alicloud_cs_kubernetes_node_pool_Terraform(Terraform)-阿里云帮助中心

Provides a ACK Nodepool resource.

This resource will help you to manage node pool in Kubernetes Cluster, see What is kubernetes node pool.

-> NOTE: Available since v1.97.0.

-> NOTE: From version 1.109.1, support managed node pools, but only for the professional managed clusters.

-> NOTE: From version 1.109.1, support remove node pool nodes.

-> NOTE: From version 1.111.0, support auto scaling node pool. For more information on how to use auto scaling node pools, see Use Terraform to create an elastic node pool. With auto-scaling is enabled, the nodes in the node pool will be labeled with k8s.aliyun.com=true to prevent system pods such as coredns, metrics-servers from being scheduled to elastic nodes, and to prevent node shrinkage from causing business abnormalities.

-> NOTE: ACK adds a new RamRole (AliyunCSManagedAutoScalerRole) for the permission control of the node pool with auto-scaling enabled. If you are using a node pool with auto scaling, please click AliyunCSManagedAutoScalerRole to complete the authorization.

-> NOTE: ACK adds a new RamRole（AliyunCSManagedNlcRole） for the permission control of the management node pool. If you use the management node pool, please click AliyunCSManagedNlcRole to complete the authorization.

-> NOTE: From version 1.123.1, supports the creation of a node pool of spot instance.

-> NOTE: It is recommended to create a cluster with zero worker nodes, and then use a node pool to manage the cluster nodes.

-> NOTE: From version 1.127.0, support for adding existing nodes to the node pool. In order to distinguish automatically created nodes, it is recommended that existing nodes be placed separately in a node pool for management.

-> NOTE: From version 1.149.0, support for specifying deploymentSet for node pools.

-> NOTE: From version 1.158.0, Support for specifying the desired size of nodes for the node pool, for more information, visit Modify the expected number of nodes in a node pool

-> NOTE: From version 1.166.0, Support configuring system disk encryption.

-> NOTE: From version 1.177.0+, Support kms_encryption_context, rds_instances, system_disk_snapshot_policy_id and cpu_policy, add spot strategy SpotAsPriceGo and NoSpot.

-> NOTE: From version 1.180.0+, Support worker nodes customized kubelet parameters by field kubelet_configuration and rollout_policy.

-> NOTE: From version 1.185.0+, Field rollout_policy will be deprecated and please use field rolling_policy instead.

For information about ACK Nodepool and how to use it, see What is Nodepool.

Example Usage

Basic Usage

resource "random_integer" "default" {
  max = 99999
  min = 10000
}

variable "name" {
  default = "terraform-example"
}

data "alicloud_enhanced_nat_available_zones" "enhanced" {
}

data "alicloud_instance_types" "cloud_efficiency" {
  availability_zone    = data.alicloud_enhanced_nat_available_zones.enhanced.zones.0.zone_id
  cpu_core_count       = 4
  memory_size          = 8
  kubernetes_node_role = "Worker"
  system_disk_category = "cloud_efficiency"
}

resource "alicloud_vpc" "default" {
  vpc_name   = var.name
  cidr_block = "10.4.0.0/16"
}
resource "alicloud_vswitch" "default" {
  vswitch_name = var.name
  cidr_block   = "10.4.0.0/24"
  vpc_id       = alicloud_vpc.default.id
  zone_id      = data.alicloud_enhanced_nat_available_zones.enhanced.zones.0.zone_id
}

resource "alicloud_cs_managed_kubernetes" "default" {
  name_prefix          = "terraform-example-${random_integer.default.result}"
  cluster_spec         = "ack.pro.small"
  worker_vswitch_ids   = [alicloud_vswitch.default.id]
  new_nat_gateway      = true
  pod_cidr             = cidrsubnet("10.0.0.0/8", 8, 36)
  service_cidr         = cidrsubnet("172.16.0.0/16", 4, 7)
  slb_internet_enabled = true
  enable_rrsa          = true
}

resource "alicloud_key_pair" "default" {
  key_pair_name = "terraform-example-${random_integer.default.result}"
}

resource "alicloud_cs_kubernetes_node_pool" "default" {
  node_pool_name       = var.name
  cluster_id           = alicloud_cs_managed_kubernetes.default.id
  vswitch_ids          = [alicloud_vswitch.default.id]
  instance_types       = [data.alicloud_instance_types.cloud_efficiency.instance_types.0.id]
  system_disk_category = "cloud_efficiency"
  system_disk_size     = 40
  key_name             = alicloud_key_pair.default.key_pair_name
  // define with multi-labels by defining with labels blocks
  labels {
    key   = "test1"
    value = "nodepool"
  }
  labels {
    key   = "test2"
    value = "nodepool"
  }
  // define with multi-taints by defining with taints blocks
  taints {
    key    = "tf"
    effect = "NoSchedule"
    value  = "example"
  }
  taints {
    key    = "tf2"
    effect = "NoSchedule"
    value  = "example2"
  }
}

