kubernetes.core.k8s_drain module – Drain, Cordon, or Uncordon node in k8s cluster

Note

This module is part of the kubernetes.core collection (version 5.0.0).

You might already have this collection installed if you are using the ansible package. It is not included in ansible-core. To check whether it is installed, run ansible-galaxy collection list.

To install it, use: ansible-galaxy collection install kubernetes.core. You need further requirements to be able to use this module, see Requirements for details.

To use it in a playbook, specify: kubernetes.core.k8s_drain.

New in kubernetes.core 2.2.0

Synopsis

  • Drain node in preparation for maintenance same as kubectl drain.

  • Cordon will mark the node as unschedulable.

  • Uncordon will mark the node as schedulable.

  • The given node will be marked unschedulable to prevent new pods from arriving.

  • Then drain deletes all pods except mirror pods (which cannot be deleted through the API server).

Note

This module has a corresponding action plugin.

Requirements

The below requirements are needed on the host that executes this module.

  • python >= 3.9

  • kubernetes >= 24.2.0

Parameters

Parameter

Comments

api_key

string

Token used to authenticate with the API. Can also be specified via K8S_AUTH_API_KEY environment variable.

ca_cert

aliases: ssl_ca_cert

path

Path to a CA certificate used to authenticate with the API. The full certificate chain must be provided to avoid certificate validation errors. Can also be specified via K8S_AUTH_SSL_CA_CERT environment variable.

client_cert

aliases: cert_file

path

Path to a certificate used to authenticate with the API. Can also be specified via K8S_AUTH_CERT_FILE environment variable.

client_key

aliases: key_file

path

Path to a key file used to authenticate with the API. Can also be specified via K8S_AUTH_KEY_FILE environment variable.

context

string

The name of a context found in the config file. Can also be specified via K8S_AUTH_CONTEXT environment variable.

delete_options

dictionary

Specify options to delete pods.

This option has effect only when state is set to drain.

Default: {}

delete_emptydir_data

boolean

added in kubernetes.core 2.3.0

Continue even if there are pods using emptyDir (local data that will be deleted when the node is drained).

Choices:

  • false ← (default)

  • true

disable_eviction

boolean

Forces drain to use delete rather than evict.

Choices:

  • false ← (default)

  • true

force

boolean

Continue even if there are pods not managed by a ReplicationController, Job, or DaemonSet.

Choices:

  • false ← (default)

  • true

ignore_daemonsets

boolean

Ignore DaemonSet-managed pods.

Choices:

  • false ← (default)

  • true

terminate_grace_period

integer

Specify how many seconds to wait before forcefully terminating.

If not specified, the default grace period for the object type will be used.

The value zero indicates delete immediately.

wait_sleep

integer

Number of seconds to sleep between checks.

Ignored if wait_timeout is not set.

Default: 5

wait_timeout

integer

The length of time to wait in seconds for pod to be deleted before giving up, zero means infinite.

host

string

Provide a URL for accessing the API. Can also be specified via K8S_AUTH_HOST environment variable.

impersonate_groups

list / elements=string

added in kubernetes.core 2.3.0

Group(s) to impersonate for the operation.

Can also be specified via K8S_AUTH_IMPERSONATE_GROUPS environment. Example: Group1,Group2

impersonate_user

string

added in kubernetes.core 2.3.0

Username to impersonate for the operation.

Can also be specified via K8S_AUTH_IMPERSONATE_USER environment.

kubeconfig

any

Path to an existing Kubernetes config file. If not provided, and no other connection options are provided, the Kubernetes client will attempt to load the default configuration file from ~/.kube/config. Can also be specified via K8S_AUTH_KUBECONFIG environment variable.

Multiple Kubernetes config file can be provided using separator ‘;’ for Windows platform or ‘:’ for others platforms.

The kubernetes configuration can be provided as dictionary. This feature requires a python kubernetes client version >= 17.17.0. Added in version 2.2.0.

name

string / required

The name of the node.

no_proxy

string

added in kubernetes.core 2.3.0

The comma separated list of hosts/domains/IP/CIDR that shouldn’t go through proxy. Can also be specified via K8S_AUTH_NO_PROXY environment variable.

Please note that this module does not pick up typical proxy settings from the environment (e.g. NO_PROXY).

