Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Automated cherry pick of #7925: Extract the list of instance groups earlier in validation #8039: Vendor github.com/stretchr/testify/require #8159: Determine node role from instancegroup spec #8600: Fail cluster validation if a master missing #8606

Merged
merged 8 commits into from
Feb 25, 2020
Merged

Automated cherry pick of #7925: Extract the list of instance groups earlier in validation #8039: Vendor github.com/stretchr/testify/require #8159: Determine node role from instancegroup spec #8600: Fail cluster validation if a master missing #8606

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
26b246d
Extract the list of instance groups earlier in validation
johngmyers Nov 11, 2019
b33202e
Pass the cloud object to validator from caller
johngmyers Nov 11, 2019
d88a32c
Vendor github.com/stretchr/testify/require
johngmyers Nov 16, 2019
964229e
Test validation through the public interface
johngmyers Nov 14, 2019
c11833c
Make validation assertions more concise
johngmyers Nov 16, 2019
04274b9
Increase validation test coverage
johngmyers Nov 17, 2019
206782a
Determine node role from instancegroup spec
johngmyers Dec 19, 2019
eea95d6
Fail cluster validation if a master missing kube-controller-manager
johngmyers Feb 21, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion cmd/kops/rollingupdatecluster.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -392,7 +392,7 @@ func RunRollingUpdateCluster(f *util.Factory, out io.Writer, options *RollingUpd
if featureflag.DrainAndValidateRollingUpdate.Enabled() {
klog.V(2).Infof("Rolling update with drain and validate enabled.")
if !options.CloudOnly {
clusterValidator, err = validation.NewClusterValidator(cluster, list, k8sClient)
clusterValidator, err = validation.NewClusterValidator(cluster, cloud, list, k8sClient)
if err != nil {
return fmt.Errorf("cannot create cluster validator: %v", err)
}
Expand Down
9 changes: 8 additions & 1 deletion cmd/kops/validate_cluster.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,8 @@ import (
"strings"
"time"

"k8s.io/kops/upup/pkg/fi/cloudup"

"github.com/ghodss/yaml"
"github.com/spf13/cobra"
v1 "k8s.io/api/core/v1"
Expand Down Expand Up @@ -93,6 +95,11 @@ func RunValidateCluster(f *util.Factory, cmd *cobra.Command, args []string, out
return nil, err
}

cloud, err := cloudup.BuildCloud(cluster)
if err != nil {
return nil, err
}

clientSet, err := f.Clientset()
if err != nil {
return nil, err
Expand Down Expand Up @@ -134,7 +141,7 @@ func RunValidateCluster(f *util.Factory, cmd *cobra.Command, args []string, out
timeout := time.Now().Add(options.wait)
pollInterval := 10 * time.Second

validator, err := validation.NewClusterValidator(cluster, list, k8sClient)
validator, err := validation.NewClusterValidator(cluster, cloud, list, k8sClient)
if err != nil {
return nil, fmt.Errorf("unexpected error creating validatior: %v", err)
}
Expand Down
9 changes: 6 additions & 3 deletions pkg/validation/BUILD.bazel
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,10 +10,9 @@ go_library(
visibility = ["//visibility:public"],
deps = [
"//pkg/apis/kops:go_default_library",
"//pkg/apis/kops/util:go_default_library",
"//pkg/cloudinstances:go_default_library",
"//pkg/dns:go_default_library",
"//upup/pkg/fi/cloudup:go_default_library",
"//upup/pkg/fi:go_default_library",
"//vendor/k8s.io/api/core/v1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/apis/meta/v1:go_default_library",
"//vendor/k8s.io/client-go/kubernetes:go_default_library",
Expand All @@ -29,9 +28,13 @@ go_test(
deps = [
"//pkg/apis/kops:go_default_library",
"//pkg/cloudinstances:go_default_library",
"//upup/pkg/fi:go_default_library",
"//upup/pkg/fi/cloudup/awsup:go_default_library",
"//vendor/github.com/stretchr/testify/assert:go_default_library",
"//vendor/github.com/stretchr/testify/require:go_default_library",
"//vendor/k8s.io/api/core/v1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/apis/meta/v1:go_default_library",
"//vendor/k8s.io/client-go/kubernetes:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/runtime:go_default_library",
"//vendor/k8s.io/client-go/kubernetes/fake:go_default_library",
],
)
103 changes: 60 additions & 43 deletions pkg/validation/validate_cluster.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,16 +22,16 @@ import (
"net/url"
"strings"

