Configure OIDC, make idempotent, fix bugs. Claude.ai

This commit is contained in:
2026-02-25 13:20:12 -05:00
parent 995b7c4070
commit d981b69669
23 changed files with 2269 additions and 760 deletions

View File

@@ -0,0 +1,542 @@
---
# Install cert-manager operator and configure LetsEncrypt certificates.
#
# Installs the Red Hat cert-manager operator via OLM, creates a ClusterIssuer
# for LetsEncrypt with DNS-01 challenges via DNS Made Easy, and provisions
# certificates for the ingress wildcard and API server.
# ------------------------------------------------------------------
# Step 1: Install cert-manager operator via OLM
# ------------------------------------------------------------------
- name: Ensure cert-manager-operator namespace exists
kubernetes.core.k8s:
state: present
definition:
apiVersion: v1
kind: Namespace
metadata:
name: cert-manager-operator
- name: Create OperatorGroup for cert-manager-operator
kubernetes.core.k8s:
state: present
definition:
apiVersion: operators.coreos.com/v1
kind: OperatorGroup
metadata:
name: cert-manager-operator
namespace: cert-manager-operator
spec:
targetNamespaces:
- cert-manager-operator
- name: Subscribe to cert-manager operator
kubernetes.core.k8s:
state: present
definition:
apiVersion: operators.coreos.com/v1alpha1
kind: Subscription
metadata:
name: openshift-cert-manager-operator
namespace: cert-manager-operator
spec:
channel: "{{ sno_deploy_certmanager_channel }}"
installPlanApproval: Automatic
name: openshift-cert-manager-operator
source: "{{ sno_deploy_certmanager_source }}"
sourceNamespace: openshift-marketplace
# ------------------------------------------------------------------
# Step 2: Wait for cert-manager to be ready
# ------------------------------------------------------------------
- name: Wait for cert-manager CRDs to be available
kubernetes.core.k8s_info:
api_version: apiextensions.k8s.io/v1
kind: CustomResourceDefinition
name: certificates.cert-manager.io
register: __sno_deploy_certmanager_crd
until: __sno_deploy_certmanager_crd.resources | length > 0
retries: "{{ (sno_deploy_certmanager_wait_timeout / 10) | int }}"
delay: 10
- name: Wait for cert-manager deployment to be ready
kubernetes.core.k8s_info:
api_version: apps/v1
kind: Deployment
namespace: cert-manager
name: cert-manager
register: __sno_deploy_certmanager_deploy
until: >-
__sno_deploy_certmanager_deploy.resources | length > 0 and
(__sno_deploy_certmanager_deploy.resources[0].status.readyReplicas | default(0)) >= 1
retries: "{{ (sno_deploy_certmanager_wait_timeout / 10) | int }}"
delay: 10
# ------------------------------------------------------------------
# Step 3: Create DNS Made Easy API credentials for DNS-01 challenges
# ------------------------------------------------------------------
- name: Create DNS Made Easy API credentials secret
kubernetes.core.k8s:
state: present
definition:
apiVersion: v1
kind: Secret
metadata:
name: dme-api-credentials
namespace: cert-manager
type: Opaque
stringData:
api-key: "{{ dme_account_key }}"
secret-key: "{{ dme_account_secret }}"
no_log: true
# ------------------------------------------------------------------
# Step 4: Deploy DNS Made Easy webhook solver
# ------------------------------------------------------------------
- name: Create webhook namespace
kubernetes.core.k8s:
state: present
definition:
apiVersion: v1
kind: Namespace
metadata:
name: cert-manager-webhook-dnsmadeeasy
- name: Create webhook ServiceAccount
kubernetes.core.k8s:
state: present
definition:
apiVersion: v1
kind: ServiceAccount
metadata:
name: cert-manager-webhook-dnsmadeeasy
namespace: cert-manager-webhook-dnsmadeeasy
- name: Create webhook ClusterRole
kubernetes.core.k8s:
state: present
definition:
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
name: cert-manager-webhook-dnsmadeeasy
rules:
- apiGroups: [""]
resources: ["secrets"]
verbs: ["get", "list", "watch"]
- apiGroups: ["flowcontrol.apiserver.k8s.io"]
resources: ["flowschemas", "prioritylevelconfigurations"]
verbs: ["list", "watch"]
- name: Create webhook ClusterRoleBinding
kubernetes.core.k8s:
state: present
definition:
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
name: cert-manager-webhook-dnsmadeeasy
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: ClusterRole
name: cert-manager-webhook-dnsmadeeasy
subjects:
- kind: ServiceAccount
name: cert-manager-webhook-dnsmadeeasy
namespace: cert-manager-webhook-dnsmadeeasy
- name: Create auth-delegator ClusterRoleBinding for webhook
kubernetes.core.k8s:
state: present
definition:
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
name: cert-manager-webhook-dnsmadeeasy:auth-delegator
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: ClusterRole
name: system:auth-delegator
subjects:
- kind: ServiceAccount
name: cert-manager-webhook-dnsmadeeasy
namespace: cert-manager-webhook-dnsmadeeasy
