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

Jump to bottom

🌱 Implement topology validations for topology kubernetes version upgrades #10063

Merged
merged 2 commits into from
Feb 20, 2024
Merged

🌱 Implement topology validations for topology kubernetes version upgrades #10063

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
da4fa22
topology: implement version validations (8 days ago) <Christian Schlo…
chrischdi Jan 24, 2024
09f3520
review fixes
chrischdi Feb 20, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
4 changes: 4 additions & 0 deletions api/v1beta1/common_types.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -68,6 +68,10 @@ const (
// update that disallows a pre-existing Cluster to be populated with Topology information and Class.
ClusterTopologyUnsafeUpdateClassNameAnnotation = "unsafe.topology.cluster.x-k8s.io/disable-update-class-name-check"

// ClusterTopologyUnsafeUpdateVersionAnnotation can be used to disable the webhook checks on
// update that disallows updating the .topology.spec.version on certain conditions.
ClusterTopologyUnsafeUpdateVersionAnnotation = "unsafe.topology.cluster.x-k8s.io/disable-update-version-check"
chrischdi marked this conversation as resolved.
Show resolved Hide resolved

// ProviderNameLabel is the label set on components in the provider manifest.
// This label allows to easily identify all the components belonging to a provider; the clusterctl
// tool uses this label for implementing provider's lifecycle operations.
Expand Down
5 changes: 5 additions & 0 deletions controllers/remote/cluster_cache_tracker.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -189,6 +189,11 @@ func (t *ClusterCacheTracker) GetClient(ctx context.Context, cluster client.Obje
return accessor.client, nil
}

// GetReader returns a cached read-only client for the given cluster.
func (t *ClusterCacheTracker) GetReader(ctx context.Context, cluster client.ObjectKey) (client.Reader, error) {
sbueringer marked this conversation as resolved.
Show resolved Hide resolved
return t.GetClient(ctx, cluster)
}

// GetRESTConfig returns a cached REST config for the given cluster.
func (t *ClusterCacheTracker) GetRESTConfig(ctc context.Context, cluster client.ObjectKey) (*rest.Config, error) {
accessor, err := t.getClusterAccessor(ctc, cluster, t.indexes...)
Expand Down
5 changes: 3 additions & 2 deletions docs/book/src/reference/labels_and_annotations.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@


| Label | Note |
| :---------------------------------------- | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|:------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| cluster.x-k8s.io/cluster-name | It is set on machines linked to a cluster and external objects(bootstrap and infrastructure providers). |
| topology.cluster.x-k8s.io/owned | It is set on all the object which are managed as part of a ClusterTopology. |
| topology.cluster.x-k8s.io/deployment-name | It is set on the generated MachineDeployment objects to track the name of the MachineDeployment topology it represents. |
Expand All @@ -21,11 +21,12 @@
**Supported Annotations:**

| Annotation | Note |
| :--------------------------------------------------------------- | :---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|:-----------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| clusterctl.cluster.x-k8s.io/skip-crd-name-preflight-check | Can be placed on provider CRDs, so that clusterctl doesn't emit an error if the CRD doesn't comply with Cluster APIs naming scheme. Only CRDs that are referenced by core Cluster API CRDs have to comply with the naming scheme. |
| clusterctl.cluster.x-k8s.io/delete-for-move | DeleteForMoveAnnotation will be set to objects that are going to be deleted from the source cluster after being moved to the target cluster during the clusterctl move operation. It will help any validation webhook to take decision based on it. |
| clusterctl.cluster.x-k8s.io/block-move | BlockMoveAnnotation prevents the cluster move operation from starting if it is defined on at least one of the objects in scope. Provider controllers are expected to set the annotation on resources that cannot be instantaneously paused and remove the annotation when the resource has been actually paused. |
| unsafe.topology.cluster.x-k8s.io/disable-update-class-name-check | It can be used to disable the webhook check on update that disallows a pre-existing Cluster to be populated with Topology information and Class. |
| unsafe.topology.cluster.x-k8s.io/disable-update-version-check | It can be used to disable the webhook checks on update that disallows updating the `.topology.spec.version` on certain conditions. |
| cluster.x-k8s.io/cluster-name | It is set on nodes identifying the name of the cluster the node belongs to. |
| cluster.x-k8s.io/cluster-namespace | It is set on nodes identifying the namespace of the cluster the node belongs to. |
| cluster.x-k8s.io/machine | It is set on nodes identifying the machine the node belongs to. |
Expand Down
56 changes: 7 additions & 49 deletions internal/controllers/topology/cluster/scope/state.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,16 +18,14 @@ package scope

import (
"context"
"fmt"

