Fixes #40819

Start looking up the virtual machine by it's UUID in vSphere again. Looking up by IP address is problematic and can either not return a VM entirely, or could return the wrong VM.

Retrieves the VM's UUID in one of two methods - either by a `vm-uuid` entry in the cloud config file on the VM, or via sysfs. The sysfs route requires root access, but restores the previous functionality.

Multiple VMs in a vCenter cluster can share an IP address - for example, if you have multiple VM networks, but they're all isolated and use the same address range. Additionally, flannel network address ranges can overlap.

vSphere seems to have a limitation of reporting no more than 16 interfaces from a virtual machine, so it's possible that the IP address list on a VM is completely untrustworthy anyhow - it can either be empty (because the 16 interfaces it found were veth interfaces with no IP address), or it can report the flannel IP.

pkg/cloudprovider/providers/vsphere/vsphere.go

@@ -20,7 +20,7 @@ import (
 	"errors"
 	"fmt"
 	"io"
-	"net"
+	"io/ioutil"
 	"net/url"
 	"path"
 	"path/filepath"
@@ -122,12 +122,17 @@ type VSphereConfig struct {
 		WorkingDir string `gcfg:"working-dir"`
 		// Soap round tripper count (retries = RoundTripper - 1)
 		RoundTripperCount uint `gcfg:"soap-roundtrip-count"`
+		// VMUUID is the VM Instance UUID of virtual machine which can be retrieved from instanceUuid
+		// property in VmConfigInfo, or also set as vc.uuid in VMX file.
+		// If not set, will be fetched from the machine via sysfs (requires root)
+		VMUUID string `gcfg:"vm-uuid"`
 	}
 
 	Network struct {
 		// PublicNetwork is name of the network the VMs are joined to.
 		PublicNetwork string `gcfg:"public-network"`
 	}
+
 	Disk struct {
 		// SCSIControllerType defines SCSI controller to be used.
 		SCSIControllerType string `dcfg:"scsicontrollertype"`
@@ -201,14 +206,27 @@ func init() {
 
 // Returns the name of the VM on which this code is running.
 // Prerequisite: this code assumes VMWare vmtools or open-vm-tools to be installed in the VM.
+// Will attempt to determine the machine's name via it's UUID in this precedence order, failing if neither have a UUID:
+// * cloud config value VMUUID
+// * sysfs entry
 func getVMName(client *govmomi.Client, cfg *VSphereConfig) (string, error) {
-	addrs, err := net.InterfaceAddrs()
+	var vmUUID string
+
+	if cfg.Global.VMUUID != "" {
+		vmUUID = cfg.Global.VMUUID
+	} else {
+		// This needs root privileges on the host, and will fail otherwise.
+		vmUUIDbytes, err := ioutil.ReadFile("/sys/devices/virtual/dmi/id/product_uuid")
 		if err != nil {
 			return "", err
 		}
 
-	if len(addrs) == 0 {
-		return "", fmt.Errorf("unable to retrieve Instance ID")
+		vmUUID = string(vmUUIDbytes)
+		cfg.Global.VMUUID = vmUUID
+	}
+
+	if vmUUID == "" {
+		return "", fmt.Errorf("unable to determine machine ID from cloud configuration or sysfs")
 	}
 
 	// Create context
@@ -227,28 +245,17 @@ func getVMName(client *govmomi.Client, cfg *VSphereConfig) (string, error) {
 
 	s := object.NewSearchIndex(client.Client)
 
-	var svm object.Reference
-	for _, v := range addrs {
-		ip, _, err := net.ParseCIDR(v.String())
-		if err != nil {
-			return "", fmt.Errorf("unable to parse cidr from ip")
-		}
-
-		// Finds a virtual machine or host by IP address.
-		svm, err = s.FindByIp(ctx, dc, ip.String(), true)
-		if err == nil && svm != nil {
-			break
-		}
-	}
-	if svm == nil {
-		return "", fmt.Errorf("unable to retrieve vm reference from vSphere")
-	}
-
-	var vm mo.VirtualMachine
-	err = s.Properties(ctx, svm.Reference(), []string{"name", "resourcePool"}, &vm)
+	svm, err := s.FindByUuid(ctx, dc, strings.ToLower(strings.TrimSpace(vmUUID)), true, nil)
 	if err != nil {
 		return "", err
 	}
+
+	var vm mo.VirtualMachine
+	err = s.Properties(ctx, svm.Reference(), []string{"name"}, &vm)
+	if err != nil {
+		return "", err
+	}
+
 	return vm.Name, nil
 }
 

pkg/cloudprovider/providers/vsphere/vsphere_test.go

@@ -71,6 +71,7 @@ user = user
 password = password
 insecure-flag = true
 datacenter = us-west
+vm-uuid = 1234
 `))
 	if err != nil {
 		t.Fatalf("Should succeed when a valid config is provided: %s", err)
@@ -87,6 +88,10 @@ datacenter = us-west
 	if cfg.Global.Datacenter != "us-west" {
 		t.Errorf("incorrect datacenter: %s", cfg.Global.Datacenter)
 	}
+
+	if cfg.Global.VMUUID != "1234" {
+		t.Errorf("incorrect vm-uuid: %s", cfg.Global.VMUUID)
+	}
 }
 
 func TestNewVSphere(t *testing.T) {