- name: Create authentication-reader RoleBinding for webhook
kubernetes.core.k8s:
state: present
definition:
apiVersion: rbac.authorization.k8s.io/v1
kind: RoleBinding
metadata:
name: cert-manager-webhook-dnsmadeeasy:webhook-authentication-reader
namespace: kube-system
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: Role
name: extension-apiserver-authentication-reader
subjects:
- kind: ServiceAccount
name: cert-manager-webhook-dnsmadeeasy
namespace: cert-manager-webhook-dnsmadeeasy
- name: Create domain-solver ClusterRole for cert-manager
kubernetes.core.k8s:
state: present
definition:
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
name: cert-manager-webhook-dnsmadeeasy:domain-solver
rules:
- apiGroups: ["{{ sno_deploy_webhook_group_name }}"]
resources: ["*"]
verbs: ["create"]
- name: Bind domain-solver to cert-manager ServiceAccount
kubernetes.core.k8s:
state: present
definition:
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
name: cert-manager-webhook-dnsmadeeasy:domain-solver
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: ClusterRole
name: cert-manager-webhook-dnsmadeeasy:domain-solver
subjects:
- kind: ServiceAccount
name: cert-manager
namespace: cert-manager
- name: Create self-signed Issuer for webhook TLS
kubernetes.core.k8s:
state: present
definition:
apiVersion: cert-manager.io/v1
kind: Issuer
metadata:
name: cert-manager-webhook-dnsmadeeasy-selfsign
namespace: cert-manager-webhook-dnsmadeeasy
spec:
selfSigned: {}
- name: Create webhook TLS certificate
kubernetes.core.k8s:
state: present
definition:
apiVersion: cert-manager.io/v1
kind: Certificate
metadata:
name: cert-manager-webhook-dnsmadeeasy-tls
namespace: cert-manager-webhook-dnsmadeeasy
spec:
secretName: cert-manager-webhook-dnsmadeeasy-tls
duration: 8760h
renewBefore: 720h
issuerRef:
name: cert-manager-webhook-dnsmadeeasy-selfsign
kind: Issuer
dnsNames:
- cert-manager-webhook-dnsmadeeasy
- cert-manager-webhook-dnsmadeeasy.cert-manager-webhook-dnsmadeeasy
- cert-manager-webhook-dnsmadeeasy.cert-manager-webhook-dnsmadeeasy.svc
- name: Deploy webhook solver
kubernetes.core.k8s:
state: present
definition:
apiVersion: apps/v1
kind: Deployment
metadata:
name: cert-manager-webhook-dnsmadeeasy
namespace: cert-manager-webhook-dnsmadeeasy
spec:
replicas: 1
selector:
matchLabels:
app: cert-manager-webhook-dnsmadeeasy
template:
metadata:
labels:
app: cert-manager-webhook-dnsmadeeasy
spec:
serviceAccountName: cert-manager-webhook-dnsmadeeasy
containers:
- name: webhook
image: "{{ sno_deploy_webhook_image }}"
args:
- --tls-cert-file=/tls/tls.crt
- --tls-private-key-file=/tls/tls.key
- --secure-port=8443
ports:
- containerPort: 8443
name: https
protocol: TCP
env:
- name: GROUP_NAME
value: "{{ sno_deploy_webhook_group_name }}"
livenessProbe:
httpGet:
path: /healthz
port: https
scheme: HTTPS
initialDelaySeconds: 5
periodSeconds: 10
readinessProbe:
httpGet:
path: /healthz
port: https
scheme: HTTPS
initialDelaySeconds: 5
periodSeconds: 10
resources:
requests:
cpu: 10m
memory: 32Mi
limits:
memory: 64Mi
volumeMounts:
- name: certs
mountPath: /tls
readOnly: true
volumes:
- name: certs
secret:
secretName: cert-manager-webhook-dnsmadeeasy-tls
- name: Create webhook Service
kubernetes.core.k8s:
state: present
definition:
apiVersion: v1
kind: Service
metadata:
name: cert-manager-webhook-dnsmadeeasy
namespace: cert-manager-webhook-dnsmadeeasy
spec:
type: ClusterIP
ports:
- port: 443
targetPort: https
protocol: TCP
name: https
selector:
app: cert-manager-webhook-dnsmadeeasy
- name: Register webhook APIService
kubernetes.core.k8s:
state: present
definition:
apiVersion: apiregistration.k8s.io/v1
kind: APIService
metadata:
name: "v1alpha1.{{ sno_deploy_webhook_group_name }}"
spec:
group: "{{ sno_deploy_webhook_group_name }}"
groupPriorityMinimum: 1000
versionPriority: 15
service:
name: cert-manager-webhook-dnsmadeeasy
namespace: cert-manager-webhook-dnsmadeeasy
version: v1alpha1
insecureSkipTLSVerify: true
- name: Wait for webhook deployment to be ready
kubernetes.core.k8s_info:
api_version: apps/v1
kind: Deployment
namespace: cert-manager-webhook-dnsmadeeasy
name: cert-manager-webhook-dnsmadeeasy
register: __sno_deploy_webhook_deploy
until: >-
__sno_deploy_webhook_deploy.resources | length > 0 and
(__sno_deploy_webhook_deploy.resources[0].status.readyReplicas | default(0)) >= 1
retries: 30
delay: 10
# ------------------------------------------------------------------
# Step 5: Create ClusterIssuer for LetsEncrypt
# ------------------------------------------------------------------
- name: Create LetsEncrypt ClusterIssuer
kubernetes.core.k8s:
state: present
definition:
apiVersion: cert-manager.io/v1
kind: ClusterIssuer
metadata:
name: letsencrypt-production
spec:
acme:
email: "{{ sno_deploy_letsencrypt_email }}"
server: "{{ __sno_deploy_letsencrypt_server_url }}"
