constellation/bootstrapper/internal/kubernetes/kubernetes.go
2024-02-08 14:20:01 +00:00

319 lines
12 KiB
Go

/*
Copyright (c) Edgeless Systems GmbH
SPDX-License-Identifier: AGPL-3.0-only
*/
// Package kubernetes provides functionality to bootstrap a Kubernetes cluster, or join an exiting one.
package kubernetes
import (
"context"
"fmt"
"log/slog"
"net"
"regexp"
"strings"
"time"
"github.com/edgelesssys/constellation/v2/bootstrapper/internal/kubernetes/k8sapi"
"github.com/edgelesssys/constellation/v2/bootstrapper/internal/kubernetes/kubewaiter"
"github.com/edgelesssys/constellation/v2/internal/cloud/cloudprovider"
"github.com/edgelesssys/constellation/v2/internal/constants"
"github.com/edgelesssys/constellation/v2/internal/kubernetes"
"github.com/edgelesssys/constellation/v2/internal/role"
"github.com/edgelesssys/constellation/v2/internal/versions/components"
corev1 "k8s.io/api/core/v1"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
kubeadm "k8s.io/kubernetes/cmd/kubeadm/app/apis/kubeadm/v1beta3"
)
var validHostnameRegex = regexp.MustCompile(`^[a-z0-9]([-a-z0-9]*[a-z0-9])?(\.[a-z0-9]([-a-z0-9]*[a-z0-9])?)*$`)
// configurationProvider provides kubeadm init and join configuration.
type configurationProvider interface {
InitConfiguration(externalCloudProvider bool, k8sVersion string) k8sapi.KubeadmInitYAML
JoinConfiguration(externalCloudProvider bool) k8sapi.KubeadmJoinYAML
}
type kubeAPIWaiter interface {
Wait(ctx context.Context, kubernetesClient kubewaiter.KubernetesClient) error
}
// KubeWrapper implements Cluster interface.
type KubeWrapper struct {
cloudProvider string
clusterUtil clusterUtil
kubeAPIWaiter kubeAPIWaiter
configProvider configurationProvider
client k8sapi.Client
providerMetadata ProviderMetadata
getIPAddr func() (string, error)
}
// New creates a new KubeWrapper with real values.
func New(cloudProvider string, clusterUtil clusterUtil, configProvider configurationProvider, client k8sapi.Client,
providerMetadata ProviderMetadata, kubeAPIWaiter kubeAPIWaiter,
) *KubeWrapper {
return &KubeWrapper{
cloudProvider: cloudProvider,
clusterUtil: clusterUtil,
kubeAPIWaiter: kubeAPIWaiter,
configProvider: configProvider,
client: client,
providerMetadata: providerMetadata,
getIPAddr: getIPAddr,
}
}
// InitCluster initializes a new Kubernetes cluster and applies pod network provider.
func (k *KubeWrapper) InitCluster(
ctx context.Context, versionString, clusterName string, conformanceMode bool, kubernetesComponents components.Components, apiServerCertSANs []string, serviceCIDR string, log *slog.Logger,
) ([]byte, error) {
log.With(slog.String("version", versionString)).Info("Installing Kubernetes components")
if err := k.clusterUtil.InstallComponents(ctx, kubernetesComponents); err != nil {
return nil, err
}
var validIPs []net.IP
// Step 1: retrieve cloud metadata for Kubernetes configuration
log.Info("Retrieving node metadata")
instance, err := k.providerMetadata.Self(ctx)
if err != nil {
return nil, fmt.Errorf("retrieving own instance metadata: %w", err)
}
if instance.VPCIP != "" {
validIPs = append(validIPs, net.ParseIP(instance.VPCIP))
}
nodeName, err := k8sCompliantHostname(instance.Name)
if err != nil {
return nil, fmt.Errorf("generating node name: %w", err)
}
nodeIP := instance.VPCIP
subnetworkPodCIDR := instance.SecondaryIPRange
// this is the endpoint in "kubeadm init --control-plane-endpoint=<IP/DNS>:<port>"
// TODO(malt3): switch over to DNS name on AWS and Azure
// soon as every apiserver certificate of every control-plane node
// has the dns endpoint in its SAN list.
controlPlaneHost, controlPlanePort, err := k.providerMetadata.GetLoadBalancerEndpoint(ctx)
if err != nil {
return nil, fmt.Errorf("retrieving load balancer endpoint: %w", err)
}
certSANs := []string{nodeIP}
certSANs = append(certSANs, apiServerCertSANs...)
