forked from kubernetes/kubernetes
-
Notifications
You must be signed in to change notification settings - Fork 0
/
available_controller.go
441 lines (381 loc) · 14.6 KB
/
available_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
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
/*
Copyright 2017 The Kubernetes 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 apiserver
import (
"crypto/tls"
"fmt"
"net/http"
"net/url"
"time"
"github.com/golang/glog"
"k8s.io/api/core/v1"
apierrors "k8s.io/apimachinery/pkg/api/errors"
"k8s.io/apimachinery/pkg/api/meta"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/labels"
"k8s.io/apimachinery/pkg/runtime"
utilruntime "k8s.io/apimachinery/pkg/util/runtime"
"k8s.io/apimachinery/pkg/util/wait"
v1informers "k8s.io/client-go/informers/core/v1"
v1listers "k8s.io/client-go/listers/core/v1"
"k8s.io/client-go/tools/cache"
"k8s.io/client-go/util/workqueue"
"k8s.io/kube-aggregator/pkg/apis/apiregistration"
apiregistrationclient "k8s.io/kube-aggregator/pkg/client/clientset_generated/internalclientset/typed/apiregistration/internalversion"
informers "k8s.io/kube-aggregator/pkg/client/informers/internalversion/apiregistration/internalversion"
listers "k8s.io/kube-aggregator/pkg/client/listers/apiregistration/internalversion"
"k8s.io/kube-aggregator/pkg/controllers"
)
type ServiceResolver interface {
ResolveEndpoint(namespace, name string) (*url.URL, error)
}
type AvailableConditionController struct {
apiServiceClient apiregistrationclient.APIServicesGetter
apiServiceLister listers.APIServiceLister
apiServiceSynced cache.InformerSynced
// serviceLister is used to get the IP to create the transport for
serviceLister v1listers.ServiceLister
servicesSynced cache.InformerSynced
endpointsLister v1listers.EndpointsLister
endpointsSynced cache.InformerSynced
discoveryClient *http.Client
serviceResolver ServiceResolver
// To allow injection for testing.
syncFn func(key string) error
queue workqueue.RateLimitingInterface
}
func NewAvailableConditionController(
apiServiceInformer informers.APIServiceInformer,
serviceInformer v1informers.ServiceInformer,
endpointsInformer v1informers.EndpointsInformer,
apiServiceClient apiregistrationclient.APIServicesGetter,
proxyTransport *http.Transport,
serviceResolver ServiceResolver,
) *AvailableConditionController {
c := &AvailableConditionController{
apiServiceClient: apiServiceClient,
apiServiceLister: apiServiceInformer.Lister(),
apiServiceSynced: apiServiceInformer.Informer().HasSynced,
serviceLister: serviceInformer.Lister(),
servicesSynced: serviceInformer.Informer().HasSynced,
endpointsLister: endpointsInformer.Lister(),
endpointsSynced: endpointsInformer.Informer().HasSynced,
serviceResolver: serviceResolver,
queue: workqueue.NewNamedRateLimitingQueue(workqueue.DefaultControllerRateLimiter(), "AvailableConditionController"),
}
// construct an http client that will ignore TLS verification (if someone owns the network and messes with your status
// that's not so bad) and sets a very short timeout.
discoveryClient := &http.Client{
Transport: &http.Transport{
TLSClientConfig: &tls.Config{InsecureSkipVerify: true},
},
// the request should happen quickly.
Timeout: 5 * time.Second,
}
if proxyTransport != nil {
discoveryClient.Transport = proxyTransport
}
c.discoveryClient = discoveryClient
// resync on this one because it is low cardinality and rechecking the actual discovery
// allows us to detect health in a more timely fashion when network connectivity to
// nodes is snipped, but the network still attempts to route there. See
// https://github.com/openshift/origin/issues/17159#issuecomment-341798063
apiServiceInformer.Informer().AddEventHandlerWithResyncPeriod(
cache.ResourceEventHandlerFuncs{
AddFunc: c.addAPIService,
UpdateFunc: c.updateAPIService,
DeleteFunc: c.deleteAPIService,
},
30*time.Second)
serviceInformer.Informer().AddEventHandler(cache.ResourceEventHandlerFuncs{
AddFunc: c.addService,
UpdateFunc: c.updateService,
DeleteFunc: c.deleteService,
})
endpointsInformer.Informer().AddEventHandler(cache.ResourceEventHandlerFuncs{
AddFunc: c.addEndpoints,
UpdateFunc: c.updateEndpoints,
DeleteFunc: c.deleteEndpoints,
})
c.syncFn = c.sync
return c
}
func (c *AvailableConditionController) sync(key string) error {
inAPIService, err := c.apiServiceLister.Get(key)
if apierrors.IsNotFound(err) {
return nil
}
if err != nil {
return err
}
apiService := inAPIService.DeepCopy()
availableCondition := apiregistration.APIServiceCondition{
Type: apiregistration.Available,
Status: apiregistration.ConditionTrue,
LastTransitionTime: metav1.Now(),
}
// local API services are always considered available
if apiService.Spec.Service == nil {
apiregistration.SetAPIServiceCondition(apiService, apiregistration.NewLocalAvailableAPIServiceCondition())
_, err := c.apiServiceClient.APIServices().UpdateStatus(apiService)
return err
}