#The parameter `node_count` is deprecated from version 1.158.0. Please use the new parameter `desired_size` instead, you can update it as follows.
resource "alicloud_cs_kubernetes_node_pool" "desired_size" {
  node_pool_name       = "desired_size"
  cluster_id           = alicloud_cs_managed_kubernetes.default.id
  vswitch_ids          = [alicloud_vswitch.default.id]
  instance_types       = [data.alicloud_instance_types.cloud_efficiency.instance_types.0.id]
  system_disk_category = "cloud_efficiency"
  system_disk_size     = 40
  key_name             = alicloud_key_pair.default.key_pair_name
  desired_size         = 0
}

# Create a managed node pool. If you need to enable maintenance window, you need to set the maintenance window in `alicloud_cs_managed_kubernetes`.
resource "alicloud_cs_kubernetes_node_pool" "maintenance" {
  node_pool_name       = "maintenance"
  cluster_id           = alicloud_cs_managed_kubernetes.default.id
  vswitch_ids          = [alicloud_vswitch.default.id]
  instance_types       = [data.alicloud_instance_types.cloud_efficiency.instance_types.0.id]
  system_disk_category = "cloud_efficiency"
  system_disk_size     = 40

  # only key_name is supported in the management node pool
  key_name = alicloud_key_pair.default.key_pair_name

  # you need to specify the number of nodes in the node pool, which can be zero
  desired_size = 1

  # management node pool configuration.
  management {
    enable      = true
    auto_repair = true
    auto_repair_policy {
      restart_node = true
    }
    auto_upgrade = true
    auto_upgrade_policy {
      auto_upgrade_kubelet = true
    }
    auto_vul_fix = true
    auto_vul_fix_policy {
      vul_level    = "asap"
      restart_node = true
    }
    max_unavailable = 1
  }

  # Enable with automatic scaling node pool configuration.
  # With auto-scaling is enabled, the nodes in the node pool will be labeled with `k8s.aliyun.com=true` to prevent system pods such as coredns, metrics-servers from being scheduled to elastic nodes, and to prevent node shrinkage from causing business abnormalities.
  #  scaling_config {
  #    min_size = 1
  #    max_size = 10
  #    type     = "cpu"
  #  }
}

#Create a node pool with spot instance.
resource "alicloud_cs_kubernetes_node_pool" "spot_instance" {
  node_pool_name       = "spot_instance"
  cluster_id           = alicloud_cs_managed_kubernetes.default.id
  vswitch_ids          = [alicloud_vswitch.default.id]
  instance_types       = [data.alicloud_instance_types.cloud_efficiency.instance_types.0.id, data.alicloud_instance_types.cloud_efficiency.instance_types.1.id]
  system_disk_category = "cloud_efficiency"
  system_disk_size     = 40
  key_name             = alicloud_key_pair.default.key_pair_name

  # you need to specify the number of nodes in the node pool, which can be 0
  desired_size = 1

  # spot config
  spot_strategy = "SpotWithPriceLimit"
  spot_price_limit {
    instance_type = data.alicloud_instance_types.cloud_efficiency.instance_types.0.id
    # Different instance types have different price caps
    price_limit = "0.70"
  }
  // define with multi-spot_price_limit by defining with spot_price_limit blocks
  spot_price_limit {
    instance_type = data.alicloud_instance_types.cloud_efficiency.instance_types.1.id
    price_limit   = "0.72"
  }
}


#Use Spot instances to create a node pool with auto-scaling enabled
resource "alicloud_cs_kubernetes_node_pool" "spot_auto_scaling" {
  node_pool_name       = "spot_auto_scaling"
  cluster_id           = alicloud_cs_managed_kubernetes.default.id
  vswitch_ids          = [alicloud_vswitch.default.id]
  instance_types       = [data.alicloud_instance_types.cloud_efficiency.instance_types.0.id]
  system_disk_category = "cloud_efficiency"
  system_disk_size     = 40
  key_name             = alicloud_key_pair.default.key_pair_name

  # automatic scaling node pool configuration.
  scaling_config {
    min_size = 1
    max_size = 10
    type     = "spot"
  }
  # spot price config
  spot_strategy = "SpotWithPriceLimit"
  spot_price_limit {
    instance_type = data.alicloud_instance_types.cloud_efficiency.instance_types.0.id
    price_limit   = "0.70"
  }
}

