Skip to content

Latest commit

 

History

History
232 lines (173 loc) · 10.5 KB

installation.md

File metadata and controls

232 lines (173 loc) · 10.5 KB
Raw

documentation / Installation

Installation

The Kyverno policy engine runs as an admission webhook and requires a CA-signed certificate and key to setup secure TLS communication with the kube-apiserver (the CA can be self-signed).

There are 2 ways to configure the secure communications link between Kyverno and the kube-apiserver.

Option 1: Use kube-controller-manager to generate a CA-signed certificate

Kyverno can request a CA signed certificate-key pair from kube-controller-manager. This method requires that the kube-controller-manager is configured to act as a certificate signer. To verify that this option is enabled for your cluster, check the command-line args for the kube-controller-manager. If --cluster-signing-cert-file and --cluster-signing-key-file are passed to the controller manager with paths to your CA's key-pair, then you can proceed to install Kyverno using this method.

Deploying on EKS requires enabling a command-line argument --fqdn-as-cn in the 'kyverno' container in the deployment, due to a current limitation with the certificates returned by EKS for CSR(bug: awslabs/amazon-eks-ami#341)

To install Kyverno in a cluster that supports certificate signing, run the following command on a host with kubectl cluster-admin access:

kubectl create -f https://github.com/nirmata/kyverno/raw/master/definitions/install.yaml

Note that the above command will install the last released (stable) version of Kyverno. If you want to install the latest version, you can edit the install.yaml and update the image tag.

To check the Kyverno controller status, run the command:

kubectl get pods -n kyverno

If the Kyverno controller is not running, you can check its status and logs for errors:

kubectl describe pod <kyverno-pod-name> -n kyverno
kubectl logs <kyverno-pod-name> -n kyverno

Option 2: Use your own CA-signed certificate

You can install your own CA-signed certificate, or generate a self-signed CA and use it to sign a certifcate. Once you have a CA and X.509 certificate-key pair, you can install these as Kubernetes secrets in your cluster. If Kyverno finds these secrets, it uses them. Otherwise it will request the kube-controller-manager to generate a certificate (see Option 1 above).

1. Generate a self-signed CA and signed certificate-key pair

Note: using a separate self-signed root CA is difficult to manage and not recommeded for production use.

If you already have a CA and a signed certificate, you can directly proceed to Step 2.

Here are the commands to create a self-signed root CA, and generate a signed certificate and key using openssl (you can customize the certificate attributes for your deployment):

openssl genrsa -out rootCA.key 4096
openssl req -x509 -new -nodes -key rootCA.key -sha256 -days 1024 -out rootCA.crt  -subj "/C=US/ST=test/L=test /O=test /OU=PIB/CN=*.kyverno.svc/emailAddress=test@test.com"
openssl genrsa -out webhook.key 4096
openssl req -new -key webhook.key -out webhook.csr  -subj "/C=US/ST=test /L=test /O=test /OU=PIB/CN=kyverno-svc.kyverno.svc/emailAddress=test@test.com"
openssl x509 -req -in webhook.csr -CA rootCA.crt -CAkey rootCA.key -CAcreateserial -out webhook.crt -days 1024 -sha256

Among the files that will be generated, you can use the following files to create Kubernetes secrets:

  • rootCA.crt
  • webhooks.crt
  • webhooks.key

2. Configure secrets for the CA and TLS certificate-key pair

To create the required secrets, use the following commands (do not change the secret names):

kubectl create ns kyverno
kubectl -n kyverno create secret tls kyverno-svc.kyverno.svc.kyverno-tls-pair --cert=webhook.crt --key=webhook.key
kubectl annotate secret kyverno-svc.kyverno.svc.kyverno-tls-pair -n kyverno self-signed-cert=true
kubectl -n kyverno create secret generic kyverno-svc.kyverno.svc.kyverno-tls-ca --from-file=rootCA.crt

NOTE: The annotation on the TLS pair secret is used by Kyverno to identify the use of self-signed certificates and checks for the required root CA secret

Secret Data Content
kyverno-svc.kyverno.svc.kyverno-tls-pair rootCA.crt root CA used to sign the certificate
kyverno-svc.kyverno.svc.kyverno-tls-ca tls.key & tls.crt key and signed certificate

Kyverno uses secrets created above to setup TLS communication with the kube-apiserver and specify the CA bundle to be used to validate the webhook server's certificate in the admission webhook configurations.

3. Configure Kyverno Role

Kyverno, in foreground mode, leverages admission webhooks to manage incoming api-requests, and background mode applies the policies on existing resources. It uses ServiceAccount kyverno-service-account, which is bound to multiple ClusterRole, which defines the default resources and operations that are permitted.

