Managing CLI profiles

A CLI configuration file allows you to configure different profiles, or contexts, for use with the CLI tools overview. A context consists of user authentication and OKD server information associated with a nickname.

About switches between CLI profiles

Contexts allow you to easily switch between multiple users across multiple OKD servers, or clusters, when using CLI operations. Nicknames make managing CLI configurations easier by providing short-hand references to contexts, user credentials, and cluster details. After logging in with the CLI for the first time, OKD creates a ~/.kube/config file if one does not already exist. As more authentication and connection details are provided to the CLI, either automatically during an oc login operation or by manually configuring CLI profiles, the updated information is stored in the configuration file:

CLI config file

apiVersion: v1
clusters: (1)
- cluster:
    insecure-skip-tls-verify: true
    server: https://openshift1.example.com:8443
  name: openshift1.example.com:8443
- cluster:
    insecure-skip-tls-verify: true
    server: https://openshift2.example.com:8443
  name: openshift2.example.com:8443
contexts: (2)
- context:
    cluster: openshift1.example.com:8443
    namespace: alice-project
    user: alice/openshift1.example.com:8443
  name: alice-project/openshift1.example.com:8443/alice
- context:
    cluster: openshift1.example.com:8443
    namespace: joe-project
    user: alice/openshift1.example.com:8443
  name: joe-project/openshift1/alice
current-context: joe-project/openshift1.example.com:8443/alice (3)
kind: Config
preferences: {}
users: (4)
- name: alice/openshift1.example.com:8443
  user:
    token: xZHd2piv5_9vQrg-SKXRJ2Dsl9SceNJdhNTljEKTb8k

The CLI can support multiple configuration files which are loaded at runtime and merged together along with any override options specified from the command line. After you are logged in, you can use the oc status or oc project command to verify your current working environment:

Verify the current working environment

$ oc status

Example output

oc status
In project Joe's Project (joe-project)
service database (172.30.43.12:5434 -> 3306)
  database deploys docker.io/openshift/mysql-55-centos7:latest
    #1 deployed 25 minutes ago - 1 pod
service frontend (172.30.159.137:5432 -> 8080)
  frontend deploys origin-ruby-sample:latest <-
    builds https://github.com/openshift/ruby-hello-world with joe-project/ruby-20-centos7:latest
    #1 deployed 22 minutes ago - 2 pods
To see more information about a service or deployment, use 'oc describe service <name>' or 'oc describe dc <name>'.
You can use 'oc get all' to see lists of each of the types described in this example.

List the current project

$ oc project

Example output

Using project "joe-project" from context named "joe-project/openshift1.example.com:8443/alice" on server "https://openshift1.example.com:8443".

You can run the oc login command again and supply the required information during the interactive process, to log in using any other combination of user credentials and cluster details. A context is constructed based on the supplied information if one does not already exist. If you are already logged in and want to switch to another project the current user already has access to, use the oc project command and enter the name of the project:

$ oc project alice-project

Example output

Now using project "alice-project" on server "https://openshift1.example.com:8443".

At any time, you can use the oc config view command to view your current CLI configuration, as seen in the output. Additional CLI configuration commands are also available for more advanced usage.

$ oc login -u system:admin -n default

Manual configuration of CLI profiles

If you want to manually configure your CLI config files, you can use the oc config command instead of directly modifying the files. The oc config command includes a number of helpful sub-commands for this purpose:

$ oc config set-cluster <cluster_nickname> [—server=<master_ip_or_fqdn>]
[—certificate-authority=<path/to/certificate/authority>]
[—api-version=<apiversion>] [—insecure-skip-tls-verify=true]

$ oc config set-context <context_nickname> [—cluster=<cluster_nickname>]
[—user=<user_nickname>] [—namespace=<namespace>]

$ oc config use-context <context_nickname>

$ oc config set <property_name> <property_value>

$ oc config unset <property_name>

$ oc config view

$ oc config view —config=<specific_filename>

Example usage

• Log in as a user that uses an access token. This token is used by the alice user:

$ oc login https://openshift1.example.com --token=ns7yVhuRNpDM9cgzfhhxQ7bM5s7N2ZVrkZepSRf4LC0

• View the cluster entry automatically created:

$ oc config view

Example output

apiVersion: v1
clusters:
- cluster:
    insecure-skip-tls-verify: true
    server: https://openshift1.example.com
  name: openshift1-example-com
contexts:
- context:
    cluster: openshift1-example-com
    namespace: default
    user: alice/openshift1-example-com
  name: default/openshift1-example-com/alice
current-context: default/openshift1-example-com/alice
kind: Config
preferences: {}
users:
- name: alice/openshift1.example.com
  user:
    token: ns7yVhuRNpDM9cgzfhhxQ7bM5s7N2ZVrkZepSRf4LC0

• Update the current context to have users log in to the desired namespace:

$ oc config set-context `oc config current-context` --namespace=<project_name>

• Examine the current context, to confirm that the changes are implemented:

$ oc whoami -c

All subsequent CLI operations uses the new context, unless otherwise specified by overriding CLI options or until the context is switched.

Load and merge rules

You can follow these rules, when issuing CLI operations for the loading and merging order for the CLI configuration:

• CLI config files are retrieved from your workstation, using the following hierarchy and merge rules:

  • If the --config option is set, then only that file is loaded. The flag is set once and no merging takes place.

  • If the $KUBECONFIG environment variable is set, then it is used. The variable can be a list of paths, and if so the paths are merged together. When a value is modified, it is modified in the file that defines the stanza. When a value is created, it is created in the first file that exists. If no files in the chain exist, then it creates the last file in the list.

  • Otherwise, the *~/.kube/config* file is used and no merging takes place.

• The context to use is determined based on the first match in the following flow:

  • The value of the --context option.

  • The current-context value from the CLI config file.

  • An empty value is allowed at this stage.

• The user and cluster to use is determined. At this point, you may or may not have a context; they are built based on the first match in the following flow, which is run once for the user and once for the cluster:

  • The value of the --user for user name and --cluster option for cluster name.

  • If the --context option is present, then use the context’s value.

  • An empty value is allowed at this stage.

• The actual cluster information to use is determined. At this point, you may or may not have cluster information. Each piece of the cluster information is built based on the first match in the following flow:

  • The values of any of the following command line options:

    • --server,

    • --api-version

    • --certificate-authority

    • --insecure-skip-tls-verify

  • If cluster information and a value for the attribute is present, then use it.

  • If you do not have a server location, then there is an error.

• The actual user information to use is determined. Users are built using the same rules as clusters, except that you can only have one authentication technique per user; conflicting techniques cause the operation to fail. Command line options take precedence over config file values. Valid command line options are:

  • --auth-path

  • --client-certificate

  • --client-key

  • --token

• For any information that is still missing, default values are used and prompts are given for additional information.