-
Notifications
You must be signed in to change notification settings - Fork 86
/
controller.go
290 lines (250 loc) · 9.08 KB
/
controller.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
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
/*
* Copyright 2021 The Multicluster-Scheduler Authors.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package service
import (
"context"
"fmt"
"reflect"
"strings"
"time"
agentconfig "admiralty.io/multicluster-scheduler/pkg/config/agent"
corev1 "k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/api/errors"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/labels"
utilruntime "k8s.io/apimachinery/pkg/util/runtime"
coreinformers "k8s.io/client-go/informers/core/v1"
"k8s.io/client-go/kubernetes"
corelisters "k8s.io/client-go/listers/core/v1"
"k8s.io/client-go/tools/cache"
"admiralty.io/multicluster-scheduler/pkg/common"
"admiralty.io/multicluster-scheduler/pkg/controller"
"admiralty.io/multicluster-scheduler/pkg/model/delegatepod"
"admiralty.io/multicluster-scheduler/pkg/model/proxypod"
)
type reconciler struct {
target agentconfig.Target
kubeclientset kubernetes.Interface
remoteClient kubernetes.Interface
svcLister corelisters.ServiceLister
podLister corelisters.PodLister
remoteSvcLister corelisters.ServiceLister
serviceRerouteEnabled bool
}
func NewController(
target agentconfig.Target,
kubeclientset kubernetes.Interface,
remoteClient kubernetes.Interface,
epInformer coreinformers.EndpointsInformer,
svcInformer coreinformers.ServiceInformer,
podInformer coreinformers.PodInformer,
remoteSvcInformer coreinformers.ServiceInformer) *controller.Controller {
r := &reconciler{
target: target,
kubeclientset: kubeclientset,
remoteClient: remoteClient,
svcLister: svcInformer.Lister(),
podLister: podInformer.Lister(),
remoteSvcLister: remoteSvcInformer.Lister(),
serviceRerouteEnabled: true, // TODO configurable
}
c := controller.New(
"services-follow",
r,
epInformer.Informer().HasSynced,
svcInformer.Informer().HasSynced,
podInformer.Informer().HasSynced,
remoteSvcInformer.Informer().HasSynced,
)
svcInformer.Informer().AddEventHandler(controller.HandleAddUpdateWith(c.EnqueueObject))
remoteSvcInformer.Informer().AddEventHandler(controller.HandleAllWith(c.EnqueueRemoteController(
"Service",
func(namespace, name string) (metav1.Object, error) {
return r.svcLister.Services(namespace).Get(name)
},
)))
epInformer.Informer().AddEventHandler(controller.HandleAddUpdateWith(c.EnqueueObject))
// no need to watch pods, as pod events will update endpoints
// we just need the lister to see if pods are delegates in shouldFollow()
return c
}
func (r reconciler) Handle(obj interface{}) (requeueAfter *time.Duration, err error) {
ctx := context.Background()
key := obj.(string)
namespace, name, err := cache.SplitMetaNamespaceKey(key)
utilruntime.Must(err)
svc, err := r.svcLister.Services(namespace).Get(name)
if err != nil {
if errors.IsNotFound(err) {
return nil, nil
}
return nil, err
}
if _, ok := svc.Annotations[common.AnnotationKeyIsDelegate]; ok {
return nil, nil
}
terminating := svc.DeletionTimestamp != nil
hasFinalizer, j := controller.HasFinalizer(svc.Finalizers, r.target.GetFinalizer())
shouldFollow, originalSelector, err := r.shouldFollow(svc)
if err != nil {
return nil, err
}
// get remote owned services
// eponymous services that aren't owned are not included (because we don't want to delete them, see below)
remoteSvc, err := r.remoteSvcLister.Services(namespace).Get(name)
if err != nil && !errors.IsNotFound(err) {
return nil, err
}
if remoteSvc != nil && !controller.ParentControlsChild(remoteSvc, svc) {
return nil, nil
}
if terminating {
if remoteSvc != nil {
if err := r.remoteClient.CoreV1().Services(namespace).Delete(ctx, name, metav1.DeleteOptions{}); err != nil && !errors.IsNotFound(err) {
return nil, err
}
} else if hasFinalizer {
if _, err = r.removeFinalizer(ctx, svc, j); err != nil {
return nil, err
}
}
} else if shouldFollow {
svcCopy := svc.DeepCopy()
needUpdateLocal := false
if !hasFinalizer {
needUpdateLocal = true
r.addFinalizer(svcCopy)
}
if svcCopy.Annotations == nil {
svcCopy.Annotations = map[string]string{}
}
if svcCopy.Annotations[common.AnnotationKeyCiliumGlobalService] != "true" {
needUpdateLocal = true
svcCopy.Annotations[common.AnnotationKeyCiliumGlobalService] = "true"
