Tencent Cloud

Recent Pages

Creating a Cluster

Last updated: 2023-07-14 16:30:12

This document is currently invalid. Please refer to the documentation page of the product.

﻿
This document describes how to create an edge cluster to use TKE Edge from the Tencent Cloud TKE console.
﻿
Container Network Explanation
TKE Edge Container Service will utilize the node-side network to build an overlay network. Please ensure that the cluster network (control plane in VPC network) and container network (Pod CIDR/Service CIDR) do not conflict with the internal network of the edge nodes.
Operation Steps
1. Log in to the Tencent Cloud TKE console and click Edge Clusters in the left sidebar.
2. On the Cluster Management page, click New.
3. Select Edge Cluster, then click 'Create'.
﻿
﻿
﻿
4. On the Create Edge Cluster page, create an edge cluster according to the following information, as shown in the figure below:
﻿
﻿
﻿
Cluster name: indicates the name of the edge cluster to be created, with a maximum length of 60 characters.
Kubernetes version: Kubernetes version 1.16 is currently supported. This version will be updated when a newer Kubernetes version is published by the Kubernetes community.
Runtime Components: Supports Docker runtime and Containerd runtime. For more information, please refer to How to Choose Between Containerd and Docker.
Region: select the region that is closest to your location to minimize access latency and improve the download speed.
Billing Method: The current Edge Cluster supports pay-as-you-go billing.
Cluster Specifications: You can choose different cluster specifications according to your needs. Different specifications have restrictions on the number of access nodes, as well as the number of pods, configmaps, and CRDs. For more details, please refer to the Purchase Cluster Quota Limitations.﻿
Auto Upgrade Cluster Specifications: If you enable this feature, when the cluster resources exceed the maximum limit of the current specification, the cluster specification will be automatically upgraded to a higher level and billed accordingly. Later on, you can reclaim resources, but the platform will not automatically downgrade. You can manually adjust the specifications to a more appropriate lower level. If this feature is not enabled, when the cluster resources exceed the maximum limit of the current specification, the newly added nodes will be in a "locked" state, not allowing application scheduling to that node. You will have to manually upgrade the cluster specification for the nodes to automatically return to normal.
Cluster network: assign a network for the cluster according to the internal network management of edge servers.
Pod CIDR: you need to assign a container network for the cluster according to the internal network management of edge servers. Therefore, plan the cluster size in advance to assign an IP range with sufficient IP addresses for the container network. The pod CIDR block cannot overlap with IP ranges used by a VPC instance and existing Kubernetes clusters in the VPC instance. In addition, it cannot be modified once created.
Service CIDR: you need to assign a service network for the cluster according to the internal network management of edge servers. Therefore, plan the cluster size in advance to assign an IP range with sufficient IP addresses for the service network. The service CIDR block cannot overlap with IP ranges used by a VPC instance and existing Kubernetes clusters in the VPC instance. In addition, it cannot be modified once created.
Kube-proxy Mode: You can choose between iptables mode and ipvs mode.
Cluster description: indicates information about the cluster, which is displayed on the Cluster Information page.
5. Click Done to finish creating the Master components of the cluster. You can check the progress of cluster creation on the "Edge Clusters" page.
Node Management
Note：
After the cluster is created, you need to enable the cluster's internal/external access in order to provide services to the outside world properly. For more details, please refer to Enabling Internal and External Access for the Cluster.
﻿