"k8s.io/kops/upup/pkg/fi"

v1 "k8s.io/api/core/v1"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/client-go/kubernetes"
"k8s.io/client-go/tools/clientcmd"
"k8s.io/klog"
"k8s.io/kops/pkg/apis/kops"
"k8s.io/kops/pkg/apis/kops/util"
"k8s.io/kops/pkg/cloudinstances"
"k8s.io/kops/pkg/dns"
"k8s.io/kops/upup/pkg/fi/cloudup"
)

// ValidationCluster uses a cluster to validate.
Expand All @@ -54,9 +54,10 @@ type ClusterValidator interface {
}

type clusterValidatorImpl struct {
cluster *kops.Cluster
instanceGroupList *kops.InstanceGroupList
k8sClient kubernetes.Interface
cluster *kops.Cluster
cloud fi.Cloud
instanceGroups []*kops.InstanceGroup
k8sClient kubernetes.Interface
}

func (v *ValidationCluster) addError(failure *ValidationError) {
Expand Down Expand Up @@ -100,23 +101,30 @@ func hasPlaceHolderIP(clusterName string) (bool, error) {
return false, nil
}

func NewClusterValidator(cluster *kops.Cluster, instanceGroupList *kops.InstanceGroupList, k8sClient kubernetes.Interface) (ClusterValidator, error) {
func NewClusterValidator(cluster *kops.Cluster, cloud fi.Cloud, instanceGroupList *kops.InstanceGroupList, k8sClient kubernetes.Interface) (ClusterValidator, error) {
var instanceGroups []*kops.InstanceGroup

for i := range instanceGroupList.Items {
ig := &instanceGroupList.Items[i]
instanceGroups = append(instanceGroups, ig)
}

if len(instanceGroups) == 0 {
return nil, fmt.Errorf("no InstanceGroup objects found")
}

return &clusterValidatorImpl{
cluster: cluster,
instanceGroupList: instanceGroupList,
k8sClient: k8sClient,
cluster: cluster,
cloud: cloud,
instanceGroups: instanceGroups,
k8sClient: k8sClient,
}, nil
}

func (v *clusterValidatorImpl) Validate() (*ValidationCluster, error) {
return validateCluster(v.cluster, v.instanceGroupList, v.k8sClient)
}
clusterName := v.cluster.Name

// validateCluster validates a k8s cluster with a provided instance group list
func validateCluster(cluster *kops.Cluster, instanceGroupList *kops.InstanceGroupList, k8sClient kubernetes.Interface) (*ValidationCluster, error) {
clusterName := cluster.Name

v := &ValidationCluster{}
validation := &ValidationCluster{}

// Do not use if we are running gossip
if !dns.IsGossipHostname(clusterName) {
Expand All @@ -134,52 +142,36 @@ func validateCluster(cluster *kops.Cluster, instanceGroupList *kops.InstanceGrou
" Please wait about 5-10 minutes for a master to start, dns-controller to launch, and DNS to propagate." +
" The protokube container and dns-controller deployment logs may contain more diagnostic information." +
" Etcd and the API DNS entries must be updated for a kops Kubernetes cluster to start."
v.addError(&ValidationError{
validation.addError(&ValidationError{
Kind: "dns",
Name: "apiserver",
Message: message,
})
return v, nil
return validation, nil
}
}