privateKeySecretRef:
name: letsencrypt-production-account-key
solvers:
- dns01:
webhook:
groupName: "{{ sno_deploy_webhook_group_name }}"
solverName: dnsmadeeasy
config:
apiKeySecretRef:
name: dme-api-credentials
key: api-key
secretKeySecretRef:
name: dme-api-credentials
key: secret-key
- name: Wait for ClusterIssuer to be ready
kubernetes.core.k8s_info:
api_version: cert-manager.io/v1
kind: ClusterIssuer
name: letsencrypt-production
register: __sno_deploy_clusterissuer
until: >-
__sno_deploy_clusterissuer.resources | length > 0 and
(__sno_deploy_clusterissuer.resources[0].status.conditions | default([])
| selectattr('type', '==', 'Ready')
| selectattr('status', '==', 'True') | list | length > 0)
retries: 12
delay: 10
# ------------------------------------------------------------------
# Step 6: Create Certificate resources
# ------------------------------------------------------------------
- name: Create apps wildcard certificate
kubernetes.core.k8s:
state: present
definition:
apiVersion: cert-manager.io/v1
kind: Certificate
metadata:
name: apps-wildcard-cert
namespace: openshift-ingress
spec:
secretName: apps-wildcard-tls
issuerRef:
name: letsencrypt-production
kind: ClusterIssuer
dnsNames:
- "{{ __sno_deploy_apps_wildcard }}"
duration: 2160h
renewBefore: 720h
- name: Create API server certificate
kubernetes.core.k8s:
state: present
definition:
apiVersion: cert-manager.io/v1
kind: Certificate
metadata:
name: api-server-cert
namespace: openshift-config
spec:
secretName: api-server-tls
issuerRef:
name: letsencrypt-production
kind: ClusterIssuer
dnsNames:
- "{{ __sno_deploy_api_hostname }}"
duration: 2160h
renewBefore: 720h
# ------------------------------------------------------------------
# Step 7: Wait for certificates to be issued
# ------------------------------------------------------------------
- name: Wait for apps wildcard certificate to be ready
kubernetes.core.k8s_info:
api_version: cert-manager.io/v1
kind: Certificate
namespace: openshift-ingress
name: apps-wildcard-cert
register: __sno_deploy_apps_cert
until: >-
__sno_deploy_apps_cert.resources | length > 0 and
(__sno_deploy_apps_cert.resources[0].status.conditions | default([])
| selectattr('type', '==', 'Ready')
| selectattr('status', '==', 'True') | list | length > 0)
retries: "{{ (sno_deploy_certificate_wait_timeout / 10) | int }}"
delay: 10
- name: Wait for API server certificate to be ready
kubernetes.core.k8s_info:
api_version: cert-manager.io/v1
kind: Certificate
namespace: openshift-config
name: api-server-cert
register: __sno_deploy_api_cert
until: >-
__sno_deploy_api_cert.resources | length > 0 and
(__sno_deploy_api_cert.resources[0].status.conditions | default([])
| selectattr('type', '==', 'Ready')
| selectattr('status', '==', 'True') | list | length > 0)
retries: "{{ (sno_deploy_certificate_wait_timeout / 10) | int }}"
delay: 10
# ------------------------------------------------------------------
# Step 8: Patch IngressController and APIServer to use the certs
# ------------------------------------------------------------------
- name: Patch default IngressController to use LetsEncrypt cert
kubernetes.core.k8s:
state: present
merge_type: merge
definition:
apiVersion: operator.openshift.io/v1
kind: IngressController
metadata:
name: default
namespace: openshift-ingress-operator
spec:
defaultCertificate:
name: apps-wildcard-tls
- name: Patch APIServer to use LetsEncrypt cert
kubernetes.core.k8s:
state: present
merge_type: merge
definition:
apiVersion: config.openshift.io/v1
kind: APIServer
metadata:
name: cluster
spec:
servingCerts:
namedCertificates:
- names:
- "{{ __sno_deploy_api_hostname }}"
servingCertificate:
name: api-server-tls
# ------------------------------------------------------------------
# Step 9: Wait for rollouts
# ------------------------------------------------------------------
- name: Wait for API server to begin restart
ansible.builtin.pause:
seconds: 30
- name: Wait for router pods to restart with new cert
kubernetes.core.k8s_info:
api_version: apps/v1
kind: Deployment
namespace: openshift-ingress
name: router-default
register: __sno_deploy_router
until: >-
__sno_deploy_router.resources is defined and
__sno_deploy_router.resources | length > 0 and
(__sno_deploy_router.resources[0].status.updatedReplicas | default(0)) ==
(__sno_deploy_router.resources[0].status.replicas | default(1)) and
(__sno_deploy_router.resources[0].status.readyReplicas | default(0)) ==
(__sno_deploy_router.resources[0].status.replicas | default(1))
retries: 60
delay: 10
- name: Display cert-manager configuration summary
ansible.builtin.debug:
msg:
- "cert-manager configuration complete!"
- " ClusterIssuer : letsencrypt-production"
- " Apps wildcard : {{ __sno_deploy_apps_wildcard }}"
- " API cert : {{ __sno_deploy_api_hostname }}"
verbosity: 1

View File

