Merge pull request #4998 from serngawy/mgmtConfig

✨ ROSA: Add upgrade config to ROSAMachinePool
kubernetes-sigs · Jun 10, 2024 · f04cb3e · f04cb3e
2 parents 3a28a4d + 6d42d41
commit f04cb3e
Show file tree

Hide file tree

Showing 7 changed files with 195 additions and 3 deletions.
diff --git a/config/crd/bases/infrastructure.cluster.x-k8s.io_rosamachinepools.yaml b/config/crd/bases/infrastructure.cluster.x-k8s.io_rosamachinepools.yaml
@@ -161,6 +161,59 @@ spec:
                 items:
                   type: string
                 type: array
+              updateConfig:
+                description: UpdateConfig specifies update configurations.
+                properties:
+                  rollingUpdate:
+                    description: RollingUpdate specifies MaxUnavailable & MaxSurge
+                      number of nodes during update.
+                    properties:
+                      maxSurge:
+                        anyOf:
+                        - type: integer
+                        - type: string
+                        default: 1
+                        description: |-
+                          MaxSurge is the maximum number of nodes that can be provisioned above the desired number of nodes.
+                          Value can be an absolute number (ex: 5) or a percentage of desired nodes (ex: 10%).
+                          Absolute number is calculated from percentage by rounding up.
+
+
+                          MaxSurge can not be 0 if MaxUnavailable is 0, default is 1.
+                          Both MaxSurge & MaxUnavailable must use the same units (absolute value or percentage).
+
+
+                          Example: when MaxSurge is set to 30%, new nodes can be provisioned immediately
+                          when the rolling update starts, such that the total number of old and new
+                          nodes do not exceed 130% of desired nodes. Once old nodes have been
+                          deleted, new nodes can be provisioned, ensuring that total number of nodes
+                          running at any time during the update is at most 130% of desired nodes.
+                        pattern: ^((100|[0-9]{1,2})%|[0-9]+)$
+                        x-kubernetes-int-or-string: true
+                      maxUnavailable:
+                        anyOf:
+                        - type: integer
+                        - type: string
+                        default: 0
+                        description: |-
+                          MaxUnavailable is the maximum number of nodes that can be unavailable during the update.
+                          Value can be an absolute number (ex: 5) or a percentage of desired nodes (ex: 10%).
+                          Absolute number is calculated from percentage by rounding down.
+
+
+                          MaxUnavailable can not be 0 if MaxSurge is 0, default is 0.
+                          Both MaxUnavailable & MaxSurge must use the same units (absolute value or percentage).
+
+
+                          Example: when MaxUnavailable is set to 30%, old nodes can be deleted down to 70% of
+                          desired nodes immediately when the rolling update starts. Once new nodes
+                          are ready, more old nodes be deleted, followed by provisioning new nodes,
+                          ensuring that the total number of nodes available at all times during the
+                          update is at least 70% of desired nodes.
+                        pattern: ^((100|[0-9]{1,2})%|[0-9]+)$
+                        x-kubernetes-int-or-string: true
+                    type: object
+                type: object
               version:
                 description: |-
                   Version specifies the OpenShift version of the nodes associated with this machinepool.

diff --git a/exp/api/v1beta2/rosamachinepool_types.go b/exp/api/v1beta2/rosamachinepool_types.go
@@ -19,6 +19,7 @@ package v1beta2
 import (
 	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/util/intstr"
 
 	infrav1 "sigs.k8s.io/cluster-api-provider-aws/v2/api/v1beta2"
 	clusterv1 "sigs.k8s.io/cluster-api/api/v1beta1"
@@ -105,6 +106,11 @@ type RosaMachinePoolSpec struct {
 	//
 	// +optional
 	NodeDrainGracePeriod *metav1.Duration `json:"nodeDrainGracePeriod,omitempty"`
+
+	// UpdateConfig specifies update configurations.
+	//
+	// +optional
+	UpdateConfig *RosaUpdateConfig `json:"updateConfig,omitempty"`
 }
 
 // RosaTaint represents a taint to be applied to a node.
@@ -134,6 +140,55 @@ type RosaMachinePoolAutoScaling struct {
 	MaxReplicas int `json:"maxReplicas,omitempty"`
 }
 