#Create a `PrePaid` node pool.
resource "alicloud_cs_kubernetes_node_pool" "prepaid_node" {
  node_pool_name       = "prepaid_node"
  cluster_id           = alicloud_cs_managed_kubernetes.default.id
  vswitch_ids          = [alicloud_vswitch.default.id]
  instance_types       = [data.alicloud_instance_types.cloud_efficiency.instance_types.0.id]
  system_disk_category = "cloud_efficiency"
  system_disk_size     = 40
  key_name             = alicloud_key_pair.default.key_pair_name
  # use PrePaid
  instance_charge_type = "PrePaid"
  period               = 1
  period_unit          = "Month"
  auto_renew           = true
  auto_renew_period    = 1

  # open cloud monitor
  install_cloud_monitor = true
}

##Create a node pool with customized kubelet parameters
resource "alicloud_cs_kubernetes_node_pool" "customized_kubelet" {
  node_pool_name       = "customized_kubelet"
  cluster_id           = alicloud_cs_managed_kubernetes.default.id
  vswitch_ids          = [alicloud_vswitch.default.id]
  instance_types       = [data.alicloud_instance_types.cloud_efficiency.instance_types.0.id]
  system_disk_category = "cloud_efficiency"
  system_disk_size     = 40
  instance_charge_type = "PostPaid"
  desired_size         = 0

  # kubelet configuration parameters
  kubelet_configuration {
    registry_pull_qps     = 10
    registry_burst        = 5
    event_record_qps      = 10
    event_burst           = 5
    serialize_image_pulls = true
    eviction_hard = {
      "memory.available"  = "1024Mi"
      "nodefs.available"  = "10%"
      "nodefs.inodesFree" = "5%"
      "imagefs.available" = "10%"
    }
    system_reserved = {
      "cpu"               = "1"
      "memory"            = "1Gi"
      "ephemeral-storage" = "10Gi"
    }
    kube_reserved = {
      "cpu"    = "500m"
      "memory" = "1Gi"
    }
    container_log_max_size  = "200Mi"
    container_log_max_files = 3
    max_pods                = 100
    read_only_port          = 0
    allowed_unsafe_sysctls  = ["net.ipv4.route.min_pmtu"]
  }

  # rolling policy: works when updating
  rolling_policy {
    max_parallelism = 1
  }
}

Argument Reference

The following arguments are supported:

auto_renew - (Optional) Whether to enable automatic renewal for nodes in the node pool takes effect only when instance_charge_type is set to PrePaid. Default value: false. Valid values:
- true: Automatic renewal.
- false: Do not renew automatically.
auto_renew_period - (Optional, Int) The automatic renewal period of nodes in the node pool takes effect only when you select Prepaid and Automatic Renewal, and is a required value. When PeriodUnit = Month, the value range is {1, 2, 3, 6, 12}. Default value: 1.
cis_enabled - (Optional, ForceNew, Deprecated since v1.223.1) Whether enable worker node to support cis security reinforcement, its valid value true or false. Default to false and apply to AliyunLinux series. Use security_hardening_os instead.
cluster_id - (Required, ForceNew) The id of kubernetes cluster.
compensate_with_on_demand - (Optional) Specifies whether to automatically create pay-as-you-go instances to meet the required number of ECS instances if preemptible instances cannot be created due to reasons such as cost or insufficient inventory. This parameter takes effect when you set multi_az_policy to COST_OPTIMIZED. Valid values: true: automatically creates pay-as-you-go instances to meet the required number of ECS instances if preemptible instances cannot be created. false: does not create pay-as-you-go instances to meet the required number of ECS instances if preemptible instances cannot be created.
cpu_policy - (Optional, Computed) Node CPU management policies. Default value: none. When the cluster version is 1.12.6 or later, the following two policies are supported:
- static: allows pods with certain resource characteristics on the node to enhance its CPU affinity and exclusivity.
- none: Enables the existing default CPU affinity scheme.
data_disks - (Optional, List) Configure the data disk of the node in the node pool. See data_disks below.
deployment_set_id - (Optional, ForceNew) The deployment set of node pool. Specify the deploymentSet to ensure that the nodes in the node pool can be distributed on different physical machines.
desired_size - (Optional, Int) Number of expected nodes in the node pool.
force_delete - (Optional) Whether to force deletion.
image_id - (Optional, Computed) The custom image ID. The system-provided image is used by default.
image_type - (Optional, Computed) The operating system image type and the platform parameter can be selected from the following values:
- AliyunLinux : Alinux2 image.
- AliyunLinux3 : Alinux3 image.
- AliyunLinux3Arm64 : Alinux3 mirror ARM version.
- AliyunLinuxUEFI : Alinux2 Image UEFI version.
- CentOS : CentOS image.
- Windows : Windows image.
- WindowsCore : WindowsCore image.
- ContainerOS : container-optimized image.
install_cloud_monitor - (Optional) Whether to install cloud monitoring on the ECS node. After installation, you can view the monitoring information of the created ECS instance in the cloud monitoring console and recommend enable it. Default value: false. Valid values:
- true : install cloud monitoring on the ECS node.
- false : does not install cloud monitoring on the ECS node.
instance_charge_type - (Optional, Computed) Node payment type. Valid values: PostPaid, PrePaid, default is PostPaid. If value is PrePaid, the arguments period, period_unit, auto_renew and auto_renew_period are required.
instance_types - (Required, List) In the node instance specification list, you can select multiple instance specifications as alternatives. When each node is created, it will try to purchase from the first specification until it is created successfully. The final purchased instance specifications may vary with inventory changes.
internet_charge_type - (Optional) The billing method for network usage. Valid values PayByBandwidth and PayByTraffic. Conflict with eip_internet_charge_type, EIP and public network IP can only choose one.
internet_max_bandwidth_out - (Optional, Int) The maximum bandwidth of the public IP address of the node. The unit is Mbps(Mega bit per second). The value range is:[1,100]
key_name - (Optional) The name of the key pair. When the node pool is a managed node pool, only key_name is supported.
kubelet_configuration - (Optional, List) Kubelet configuration parameters for worker nodes. See kubelet_configuration below. More information in Kubelet Configuration. See kubelet_configuration below.
labels - (Optional, List) A List of Kubernetes labels to assign to the nodes . Only labels that are applied with the ACK API are managed by this argument. Detailed below. More information in Labels. See labels below.
login_as_non_root - (Optional, ForceNew) Whether the ECS instance is logged on as a ecs-user user. Valid value: true and false.
management - (Optional, Computed, List) Managed node pool configuration. See management below.