@@ -0,0 +1,145 @@
---
# Configure OpenShift OAuth with Keycloak OIDC.
#
# Prerequisites:
# - SNO cluster installed and accessible
# - Keycloak OIDC client created (Play 5 in deploy_openshift.yml)
# - KUBECONFIG environment variable set or oc_kubeconfig defined
# ------------------------------------------------------------------
# Secret: Keycloak client secret in openshift-config namespace
# ------------------------------------------------------------------
- name: Set OIDC client secret value
ansible.builtin.set_fact:
__sno_deploy_oidc_client_secret_value: >-
{{ hostvars[inventory_hostname]['__oidc_client_secret']
| default(vault_oidc_client_secret) }}
no_log: true
- name: Create Keycloak client secret in openshift-config
kubernetes.core.k8s:
state: present
definition:
apiVersion: v1
kind: Secret
metadata:
name: "{{ __sno_deploy_oidc_secret_name }}"
namespace: openshift-config
type: Opaque
stringData:
clientSecret: "{{ __sno_deploy_oidc_client_secret_value }}"
no_log: false
# ------------------------------------------------------------------
# CA bundle: only needed when Keycloak uses a private/internal CA
# ------------------------------------------------------------------
- name: Create CA bundle ConfigMap for Keycloak TLS
kubernetes.core.k8s:
state: present
definition:
apiVersion: v1
kind: ConfigMap
metadata:
name: "{{ __sno_deploy_oidc_ca_configmap_name }}"
namespace: openshift-config
data:
ca.crt: "{{ lookup('ansible.builtin.file', oidc_ca_cert_file) }}"
when: oidc_ca_cert_file | default('') | length > 0
# ------------------------------------------------------------------
# OAuth cluster resource: add/replace Keycloak IdP entry
# ------------------------------------------------------------------
- name: Get current OAuth cluster configuration
kubernetes.core.k8s_info:
api_version: config.openshift.io/v1
kind: OAuth
name: cluster
register: __sno_deploy_current_oauth
- name: Build Keycloak OIDC identity provider definition
ansible.builtin.set_fact:
__sno_deploy_new_idp: >-
{{
{
'name': oidc_provider_name,
'mappingMethod': 'claim',
'type': 'OpenID',
'openID': (
{
'clientID': oidc_client_id,
'clientSecret': {'name': __sno_deploy_oidc_secret_name},
'issuer': __sno_deploy_oidc_issuer,
'claims': {
'preferredUsername': ['preferred_username'],
'name': ['name'],
'email': ['email'],
'groups': ['groups']
}
} | combine(
(oidc_ca_cert_file | default('') | length > 0) | ternary(
{'ca': {'name': __sno_deploy_oidc_ca_configmap_name}}, {}
)
)
)
}
}}
- name: Build updated identity providers list
ansible.builtin.set_fact:
__sno_deploy_updated_idps: >-
{{
(__sno_deploy_current_oauth.resources[0].spec.identityProviders | default([])
| selectattr('name', '!=', oidc_provider_name) | list)
+ [__sno_deploy_new_idp]
}}
- name: Apply updated OAuth cluster configuration
kubernetes.core.k8s:
state: present
merge_type: merge
definition:
apiVersion: config.openshift.io/v1
kind: OAuth
metadata:
name: cluster
spec:
identityProviders: "{{ __sno_deploy_updated_idps }}"
- name: Wait for OAuth deployment to roll out
ansible.builtin.command:
cmd: "{{ __sno_deploy_oc }} rollout status deployment/oauth-openshift -n openshift-authentication --timeout=300s --insecure-skip-tls-verify"
changed_when: false
# ------------------------------------------------------------------
# ClusterRoleBinding: grant cluster-admin to OIDC admin groups
# ------------------------------------------------------------------
- name: Create ClusterRoleBinding for OIDC admin groups
kubernetes.core.k8s:
state: present
definition:
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
name: "oidc-{{ item | regex_replace('[^a-zA-Z0-9-]', '-') }}-cluster-admin"
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: ClusterRole
name: cluster-admin
subjects:
- apiGroup: rbac.authorization.k8s.io
kind: Group
name: "{{ item }}"
loop: "{{ oidc_admin_groups }}"
when: oidc_admin_groups | length > 0
- name: Display post-configuration summary
ansible.builtin.debug:
msg:
- "OpenShift OIDC configuration complete!"
- " Provider : {{ oidc_provider_name }}"
- " Issuer : {{ __sno_deploy_oidc_issuer }}"
- " Console : https://console-openshift-console.apps.{{ ocp_cluster_name }}.{{ ocp_base_domain }}"
- " Login : https://oauth-openshift.apps.{{ ocp_cluster_name }}.{{ ocp_base_domain }}"
- ""
- "Note: OAuth pods are restarting — login may be unavailable for ~2 minutes."
verbosity: 1

View File

