Downloading the runtime configuration

Every custom image that you create must be based on one of the standard images included in IBM Watson Studio. You can get the name of the image that is used by a runtime from its runtime definition.

Note that the GPU runtime definitions exist only if the Jupyter notebooks with Python for GPU service was installed on the IBM Cloud Pak for Data platform for the required Python version.

The following runtime definitions exist for the standard Jupyter notebook images included in Watson Studio:

Table 1. Runtime definitions for standard Jupyter notebook images included in Watson Studio
Runtime definition name	Image description
`jupyter-231l-py`	JupyterLab with Python 3.10
`jupyter-231n-py`	Jupyter Notebook with Python 3.10
`jupyter-222n-py`	Jupyter Notebook with Python 3.10
`jupyter-231l-pygpu`	JupyterLab with Python 3.10 with GPU support
`jupyter-231n-pygpu`	Jupyter Notebook with Python 3.10 with GPU support
`jupyter-222n-pygpu`	Jupyter Notebook with Python 3.10 with GPU support
`jupyter-222l-py`	JupyterLab with Python 3.10
`jupyter-222l-pygpu`	JupyterLab with Python 3.10 with GPU support
`jupyter-231n-r`	Jupyter Notebook with R 4.2
`jupyter-222n-r`	Jupyter Notebook with R 4.2

Note: Starting with the 4.8.4 release of Watson Studio, Runtime 22.2 is deprecated and will be removed in a future release. We recommend that you switch to Runtime 23.1.

For RStudio, use the following runtime definitions:

Table 2. Runtime definitions for the RStudio image included in Watson Studio
Runtime definition name	Image description
`rstudio-r42`	RStudio with Runtime 22.2 on R 4.2
`rstudio-rt3`	RStudio with Runtime 23.1 on R 4.2

For SPSS Modeler, use the following runtime definition:

Table 3. Runtime definitions for the SPSS Modeler image included in Watson Studio
Runtime definition name	Image description
`spss-modeler`	SPSS Modeler

Important:

A runtime definition contains an entry for the registry where the image is stored.

Retrieve the runtime definition by using the Cloud Pak for Data API and store it as a json file:

Download the runtime definition by using cURL in a terminal:
1. Get the required platform access token.
  1. Log in to Cloud Pak for Data as an administrator.
  2. Get the CPD platform URL (CPD_URL) by running:
```
cpd-cli manage get-cpd-instance-details --cpd_instance_ns=${PROJECT_CPD_INST_OPERANDS}
```
  3. Get an authorization token. For more information, see Generating an API authorization token.
  You can also set the bearer token manually by omitting the jq command and copying the output of the accessToken field to the MY_TOKEN variable. Note the usage of -k because of certificate issues. -k disables certificate validation. You might not have to do this.
2. List the names of all available runtime definitions on the cluster
```
curl -k  -X GET -H "Authorization: ZenApiKey ${MY_TOKEN}" "${CPD_URL}/v2/runtime_definitions" | jq -r ".resources[].entity.name"
```
3. Download the runtime definition. Replace <runtime-definition-name> with the name of the runtime definition that you need for your customization.
```
myRuntimeDefinition=<runtime-definition-name>; curl -k -X GET -H "Authorization: ZenApiKey ${MY_TOKEN}" "${CPD_URL}/v2/runtime_definitions?include=launch_configuration" | jq '.resources[] | select(.entity.name=="'${myRuntimeDefinition}'") | .entity'  > ${myRuntimeDefinition}.json
```

Parent topic: Preparing to build an image