-
If you haven't already, set up a Go workspace according to the Go docs.
-
Install the Go SDK. Normally this is done with "go get".
-
Setup GOPATH environment variable
Then:
mkdir -p $GOPATH/src/github.com/vmware cd $GOPATH/src/github.com/vmware git clone (github.com/vmware or gerrit)/photon-controller-go-sdk
Here's a quick sample app that will retrieve Photon Controller status from a [local devbox]. In this example, it's under $GOPATH/src/sdkexample/main.go:
package main
import (
"fmt"
"github.com/vmware/photon-controller-go-sdk/photon"
"log"
)
func main() {
client := photon.NewClient("http://localhost:9080", nil, nil)
status, err := client.Status.Get()
if err != nil {
log.Fatal(err)
}
fmt.Print(status)
}
Then build it and run it:
cd $GOPATH/src/sdkexample
go build ./...
./sdkexample
And the output should look something like this:
&{READY [{PHOTON_CONTROLLER READY}]}
Most Photon Controller APIs use a task model. The API will return a task object, which will indicate the state of the task (such as queued, completed, error, etc). These tasks return immediately and the caller must poll to find out when the task has been completed.The Go SDK provides a tasks API to do this for you, with built-in retry and error handling.
Let's expand the sample app to create a new tenant:
package main
import (
"fmt"
"github.com/vmware/photon-controller-go-sdk/photon"
"log"
)
func main() {
client := photon.NewClient("http://localhost:9080", nil, nil)
status, err := client.Status.Get()
if err != nil {
log.Fatal(err)
}
fmt.Println(status)
// Let's create a new tenant
tenantSpec := &photon.TenantCreateSpec{Name: "new-tenant"}
task, err := client.Tenants.Create(tenantSpec)
if err != nil {
log.Fatal(err)
}
// Wait for task completion
task, err = client.Tasks.Wait(task.ID)
if err != nil {
log.Fatal(err)
}
fmt.Printf("ID of new tenant is: %s\n", task.Entity.ID)
}
It should now output this:
&{READY [{PHOTON_CONTROLLER READY}]}
ID of new tenant is: c8989a40-0fa4-4d9a-8e73-2fe4d28d0065