website/docs/admin/kubelet-tls-bootstrapping.md

assignees, title

assignees	title
ericchiang mikedanese jcbsmpsn	TLS bootstrapping

• TOC {:toc}

Overview

This document describes how to set up TLS client certificate bootstrapping for kubelets. Kubernetes 1.4 introduced an API for requesting certificates from a cluster-level Certificate Authority (CA). The original intent of this API is to enable provisioning of TLS client certificates for kubelets. The proposal can be found here and progress on the feature is being tracked as feature #43.

kube-apiserver configuration

You must provide a token file which specifies at least one "bootstrap token" assigned to a kubelet bootstrap-specific group. This group will later be used in the controller-manager configuration to scope approvals in the default approval controller. As this feature matures, you should ensure tokens are bound to a Role-Based Access Control (RBAC) policy which limits requests (using the bootstrap token) strictly to client requests related to certificate provisioning. With RBAC in place, scoping the tokens to a group allows for great flexibility (e.g. you could disable a particular bootstrap group's access when you are done provisioning the nodes).

Token authentication file

Tokens are arbitrary but should represent at least 128 bits of entropy derived from a secure random number generator (such as /dev/urandom on most modern systems). There are multiple ways you can generate a token. For example:

head -c 16 /dev/urandom | od -An -t x | tr -d ' '

will generate tokens that look like 02b50b05283e98dd0fd71db496ef01e8

The token file should look like the following example, where the first three values can be anything and the quoted group name should be as depicted:

02b50b05283e98dd0fd71db496ef01e8,kubelet-bootstrap,10001,"system:kubelet-bootstrap"

Add the --token-auth-file=FILENAME flag to the kube-apiserver command (in your systemd unit file perhaps) to enable the token file. See docs here for further details.

Client certificate CA bundle

Add the --client-ca-file=FILENAME flag to the kube-apiserver command to enable client certificate authentication, referencing a certificate authority bundle containing the signing certificate (e.g. --client-ca-file=/var/lib/kubernetes/ca.pem).

kube-controller-manager configuration

The API for requesting certificates adds a certificate-issuing control loop to the Kubernetes Controller Manager. This takes the form of a cfssl local signer using assets on disk. Currently, all certificates issued have one year validity and a default set of key usages.

Signing assets

You must provide a Certificate Authority in order to provide the cryptographic materials necessary to issue certificates. This CA should be trusted by kube-apiserver for authentication with the --client-ca-file=FILENAME flag. The management of the CA is beyond the scope of this document but it is recommended that you generate a dedicated CA for Kubernetes. Both certificate and key are assumed to be PEM-encoded.

The kube-controller-manager flags are:

--cluster-signing-cert-file="/etc/path/to/kubernetes/ca/ca.crt" --cluster-signing-key-file="/etc/path/to/kubernetes/ca/ca.key"

Approval controller

In 1.7 the experimental "group auto approver" controller is dropped in favor of the new csrapproving controller that ships as part of kube-controller-manager and is enabled by default. The controller uses the SubjectAccessReview API to determine if a given user is authorized to request a CSR, then approves based on the authorization outcome. To prevent conflicts with other approvers, the builtin approver doesn't explicitly deny CSRs, only ignoring unauthorized requests.

The controller categorizes CSRs into three subresources:

1. nodeclient - a request by a user for a client certificate with O=system:nodes and CN=system:node:(node name).
2. selfnodeclient - a node renewing a client certificate with the same O and CN.
3. selfnodeserver - a node renewing a serving certificate. (ALPHA, requires feature gate)

The checks to determine if a CSR is a selfnodeserver request is currently tied to the kubelet's credential rotation implementation, an alpha feature. As such, the definition of selfnodeserver will likely change in a future and requires the RotateKubeletServerCertificate feature gate on the controller manager. The feature progress can be tracked at kubernetes/features#267.

--feature-gates=RotateKubeletServerCertificate=true

The following RBAC ClusterRoles represent the nodeclient, selfnodeclient, and selfnodeserver capabilities. Similar roles may be automatically created in future releases.

