Kubernetes

Introduction

In this section, we explain how to deploy Kubernetes workloads on the TFGrid using tfcmd.

Deploy

tfcmd deploy kubernetes [flags]

Required Flags

  • name: name for the master node deployment also used for canceling the cluster deployment. must be unique.
  • ssh: path to public ssh key to set in the master node.

Optional Flags

  • master-node: node id master should be deployed on.
  • master-farm: farm id master should be deployed on, if set choose available node from farm that fits master specs (default 1). note: master-node and master-farm flags cannot be set both.
  • workers-nodes: array of nodes ids workers should be deployed on and the remaining unassigned workers will be randomly assigned to nodes that meet the specifications.
  • workers-farm: farm id workers should be deployed on, if set choose available node from farm that fits master specs (default 1). note: workers-nodes and workers-farm flags cannot be set both.
  • ipv4: assign public ipv4 for master node (default false).
  • ipv6: assign public ipv6 for master node (default false).
  • ygg: assign yggdrasil ip for master node (default true).
  • mycelium: assign mycelium ip for master node (default true).
  • master-cpu: number of cpu units for master node (default 1).
  • master-memory: master node memory size in GB (default 1).
  • master-disk: master node disk size in GB (default 2).
  • workers-number: number of workers nodes (default 0).
  • workers-ipv4: assign public ipv4 for each worker node (default false)
  • workers-ipv6: assign public ipv6 for each worker node (default false)
  • workers-ygg: assign yggdrasil ip for each worker node (default true)
  • workers-mycelium: assign mycelium ip for each worker node (default true)
  • workers-cpu: number of cpu units for each worker node (default 1).
  • workers-memory: memory size for each worker node in GB (default 1).
  • workers-disk: disk size in GB for each worker node (default 2).

Example:

$ tfcmd deploy kubernetes -n kube --ssh ~/.ssh/id_rsa.pub --master-node 14 --workers-number 2 --workers-nodes 14,1
11:43AM INF starting peer session=tf-1510734 twin=192
11:43AM INF deploying network
11:43AM INF deploying cluster
11:43AM INF master wireguard ip: 10.20.2.2
11:43AM INF master planetary ip: 300:e9c4:9048:57cf:d73d:eb4c:7a6d:503
11:43AM INF master mycelium ip: 423:16f5:ca74:b600:ff0f:9ee:d57e:827a
11:43AM INF worker1 wireguard ip: 10.20.2.3
11:43AM INF worker0 wireguard ip: 10.20.2.3
11:43AM INF worker1 planetary ip: 300:e9c4:9048:57cf:3c4f:d477:b4a5:890b
11:43AM INF worker0 planetary ip: 300:e9c4:9048:57cf:77ca:5424:21da:4fff
11:43AM INF worker1 mycelium ip: 423:16f5:ca74:b600:ff0f:e02f:1ad7:d74e
11:43AM INF worker0 mycelium ip: 423:16f5:ca74:b600:ff0f:bb5a:6036:56f6

Get

tfcmd get kubernetes <kubernetes>

kubernetes is the name used when deploying kubernetes cluster using tfcmd.

Example:

$ tfcmd get kubernetes examplevm
3:14PM INF k8s cluster:
{
        "Master": {
                "Name": "kube",
                "Node": 14,
                "DiskSize": 2,
                "PublicIP": false,
                "PublicIP6": false,
                "Planetary": true,
                "Flist": "https://hub.grid.tf/tf-official-apps/threefoldtech-k3s-latest.flist",
                "FlistChecksum": "c87cf57e1067d21a3e74332a64ef9723",
                "ComputedIP": "",
                "ComputedIP6": "",
                "YggIP": "300:e9c4:9048:57cf:e8a0:662b:4e66:8faa",
                "IP": "10.20.2.2",
                "CPU": 1,
                "Memory": 1024
        },
        "Workers": [
                {
                        "Name": "worker1",
                        "Node": 14,
                        "DiskSize": 2,
                        "PublicIP": false,
                        "PublicIP6": false,
                        "Planetary": true,
                        "Flist": "https://hub.grid.tf/tf-official-apps/threefoldtech-k3s-latest.flist",
                        "FlistChecksum": "c87cf57e1067d21a3e74332a64ef9723",
                        "ComputedIP": "",
                        "ComputedIP6": "",
                        "YggIP": "300:e9c4:9048:57cf:66d0:3ee4:294e:d134",
                        "IP": "10.20.2.2",
                        "CPU": 1,
                        "Memory": 1024
                },
                {
                        "Name": "worker0",
                        "Node": 14,
                        "DiskSize": 2,
                        "PublicIP": false,
                        "PublicIP6": false,
                        "Planetary": true,
                        "Flist": "https://hub.grid.tf/tf-official-apps/threefoldtech-k3s-latest.flist",
                        "FlistChecksum": "c87cf57e1067d21a3e74332a64ef9723",
                        "ComputedIP": "",
                        "ComputedIP6": "",
                        "YggIP": "300:e9c4:9048:57cf:1ae5:cc51:3ffc:81e",
                        "IP": "10.20.2.2",
                        "CPU": 1,
                        "Memory": 1024
                }
        ],
        "Token": "",
        "NetworkName": "",
        "SolutionType": "kube",
        "SSHKey": "",
        "NodesIPRange": null,
        "NodeDeploymentID": {
                "14": 22743
        }
}

Cancel

tfcmd cancel <deployment-name>

deployment-name is the name of the deployment specified in while deploying using tfcmd.

Example:

$ tfcmd cancel kube
3:37PM INF canceling contracts for project kube
3:37PM INF kube canceled
Last change: 2024-11-05