@@ -0,0 +1,82 @@
---
# Create a Proxmox VM for Single Node OpenShift.
# Uses q35 machine type with UEFI (required for SNO / RHCOS).
# An empty ide2 CD-ROM slot is created for the agent installer ISO.
- name: Build net0 string
ansible.builtin.set_fact:
__sno_deploy_net0: >-
virtio{{
'=' + sno_mac if sno_mac | length > 0 else ''
}},bridge={{ sno_bridge }},tag={{ sno_vlan }}
- name: Create SNO VM in Proxmox
community.proxmox.proxmox_kvm:
api_host: "{{ hostvars['proxmox_api']['ansible_host'] }}"
api_user: "{{ proxmox_api_user }}"
api_port: "{{ hostvars['proxmox_api']['ansible_port'] }}"
api_token_id: "{{ proxmox_api_token_id }}"
api_token_secret: "{{ proxmox_api_token_secret }}"
validate_certs: "{{ proxmox_validate_certs }}"
node: "{{ proxmox_node }}"
vmid: "{{ sno_vm_id | default(omit, true) }}"
name: "{{ sno_vm_name }}"
cores: "{{ sno_cpu }}"
memory: "{{ sno_memory_mb }}"
cpu: host
numa_enabled: true
machine: q35
bios: ovmf
efidisk0:
storage: "{{ proxmox_storage }}"
format: raw
efitype: 4m
pre_enrolled_keys: false
scsi:
scsi0: "{{ proxmox_storage }}:{{ sno_disk_gb }},format=raw,iothread=1,cache=writeback"
scsihw: virtio-scsi-single
ide:
ide2: none,media=cdrom
net:
net0: "{{ __sno_deploy_net0 }}"
boot: "order=scsi0;ide2"
onboot: true
state: present
register: __sno_deploy_vm_result
- name: Retrieve VM info
community.proxmox.proxmox_vm_info:
api_host: "{{ hostvars['proxmox_api']['ansible_host'] }}"
api_user: "{{ proxmox_api_user }}"
api_port: "{{ hostvars['proxmox_api']['ansible_port'] }}"
api_token_id: "{{ proxmox_api_token_id }}"
api_token_secret: "{{ proxmox_api_token_secret }}"
validate_certs: "{{ proxmox_validate_certs }}"
node: "{{ proxmox_node }}"
name: "{{ sno_vm_name }}"
type: qemu
config: current
register: __sno_deploy_vm_info
retries: 5
- name: Set VM ID fact for subsequent plays
ansible.builtin.set_fact:
sno_vm_id: "{{ __sno_deploy_vm_info.proxmox_vms[0].vmid }}"
cacheable: true
- name: Extract MAC address from VM config
ansible.builtin.set_fact:
sno_mac: >-
{{ __sno_deploy_vm_info.proxmox_vms[0].config.net0
| regex_search('([0-9A-Fa-f]{2}(?::[0-9A-Fa-f]{2}){5})', '\1')
| first }}
cacheable: true
when: sno_mac | length == 0
- name: Display VM details
ansible.builtin.debug:
msg:
- "VM Name : {{ sno_vm_name }}"
- "VM ID : {{ sno_vm_id }}"
- "MAC : {{ sno_mac }}"
verbosity: 1

View File

@@ -0,0 +1,52 @@
---
# Delete the kubeadmin user after OIDC is configured and admin groups
# have cluster-admin. This is a security best practice.
#
# Safety checks:
# 1. Verify at least one group in oidc_admin_groups is configured
# 2. Verify ClusterRoleBindings exist for those groups
# 3. Verify the OAuth deployment is ready (OIDC login is available)
# 4. Only then delete the kubeadmin secret
- name: Fail if no admin groups are configured
ansible.builtin.fail:
msg: >-
Cannot delete kubeadmin: oidc_admin_groups is empty.
At least one OIDC group must have cluster-admin before kubeadmin can be removed.
when: oidc_admin_groups | length == 0
- name: Verify OIDC admin ClusterRoleBindings exist
kubernetes.core.k8s_info:
api_version: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
name: "oidc-{{ item | regex_replace('[^a-zA-Z0-9-]', '-') }}-cluster-admin"
loop: "{{ oidc_admin_groups }}"
register: __sno_deploy_admin_crbs
failed_when: __sno_deploy_admin_crbs.resources | length == 0
- name: Verify OAuth deployment is ready
kubernetes.core.k8s_info:
api_version: apps/v1
kind: Deployment
namespace: openshift-authentication
name: oauth-openshift
register: __sno_deploy_oauth_status
failed_when: >-
__sno_deploy_oauth_status.resources | length == 0 or
(__sno_deploy_oauth_status.resources[0].status.readyReplicas | default(0)) < 1
- name: Delete kubeadmin secret
kubernetes.core.k8s:
api_version: v1
kind: Secret
namespace: kube-system
name: kubeadmin
state: absent
register: __sno_deploy_kubeadmin_deleted
- name: Display kubeadmin deletion result
ansible.builtin.debug:
msg: >-
{{ 'kubeadmin user deleted successfully. Login is now only available via OIDC.'
if __sno_deploy_kubeadmin_deleted.changed
else 'kubeadmin was already deleted.' }}

View File

@@ -0,0 +1,389 @@
---
# Generate Agent ISO and deploy SNO (agent-based installer).
#
# Uses `openshift-install agent create image` — no SaaS API, no SSO required.
# The pull secret is the only Red Hat credential needed.
# Credentials (kubeconfig, kubeadmin-password) are generated locally under
# sno_install_dir/auth/ by openshift-install itself.
#
# Idempotency: If the cluster API is already responding, all install steps
# are skipped. Credentials on Proxmox host are never overwritten once saved.
# ------------------------------------------------------------------
# Step 0: Ensure sno_vm_id and sno_mac are populated.
# These are set as cacheable facts by create_vm.yml, but in ephemeral
# EEs or when running --tags sno_deploy_install alone the cache is empty.
# ------------------------------------------------------------------
- name: Retrieve VM info from Proxmox (needed when fact cache is empty)
community.proxmox.proxmox_vm_info:
api_host: "{{ hostvars['proxmox_api']['ansible_host'] }}"
api_user: "{{ proxmox_api_user }}"
api_port: "{{ hostvars['proxmox_api']['ansible_port'] }}"
api_token_id: "{{ proxmox_api_token_id }}"
api_token_secret: "{{ proxmox_api_token_secret }}"
validate_certs: "{{ proxmox_validate_certs }}"
node: "{{ proxmox_node }}"
name: "{{ sno_vm_name }}"
type: qemu
config: current
register: __sno_deploy_vm_info
when: (sno_vm_id | default('')) == '' or (sno_mac | default('')) == ''
- name: Set sno_vm_id and sno_mac from live Proxmox query
ansible.builtin.set_fact:
sno_vm_id: "{{ __sno_deploy_vm_info.proxmox_vms[0].vmid }}"
sno_mac: >-
{{ __sno_deploy_vm_info.proxmox_vms[0].config.net0
| regex_search('([0-9A-Fa-f]{2}(?::[0-9A-Fa-f]{2}){5})', '\1')
