detached from the terminal. In that case use the following command on your Linux or Mac machine:

nohup <complete CLI command> > <stdout-file> 2> <stderr-file> &

The command prints out the Process ID (PID) as well as appends the logs to the stdout and stderr file locations.

The process can be interrupted by sending a SIGINT to the PID returned by using the kill -INT <PID> command.

SSH timeout issues

If an SSH timeout occurs during the project migration or the creation of Models, Sessions or Jobs, increase the ServerAliveInterval value in the ~/.ssh/config command of the bastion host or third computer. The following example is a sample command to increase the SSH timeout in the ~/.ssh/config command:

  ServerAliveInterval 60 
   ServerAliveCountMax 30

'`rsync` connection closed' issues on the bastion host

The third computer's, or bastion host's rsync might stop with the connection broken error. In that case, make sure that the rsync version is upgraded to 3.X as rsync 3.X and higher versions have a retry functionality inherently implemented when stopped.

Remote host identification has changed or Host key verification failed error messages

Remove the host key entry from the .ssh/known_hosts file by using the ssh-keygen -R server-hostname-or-IP command.

For example: ssh-keygen -R '[localhost]:5104'

SSL certificate_verify_failed error message

Create a bundle of all root or intermediate Certification Authorities (CAs) in .pem format. Save this bundle file and provide its path (ca_path) inside the export/import-config.ini file. This ensures that Python can locate the trusted certificate and prevent the error from occurring.

`configparser.NoSectionError: No section: ‘sample_project` error message

The command-line utility tool cannot find a section for the sample_project line in the export-config.ini or the import-config.ini file. Add the section to the files, with the section name being the same as the Project name.