+// RosaUpdateConfig specifies update configuration
+type RosaUpdateConfig struct {
+	// RollingUpdate specifies MaxUnavailable & MaxSurge number of nodes during update.
+	//
+	// +optional
+	RollingUpdate *RollingUpdate `json:"rollingUpdate,omitempty"`
+}
+
+// RollingUpdate specifies MaxUnavailable & MaxSurge number of nodes during update.
+type RollingUpdate struct {
+	// MaxUnavailable is the maximum number of nodes that can be unavailable during the update.
+	// Value can be an absolute number (ex: 5) or a percentage of desired nodes (ex: 10%).
+	// Absolute number is calculated from percentage by rounding down.
+	//
+	// MaxUnavailable can not be 0 if MaxSurge is 0, default is 0.
+	// Both MaxUnavailable & MaxSurge must use the same units (absolute value or percentage).
+	//
+	// Example: when MaxUnavailable is set to 30%, old nodes can be deleted down to 70% of
+	// desired nodes immediately when the rolling update starts. Once new nodes
+	// are ready, more old nodes be deleted, followed by provisioning new nodes,
+	// ensuring that the total number of nodes available at all times during the
+	// update is at least 70% of desired nodes.
+	//
+	// +kubebuilder:validation:Pattern="^((100|[0-9]{1,2})%|[0-9]+)$"
+	// +kubebuilder:validation:XIntOrString
+	// +kubebuilder:default=0
+	// +optional
+	MaxUnavailable *intstr.IntOrString `json:"maxUnavailable,omitempty"`
+
+	// MaxSurge is the maximum number of nodes that can be provisioned above the desired number of nodes.
+	// Value can be an absolute number (ex: 5) or a percentage of desired nodes (ex: 10%).
+	// Absolute number is calculated from percentage by rounding up.
+	//
+	// MaxSurge can not be 0 if MaxUnavailable is 0, default is 1.
+	// Both MaxSurge & MaxUnavailable must use the same units (absolute value or percentage).
+	//
+	// Example: when MaxSurge is set to 30%, new nodes can be provisioned immediately
+	// when the rolling update starts, such that the total number of old and new
+	// nodes do not exceed 130% of desired nodes. Once old nodes have been
+	// deleted, new nodes can be provisioned, ensuring that total number of nodes
+	// running at any time during the update is at most 130% of desired nodes.
+	//
+	// +kubebuilder:validation:Pattern="^((100|[0-9]{1,2})%|[0-9]+)$"
+	// +kubebuilder:validation:XIntOrString
+	// +kubebuilder:default=1
+	// +optional
+	MaxSurge *intstr.IntOrString `json:"maxSurge,omitempty"`
+}
+
 // RosaMachinePoolStatus defines the observed state of RosaMachinePool.
 type RosaMachinePoolStatus struct {
 	// Ready denotes that the RosaMachinePool nodepool has joined

diff --git a/exp/api/v1beta2/zz_generated.deepcopy.go b/exp/api/v1beta2/zz_generated.deepcopy.go
diff --git a/exp/controllers/rosamachinepool_controller.go b/exp/controllers/rosamachinepool_controller.go
@@ -17,6 +17,7 @@ import (
 	apierrors "k8s.io/apimachinery/pkg/api/errors"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/runtime/schema"
+	"k8s.io/apimachinery/pkg/util/intstr"
 	"k8s.io/client-go/tools/record"
 	"k8s.io/klog/v2"
 	"k8s.io/utils/ptr"
@@ -477,6 +478,21 @@ func nodePoolBuilder(rosaMachinePoolSpec expinfrav1.RosaMachinePoolSpec, machine
 		npBuilder.NodeDrainGracePeriod(valueBuilder)
 	}
 
+	if rosaMachinePoolSpec.UpdateConfig != nil {
+		configMgmtBuilder := cmv1.NewNodePoolManagementUpgrade()
+
+		if rosaMachinePoolSpec.UpdateConfig.RollingUpdate != nil {
+			if rosaMachinePoolSpec.UpdateConfig.RollingUpdate.MaxSurge != nil {
+				configMgmtBuilder = configMgmtBuilder.MaxSurge(rosaMachinePoolSpec.UpdateConfig.RollingUpdate.MaxSurge.String())
+			}
+			if rosaMachinePoolSpec.UpdateConfig.RollingUpdate.MaxUnavailable != nil {
+				configMgmtBuilder = configMgmtBuilder.MaxUnavailable(rosaMachinePoolSpec.UpdateConfig.RollingUpdate.MaxUnavailable.String())
+			}
+		}
+
+		npBuilder = npBuilder.ManagementUpgrade(configMgmtBuilder)
+	}
+
 	return npBuilder
 }
 