"github.com/pkg/errors"
corev1 "k8s.io/api/core/v1"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/apis/meta/v1/unstructured"
"sigs.k8s.io/controller-runtime/pkg/client"

clusterv1 "sigs.k8s.io/cluster-api/api/v1beta1"
expv1 "sigs.k8s.io/cluster-api/exp/api/v1beta1"
"sigs.k8s.io/cluster-api/internal/topology/check"
)

// ClusterState holds all the objects representing the state of a managed Cluster topology.
Expand Down Expand Up @@ -68,7 +66,7 @@ type MachineDeploymentsStateMap map[string]*MachineDeploymentState

// Upgrading returns the list of the machine deployments
// that are upgrading.
func (mds MachineDeploymentsStateMap) Upgrading(ctx context.Context, c client.Client) ([]string, error) {
func (mds MachineDeploymentsStateMap) Upgrading(ctx context.Context, c client.Reader) ([]string, error) {
names := []string{}
for _, md := range mds {
upgrading, err := md.IsUpgrading(ctx, c)
Expand Down Expand Up @@ -101,40 +99,16 @@ type MachineDeploymentState struct {
// IsUpgrading determines if the MachineDeployment is upgrading.
// A machine deployment is considered upgrading if at least one of the Machines of this
// MachineDeployment has a different version.
func (md *MachineDeploymentState) IsUpgrading(ctx context.Context, c client.Client) (bool, error) {
// If the MachineDeployment has no version there is no definitive way to check if it is upgrading. Therefore, return false.
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This code got moved to internal/topology/check/upgrade.go

// Note: This case should not happen.
if md.Object.Spec.Template.Spec.Version == nil {
return false, nil
}
selectorMap, err := metav1.LabelSelectorAsMap(&md.Object.Spec.Selector)
if err != nil {
return false, errors.Wrapf(err, "failed to check if MachineDeployment %s is upgrading: failed to convert label selector to map", md.Object.Name)
}
machines := &clusterv1.MachineList{}
if err := c.List(ctx, machines, client.InNamespace(md.Object.Namespace), client.MatchingLabels(selectorMap)); err != nil {
return false, errors.Wrapf(err, "failed to check if MachineDeployment %s is upgrading: failed to list Machines", md.Object.Name)
}
mdVersion := *md.Object.Spec.Template.Spec.Version
// Check if the versions of the all the Machines match the MachineDeployment version.
for i := range machines.Items {
machine := machines.Items[i]
if machine.Spec.Version == nil {
return false, fmt.Errorf("failed to check if MachineDeployment %s is upgrading: Machine %s has no version", md.Object.Name, machine.Name)
}
if *machine.Spec.Version != mdVersion {
return true, nil
}
}
return false, nil
func (md *MachineDeploymentState) IsUpgrading(ctx context.Context, c client.Reader) (bool, error) {
return check.IsMachineDeploymentUpgrading(ctx, c, md.Object)
}

// MachinePoolsStateMap holds a collection of MachinePool states.
type MachinePoolsStateMap map[string]*MachinePoolState

// Upgrading returns the list of the machine pools
// that are upgrading.
func (mps MachinePoolsStateMap) Upgrading(ctx context.Context, c client.Client) ([]string, error) {
func (mps MachinePoolsStateMap) Upgrading(ctx context.Context, c client.Reader) ([]string, error) {
names := []string{}
for _, mp := range mps {
upgrading, err := mp.IsUpgrading(ctx, c)
Expand Down Expand Up @@ -163,22 +137,6 @@ type MachinePoolState struct {
// IsUpgrading determines if the MachinePool is upgrading.
// A machine pool is considered upgrading if at least one of the Machines of this
// MachinePool has a different version.
func (mp *MachinePoolState) IsUpgrading(ctx context.Context, c client.Client) (bool, error) {
// If the MachinePool has no version there is no definitive way to check if it is upgrading. Therefore, return false.
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This code got moved to internal/topology/check/upgrade.go

