-
Notifications
You must be signed in to change notification settings - Fork 201
/
sync.go
235 lines (215 loc) · 6.81 KB
/
sync.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
package operator
import (
"fmt"
"time"
"github.com/golang/glog"
appsv1 "k8s.io/api/apps/v1"
corev1 "k8s.io/api/core/v1"
apierrors "k8s.io/apimachinery/pkg/api/errors"
"k8s.io/apimachinery/pkg/api/resource"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/util/wait"
"k8s.io/utils/pointer"
osev1 "github.com/openshift/api/config/v1"
"github.com/openshift/cluster-version-operator/lib/resourceapply"
)
const (
deploymentRolloutPollInterval = time.Second
deploymentRolloutTimeout = 5 * time.Minute
)
func (optr *Operator) syncAll(config *OperatorConfig) error {
if err := optr.statusProgressing(); err != nil {
glog.Errorf("Error syncing ClusterOperatorStatus: %v", err)
return fmt.Errorf("error syncing ClusterOperatorStatus: %v", err)
}
if config.Controllers.Provider != clusterAPIControllerNoOp {
if err := optr.syncClusterAPIController(config); err != nil {
if err := optr.statusDegraded(err.Error()); err != nil {
// Just log the error here. We still want to
// return the outer error.
glog.Errorf("Error syncing ClusterOperatorStatus: %v", err)
}
glog.Errorf("Error syncing cluster api controller: %v", err)
return err
}
glog.V(3).Info("Synced up all components")
}
if err := optr.statusAvailable(); err != nil {
glog.Errorf("Error syncing ClusterOperatorStatus: %v", err)
return fmt.Errorf("error syncing ClusterOperatorStatus: %v", err)
}
return nil
}
func (optr *Operator) syncClusterAPIController(config *OperatorConfig) error {
// Fetch the Feature
featureGate, err := optr.featureGateLister.Get(MachineAPIFeatureGateName)
var featureSet osev1.FeatureSet
if err != nil {
if !apierrors.IsNotFound(err) {
return err
}
glog.V(2).Infof("Failed to find feature gate %q, will use default feature set", MachineAPIFeatureGateName)
featureSet = osev1.Default
} else {
featureSet = featureGate.Spec.FeatureSet
}
features, err := generateFeatureMap(featureSet)
if err != nil {
return err
}
controller := newDeployment(config, features)
_, updated, err := resourceapply.ApplyDeployment(optr.kubeClient.AppsV1(), controller)
if err != nil {
return err
}
if updated {
return optr.waitForDeploymentRollout(controller)
}
return nil
}
func (optr *Operator) waitForDeploymentRollout(resource *appsv1.Deployment) error {
return wait.Poll(deploymentRolloutPollInterval, deploymentRolloutTimeout, func() (bool, error) {
d, err := optr.deployLister.Deployments(resource.Namespace).Get(resource.Name)
if apierrors.IsNotFound(err) {
return false, nil
}
if err != nil {
// Do not return error here, as we could be updating the API Server itself, in which case we
// want to continue waiting.
glog.Errorf("Error getting Deployment %q during rollout: %v", resource.Name, err)
return false, nil
}
if d.DeletionTimestamp != nil {
return false, fmt.Errorf("deployment %q is being deleted", resource.Name)
}
if d.Generation <= d.Status.ObservedGeneration && d.Status.UpdatedReplicas == d.Status.Replicas && d.Status.UnavailableReplicas == 0 {
return true, nil
}
glog.V(4).Infof("Deployment %q is not ready. status: (replicas: %d, updated: %d, ready: %d, unavailable: %d)", d.Name, d.Status.Replicas, d.Status.UpdatedReplicas, d.Status.ReadyReplicas, d.Status.UnavailableReplicas)
return false, nil
})
}
func newDeployment(config *OperatorConfig, features map[string]bool) *appsv1.Deployment {
replicas := int32(1)
template := newPodTemplateSpec(config, features)
return &appsv1.Deployment{
ObjectMeta: metav1.ObjectMeta{
Name: "machine-api-controllers",
Namespace: config.TargetNamespace,
Labels: map[string]string{
"api": "clusterapi",
"k8s-app": "controller",
},
},
Spec: appsv1.DeploymentSpec{
Replicas: &replicas,
Selector: &metav1.LabelSelector{
MatchLabels: map[string]string{
"api": "clusterapi",
"k8s-app": "controller",
},
},
Template: *template,
},
}
}
func newPodTemplateSpec(config *OperatorConfig, features map[string]bool) *corev1.PodTemplateSpec {
containers := newContainers(config, features)
tolerations := []corev1.Toleration{
{
Key: "node-role.kubernetes.io/master",
Effect: corev1.TaintEffectNoSchedule,
},
{
Key: "CriticalAddonsOnly",
Operator: corev1.TolerationOpExists,
},
{
Key: "node.kubernetes.io/not-ready",
Effect: corev1.TaintEffectNoExecute,
Operator: corev1.TolerationOpExists,
TolerationSeconds: pointer.Int64Ptr(120),
},
{
Key: "node.kubernetes.io/unreachable",
Effect: corev1.TaintEffectNoExecute,
Operator: corev1.TolerationOpExists,
TolerationSeconds: pointer.Int64Ptr(120),
},
}
return &corev1.PodTemplateSpec{
ObjectMeta: metav1.ObjectMeta{
Labels: map[string]string{
"api": "clusterapi",
"k8s-app": "controller",
},
},
Spec: corev1.PodSpec{
Containers: containers,
PriorityClassName: "system-node-critical",
NodeSelector: map[string]string{"node-role.kubernetes.io/master": ""},
SecurityContext: &corev1.PodSecurityContext{
RunAsNonRoot: pointer.BoolPtr(true),
RunAsUser: pointer.Int64Ptr(65534),
},
ServiceAccountName: "machine-api-controllers",
Tolerations: tolerations,
},
}
}
func newContainers(config *OperatorConfig, features map[string]bool) []corev1.Container {
resources := corev1.ResourceRequirements{
Requests: map[corev1.ResourceName]resource.Quantity{
corev1.ResourceMemory: resource.MustParse("20Mi"),
corev1.ResourceCPU: resource.MustParse("10m"),
},
}
args := []string{
"--logtostderr=true",
"--v=3",
fmt.Sprintf("--namespace=%s", config.TargetNamespace),
}
containers := []corev1.Container{
{
Name: "controller-manager",
Image: config.Controllers.Provider,
Command: []string{"/manager"},
Args: args,
Resources: resources,
},
{
Name: "machine-controller",
Image: config.Controllers.Provider,
Command: []string{"/machine-controller-manager"},
Args: args,
Env: []corev1.EnvVar{
{
Name: "NODE_NAME",
ValueFrom: &corev1.EnvVarSource{
FieldRef: &corev1.ObjectFieldSelector{
FieldPath: "spec.nodeName",
},
},
},
},
},
{
Name: "nodelink-controller",
Image: config.Controllers.NodeLink,
Command: []string{"/nodelink-controller"},
Args: args,
Resources: resources,
},
}
// add machine-health-check controller container if it exists and enabled under feature gates
if enabled, ok := features[FeatureGateMachineHealthCheck]; ok && enabled {
containers = append(containers, corev1.Container{
Name: "machine-healthcheck-controller",
Image: config.Controllers.MachineHealthCheck,
Command: []string{"/machine-healthcheck"},
Args: args,
Resources: resources,
})
}
return containers
}