This feature requires kubernetes>=19.15.0. When kubernetes library is less than 19.15.0, it fails even no_proxy set in correct.

example value is “localhost,.local,.example.com,127.0.0.1,127.0.0.0/8,10.0.0.0/8,172.16.0.0/12,192.168.0.0/16”

password

string

Provide a password for authenticating with the API. Can also be specified via K8S_AUTH_PASSWORD environment variable.

Please read the description of the username option for a discussion of when this option is applicable.

persist_config

boolean

Whether or not to save the kube config refresh tokens. Can also be specified via K8S_AUTH_PERSIST_CONFIG environment variable.

When the k8s context is using a user credentials with refresh tokens (like oidc or gke/gcloud auth), the token is refreshed by the k8s python client library but not saved by default. So the old refresh token can expire and the next auth might fail. Setting this flag to true will tell the k8s python client to save the new refresh token to the kube config file.

Default to false.

Please note that the current version of the k8s python client library does not support setting this flag to True yet.

The fix for this k8s python library is here: https://github.com/kubernetes-client/python-base/pull/169

Choices:

  • false

  • true

pod_selectors

aliases: label_selectors

list / elements=string

added in kubernetes.core 2.5.0

Label selector to filter pods on the node.

This option has effect only when state is set to drain.

proxy

string

The URL of an HTTP proxy to use for the connection. Can also be specified via K8S_AUTH_PROXY environment variable.

Please note that this module does not pick up typical proxy settings from the environment (e.g. HTTP_PROXY).

proxy_headers

dictionary

added in kubernetes.core 2.0.0

The Header used for the HTTP proxy.

Documentation can be found here https://urllib3.readthedocs.io/en/latest/reference/urllib3.util.html?highlight=proxy_headers#urllib3.util.make_headers.

basic_auth

string

Colon-separated username:password for basic authentication header.

Can also be specified via K8S_AUTH_PROXY_HEADERS_BASIC_AUTH environment.

proxy_basic_auth

string

Colon-separated username:password for proxy basic authentication header.

Can also be specified via K8S_AUTH_PROXY_HEADERS_PROXY_BASIC_AUTH environment.

user_agent

string

String representing the user-agent you want, such as foo/1.0.

Can also be specified via K8S_AUTH_PROXY_HEADERS_USER_AGENT environment.

state

string

Determines whether to drain, cordon, or uncordon node.

Choices:

  • "cordon"

  • "drain" ← (default)

  • "uncordon"

username

string

Provide a username for authenticating with the API. Can also be specified via K8S_AUTH_USERNAME environment variable.

Please note that this only works with clusters configured to use HTTP Basic Auth. If your cluster has a different form of authentication (e.g. OAuth2 in OpenShift), this option will not work as expected and you should look into the community.okd.k8s_auth module, as that might do what you need.

validate_certs

aliases: verify_ssl

boolean

Whether or not to verify the API server’s SSL certificates. Can also be specified via K8S_AUTH_VERIFY_SSL environment variable.

Choices:

  • false

  • true

Notes

Note

  • To avoid SSL certificate validation errors when validate_certs is True, the full certificate chain for the API server must be provided via ca_cert or in the kubeconfig file.

Examples

- name: Drain node "foo", even if there are pods not managed by a ReplicationController, Job, or DaemonSet on it.
  kubernetes.core.k8s_drain:
    state: drain
    name: foo
    force: yes

- name: Drain node "foo", but abort if there are pods not managed by a ReplicationController, Job, or DaemonSet, and use a grace period of 15 minutes.
  kubernetes.core.k8s_drain:
    state: drain
    name: foo
    delete_options:
      terminate_grace_period: 900

- name: Mark node "foo" as schedulable.
  kubernetes.core.k8s_drain:
    state: uncordon
    name: foo

- name: Mark node "foo" as unschedulable.
  kubernetes.core.k8s_drain:
    state: cordon
    name: foo

- name: Drain node "foo" using label selector to filter the list of pods to be drained.
  kubernetes.core.k8s_drain:
    state: drain
    name: foo
    pod_selectors:
    - 'app!=csi-attacher'
    - 'app!=csi-provisioner'

Return Values

Common return values are documented here, the following are the fields unique to this module:

Key

Description

result

string

The node status and the number of pods deleted.

Returned: success

Authors

  • Aubin Bikouo (@abikouo)