multi_az_policy - (Optional, Computed) The scaling policy for ECS instances in a multi-zone scaling group. Valid value: PRIORITY, COST_OPTIMIZED and BALANCE. PRIORITY: scales the capacity according to the virtual switches you define (VSwitchIds.N). When an ECS instance cannot be created in the zone where the higher-priority vSwitch is located, the next-priority vSwitch is automatically used to create an ECS instance. COST_OPTIMIZED: try to create by vCPU unit price from low to high. When the scaling configuration is configured with multiple instances of preemptible billing, preemptible instances are created first. You can continue to use the CompensateWithOnDemand parameter to specify whether to automatically try to create a preemptible instance by paying for it. It takes effect only when the scaling configuration has multi-instance specifications or preemptible instances. BALANCE: distributes ECS instances evenly among the multi-zone specified by the scaling group. If the zones become unbalanced due to insufficient inventory, you can use the API [RebalanceInstances](~~ 71516 ~~) to balance resources.
node_name_mode - (Optional, ForceNew, Computed) Each node name consists of a prefix, its private network IP, and a suffix, separated by commas. The input format is customized,,ip,.
- The prefix and suffix can be composed of one or more parts separated by '.', each part can use lowercase letters, numbers and '-', and the beginning and end of the node name must be lowercase letters and numbers.
- The node IP address is the complete private IP address of the node.
- For example, if the string customized,aliyun,ip,com is passed in (where 'customized' and 'ip' are fixed strings, 'aliyun' is the prefix, and 'com' is the suffix), the name of the node is aliyun192.168.xxx.xxxcom.
node_pool_name - (Optional) The name of node pool.
on_demand_base_capacity - (Optional, Int) The minimum number of pay-as-you-go instances that must be kept in the scaling group. Valid values: 0 to 1000. If the number of pay-as-you-go instances is less than the value of this parameter, Auto Scaling preferably creates pay-as-you-go instances.
on_demand_percentage_above_base_capacity - (Optional, Int) The percentage of pay-as-you-go instances among the extra instances that exceed the number specified by on_demand_base_capacity. Valid values: 0 to 100.
password - (Optional) The password of ssh login. You have to specify one of password and key_name fields. The password rule is 8 to 30 characters and contains at least three items (upper and lower case letters, numbers, and special symbols).
period - (Optional, Int) Node payment period. Its valid value is one of {1, 2, 3, 6, 12}.
period_unit - (Optional) Node payment period unit, valid value: Month. Default is Month.
platform - (Optional, Computed, Deprecated since v1.145.0) Operating system release, using image_type instead.
pre_user_data - (Optional, Available since v1.232.0) Node pre custom data, base64-encoded, the script executed before the node is initialized.
private_pool_options - (Optional, List) Private node pool configuration. See private_pool_options below.
rds_instances - (Optional, List) The list of RDS instances.
resource_group_id - (Optional, Computed) The ID of the resource group
rolling_policy - (Optional, List) Rotary configuration. See rolling_policy below.
runtime_name - (Optional, Computed) The runtime name of containers. If not set, the cluster runtime will be used as the node pool runtime. If you select another container runtime, see Comparison of Docker, containerd, and Sandboxed-Container.
runtime_version - (Optional, Computed) The runtime version of containers. If not set, the cluster runtime will be used as the node pool runtime.
scaling_config - (Optional, Computed, List) Automatic scaling configuration. See scaling_config below.
scaling_policy - (Optional, Computed) Scaling group mode, default value: release. Valid values:
- release: in the standard mode, scaling is performed by creating and releasing ECS instances based on the usage of the application resource value.
- recycle: in the speed mode, scaling is performed through creation, shutdown, and startup to increase the speed of scaling again (computing resources are not charged during shutdown, only storage fees are charged, except for local disk models).
security_group_id - (Optional, ForceNew, Computed, Deprecated since v1.145.0) The security group ID of the node pool. This field has been replaced by security_group_ids, please use the security_group_ids field instead.
security_group_ids - (Optional, ForceNew, Computed, List) Multiple security groups can be configured for a node pool. If both security_group_ids and security_group_id are configured, security_group_ids takes effect. This field cannot be modified.
security_hardening_os - (Optional, ForceNew) Alibaba Cloud OS security reinforcement. Default value: false. Value: -true: enable Alibaba Cloud OS security reinforcement. -false: does not enable Alibaba Cloud OS security reinforcement.
soc_enabled - (Optional, ForceNew) Whether enable worker node to support soc security reinforcement, its valid value true or false. Default to false and apply to AliyunLinux series. See SOC Reinforcement.