ClusterRoles used by kyverno:

  • kyverno:webhook
  • kyverno:userinfo
  • kyverno:customresources
  • kyverno:policycontroller
  • kyverno:generatecontroller

The generate rule creates a new resource, and to allow kyverno to create resource kyverno ClusterRole needs permissions to create/update/delete. This can be done by adding the resource to the ClusterRole kyverno:generatecontroller used by kyverno or by creating a new ClusterRole and a ClusterRoleBinding to kyverno's default ServiceAccount.

apiVersion: rbac.authorization.k8s.io/v1beta1
kind: ClusterRole
metadata:
  name: kyverno:generatecontroller
rules:
- apiGroups:
  - "*"
  resources:
  - namespaces
  - networkpolicies
  - secrets
  - configmaps
  - resourcequotas
  - limitranges
  - ResourceA # new Resource to be generated
  - ResourceB
  verbs:
  - create # generate new resources
  - get # check the contents of exiting resources
  - update # update existing resource, if required configuration defined in policy is not present
  - delete # clean-up, if the generate trigger resource is deleted
kind: ClusterRoleBinding
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
  name: kyverno-admin-generate
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: kyverno:generatecontroller # clusterRole defined above, to manage generated resources
subjects:
- kind: ServiceAccount
  name: kyverno-service-account # default kyverno serviceAccount
  namespace: kyverno

4. Install Kyverno

To install a specific version, download install.yaml and then change the image tag.

e.g., change image tag from latest to the specific tag v1.0.0.

spec:
  containers:
    - name: kyverno
      # image: nirmata/kyverno:latest
      image: nirmata/kyverno:v1.0.0

kubectl create -f ./install.yaml

To check the Kyverno controller status, run the command:

kubectl get pods -n kyverno

If the Kyverno controller is not running, you can check its status and logs for errors:

kubectl describe pod <kyverno-pod-name> -n kyverno
kubectl logs <kyverno-pod-name> -n kyverno

Here is a script that generates a self-signed CA, a TLS certificate-key pair, and the corresponding kubernetes secrets: helper script

Configure a namespace admin to access policy violations

During Kyverno installation, it creates a ClusterRole kyverno:policyviolations which has the list,get,watch operations on resource policyviolations. To grant access to a namespace admin, configure the following YAML file then apply to the cluster.

  • Replace metadata.namespace with namespace of the admin
  • Configure subjects field to bind admin's role to the ClusterRole policyviolation
apiVersion: rbac.authorization.k8s.io/v1beta1
kind: RoleBinding
metadata:
  name: policyviolation
  # change namespace below to create rolebinding for the namespace admin
  namespace: default
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: policyviolation
subjects:
# configure below to access policy violation for the namespace admin
- kind: ServiceAccount
  name: default
  namespace: default
# - apiGroup: rbac.authorization.k8s.io
#   kind: User
#   name: 
# - apiGroup: rbac.authorization.k8s.io
#   kind: Group
#   name:

Installing outside of the cluster (debug mode)

To build Kyverno in a development environment see: https://github.com/nirmata/kyverno/wiki/Building

To run controller in this mode you should prepare a TLS key/certificate pair for debug webhook, then start controller with kubeconfig and the server address.

  1. Run scripts/deploy-controller-debug.sh --service=localhost --serverIP=<server_IP>, where <server_IP> is the IP address of the host where controller runs. This scripts will generate a TLS certificate for debug webhook server and register this webhook in the cluster. It also registers a CustomResource policy.

  2. Start the controller using the following command: sudo kyverno --kubeconfig=~/.kube/config --serverIP=<server_IP>

Filter Kubernetes resources that admission webhook should not process

The admission webhook checks if a policy is applicable on all admission requests. The Kubernetes kinds that are not be processed can be filtered by adding a ConfigMap in namespace kyverno and specifying the resources to be filtered under data.resourceFilters. The default name of this ConfigMap is init-config but can be changed by modifying the value of the environment variable INIT_CONFIG in the kyverno deployment dpec. data.resourceFilters must be a sequence of one or more [<Kind>,<Namespace>,<Name>] entries with * as wildcard. Thus, an item [Node,*,*] means that admissions of Node in any namespace and with any name will be ignored.

By default we have specified Nodes, Events, APIService & SubjectAccessReview as the kinds to be skipped in the default configuration install.yaml.

apiVersion: v1
kind: ConfigMap
metadata:
  name: init-config
  namespace: kyverno
data:
  # resource types to be skipped by kyverno policy engine
  resourceFilters: "[Event,*,*][*,kube-system,*][*,kube-public,*][*,kube-node-lease,*][Node,*,*][APIService,*,*][TokenReview,*,*][SubjectAccessReview,*,*][*,kyverno,*]"

To modify the ConfigMap, either directly edit the ConfigMap init-config in the default configuration install.yaml and redeploy it or modify the ConfigMap use kubectl. Changes to the ConfigMap through kubectl will automatically be picked up at runtime.

Read Next >> Writing Policies