@@ -518,6 +534,17 @@ func nodePoolToRosaMachinePoolSpec(nodePool *cmv1.NodePool) expinfrav1.RosaMachi
 			Duration: time.Minute * time.Duration(nodePool.NodeDrainGracePeriod().Value()),
 		}
 	}
+	if nodePool.ManagementUpgrade() != nil {
+		spec.UpdateConfig = &expinfrav1.RosaUpdateConfig{
+			RollingUpdate: &expinfrav1.RollingUpdate{},
+		}
+		if nodePool.ManagementUpgrade().MaxSurge() != "" {
+			spec.UpdateConfig.RollingUpdate.MaxSurge = ptr.To(intstr.Parse(nodePool.ManagementUpgrade().MaxSurge()))
+		}
+		if nodePool.ManagementUpgrade().MaxUnavailable() != "" {
+			spec.UpdateConfig.RollingUpdate.MaxSurge = ptr.To(intstr.Parse(nodePool.ManagementUpgrade().MaxUnavailable()))
+		}
+	}
 
 	return spec
 }

diff --git a/exp/controllers/rosamachinepool_controller_test.go b/exp/controllers/rosamachinepool_controller_test.go
@@ -6,6 +6,7 @@ import (
 
 	. "github.com/onsi/gomega"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/util/intstr"
 	"k8s.io/utils/ptr"
 
 	expinfrav1 "sigs.k8s.io/cluster-api-provider-aws/v2/exp/api/v1beta2"
@@ -25,6 +26,11 @@ func TestNodePoolToRosaMachinePoolSpec(t *testing.T) {
 		NodeDrainGracePeriod: &metav1.Duration{
 			Duration: time.Minute * 10,
 		},
+		UpdateConfig: &expinfrav1.RosaUpdateConfig{
+			RollingUpdate: &expinfrav1.RollingUpdate{
+				MaxSurge: &intstr.IntOrString{IntVal: 3},
+			},
+		},
 	}
 
 	machinePoolSpec := expclusterv1.MachinePoolSpec{

diff --git a/go.mod b/go.mod
@@ -32,7 +32,7 @@ require (
 	github.com/onsi/ginkgo/v2 v2.17.1
 	github.com/onsi/gomega v1.32.0
 	github.com/openshift-online/ocm-common v0.0.0-20240129111424-ff8c6c11d909
-	github.com/openshift-online/ocm-sdk-go v0.1.414
+	github.com/openshift-online/ocm-sdk-go v0.1.422
 	github.com/openshift/rosa v1.2.35-rc1.0.20240301152457-ad986cecd364
 	github.com/pkg/errors v0.9.1
 	github.com/prometheus/client_golang v1.19.0

diff --git a/go.sum b/go.sum
@@ -383,8 +383,8 @@ github.com/opencontainers/image-spec v1.1.0-rc2.0.20221005185240-3a7f492d3f1b h1
 github.com/opencontainers/image-spec v1.1.0-rc2.0.20221005185240-3a7f492d3f1b/go.mod h1:3OVijpioIKYWTqjiG0zfF6wvoJ4fAXGbjdZuI2NgsRQ=
 github.com/openshift-online/ocm-common v0.0.0-20240129111424-ff8c6c11d909 h1:WV67GNazQuGDaLX3kBbz0859NYPOQCsDCY5XUScF85M=
 github.com/openshift-online/ocm-common v0.0.0-20240129111424-ff8c6c11d909/go.mod h1:7FaAb07S63RF4sFMLSLtQaJLvPdaRnhAT4dBLD8/5kM=
-github.com/openshift-online/ocm-sdk-go v0.1.414 h1:pvsczJlartURjMOhHYxC6idsSCrixwMJZRuBQWDAIOM=
-github.com/openshift-online/ocm-sdk-go v0.1.414/go.mod h1:CiAu2jwl3ITKOxkeV0Qnhzv4gs35AmpIzVABQLtcI2Y=
+github.com/openshift-online/ocm-sdk-go v0.1.422 h1:NWXLNTg7sLgUJRM3tyuk/QuVbUCRuMH+aLlbCKNzXWc=
+github.com/openshift-online/ocm-sdk-go v0.1.422/go.mod h1:CiAu2jwl3ITKOxkeV0Qnhzv4gs35AmpIzVABQLtcI2Y=
 github.com/openshift/rosa v1.2.35-rc1.0.20240301152457-ad986cecd364 h1:j1aGLgZhO5xXpYgGAjmraioHTvCK7+gXZXoN9cnpnkw=
 github.com/openshift/rosa v1.2.35-rc1.0.20240301152457-ad986cecd364/go.mod h1:kSNsBW8P9KfLCsZYGIrr/aKbLDct8I5gW0e4cCRrr0o=
 github.com/pelletier/go-toml v1.9.4/go.mod h1:u1nR/EPcESfeI/szUZKdtJ0xRNbUoANCkoOuaOx1Y+c=