bennu / terraform-kubernetes-yagan

Yagan kubernetes

Requirements

Name	Version
terraform	>= 1.3.7

Providers

Name	Version
rke	1.3.4
helm	2.8.0
kubernetes	2.16.1
local	2.3.0
random	3.4.3

Usage

module "cluster" {
  source = "[email protected]:bennu/terraform-kubernetes-yagan.git"

  private_key           = file("/path/to/privatekey.pem")
  node_user             = "root"
  nodes = {
    node-name-1 = [{
      ip     = "1.1.1.1"
      type   = ["controlplane", "etcd"]
      labels = {}
      taints = []
    }],
    node-name-2 = [{
      ip     = "2.2.2.2"
      type   = ["worker"]
      labels = {}
      taints = []
    }]
  }
  # Choose one of the above CNI to install.
  install_calico           = false
  cluster_cidr             = "10.42.0.0/16"
  service_cluster_ip_range = "10.43.0.0/16"
}

Some considerations

always_pull_images: Enable always pull images admission controler in the api-server

nodes: A map of objects containing a list of node names and a IPs for each type (See: yagan byoi example

Inputs

Name	Description	Type	Default	Required
support_version	Supported version of rke kubernetes : v1.24.4-rancher1-1 - v1.19.16-rancher2-1	string	v1.24.4-rancher1-1	yes
nodes	A map of objects containing a list of node names and a IPs for each type	any	n/a	yes
private_key	Default private ssh key for nodes	any	n/a	yes
addon_job_timeout	Timeout for addons deployment in seconds	number	120	no
addons_include	URLs and/or local files to deploy withing RKE bootstrapping process	list	[]	no
always_pull_images	Enable always pull images admission controler	bool	true	no
api_server_lb	List of IPs on loadbalancer in front of kube-api-sever(s)	list	[]	no
cgroup_driver	Driver that the kubelet uses to manipulate cgroups on the host	string	"cgroupfs"	no
cloud_provider	Cloud provider to deploy	string	"none"	no
cloud_provider_vsphere_in_tree	vSphere Cloud Provider in-tree configuration, list of maps	list(map(string))	[]	no
cluster_cidr	Cluster CIDR for pods IP allocation	string	"10.42.0.0/16"	no
cluster_domain	Domain for cluster-wide service discovery	string	"cluster.local"	no
delete_local_data_on_drain	Delete local data on node drain	bool	true	no
dns_provider	Cluster DNS service provider	string	"coredns"	no
drain_grace_period	Grace period to wait for node to drain	string	"-1"	no
drain_on_upgrade	Do drain operations on upgrades	bool	true	no
drain_timeout	Time to wait for node to drain	number	60	no
enable_cri_dockerd	Enable/Disable CRI dockerd for kubelet (Required on K8s v1.24+)	bool	true	no
enforce_node_allocatable	Enforce allocatable resources	string	"pods,system-reserved,kube-reserved"	no
etcd_backup_interval_hours	Interval hours for etcd backups	number	8	no
etcd_backup_retention	Amount of backups to keep in parallel	number	6	no
etcd_extra_args	A map of extra args for etcd	map	{}	no
etcd_extra_binds	A list of host volumes to bind to etcd	list	[]	no
etcd_extra_env	A list of env vars to prepend to etcd	list	[]	no
etcd_s3_access_key	S3 account access key for storing etcd backups	string	""	no
etcd_s3_bucket_name	S3 bucket for storing etcd backups	string	""	no
etcd_s3_endpoint	Endpoint for S3 and S3 compatible services for storing etcd backups	string	""	no
etcd_s3_folder	S3 folder for storing etcd backups	string	""	no
etcd_s3_region	S3 region for storing etcd backups	string	"us-east-1"	no
etcd_s3_secret_key	S3 account secret for storing etcd backups	string	""	no
eviction_hard	Params for eviction	string	"memory.available<15%,nodefs.available<10%,nodefs.inodesFree<5%,imagefs.available<15%,imagefs.inodesFree<20%"	no
fail_swap_on	Do not allow to deploy kubernetes on systems with swap partitions enabled	bool	true	no
force_drain	Force drain on upgrades	bool	true	no
generate_serving_certificate	Generate serving certificate	bool	true	no
hubble_enabled	Enable hubble	bool	true	no
hubble_metrics	Metrics to be fetched by hubble	string	"{dns,drop,tcp,flow,port-distribution,icmp,http}"	no
hubble_relay_enabled	Enable hubble releay	bool	true	no
hubble_ui_enabled	Enable hubble UI	bool	true	no
ignore_daemon_sets_on_drain	Drain despite of daemonset	bool	true	no