}
if svcCopy.Annotations[common.AnnotationKeyGlobal] != "true" {
needUpdateLocal = true
svcCopy.Annotations[common.AnnotationKeyGlobal] = "true"
}
if originalSelector != svcCopy.Annotations[common.AnnotationKeyOriginalSelector] {
needUpdateLocal = true
svcCopy.Annotations[common.AnnotationKeyOriginalSelector] = originalSelector
}
if r.serviceRerouteEnabled {
selector, changed := delegatepod.ChangeLabels(svcCopy.Spec.Selector)
if changed {
needUpdateLocal = true
svcCopy.Spec.Selector = selector
}
}
if needUpdateLocal {
if svc, err = r.kubeclientset.CoreV1().Services(svcCopy.Namespace).Update(ctx, svcCopy, metav1.UpdateOptions{}); err != nil {
return nil, err
}
}
if remoteSvc == nil {
gold := makeRemoteService(svc) // at this point, svc includes updates from above (including reroute and cilium)
_, err := r.remoteClient.CoreV1().Services(namespace).Create(ctx, gold, metav1.CreateOptions{})
if err != nil && !errors.IsAlreadyExists(err) {
return nil, err
}
} else {
spec := svc.Spec.DeepCopy()
spec.ClusterIP = remoteSvc.Spec.ClusterIP
if !reflect.DeepEqual(&remoteSvc.Spec, spec) {
remoteCopy := remoteSvc.DeepCopy()
remoteCopy.Spec = *spec.DeepCopy()
_, err := r.remoteClient.CoreV1().Services(namespace).Update(ctx, remoteCopy, metav1.UpdateOptions{})
if err != nil {
return nil, err
}
}
}
}
// TODO? cleanup remote services that shouldn't follow anymore
return nil, nil
}
func (r reconciler) shouldFollow(service *corev1.Service) (bool, string, error) {
// an empty selector would list everything!
// when it actually means the service doesn't select pods (e.g., uses custom Endpoints or external name)
if len(service.Spec.Selector) == 0 {
return false, "", nil
}
// delegate pods may not be in the same cluster, so we need to select proxy pods
// if service was never rerouted, we can use the selector as is
// if service was rerouted, we need to recover the original selector from annotations
// (because we can't recover original selector from transformed one:
// with current algorithm, domains are overwritten; with a future one, they may be hashed, same issue),
// unless the current selector includes labels NOT prefixed with out domain,
// which means it's been updated (or the original selector has been reapplied)
// in that case, use those labels
selector := labels.Set{}
for k, v := range service.Spec.Selector {
if !strings.HasPrefix(k, common.KeyPrefix) { // user shouldn't use our domain (which is ok by convention)
selector[k] = v
}
}
if len(selector) == 0 {
s, ok := service.Annotations[common.AnnotationKeyOriginalSelector]
if !ok {
return false, "", fmt.Errorf("original selector not found")
}
var err error
selector, err = labels.ConvertSelectorToLabelsMap(s)
if err != nil {
return false, "", fmt.Errorf("original selector is invalid (was tampered with?): %v", err)
}
}
pods, err := r.podLister.Pods(service.Namespace).List(labels.SelectorFromValidatedSet(selector))
if err != nil {
return false, "", err
}
for _, pod := range pods {
if proxypod.IsProxy(pod) {
return true, selector.String(), nil
}
}
return false, selector.String(), nil
}
func (r reconciler) addFinalizer(actualCopy *corev1.Service) {
actualCopy.Finalizers = append(actualCopy.Finalizers, r.target.GetFinalizer())
if actualCopy.Labels == nil {
actualCopy.Labels = map[string]string{}
}
actualCopy.Labels[common.LabelKeyHasFinalizer] = "true"
}
func (r reconciler) removeFinalizer(ctx context.Context, actual *corev1.Service, j int) (*corev1.Service, error) {
actualCopy := actual.DeepCopy()
actualCopy.Finalizers = append(actualCopy.Finalizers[:j], actualCopy.Finalizers[j+1:]...)
return r.kubeclientset.CoreV1().Services(actual.Namespace).Update(ctx, actualCopy, metav1.UpdateOptions{})
}
func makeRemoteService(actual *corev1.Service) *corev1.Service {
gold := &corev1.Service{}
gold.Name = actual.Name
gold.Labels = make(map[string]string, len(actual.Labels))
for k, v := range actual.Labels {
gold.Labels[k] = v
}
gold.Annotations = make(map[string]string, len(actual.Annotations))
for k, v := range actual.Annotations {
gold.Annotations[k] = v
}
gold.Annotations[common.AnnotationKeyIsDelegate] = ""
controller.AddRemoteControllerReference(gold, actual)
gold.Spec = *actual.Spec.DeepCopy()
gold.Spec.ClusterIP = "" // cluster IP given by each cluster (not really a top-level spec)
return gold
}