-> NOTE: It is forbidden to set both security_hardening_os and soc_enabled to true at the same time.

spot_instance_pools - (Optional, Int) The number of instance types that are available. Auto Scaling creates preemptible instances of multiple instance types that are available at the lowest cost. Valid values: 1 to 10.
spot_instance_remedy - (Optional) Specifies whether to supplement preemptible instances when the number of preemptible instances drops below the specified minimum number. If you set the value to true, Auto Scaling attempts to create a new preemptible instance when the system notifies that an existing preemptible instance is about to be reclaimed. Valid values: true: enables the supplementation of preemptible instances. false: disables the supplementation of preemptible instances.
spot_price_limit - (Optional, List) The current single preemptible instance type market price range configuration. See spot_price_limit below.
spot_strategy - (Optional, Computed) The preemptible instance type. Value:
- NoSpot : Non-preemptible instance.
- SpotWithPriceLimit : Set the upper limit of the preemptible instance price.
- SpotAsPriceGo : The system automatically bids, following the actual price of the current market.
system_disk_bursting_enabled - (Optional) Specifies whether to enable the burst feature for system disks. Valid values:true: enables the burst feature. false: disables the burst feature. This parameter is supported only when system_disk_category is set to cloud_auto.
system_disk_categories - (Optional, Computed, List) The multi-disk categories of the system disk. When a high-priority disk type cannot be used, Auto Scaling automatically tries to create a system disk with the next priority disk category. Valid values: cloud: cloud disk. cloud_efficiency: a high-efficiency cloud disk. cloud_ssd:SSD cloud disk. cloud_essd: ESSD cloud disk.
system_disk_category - (Optional, Computed) The system disk category of worker node. Its valid value are cloud_ssd, cloud_efficiency, cloud_essd and cloud_auto.
system_disk_encrypt_algorithm - (Optional) The encryption algorithm used by the system disk. Value range: aes-256.
system_disk_encrypted - (Optional) Whether to encrypt the system disk. Value range: true: encryption. false: Do not encrypt.
system_disk_kms_key - (Optional) The ID of the KMS key used by the system disk.
system_disk_performance_level - (Optional) The system disk performance of the node takes effect only for the ESSD disk.
- PL0: maximum random read/write IOPS 10000 for a single disk.
- PL1: maximum random read/write IOPS 50000 for a single disk.
- PL2: highest random read/write IOPS 100000 for a single disk.
- PL3: maximum random read/write IOPS 1 million for a single disk.
system_disk_provisioned_iops - (Optional, Int) The predefined IOPS of a system disk. Valid values: 0 to min{50,000, 1,000 × Capacity - Baseline IOPS}. Baseline IOPS = min{1,800 + 50 × Capacity, 50,000}. This parameter is supported only when system_disk_category is set to cloud_auto.
system_disk_size - (Optional, Int) The system disk category of worker node. Its valid value range [40~500] in GB. Default to 120.
system_disk_snapshot_policy_id - (Optional) The ID of the automatic snapshot policy used by the system disk.
tags - (Optional, Map) Add tags only for ECS instances. The maximum length of the tag key is 128 characters. The tag key and value cannot start with aliyun or acs:, or contain https:// or http://.
taints - (Optional, List) A List of Kubernetes taints to assign to the nodes. Detailed below. More information in Taints and Toleration. See taints below.
tee_config - (Optional, ForceNew, Computed, List) The configuration about confidential computing for the cluster. See tee_config below.
unschedulable - (Optional) Whether the node after expansion can be scheduled.
update_nodes - (Optional) Synchronously update node labels and taints.
user_data - (Optional) Node custom data, base64-encoded.
vswitch_ids - (Required, List) The vswitches used by node pool workers.
kms_encrypted_password - (Optional, Available since v1.177.0) An KMS encrypts password used to a cs kubernetes. You have to specify one of password key_name kms_encrypted_password fields.
kms_encryption_context - (Optional, Available since v1.177.0) An KMS encryption context used to decrypt kms_encrypted_password before creating or updating a cs kubernetes with kms_encrypted_password. See Encryption Context. It is valid when kms_encrypted_password is set.
format_disk - (Optional, Available since v1.127.0) After you select this check box, if data disks have been attached to the specified ECS instances and the file system of the last data disk is uninitialized, the system automatically formats the last data disk to ext4 and mounts the data disk to /var/lib/docker and /var/lib/kubelet. The original data on the disk will be cleared. Make sure that you back up data in advance. If no data disk is mounted on the ECS instance, no new data disk will be purchased. Default is false.
instances - (Optional, Available since v1.127.0) The instance list. Add existing nodes under the same cluster VPC to the node pool.
node_count (Optional, Deprecated) The worker node number of the node pool. From version 1.111.0, node_count is not required.
keep_instance_name - (Optional, Available since v1.127.0) Add an existing instance to the node pool, whether to keep the original instance name. It is recommended to set to true.
rollout_policy - (Optional, Deprecated since 1.185.0) Rollout policy is used to specify the strategy when the node pool is rolling update. This field works when node pool updating. Please use rolling_policy to instead it from provider version 1.185.0. See rollout_policy below.

The following arguments will be discarded. Please use new fields as soon as possible:

name - (Deprecated since v1.219.0). Field 'name' has been deprecated from provider version 1.219.0. New field 'node_pool_name' instead.

`data_disks`

The data_disks supports the following:

auto_format - (Optional, Available since v1.229.0) Whether to automatically mount the data disk. Valid values: true and false.
auto_snapshot_policy_id - (Optional) The ID of the automatic snapshot policy that you want to apply to the system disk.
bursting_enabled - (Optional) Whether the data disk is enabled with Burst (performance Burst). This is configured when the disk type is cloud_auto.
category - (Optional) The type of the data disks. Valid values:cloud, cloud_efficiency, cloud_ssd, cloud_essd, cloud_auto.
device - (Optional) The mount target of data disk N. Valid values of N: 1 to 16. If you do not specify this parameter, the system automatically assigns a mount target when Auto Scaling creates ECS instances. The name of the mount target ranges from /dev/xvdb to /dev/xvdz.
encrypted - (Optional) Specifies whether to encrypt data disks. Valid values: true and false. Default to false.
file_system - (Optional, Available since v1.229.0) The Mount path. Works when auto_format is true.
kms_key_id - (Optional) The kms key id used to encrypt the data disk. It takes effect when encrypted is true.
mount_target - (Optional, Available since v1.229.0) The type of the mounted file system. Works when auto_format is true. Optional value: ext4, xfs.
name - (Optional, Computed) The length is 2~128 English or Chinese characters. It must start with an uppercase or lowr letter or a Chinese character and cannot start with http:// or https. Can contain numbers, colons (:), underscores (_), or dashes (-). It will be overwritten if auto_format is set.
performance_level - (Optional) Worker node data disk performance level, when category values cloud_essd, the optional values are PL0, PL1, PL2 or PL3, but the specific performance level is related to the disk capacity. For more information, see Enhanced SSDs. Default is PL1.
provisioned_iops - (Optional, Int) The read/write IOPS preconfigured for the data disk, which is configured when the disk type is cloud_auto.
size - (Optional, Int) The size of a data disk, Its valid value range [40~32768] in GB. Default to 40.
snapshot_id - (Optional) The ID of the snapshot that you want to use to create data disk N. Valid values of N: 1 to 16. If you specify this parameter, DataDisk.N.Size is ignored. The size of the disk is the same as the size of the specified snapshot. If you specify a snapshot that is created on or before July 15, 2013, the operation fails and InvalidSnapshot.TooOld is returned.

