Administer Watson Studio Local from the command line
You can automate some Watson Studio Local administration tasks using the command line interface (CLI).
Some of the administrative tasks you can automate with the command line interface include:
- Administering project jobs.
- Onboarding a new user in Watson Studio Local and associate them with the appropriate set of projects.
- Removing users and all of their associated assets from Watson Studio Local.
- Managing assets in a continuous integration/continuous delivery pipeline. For example, take assets built by a data scientist and move them through test and production environments.
- Developing custom solutions for your organization's needs.
- Monitoring the health of your clusters and notifications are provided when pods are down, such as dsx-core/usermgmt.
To use the command line interface for Watson Studio Local, download the package from https://www-01.ibm.com/marketing/iwm/iwm/web/dispatcher.do?source=mrs-ibmdselt.
Prerequisites
- Java 8 runtime environment (64-bit)
Installing the CLI
To install the command line interface:
- Unzip WSCLI.zip to extract the contents of the directory.
- From a command prompt, change to the WSCLI directory. For example, on
Windows:
cd WSCLI
- To view a list of available commands, type:
./wscli --help
which returns help for each command. For example:
$ ./wscli -h NAME wscli <command> DESCRIPTION WS command-line interface AVAILABLE COMMANDS USER MANAGEMENT add user Adds the user to Watson Studio get user Returns the user information list users Returns the list of users in Watson Studio edit user Edits the Watson Studio user delete user Deletes the user in Watson Studio add collaborator Adds the user as a collaborator to the project, with the specific role list collaborators Returns the list of collaborators in the project edit collaborator Changes the user role as a collaborator in the project delete collaborator Removes the user's access to the project as a collaborator reset password Reset the user's password. Can only be run by Admin User get token Returns the authorization token for the user
Supplying credentials for CLI
Every command has three required parameters to connect to the CLI service:
- --cluster-host (-ch)
- --auth-user (-au)
- --auth-password (-ap)
For example, these are valid values for the
cluster-host
parameter:https://cluster-host-name.domain.com
https://9.100.123.123
cluster-host-name.domain.com
9.100.123.123
If you do not want to provide these credential parameters for every command, set them using environment variables.
Option | Environment variable |
---|---|
--cluster-host -ch | WS_CLUSTER_IP |
--auth-user -au | WS_AUTH_USER |
--auth-password -ap | WS_AUTH_PWD |
To set the ENV variables:
- From a command prompt, run the
setupEnv.sh
shell command:- vi setupEnv.sh
- Edit the following fields to supply your
credentials:
export WS_CLUSTER_IP="<ws-cluster-host-or-ip>" export WS_AUTH_USER="<ws-admin-user-name>" export WS_AUTH_PWD="<ws-admin-user-password>"
- Save the changes:
source ./setupEnv.sh
Alternatively, you can retrieve a bearer token for your user credentials and store that in the
WS_AUTH_TOKEN
environment variable. When this variable is populated, CLI ignores
the values for WS_AUTH_USER
and AUTH_PWD
.
Follow these steps to retrieve and store the authentication token.
- Run this command to retrieve the
token:
WS_AUTH_TOKEN=`curl -k -X GET <CLUSTER_HOST>/v1/preauth/validateAuth -u <username>:<password> | jq -r '.accessToken'`
- Run the utility script getToken.sh to store the
token:
source ./getToken.sh
Note: Bearer tokens expire after a few hours. If the token stored in WS_AUTH_TOKEN has expired,
the command will fail and you will need to either generate a new token or clear the
WS_AUTH_TOKEN
environment variable.Using optional parameters with CLI
You can supply optional parameter that control attributes such as the format of the command
output or how much information to display for each command. Required and optional parameters can be
provided in any order. The complete list is supplied when you enter the CLI help
command:
./wscli --help
Optional parameters include:
OPTIONAL PARAMETERS:
-json --json
Provides the command output in JSON format
Default: false
Common:
-debug --debug
Provides additional debug content for IBM Support
Default: false
-h --help
Displays the command reference help
-rs --result-status
Shows the command execution status and time before the final command result output
Default: false
-v --verbose
Provides detailed command execution output messages
Default: false
Note: The
-re
parameter value dsx-scripted-ml-python3
refers to
the included Python 3 image. For earlier versions of Watson Studio Local,
the parameter value is the Jupyter with Python 3.5, Scala 2.11, R 3.4.3
image. For
Watson Studio Local 1.2.3.3, the parameter value is the Jupyter with
Python 3.7, Scala 2.11, R 3.6
image.