| first }}
cacheable: true
when: __sno_deploy_vm_info is not skipped
# ------------------------------------------------------------------
# Step 0b: Check if OpenShift is already deployed and responding.
# If the API is reachable, skip ISO generation, boot, and install.
# ------------------------------------------------------------------
- name: Check if OpenShift cluster is already responding
ansible.builtin.uri:
url: "https://api.{{ ocp_cluster_name }}.{{ ocp_base_domain }}:6443/readyz"
method: GET
validate_certs: false
status_code: [200, 401, 403]
timeout: 10
register: __sno_deploy_cluster_alive
ignore_errors: true
- name: Set cluster deployed flag
ansible.builtin.set_fact:
__sno_deploy_cluster_deployed: "{{ __sno_deploy_cluster_alive is success }}"
- name: Display cluster status
ansible.builtin.debug:
msg: >-
{{ 'OpenShift cluster is already deployed and responding — skipping install steps.'
if __sno_deploy_cluster_deployed | bool
else 'OpenShift cluster is not yet deployed — proceeding with installation.' }}
- name: Ensure local install directories exist
ansible.builtin.file:
path: "{{ item }}"
state: directory
mode: "0750"
loop:
- "{{ sno_install_dir }}"
- "{{ sno_install_dir }}/auth"
# ------------------------------------------------------------------
# Step 0c: When cluster is already deployed, ensure a valid kubeconfig
# exists so post-install tasks can authenticate to the API.
# Try in order: local file → Proxmox host backup → SSH to SNO node.
# After obtaining a kubeconfig, validate it against the API and fall
# through to the next source if credentials are expired.
# ------------------------------------------------------------------
- name: Check if local kubeconfig already exists
ansible.builtin.stat:
path: "{{ __sno_deploy_kubeconfig }}"
register: __sno_deploy_local_kubeconfig
when: __sno_deploy_cluster_deployed | bool
- name: Validate local kubeconfig against API
ansible.builtin.command:
cmd: "oc whoami --kubeconfig={{ __sno_deploy_kubeconfig }} --insecure-skip-tls-verify"
register: __sno_deploy_local_kubeconfig_valid
ignore_errors: true
changed_when: false
when:
- __sno_deploy_cluster_deployed | bool
- __sno_deploy_local_kubeconfig.stat.exists | default(false)
- name: Check if kubeconfig exists on Proxmox host
ansible.builtin.stat:
path: "{{ sno_credentials_dir }}/kubeconfig"
delegate_to: proxmox_host
register: __sno_deploy_proxmox_kubeconfig
when:
- __sno_deploy_cluster_deployed | bool
- not (__sno_deploy_local_kubeconfig.stat.exists | default(false)) or
(__sno_deploy_local_kubeconfig_valid is failed)
- name: Recover kubeconfig from Proxmox host
ansible.builtin.fetch:
src: "{{ sno_credentials_dir }}/kubeconfig"
dest: "{{ __sno_deploy_kubeconfig }}"
flat: true
delegate_to: proxmox_host
when:
- __sno_deploy_cluster_deployed | bool
- not (__sno_deploy_local_kubeconfig.stat.exists | default(false)) or
(__sno_deploy_local_kubeconfig_valid is failed)
- __sno_deploy_proxmox_kubeconfig.stat.exists | default(false)
- name: Validate recovered Proxmox kubeconfig against API
ansible.builtin.command:
cmd: "oc whoami --kubeconfig={{ __sno_deploy_kubeconfig }} --insecure-skip-tls-verify"
register: __sno_deploy_proxmox_kubeconfig_valid
ignore_errors: true
changed_when: false
when:
- __sno_deploy_cluster_deployed | bool
- not (__sno_deploy_local_kubeconfig.stat.exists | default(false)) or
(__sno_deploy_local_kubeconfig_valid is failed)
- __sno_deploy_proxmox_kubeconfig.stat.exists | default(false)
- name: Set flag - need SSH recovery
ansible.builtin.set_fact:
__sno_deploy_need_ssh_recovery: >-
{{
(__sno_deploy_cluster_deployed | bool) and
(
(not (__sno_deploy_local_kubeconfig.stat.exists | default(false)) and
not (__sno_deploy_proxmox_kubeconfig.stat.exists | default(false)))
or
((__sno_deploy_local_kubeconfig_valid | default({})) is failed and
(__sno_deploy_proxmox_kubeconfig_valid | default({})) is failed)
or
(not (__sno_deploy_local_kubeconfig.stat.exists | default(false)) and
(__sno_deploy_proxmox_kubeconfig_valid | default({})) is failed)
)
}}
- name: Recover kubeconfig from SNO node via SSH
ansible.builtin.command:
cmd: >-
ssh -o StrictHostKeyChecking=no core@{{ sno_ip }}
sudo cat /etc/kubernetes/static-pod-resources/kube-apiserver-certs/secrets/node-kubeconfigs/lb-ext.kubeconfig
register: __sno_deploy_recovered_kubeconfig
when: __sno_deploy_need_ssh_recovery | bool
- name: Write recovered kubeconfig from SNO node
ansible.builtin.copy:
content: "{{ __sno_deploy_recovered_kubeconfig.stdout }}"
dest: "{{ __sno_deploy_kubeconfig }}"
mode: "0600"
when:
- __sno_deploy_recovered_kubeconfig is not skipped
- __sno_deploy_recovered_kubeconfig.rc == 0
- name: Update kubeconfig backup on Proxmox host
ansible.builtin.copy:
src: "{{ __sno_deploy_kubeconfig }}"
dest: "{{ sno_credentials_dir }}/kubeconfig"
mode: "0600"
backup: true
delegate_to: proxmox_host
when:
- __sno_deploy_recovered_kubeconfig is not skipped
- __sno_deploy_recovered_kubeconfig.rc == 0
- name: Fail if no valid kubeconfig could be obtained
ansible.builtin.fail:
msg: >-
Cluster is deployed but no valid kubeconfig could be obtained.