`kubelet_configuration`

The kubelet_configuration supports the following:

allowed_unsafe_sysctls - (Optional, List) Allowed sysctl mode whitelist.
container_log_max_files - (Optional) The maximum number of log files that can exist in each container.
container_log_max_size - (Optional) The maximum size that can be reached before a log file is rotated.
cpu_manager_policy - (Optional) Same as cpuManagerPolicy. The name of the policy to use. Requires the CPUManager feature gate to be enabled. Valid value is none or static.
event_burst - (Optional) Same as eventBurst. The maximum size of a burst of event creations, temporarily allows event creations to burst to this number, while still not exceeding event_record_qps. It is only used when event_record_qps is greater than 0. Valid value is [0-100].
event_record_qps - (Optional) Same as eventRecordQPS. The maximum event creations per second. If 0, there is no limit enforced. Valid value is [0-50].
eviction_hard - (Optional, Map) Same as evictionHard. The map of signal names to quantities that defines hard eviction thresholds. For example: {"memory.available" = "300Mi"}.
eviction_soft - (Optional, Map) Same as evictionSoft. The map of signal names to quantities that defines soft eviction thresholds. For example: {"memory.available" = "300Mi"}.
eviction_soft_grace_period - (Optional, Map) Same as evictionSoftGracePeriod. The map of signal names to quantities that defines grace periods for each soft eviction signal. For example: {"memory.available" = "30s"}.
feature_gates - (Optional, Map) Feature switch to enable configuration of experimental features.
kube_api_burst - (Optional) Same as kubeAPIBurst. The burst to allow while talking with kubernetes api-server. Valid value is [0-100].
kube_api_qps - (Optional) Same as kubeAPIQPS. The QPS to use while talking with kubernetes api-server. Valid value is [0-50].
kube_reserved - (Optional, Map) Same as kubeReserved. The set of ResourceName=ResourceQuantity (e.g. cpu=200m,memory=150G) pairs that describe resources reserved for kubernetes system components. Currently, cpu, memory and local storage for root file system are supported. See compute resources for more details.
max_pods - (Optional) The maximum number of running pods.
read_only_port - (Optional) Read-only port number.
registry_burst - (Optional) Same as registryBurst. The maximum size of burst pulls, temporarily allows pulls to burst to this number, while still not exceeding registry_pull_qps. Only used if registry_pull_qps is greater than 0. Valid value is [0-100].
registry_pull_qps - (Optional) Same as registryPullQPS. The limit of registry pulls per second. Setting it to 0 means no limit. Valid value is [0-50].
serialize_image_pulls - (Optional) Same as serializeImagePulls. When enabled, it tells the Kubelet to pull images one at a time. We recommend not changing the default value on nodes that run docker daemon with version < 1.9 or an Aufs storage backend. Valid value is true or false.
system_reserved - (Optional, Map) Same as systemReserved. The set of ResourceName=ResourceQuantity (e.g. cpu=200m,memory=150G) pairs that describe resources reserved for non-kubernetes components. Currently, only cpu and memory are supported. See compute resources for more details.

`labels`

The labels supports the following:

key - (Required) The label key.
value - (Optional) The label value.

`management`

The management supports the following:

auto_repair - (Optional, Computed) Whether to enable automatic repair. Valid values: true: Automatic repair. false: not automatically repaired.
auto_repair_policy - (Optional, Computed, List) Automatic repair node policy. See auto_repair_policy below.
auto_upgrade - (Optional, Computed) Specifies whether to enable auto update. Valid values: true: enables auto update. false: disables auto update.
auto_upgrade_policy - (Optional, Computed, List) The auto update policy. See auto_upgrade_policy below.
auto_vul_fix - (Optional, Computed) Specifies whether to automatically patch CVE vulnerabilities. Valid values: true, false.
auto_vul_fix_policy - (Optional, Computed, List) The auto CVE patching policy. See auto_vul_fix_policy below.
enable - (Optional, Computed) Specifies whether to enable the managed node pool feature. Valid values: true: enables the managed node pool feature. false: disables the managed node pool feature. Other parameters in this section take effect only when you specify enable=true.
max_unavailable - (Optional, Int) Maximum number of unavailable nodes. Default value: 1. Value range:[1,1000].
surge - (Optional, Int, Deprecated since v1.219.0) Number of additional nodes. You have to specify one of surge, surge_percentage.
surge_percentage - (Optional, Int, Deprecated since v1.219.0) Proportion of additional nodes. You have to specify one of surge, surge_percentage.