log.With(
slog.String("nodeName", nodeName),
slog.String("providerID", instance.ProviderID),
slog.String("nodeIP", nodeIP),
slog.String("controlPlaneHost", controlPlaneHost),
slog.String("controlPlanePort", controlPlanePort),
slog.String("certSANs", strings.Join(certSANs, ",")),
slog.String("podCIDR", subnetworkPodCIDR),
).Info("Setting information for node")
// Step 2: configure kubeadm init config
ccmSupported := cloudprovider.FromString(k.cloudProvider) == cloudprovider.Azure ||
cloudprovider.FromString(k.cloudProvider) == cloudprovider.GCP ||
cloudprovider.FromString(k.cloudProvider) == cloudprovider.AWS
initConfig := k.configProvider.InitConfiguration(ccmSupported, versionString)
initConfig.SetNodeIP(nodeIP)
initConfig.SetClusterName(clusterName)
initConfig.SetCertSANs(certSANs)
initConfig.SetNodeName(nodeName)
initConfig.SetProviderID(instance.ProviderID)
initConfig.SetControlPlaneEndpoint(controlPlaneHost)
initConfig.SetServiceSubnet(serviceCIDR)
initConfigYAML, err := initConfig.Marshal()
if err != nil {
return nil, fmt.Errorf("encoding kubeadm init configuration as YAML: %w", err)
}
log.Info("Initializing Kubernetes cluster")
kubeConfig, err := k.clusterUtil.InitCluster(ctx, initConfigYAML, nodeName, clusterName, validIPs, conformanceMode, log)
if err != nil {
return nil, fmt.Errorf("kubeadm init: %w", err)
}
err = k.client.Initialize(kubeConfig)
if err != nil {
return nil, fmt.Errorf("initializing kubectl client: %w", err)
}
waitCtx, cancel := context.WithTimeout(ctx, 2*time.Minute)
defer cancel()
if err := k.kubeAPIWaiter.Wait(waitCtx, k.client); err != nil {
return nil, fmt.Errorf("waiting for Kubernetes API to be available: %w", err)
}
if err := k.client.EnforceCoreDNSSpread(ctx); err != nil {
return nil, fmt.Errorf("configuring CoreDNS deployment: %w", err)
}
// Setup the K8s components ConfigMap.
k8sComponentsConfigMap, err := k.setupK8sComponentsConfigMap(ctx, kubernetesComponents, versionString)
if err != nil {
return nil, fmt.Errorf("failed to setup k8s version ConfigMap: %w", err)
}
if cloudprovider.FromString(k.cloudProvider) == cloudprovider.GCP {
// GCP uses direct routing, so we need to set the pod CIDR of the first control-plane node for Cilium.
var nodePodCIDR string
if len(instance.AliasIPRanges) > 0 {
nodePodCIDR = instance.AliasIPRanges[0]
}
if err := k.client.PatchFirstNodePodCIDR(ctx, nodePodCIDR); err != nil {
return nil, fmt.Errorf("patching first node pod CIDR: %w", err)
}
}
// Annotate Node with the hash of the installed components
if err := k.client.AnnotateNode(ctx, nodeName,
constants.NodeKubernetesComponentsAnnotationKey, k8sComponentsConfigMap,
); err != nil {
return nil, fmt.Errorf("annotating node with Kubernetes components hash: %w", err)
}
log.Info("Setting up internal-config ConfigMap")
if err := k.setupInternalConfigMap(ctx); err != nil {
return nil, fmt.Errorf("failed to setup internal ConfigMap: %w", err)
}
return kubeConfig, nil
}
// JoinCluster joins existing Kubernetes cluster.