Tried: local file, Proxmox host ({{ sno_credentials_dir }}/kubeconfig),
and SSH to core@{{ sno_ip }}. Cannot proceed with post-install tasks.
when:
- __sno_deploy_need_ssh_recovery | bool
- __sno_deploy_recovered_kubeconfig is skipped or __sno_deploy_recovered_kubeconfig.rc != 0
# ------------------------------------------------------------------
# Step 1: Check whether a fresh ISO already exists on Proxmox
# AND the local openshift-install state dir is intact.
# ------------------------------------------------------------------
- name: Check if ISO already exists on Proxmox and is less than 24 hours old
ansible.builtin.stat:
path: "{{ proxmox_iso_dir }}/{{ sno_iso_filename }}"
get_checksum: false
delegate_to: proxmox_host
register: __sno_deploy_iso_stat
when: not __sno_deploy_cluster_deployed | bool
- name: Check if local openshift-install state directory exists
ansible.builtin.stat:
path: "{{ sno_install_dir }}/.openshift_install_state"
get_checksum: false
register: __sno_deploy_state_stat
when: not __sno_deploy_cluster_deployed | bool
- name: Set fact - skip ISO build if recent ISO exists on Proxmox and local state is intact
ansible.builtin.set_fact:
__sno_deploy_iso_fresh: >-
{{
not (__sno_deploy_cluster_deployed | bool) and
__sno_deploy_iso_stat.stat.exists | default(false) and
(now(utc=true).timestamp() | int - __sno_deploy_iso_stat.stat.mtime | default(0) | int) < 86400 and
__sno_deploy_state_stat.stat.exists | default(false)
}}
# ------------------------------------------------------------------
# Step 2: Get openshift-install binary
# Always ensure the binary is present — needed for both ISO generation
# and wait-for-install-complete regardless of __sno_deploy_iso_fresh.
# ------------------------------------------------------------------
- name: Download openshift-install tarball
ansible.builtin.get_url:
url: "https://mirror.openshift.com/pub/openshift-v4/clients/ocp/stable-{{ ocp_version }}/openshift-install-linux.tar.gz"
dest: "{{ sno_install_dir }}/openshift-install-{{ ocp_version }}.tar.gz"
mode: "0644"
checksum: "{{ ocp_install_checksum | default(omit) }}"
register: __sno_deploy_install_tarball
when: not __sno_deploy_cluster_deployed | bool
- name: Extract openshift-install binary
ansible.builtin.unarchive:
src: "{{ sno_install_dir }}/openshift-install-{{ ocp_version }}.tar.gz"
dest: "{{ sno_install_dir }}"
remote_src: false
include:
- openshift-install
when: not __sno_deploy_cluster_deployed | bool and (__sno_deploy_install_tarball.changed or not (sno_install_dir ~ '/openshift-install') is file)
- name: Download openshift-client tarball
ansible.builtin.get_url:
url: "https://mirror.openshift.com/pub/openshift-v4/clients/ocp/stable-{{ ocp_version }}/openshift-client-linux.tar.gz"
dest: "{{ sno_install_dir }}/openshift-client-{{ ocp_version }}.tar.gz"
mode: "0644"
checksum: "{{ ocp_client_checksum | default(omit) }}"
register: __sno_deploy_client_tarball
when: not __sno_deploy_cluster_deployed | bool
- name: Extract oc binary
ansible.builtin.unarchive:
src: "{{ sno_install_dir }}/openshift-client-{{ ocp_version }}.tar.gz"
dest: "{{ sno_install_dir }}"
remote_src: false
include:
- oc
when: not __sno_deploy_cluster_deployed | bool and (__sno_deploy_client_tarball.changed or not (sno_install_dir ~ '/oc') is file)
# ------------------------------------------------------------------
# Step 3: Template agent installer config files (skipped if ISO is fresh)
# ------------------------------------------------------------------
- name: Template install-config.yaml
ansible.builtin.template:
src: install-config.yaml.j2
dest: "{{ sno_install_dir }}/install-config.yaml"
mode: "0640"
when: not __sno_deploy_cluster_deployed | bool and not __sno_deploy_iso_fresh | bool
no_log: true
- name: Template agent-config.yaml
ansible.builtin.template:
src: agent-config.yaml.j2
dest: "{{ sno_install_dir }}/agent-config.yaml"
mode: "0640"
when: not __sno_deploy_cluster_deployed | bool and not __sno_deploy_iso_fresh | bool
# ------------------------------------------------------------------
# Step 4: Generate discovery ISO (skipped if ISO is fresh)
# ------------------------------------------------------------------
- name: Generate agent-based installer ISO
ansible.builtin.command:
cmd: "{{ sno_install_dir }}/openshift-install agent create image --dir {{ sno_install_dir }}"
when: not __sno_deploy_cluster_deployed | bool and not __sno_deploy_iso_fresh | bool
# ------------------------------------------------------------------
# Step 5: Upload ISO to Proxmox and attach to VM
# ------------------------------------------------------------------
- name: Copy discovery ISO to Proxmox ISO storage
ansible.builtin.copy:
src: "{{ sno_install_dir }}/{{ sno_iso_filename }}"
dest: "{{ proxmox_iso_dir }}/{{ sno_iso_filename }}"
mode: "0644"