`management-auto_repair_policy`

The management-auto_repair_policy supports the following:

restart_node - (Optional, Computed) Whether to allow node restart.

`management-auto_upgrade_policy`

The management-auto_upgrade_policy supports the following:

auto_upgrade_kubelet - (Optional, Computed) Specifies whether to automatically update the kubelet. Valid values: true: yes; false: no.

`management-auto_vul_fix_policy`

The management-auto_vul_fix_policy supports the following:

restart_node - (Optional, Computed) Specifies whether to automatically restart nodes after patching CVE vulnerabilities. Valid values: true, false.
vul_level - (Optional, Computed) The severity levels of vulnerabilities that is allowed to automatically patch. Multiple severity levels are separated by commas (,).

`private_pool_options`

The private_pool_options supports the following:

private_pool_options_id - (Optional) The ID of the private node pool.
private_pool_options_match_criteria - (Optional) The type of private node pool. This parameter specifies the type of the private pool that you want to use to create instances. A private node pool is generated when an elasticity assurance or a capacity reservation service takes effect. The system selects a private node pool to launch instances. Valid values: Open: specifies an open private node pool. The system selects an open private node pool to launch instances. If no matching open private node pool is available, the resources in the public node pool are used. Target: specifies a private node pool. The system uses the resources of the specified private node pool to launch instances. If the specified private node pool is unavailable, instances cannot be started. None: no private node pool is used. The resources of private node pools are not used to launch the instances.

`rolling_policy`

The rolling_policy supports the following:

max_parallelism - (Optional, Int) The maximum number of unusable nodes.

`scaling_config`

The scaling_config supports the following:

eip_bandwidth - (Optional, Int) Peak EIP bandwidth. Its valid value range [1~500] in Mbps. It works if is_bond_eip=true. Default to 5.
eip_internet_charge_type - (Optional) EIP billing type. PayByBandwidth: Charged at fixed bandwidth. PayByTraffic: Billed as used traffic. Default: PayByBandwidth. It works if is_bond_eip=true, conflict with internet_charge_type. EIP and public network IP can only choose one.
enable - (Optional) Whether to enable automatic scaling. Value:
- true: enables the node pool auto-scaling function.
- false: Auto scaling is not enabled. When the value is false, other auto_scaling configuration parameters do not take effect.
is_bond_eip - (Optional) Whether to bind EIP for an instance. Default: false.
max_size - (Optional, Int) Max number of instances in a auto scaling group, its valid value range [0~1000]. max_size has to be greater than min_size.
min_size - (Optional, Int) Min number of instances in a auto scaling group, its valid value range [0~1000].
type - (Optional) Instance classification, not required. Vaild value: cpu, gpu, gpushare and spot. Default: cpu. The actual instance type is determined by instance_types.

`spot_price_limit`

The spot_price_limit supports the following:

instance_type - (Optional) The type of the preemptible instance.
price_limit - (Optional) The maximum price of a single instance.

`taints`

The taints supports the following:

effect - (Optional) The scheduling policy.
key - (Required) The key of a taint.
value - (Optional) The value of a taint.

`tee_config`

The tee_config supports the following:

tee_enable - (Optional, ForceNew) Specifies whether to enable confidential computing for the cluster.

`rollout_policy`

The rollout_policy mapping supports the following:

max_unavailable - (Optional, Deprecated since 1.185.0) Maximum number of unavailable nodes during rolling upgrade. The value of this field should be greater than 0, and if it's set to a number less than or equal to 0, the default setting will be used. Please use max_parallelism to instead it from provider version 1.185.0.

Attributes Reference

The following attributes are exported:

id - The ID of the resource supplied above.The value is formulated as <cluster_id>:<node_pool_id>.
node_pool_id - The first ID of the resource.
scaling_group_id - The ID of the scaling group.

Timeouts

The timeouts block allows you to specify timeouts for certain actions:

create - (Defaults to 90 mins) Used when create the Nodepool.
delete - (Defaults to 60 mins) Used when delete the Nodepool.
update - (Defaults to 60 mins) Used when update the Nodepool.

Import

ACK Nodepool can be imported using the id, e.g.

$ terraform import alicloud_cs_kubernetes_node_pool.example <cluster_id>:<node_pool_id>