Configuring the kernel

Config file location

The kernel can be configured via a JSON file called mariadb_config.json.

The kernel will look for this config file in one of these two locations:

1. If the JUPYTER_CONFIG_DIR environment variable is defined (non-empty) mariadb_kernel will try to read $JUPYTER_CONFIG_DIR/mariadb_config.json
2. If the env variable is empty, the kernel will try to read $HOME/.jupyter/mariadb_config.json

Config example

Here’s an example file containing some of the available options that you can pass to the kernel:

cat ~/.jupyter/mariadb_config.json
{
    "user": "root",
    "host": "localhost",
    "port": "3306",
    "password": "securepassword",
    "start_server": "True",
    "client_bin": "/usr/bin/mariadb",
    "server_bin": "/usr/bin/mariadbd"
}

By default the kernel tries to connect to a server running at localhost, on port 3306 using the user root with no password. You can change any of these options to fit your use-case. This also means you can run a notebook locally with a MariaDB kernel, and make the kernel connect to a server running in the cloud for instance.

The kernel, using the default configuration, looks for the mysql and mysqld binaries in your PATH. You can point (for example if you have a local MariaDB built from sources) the kernel to an exact location for these two binaries using the client_bin and server_bin options.

The start_server option tells the kernel to start a MariaDB Server instance for you, when the kernel is loaded, if it detects no running server given the configurations passed.

The full list of JSON options

If you suspect the documentation might not be up to date, you can check the complete list of available options at this link.