delegate_to: proxmox_host
when: not __sno_deploy_cluster_deployed | bool and not __sno_deploy_iso_fresh | bool
- name: Attach ISO to VM as CDROM
ansible.builtin.command:
cmd: "qm set {{ sno_vm_id }} --ide2 {{ proxmox_iso_storage }}:iso/{{ sno_iso_filename }},media=cdrom"
delegate_to: proxmox_host
changed_when: true
when: not __sno_deploy_cluster_deployed | bool
- name: Ensure boot order prefers disk, falls back to CDROM
ansible.builtin.command:
cmd: "qm set {{ sno_vm_id }} --boot order=scsi0;ide2"
delegate_to: proxmox_host
changed_when: true
when: not __sno_deploy_cluster_deployed | bool
# ------------------------------------------------------------------
# Step 6: Boot the VM
# ------------------------------------------------------------------
- name: Start SNO VM
community.proxmox.proxmox_kvm:
api_host: "{{ hostvars['proxmox_api']['ansible_host'] }}"
api_user: "{{ proxmox_api_user }}"
api_port: "{{ hostvars['proxmox_api']['ansible_port'] }}"
api_token_id: "{{ proxmox_api_token_id }}"
api_token_secret: "{{ proxmox_api_token_secret }}"
validate_certs: "{{ proxmox_validate_certs }}"
node: "{{ proxmox_node }}"
name: "{{ sno_vm_name }}"
state: started
when: not __sno_deploy_cluster_deployed | bool
# ------------------------------------------------------------------
# Step 7: Wait for installation to complete (~60-90 min)
# ------------------------------------------------------------------
- name: Wait for SNO installation to complete
ansible.builtin.command:
cmd: "{{ sno_install_dir }}/openshift-install agent wait-for install-complete --dir {{ sno_install_dir }} --log-level=info"
async: 5400
poll: 30
when: not __sno_deploy_cluster_deployed | bool
# ------------------------------------------------------------------
# Step 8: Persist credentials to Proxmox host
# Only copy if credentials do not already exist on the remote host,
# to prevent overwriting valid credentials on re-runs.
# ------------------------------------------------------------------
- name: Create credentials directory on Proxmox host
ansible.builtin.file:
path: "{{ sno_credentials_dir }}"
state: directory
mode: "0700"
delegate_to: proxmox_host
- name: Check if credentials already exist on Proxmox host
ansible.builtin.stat:
path: "{{ sno_credentials_dir }}/kubeadmin-password"
delegate_to: proxmox_host
register: __sno_deploy_remote_creds
- name: Copy kubeconfig to Proxmox host
ansible.builtin.copy:
src: "{{ sno_install_dir }}/auth/kubeconfig"
dest: "{{ sno_credentials_dir }}/kubeconfig"
mode: "0600"
backup: true
delegate_to: proxmox_host
when: not __sno_deploy_remote_creds.stat.exists
- name: Copy kubeadmin-password to Proxmox host
ansible.builtin.copy:
src: "{{ sno_install_dir }}/auth/kubeadmin-password"
dest: "{{ sno_credentials_dir }}/kubeadmin-password"
mode: "0600"
backup: true
delegate_to: proxmox_host
when: not __sno_deploy_remote_creds.stat.exists
# ------------------------------------------------------------------
# Step 9: Eject CDROM so the VM never boots the agent ISO again
# ------------------------------------------------------------------
- name: Eject CDROM after successful installation
ansible.builtin.command:
cmd: "qm set {{ sno_vm_id }} --ide2 none,media=cdrom"
delegate_to: proxmox_host
changed_when: true
when: not __sno_deploy_cluster_deployed | bool
- name: Display post-install info
ansible.builtin.debug:
msg:
- "SNO installation complete!"
- "API URL : https://api.{{ ocp_cluster_name }}.{{ ocp_base_domain }}:6443"
- "Console : https://console-openshift-console.apps.{{ ocp_cluster_name }}.{{ ocp_base_domain }}"
- "Kubeconfig : {{ sno_credentials_dir }}/kubeconfig (on proxmox_host)"
- "kubeadmin pass : {{ sno_credentials_dir }}/kubeadmin-password (on proxmox_host)"
verbosity: 1

View File

@@ -0,0 +1,41 @@
---
# Entry point for the sno_deploy role.
#
# Each phase is gated by tags so individual steps can be run with --tags.
# When invoked from deploy_openshift.yml, individual task files are
# called directly via include_role + tasks_from to control play ordering.
- name: Create SNO VM in Proxmox
ansible.builtin.include_tasks:
file: create_vm.yml
apply:
tags: sno_deploy_vm
tags: sno_deploy_vm
- name: Install SNO via agent-based installer
ansible.builtin.include_tasks:
file: install.yml
apply:
tags: sno_deploy_install
tags: sno_deploy_install
- name: Configure OpenShift OAuth with OIDC
ansible.builtin.include_tasks:
file: configure_oidc.yml
apply:
tags: sno_deploy_oidc
tags: sno_deploy_oidc
- name: Configure cert-manager and LetsEncrypt certificates
ansible.builtin.include_tasks:
file: configure_certmanager.yml
apply:
tags: sno_deploy_certmanager
tags: sno_deploy_certmanager
- name: Delete kubeadmin user
ansible.builtin.include_tasks:
file: delete_kubeadmin.yml
apply:
tags: sno_deploy_delete_kubeadmin
tags: sno_deploy_delete_kubeadmin