var instanceGroups []*kops.InstanceGroup

for i := range instanceGroupList.Items {
ig := &instanceGroupList.Items[i]
instanceGroups = append(instanceGroups, ig)
}

if len(instanceGroups) == 0 {
return nil, fmt.Errorf("no InstanceGroup objects found")
}

cloud, err := cloudup.BuildCloud(cluster)
if err != nil {
return nil, err
}

nodeList, err := k8sClient.CoreV1().Nodes().List(metav1.ListOptions{})
nodeList, err := v.k8sClient.CoreV1().Nodes().List(metav1.ListOptions{})
if err != nil {
return nil, fmt.Errorf("error listing nodes: %v", err)
}

warnUnmatched := false
cloudGroups, err := cloud.GetCloudGroups(cluster, instanceGroups, warnUnmatched, nodeList.Items)
cloudGroups, err := v.cloud.GetCloudGroups(v.cluster, v.instanceGroups, warnUnmatched, nodeList.Items)
if err != nil {
return nil, err
}
v.validateNodes(cloudGroups)
validation.validateNodes(cloudGroups)

if err := v.collectComponentFailures(k8sClient); err != nil {
if err := validation.collectComponentFailures(v.k8sClient); err != nil {
return nil, fmt.Errorf("cannot get component status for %q: %v", clusterName, err)
}

if err = v.collectPodFailures(k8sClient); err != nil {
if err := validation.collectPodFailures(v.k8sClient, nodeList.Items); err != nil {
return nil, fmt.Errorf("cannot get pod health for %q: %v", clusterName, err)
}

return v, nil
return validation, nil
}

func (v *ValidationCluster) collectComponentFailures(client kubernetes.Interface) error {
Expand All @@ -202,12 +194,24 @@ func (v *ValidationCluster) collectComponentFailures(client kubernetes.Interface
return nil
}

func (v *ValidationCluster) collectPodFailures(client kubernetes.Interface) error {
func (v *ValidationCluster) collectPodFailures(client kubernetes.Interface, nodes []v1.Node) error {
pods, err := client.CoreV1().Pods("kube-system").List(metav1.ListOptions{})
if err != nil {
return fmt.Errorf("error listing Pods: %v", err)
}

masterWithoutManager := map[string]bool{}
nodeByAddress := map[string]string{}
for _, node := range nodes {
labels := node.GetLabels()
if labels != nil && labels["kubernetes.io/role"] == "master" {
masterWithoutManager[node.Name] = true
}
for _, nodeAddress := range node.Status.Addresses {
nodeByAddress[nodeAddress.Address] = node.Name
}
}

for _, pod := range pods.Items {
if pod.Status.Phase == v1.PodSucceeded {
continue
Expand All @@ -234,7 +238,21 @@ func (v *ValidationCluster) collectPodFailures(client kubernetes.Interface) erro
})

}

labels := pod.GetLabels()
if pod.Namespace == "kube-system" && labels != nil && labels["k8s-app"] == "kube-controller-manager" {
delete(masterWithoutManager, nodeByAddress[pod.Status.HostIP])
}
}

for node := range masterWithoutManager {
v.addError(&ValidationError{
Kind: "Node",
Name: node,
Message: fmt.Sprintf("master %q is missing kube-controller-manager pod", node),
})
}

return nil
}

Expand Down Expand Up @@ -274,7 +292,7 @@ func (v *ValidationCluster) validateNodes(cloudGroups map[string]*cloudinstances
continue
}

role := util.GetNodeRole(node)
role := strings.ToLower(string(cloudGroup.InstanceGroup.Spec.Role))
if role == "" {
role = "node"
}
Expand All @@ -289,7 +307,6 @@ func (v *ValidationCluster) validateNodes(cloudGroups map[string]*cloudinstances

ready := isNodeReady(node)

// TODO: Use instance group role instead...
if n.Role == "master" {
if !ready {
v.addError(&ValidationError{
Expand Down
Loading