# A ClusterRole which instructs the CSR approver to approve a user requesting
# node client credentials.
kind: ClusterRole
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
  name: approve-node-client-csr
rules:
- apiGroups: ["certificates.k8s.io"]
  resources: ["certificatesigningrequests/nodeclient"]
  verbs: ["create"]
---
# A ClusterRole which instructs the CSR approver to approve a node renewing its
# own client credentials.
kind: ClusterRole
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
  name: approve-node-client-renewal-csr
rules:
- apiGroups: ["certificates.k8s.io"]
  resources: ["certificatesigningrequests/selfnodeclient"]
  verbs: ["create"]
---
# A ClusterRole which instructs the CSR approver to approve a node requesting a
# serving cert matching its client cert.
kind: ClusterRole
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
  name: approve-node-server-renewal-csr
rules:
- apiGroups: ["certificates.k8s.io"]
  resources: ["certificatesigningrequests/selfnodeserver"]
  verbs: ["create"]

These powers can be granted to credentials, such as bootstrapping tokens. For example, to replicate the behavior provided by the removed auto-approval flag, of approving all CSRs by a single group:

# REMOVED: This flag no longer works as of 1.7.
--insecure-experimental-approve-all-kubelet-csrs-for-group="kubelet-bootstrap-token"

An admin would create a ClusterRoleBinding targeting that group.

# Approve all CSRs for the group "kubelet-bootstrap-token"
kind: ClusterRoleBinding
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
  name: auto-approve-csrs-for-group
subjects:
- kind: Group
  name: kubelet-bootstrap-token
  apiGroup: rbac.authorization.k8s.io
roleRef:
  kind: ClusterRole
  name: approve-node-client-csr
  apiGroup: rbac.authorization.k8s.io

To let a node renew its own credentials, an admin can construct a ClusterRoleBinding targeting that node's credentials:

kind: ClusterRoleBinding
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
  name: node1-client-cert-renewal
subjects:
- kind: User
  name: system:node:node-1 # Let "node-1" renew its client certificate.
  apiGroup: rbac.authorization.k8s.io
roleRef:
  kind: ClusterRole
  name: approve-node-client-renewal-csr
  apiGroup: rbac.authorization.k8s.io

Deleting the binding will prevent the node from renewing its client credentials, effectively removing it from the cluster once its certificate expires.

kubelet configuration

To request a client certificate from kube-apiserver, the kubelet first needs a path to a kubeconfig file that contains the bootstrap authentication token. You can use kubectl config set-cluster, set-credentials, and set-context to build this kubeconfig. Provide the name kubelet-bootstrap to kubectl config set-credentials and include --token=<token-value> as follows:

kubectl config set-credentials kubelet-bootstrap --token=${BOOTSTRAP_TOKEN} --kubeconfig=bootstrap.kubeconfig

When starting the kubelet, if the file specified by --kubeconfig does not exist, the bootstrap kubeconfig is used to request a client certificate from the API server. On approval of the certificate request and receipt back by the kubelet, a kubeconfig file referencing the generated key and obtained certificate is written to the path specified by --kubeconfig. The certificate and key file will be placed in the directory specified by --cert-dir.

The flag to enable this bootstrapping when starting the kubelet is:

--experimental-bootstrap-kubeconfig="/path/to/bootstrap/kubeconfig"

Additionally, in 1.7 the kubelet implements alpha features for enabling rotation of both its client and/or serving certs. These can be enabled through the respective RotateKubeletClientCertificate and RotateKubeletServerCertificate feature flags on the kubelet, but may change in backward incompatible ways in future releases.

--feature-gates=RotateKubeletClientCertificate=true,RotateKubeletServerCertificate=true

RotateKubeletClientCertificate causes the kubelet to rotate its client certificates by creating new CSRs as its existing credentials expire. RotateKubeletServerCertificate causes the kubelet to both request a serving certificate after bootstrapping its client credentials and rotate the certificate. The serving cert currently does not request DNS or IP SANs.

kubectl approval

The signing controller does not immediately sign all certificate requests. Instead, it waits until they have been flagged with an "Approved" status by an appropriately-privileged user. This is intended to eventually be an automated process handled by an external approval controller, but for the alpha version of the API it can be done manually by a cluster administrator using kubectl. An administrator can list CSRs with kubectl get csr and describe one in detail with kubectl describe csr <name>. Before the 1.6 release there were no direct approve/deny commands so an approver had to update the Status field directly (rough how-to). Later versions of Kubernetes offer kubectl certificate approve <name> and kubectl certificate deny <name> commands.