go-anxcloud

Go Client for the Anexia API

Go SDK for interacting with the Anexia Engine API.

Installing

To use the SDK, just add go.anx.io/go-anxcloud <version> to your Go module.

Getting started

Before using the SDK you should familiarize yourself with the Anexia Engine API.

The library is used in our terraform provider, check it out if you want some examples how to use it.

Example

Below is a short example using the new generic client in this package. Not all APIs can already be used with it, but we are working on that. Find more examples in the docs (linked to docs for main branch, not the latest (or any) release).

package main

import (
	"context"
	"log"

	"go.anx.io/go-anxcloud/pkg/api"
	apiTypes "go.anx.io/go-anxcloud/pkg/api/types"
	"go.anx.io/go-anxcloud/pkg/client"

	// apis usable with the generic client have their own package in a location analog to this
	lbaasv1 "go.anx.io/go-anxcloud/pkg/apis/lbaas/v1"
)

func main() {
	apiClient, err := api.NewAPI(
		api.WithClientOptions(
			// Get auth token from ANEXIA_TOKEN environment variable.
			// The boolean parameter specifies if the environment variable should be unset.
			client.TokenFromEnv(false),
		),
	)
	if err != nil {
		log.Fatalf("Error creating ANX API client: %v", err)
	}

	// let's list LBaaS backends of a known LoadBalancer
	frontend := lbaasv1.Frontend{
		LoadBalancer: &lbaasv1.LoadBalancer{Identifier: "285b954fdf2a449c8fdae01cc6074025"},
	}

	var frontends apiTypes.ObjectChannel
	err = apiClient.List(context.TODO(), &frontend,
		// Listing can be done with either a page iterator or a channel, we use a channel here.
		api.ObjectChannel(&frontends),

		// Most APIs only give a very small subset when listing resources, add this flag to
		// get all attributes, at the cost of doing lots of API requests.
		api.FullObjects(true),
	)
	if err != nil {
		log.Fatalf("Error listing backends for LoadBalancer '%v': %v", frontend.LoadBalancer.Identifier, err)
	}

	for retriever := range frontends {
		// reinitialise frontend every loop to reset pointers and avoid potential overwriting of data in the next loop
		var frontend lbaasv1.Frontend
		if err := retriever(&frontend); err != nil {
			log.Fatalf("Error retrieving Frontend: %v", err)
		}

		log.Printf("Got Frontend named '%v' with mode '%v'", frontend.Name, frontend.Mode)
	}
}

This new generic client will one day be the only client in go-anxcloud. The legacy API-specific clients are deprecated and will be removed in the go-anxcloud release following the one with all APIs go-anxcloud supports usable with the generic client (so if the generic client in 0.5.0 supports at least everything there is another client for in go-anxcloud, 0.6.0 will drop the API-specific clients).

package main

import (
	"context"
	"fmt"
	"time"

	anexia "go.anx.io/go-anxcloud/pkg"
	"go.anx.io/go-anxcloud/pkg/client"
	"go.anx.io/go-anxcloud/pkg/vsphere/provisioning/vm"
)

func main() {
	vlan := "<ID of the VLAN the VM should have access to>"
	location := "<ID of the location the VM should be in>"

	// Create client using the auth token in environment variable ANEXIA_TOKEN and do not unset the environment variable.
	c, err := client.New(client.AuthFromEnv(false))
	if err != nil {
		panic(fmt.Sprintf("could not create client: %v", err))
	}

	// Get some API.
	provisioning := anexia.NewAPI(c).VSphere().Provisioning()

	// Time out after 30 minutes. Yes it really takes that long sometimes.
	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Minute)
	// Look for a free ip in the given VLAN. This IP is not reserved for you so better be quick.
	ips, err := provisioning.IPs().GetFree(ctx, location, vlan)
	defer cancel()
	if err != nil {
		panic(fmt.Sprintf("provisioning vm failed: %v", err))
	}
	if len(ips) < 1 {
		panic(fmt.Sprintf("no IPs left for testing in vlan"))
	}

	// Create a NIC for the VM and connect it to the VLAN.
	networkInterfaces := []vm.Network{{NICType: "vmxnet3", IPs: []string{ips[0].Identifier}, VLAN: vlan}}
	// Create the definition of the new VM. The ID you see here is Flatcar.
	definition := vm.NewAPI(c).NewDefinition(location, "template", "44b38284-6adb-430e-b4a4-1553e29f352f", "developersfirstvm", 2, 2048, 10, networkInterfaces)
	definition.SSH = "<your SSH pub key>"

	// Provision the VM.
	provisionResponse, err := provisioning.VM().Provision(ctx, definition)
	if err != nil {
		panic(fmt.Sprintf("provisioning vm failed: %v", err))
	}

	// Wait for the VM to be ready.
	_, err = provisioning.Progress().AwaitCompletion(ctx, provisionResponse.Identifier)
	if err != nil {
		panic(fmt.Sprintf("waiting for VM provisioning failed: %v", err))
	}
}