ignore_docker_version	Do not check docker version when deploying RKE	bool	true	no
ingress_provider	Deploy RKE built-in ingress controller	string	"none"	no
install_argocd	Decides if Argo CD operator must be installed after the cluster is deployed	bool	false	no
install_calico	Decides if Calico CNI must be installed	bool	false	no
kube_api_extra_binds	A list of host volumes to bind to api-server	list	[]	no
kube_api_extra_env	A list of env vars to prepend to api-server	list	[]	no
kube_controller_extra_args	A map of extra args for controller	map	{}	no
kube_controller_extra_binds	A list of host volumes to bind to controller	list	[]	no
kube_controller_extra_env	A list of env vars to prepend to controller	list	[]	no
kube_reserved	Resources reserved for kubernetes pods	string	"cpu=300m,memory=500Mi"	no
kube_reserved_cgroup	Cgroup for kubernetes pods	string	"/podruntime.slice"	no
kubelet_extra_args	A map of extra args for kubelet	map	{}	no
kubelet_extra_binds	A list of host volumes to bind to kubelet	list	[]	no
kubelet_extra_env	A list of env vars to prepend to kubelet	list	[]	no
kubeproxy_extra_args	A map of extra args for kube-proxy	map	{}	no
kubeproxy_extra_binds	A list of host volumes to bind to kube-proxy	list	[]	no
kubeproxy_extra_env	A list of env vars to prepend to kube-proxy	list	[]	no
kubernetes_version	RKE version to deploy	string	""	no
max_pods	Max ammount of pods to deploy per node	number	32	no
monitoring	Monitoring service for kubernetes	string	"metrics-server"	no
node_cidr_mask_size	Mask size to assign to each node based on cluster_cidr	number	26	no
node_monitor_grace_period	Grace period for node monitoring	string	"15s"	no
node_monitor_period	Period time for node monitoring	string	"2s"	no
node_status_update_frequency	Frequency to report node status to api-server	string	"4s"	no
node_user	Default user to connect to nodes as	string	"sles"	no
pod_eviction_timeout	n/a	string	"30s"	no
pod_security_policy	Deploy a permissive default set of PSP	bool	false	no
registry_activate	Able to activate registry server	bool	false	no
registry_password	Password access for Registry server	string	""	no
registry_url	Registry URL for images	string	""	no
registry_username	Username access for Registry server	string	""	no
resource_naming	An arbitrary name can be prepend to resources. If not set, a random prefix will be created instead	string	""	no
rke_authorization	RKE authorization mode	string	"rbac"	no
sans	An alternative subject alternate name (SAN) list for api-server tls certs	list	[]	no
scheduler_extra_args	A map of extra args for scheduler	map	{}	no
scheduler_extra_binds	A list of host volumes to bind to scheduler	list	[]	no
scheduler_extra_env	A list of env vars to prepend to scheduler	list	[]	no
service_cluster_ip_range	CIDR for services allocation	string	"10.43.0.0/16"	no
service_node_port_range	Range for nodeport allocation	string	"30000-32767"	no
system_reserved	Resources reserved for system tasks	string	"cpu=700m,memory=1Gi"	no
system_reserved_cgroup	Cgroup for system tasks	string	"/system.slice"	no
upgrade_max_unavailable_controlplane	Max ammount of controlplane nodes that can be unavailable during upgrades	string	"1"	no
upgrade_max_unavailable_worker	Max ammount of worker nodes that can be unavailable during upgrades	string	"10%"	no
vsphere_cluster_id	vSphere cluster ID	string	""	no
vsphere_datacenter	vSphere datacenter	string	""	no
vsphere_insecure_flag	Do not verify tls cert	bool	true	no
vsphere_password	vSphere password	string	""	no
vsphere_port	vSphere port	number	443	no
vsphere_server	vSphere server	string	""	no
vsphere_username	vSphere username	string	""	no
write_cluster_yaml	Save rke cluster yaml to a file	bool	false	no
write_kubeconfig	Save kubeconfig to a file	bool	true	no

Outputs

Name	Description
api_server_url	Kubernetes api-server endpoint
ca_crt	Kubernetes CA certificate
client_cert	Kubernetes client certificate
client_key	Kubernetes client key
cluster	Kubernetes cluster object
cluster_name	Kubernetes cluster name
kube_admin_user	Kubernetes admin user
kubeconfig	Kubernetes admin kubeconfig