Managed Master Deployment Mode
Feature Overview
Tencent Kubernetes Engine (TKE) offers a fully managed Kubernetes cluster management service, including Master and Etcd components. In this mode, the Kubernetes cluster's Master and Etcd are centrally managed and maintained by the Tencent Cloud technical team. You only need to purchase the cluster and provide the worker nodes required to run the workloads, without worrying about cluster management and maintenance.
Notes on the managed master deployment mode
TKE charges cluster management fees based on the specifications of the managed clusters, and charges cloud resources (CVM, persistent storage, and CLB) fees based on the actual usage. For more information about the billing modes and prices, see TKE Billing Overview.
Master and Etcd nodes in this mode are not user-specific resources. Therefore, you cannot modify their deployment scale and service parameters. If you need to do so, please select the Independent Master Deployment Mode.
In this mode, even if you delete all the worker nodes from the cluster, the cluster will still persistently attempt to run workloads and services that have not been deleted. This process may incur fees. If you need to stop the services and prevent cluster fees, please delete the cluster.
Independent Master Deployment Mode
Feature Overview
Tencent Cloud TKE also offers a self-deployed master deployment mode, providing you with full control over the cluster. In this mode, the Kubernetes cluster's Master and Etcd components will be deployed on the CVM instances you purchased. You will have complete management and operational control over the Kubernetes cluster.
Notes on the independent master deployment mode
This mode is only available for Kubernetes v1.10.x and later.
In this mode, you need to purchase additional resources to deploy the Master and Etcd nodes of the Kubernetes cluster.
If your cluster contains a large number of nodes, we recommend that you select a high-spec model for the CVMs. When selecting models, see the following table for reference:
Cluster Scale | Recommended Master Node Configuration | Recommended Node Quantity |
Approximately 100 nodes | 8 cores, 16 GB memory, SSD system disk | 3 or more nodes |
Approximately 500 nodes | 16 cores, 32 GB memory, SSD system disk | 3 or more nodes |
Over 1,000 nodes | 3 or more nodes |
Purchase Requirements
To ensure the high availability and performance of your cluster and services, you need to meet the following requirements in independent deployment mode:
At least three CVMs are deployed for the Master and Etcd nodes.
The model of the CVMs configured for Master and Etcd nodes contains at least 4 cores.
SSDs are deployed as system disks on the CVMs for the Master and Etcd nodes.
Supports and Limits
To ensure the stability of your cluster and improve failback efficiency, we recommend that you note the following:
In the independent Master deployment mode:
Do not delete the core components of the Master node that support the operation of the Kubernetes cluster.
Do not modify the configuration parameters of the core components of the Master node.
Do not modify or delete the core resources in the Kubernetes cluster.
Do not modify or delete the relevant certificate files (with .crt and .key extensions) of the Master node.
Unless otherwise required:
Do not modify the Docker version of any node.
Do not modify the OS components, such as kernel and nfs-utils, of any node.
Note
Core components: kube-APIserver, kube-scheduler, kube-controller-manager, tke-tools, systemd, and cluster-container-agent.
Core component configuration parameters: kube-APIserver parameters, kube-scheduler parameters, and kube-controller-manager parameters.
Core resources in the cluster (including but not limited to): hpa endpoint, master service account, kube-dns, auto-scaler, master cluster role, and master cluster role binding.