// Note: This case should not happen.
if mp.Object.Spec.Template.Spec.Version == nil {
return false, nil
}
mpVersion := *mp.Object.Spec.Template.Spec.Version
// Check if the kubelet versions of the MachinePool noderefs match the MachinePool version.
for _, nodeRef := range mp.Object.Status.NodeRefs {
node := &corev1.Node{}
if err := c.Get(ctx, client.ObjectKey{Name: nodeRef.Name}, node); err != nil {
return false, fmt.Errorf("failed to check if MachinePool %s is upgrading: failed to get Node %s", mp.Object.Name, nodeRef.Name)
}
if mpVersion != node.Status.NodeInfo.KubeletVersion {
return true, nil
}
}
return false, nil
func (mp *MachinePoolState) IsUpgrading(ctx context.Context, c client.Reader) (bool, error) {
return check.IsMachinePoolUpgrading(ctx, c, mp.Object)
}
86 changes: 0 additions & 86 deletions internal/controllers/topology/cluster/scope/state_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,92 +29,6 @@ import (
"sigs.k8s.io/cluster-api/internal/test/builder"
)

func TestIsUpgrading(t *testing.T) {
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This tests got moved to internal/topology/check/upgrade_test.go

g := NewWithT(t)
scheme := runtime.NewScheme()
g.Expect(clusterv1.AddToScheme(scheme)).To(Succeed())

tests := []struct {
name string
md *clusterv1.MachineDeployment
machines []*clusterv1.Machine
want bool
wantErr bool
}{
{
name: "should return false if all the machines of MachineDeployment have the same version as the MachineDeployment",
md: builder.MachineDeployment("ns", "md1").
WithClusterName("cluster1").
WithVersion("v1.2.3").
Build(),
machines: []*clusterv1.Machine{
builder.Machine("ns", "machine1").
WithClusterName("cluster1").
WithVersion("v1.2.3").
Build(),
builder.Machine("ns", "machine2").
WithClusterName("cluster1").
WithVersion("v1.2.3").
Build(),
},
want: false,
wantErr: false,
},
{
name: "should return true if at least one of the machines of MachineDeployment has a different version",
md: builder.MachineDeployment("ns", "md1").
WithClusterName("cluster1").
WithVersion("v1.2.3").
Build(),
machines: []*clusterv1.Machine{
builder.Machine("ns", "machine1").
WithClusterName("cluster1").
WithVersion("v1.2.3").
Build(),
builder.Machine("ns", "machine2").
WithClusterName("cluster1").
WithVersion("v1.2.2").
Build(),
},
want: true,
wantErr: false,
},
{
name: "should return false if the MachineDeployment has no machines (creation phase)",
md: builder.MachineDeployment("ns", "md1").
WithClusterName("cluster1").
WithVersion("v1.2.3").
Build(),
machines: []*clusterv1.Machine{},
want: false,
wantErr: false,
},
}

for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
g := NewWithT(t)
ctx := context.Background()
objs := []client.Object{}
objs = append(objs, tt.md)
for _, m := range tt.machines {
objs = append(objs, m)
}
fakeClient := fake.NewClientBuilder().WithScheme(scheme).WithObjects(objs...).Build()
mdState := &MachineDeploymentState{
Object: tt.md,
}
got, err := mdState.IsUpgrading(ctx, fakeClient)
if tt.wantErr {
g.Expect(err).To(HaveOccurred())
} else {
g.Expect(err).ToNot(HaveOccurred())
g.Expect(got).To(Equal(tt.want))
}
})
}
}

func TestUpgrading(t *testing.T) {
g := NewWithT(t)
scheme := runtime.NewScheme()
Expand Down