func (k *KubeWrapper) JoinCluster(ctx context.Context, args *kubeadm.BootstrapTokenDiscovery, peerRole role.Role, k8sComponents components.Components, log *slog.Logger) error {
log.With("k8sComponents", k8sComponents).Info("Installing provided kubernetes components")
if err := k.clusterUtil.InstallComponents(ctx, k8sComponents); err != nil {
return fmt.Errorf("installing kubernetes components: %w", err)
}
// Step 1: retrieve cloud metadata for Kubernetes configuration
log.Info("Retrieving node metadata")
instance, err := k.providerMetadata.Self(ctx)
if err != nil {
return fmt.Errorf("retrieving own instance metadata: %w", err)
}
providerID := instance.ProviderID
nodeInternalIP := instance.VPCIP
nodeName, err := k8sCompliantHostname(instance.Name)
if err != nil {
return fmt.Errorf("generating node name: %w", err)
}
loadBalancerHost, loadBalancerPort, err := k.providerMetadata.GetLoadBalancerEndpoint(ctx)
if err != nil {
return fmt.Errorf("retrieving own instance metadata: %w", err)
}
// override join endpoint to go over lb
args.APIServerEndpoint = net.JoinHostPort(loadBalancerHost, loadBalancerPort)
log.With(
slog.String("nodeName", nodeName),
slog.String("providerID", providerID),
slog.String("nodeIP", nodeInternalIP),
slog.String("loadBalancerHost", loadBalancerHost),
slog.String("loadBalancerPort", loadBalancerPort),
).Info("Setting information for node")
// Step 2: configure kubeadm join config
ccmSupported := cloudprovider.FromString(k.cloudProvider) == cloudprovider.Azure ||
cloudprovider.FromString(k.cloudProvider) == cloudprovider.GCP
joinConfig := k.configProvider.JoinConfiguration(ccmSupported)
joinConfig.SetAPIServerEndpoint(args.APIServerEndpoint)
joinConfig.SetToken(args.Token)
joinConfig.AppendDiscoveryTokenCaCertHash(args.CACertHashes[0])
joinConfig.SetNodeIP(nodeInternalIP)
joinConfig.SetNodeName(nodeName)
joinConfig.SetProviderID(providerID)
if peerRole == role.ControlPlane {
joinConfig.SetControlPlane(nodeInternalIP)
}
joinConfigYAML, err := joinConfig.Marshal()
if err != nil {
return fmt.Errorf("encoding kubeadm join configuration as YAML: %w", err)
}
log.With(slog.String("apiServerEndpoint", args.APIServerEndpoint)).Info("Joining Kubernetes cluster")
if err := k.clusterUtil.JoinCluster(ctx, joinConfigYAML, log); err != nil {
return fmt.Errorf("joining cluster: %v; %w ", string(joinConfigYAML), err)
}
return nil
}
// setupK8sComponentsConfigMap applies a ConfigMap (cf. server-side apply) to store the installed k8s components.
// It returns the name of the ConfigMap.
func (k *KubeWrapper) setupK8sComponentsConfigMap(ctx context.Context, components components.Components, clusterVersion string) (string, error) {
componentsConfig, err := kubernetes.ConstructK8sComponentsCM(components, clusterVersion)
if err != nil {
return "", fmt.Errorf("constructing k8s-components ConfigMap: %w", err)
}
if err := k.client.CreateConfigMap(ctx, &componentsConfig); err != nil {
return "", fmt.Errorf("apply in KubeWrapper.setupK8sVersionConfigMap(..) for components config map failed with: %w", err)
}
return componentsConfig.ObjectMeta.Name, nil
}
// setupInternalConfigMap applies a ConfigMap (cf. server-side apply) to store information that is not supposed to be user-editable.
func (k *KubeWrapper) setupInternalConfigMap(ctx context.Context) error {
config := corev1.ConfigMap{
TypeMeta: metav1.TypeMeta{
APIVersion: "v1",
Kind: "ConfigMap",
},
ObjectMeta: metav1.ObjectMeta{
Name: constants.InternalConfigMap,
Namespace: "kube-system",
},
Data: map[string]string{},
}
// We do not use the client's Apply method here since we are handling a kubernetes-native type.
// These types don't implement our custom Marshaler interface.
if err := k.client.CreateConfigMap(ctx, &config); err != nil {
return fmt.Errorf("apply in KubeWrapper.setupInternalConfigMap failed with: %w", err)
}
return nil
}
// k8sCompliantHostname transforms a hostname to an RFC 1123 compliant, lowercase subdomain as required by Kubernetes node names.
// The following regex is used by k8s for validation: /^[a-z0-9]([-a-z0-9]*[a-z0-9])?(\.[a-z0-9]([-a-z0-9]*[a-z0-9])?)*$/ .
// Only a simple heuristic is used for now (to lowercase, replace underscores).
func k8sCompliantHostname(in string) (string, error) {
hostname := strings.ToLower(in)
hostname = strings.ReplaceAll(hostname, "_", "-")
if !validHostnameRegex.MatchString(hostname) {
return "", fmt.Errorf("failed to generate a Kubernetes compliant hostname for %s", in)
}
return hostname, nil
}
// StartKubelet starts the kubelet service.
func (k *KubeWrapper) StartKubelet() error {
if err := k.clusterUtil.StartKubelet(); err != nil {
return fmt.Errorf("starting kubelet: %w", err)
}
return nil
}
// getIPAddr retrieves to default sender IP used for outgoing connection.
func getIPAddr() (string, error) {
conn, err := net.Dial("udp", "8.8.8.8:80")
if err != nil {
return "", err
}
defer conn.Close()
localAddr := conn.LocalAddr().(*net.UDPAddr)
return localAddr.IP.String(), nil
}