Once you create the configuration file, you are ready to initialize the database root directory, using the voltdb init command. You issue this command on each node of the cluster, specifying the location for the root directory, the configuration file, license, and schema and stored procedure class files. There are defaults for each argument. But if you do specify the configuration, license, schema or classes you must specify the same values on every node of the cluster. For example:

$ voltdb init --dir=~/database         \ 
              --config=deployment.xml  \ 
              --license=~/license.xml  \  
              --schema=myschema.sql    \ 
              --classes=myprocs.jar      

On the command line, you can specify up to five arguments:

When you initialize the root directory, VoltDB:

Note that you only need to initialize the root directory once. Once the root directory is initialized, you can start and stop the database as needed. VoltDB uses the root directory to manage the current configuration options and backups of the data — if those features are selected — in command logs and snapshots. If you do not specify a license on the command line, VoltDB looks for a license in the current working directory, your home directory, or in the directory where the VoltDB software is installed and copies it into the root directory if it finds one.

If the root directory already exists or has been initialized before, you cannot re-initialize the directory unless you include the --force argument. This is to protect you against accidentally deleting data from a previous database session.