service, err := c.serviceLister.Services(apiService.Spec.Service.Namespace).Get(apiService.Spec.Service.Name)
if apierrors.IsNotFound(err) {
availableCondition.Status = apiregistration.ConditionFalse
availableCondition.Reason = "ServiceNotFound"
availableCondition.Message = fmt.Sprintf("service/%s in %q is not present", apiService.Spec.Service.Name, apiService.Spec.Service.Namespace)
apiregistration.SetAPIServiceCondition(apiService, availableCondition)
_, err := c.apiServiceClient.APIServices().UpdateStatus(apiService)
return err
} else if err != nil {
availableCondition.Status = apiregistration.ConditionUnknown
availableCondition.Reason = "ServiceAccessError"
availableCondition.Message = fmt.Sprintf("service/%s in %q cannot be checked due to: %v", apiService.Spec.Service.Name, apiService.Spec.Service.Namespace, err)
apiregistration.SetAPIServiceCondition(apiService, availableCondition)
_, err := c.apiServiceClient.APIServices().UpdateStatus(apiService)
return err
}
if service.Spec.Type == v1.ServiceTypeClusterIP {
// if we have a cluster IP service, it must be listening on 443 and we can check that
foundPort := false
for _, port := range service.Spec.Ports {
if port.Port == 443 {
foundPort = true
}
}
if !foundPort {
availableCondition.Status = apiregistration.ConditionFalse
availableCondition.Reason = "ServicePortError"
availableCondition.Message = fmt.Sprintf("service/%s in %q is not listening on port 443", apiService.Spec.Service.Name, apiService.Spec.Service.Namespace)
apiregistration.SetAPIServiceCondition(apiService, availableCondition)
_, err := c.apiServiceClient.APIServices().UpdateStatus(apiService)
return err
}
endpoints, err := c.endpointsLister.Endpoints(apiService.Spec.Service.Namespace).Get(apiService.Spec.Service.Name)
if apierrors.IsNotFound(err) {
availableCondition.Status = apiregistration.ConditionFalse
availableCondition.Reason = "EndpointsNotFound"
availableCondition.Message = fmt.Sprintf("cannot find endpoints for service/%s in %q", apiService.Spec.Service.Name, apiService.Spec.Service.Namespace)
apiregistration.SetAPIServiceCondition(apiService, availableCondition)
_, err := c.apiServiceClient.APIServices().UpdateStatus(apiService)
return err
} else if err != nil {
availableCondition.Status = apiregistration.ConditionUnknown
availableCondition.Reason = "EndpointsAccessError"
availableCondition.Message = fmt.Sprintf("service/%s in %q cannot be checked due to: %v", apiService.Spec.Service.Name, apiService.Spec.Service.Namespace, err)
apiregistration.SetAPIServiceCondition(apiService, availableCondition)
_, err := c.apiServiceClient.APIServices().UpdateStatus(apiService)
return err
}
hasActiveEndpoints := false
for _, subset := range endpoints.Subsets {
if len(subset.Addresses) > 0 {
hasActiveEndpoints = true
break
}
}
if !hasActiveEndpoints {
availableCondition.Status = apiregistration.ConditionFalse
availableCondition.Reason = "MissingEndpoints"
availableCondition.Message = fmt.Sprintf("endpoints for service/%s in %q have no addresses", apiService.Spec.Service.Name, apiService.Spec.Service.Namespace)
apiregistration.SetAPIServiceCondition(apiService, availableCondition)
_, err := c.apiServiceClient.APIServices().UpdateStatus(apiService)
return err
}
}
// actually try to hit the discovery endpoint when it isn't local and when we're routing as a service.
if apiService.Spec.Service != nil && c.serviceResolver != nil {
discoveryURL, err := c.serviceResolver.ResolveEndpoint(apiService.Spec.Service.Namespace, apiService.Spec.Service.Name)
if err != nil {
return err
}
errCh := make(chan error)
go func() {
resp, err := c.discoveryClient.Get(discoveryURL.String())
if resp != nil {
resp.Body.Close()
}
errCh <- err
}()
select {
case err = <-errCh:
// we had trouble with slow dial and DNS responses causing us to wait too long.
// we added this as insurance
case <-time.After(6 * time.Second):
err = fmt.Errorf("timed out waiting for %v", discoveryURL)
}
if err != nil {
availableCondition.Status = apiregistration.ConditionFalse
availableCondition.Reason = "FailedDiscoveryCheck"
availableCondition.Message = fmt.Sprintf("no response from %v: %v", discoveryURL, err)
apiregistration.SetAPIServiceCondition(apiService, availableCondition)
_, updateErr := c.apiServiceClient.APIServices().UpdateStatus(apiService)
if updateErr != nil {
return updateErr
}
// force a requeue to make it very obvious that this will be retried at some point in the future
// along with other requeues done via service change, endpoint change, and resync
return err
}
}
availableCondition.Reason = "Passed"
availableCondition.Message = "all checks passed"
apiregistration.SetAPIServiceCondition(apiService, availableCondition)
_, err = c.apiServiceClient.APIServices().UpdateStatus(apiService)
return err
}
func (c *AvailableConditionController) Run(threadiness int, stopCh <-chan struct{}) {
defer utilruntime.HandleCrash()
defer c.queue.ShutDown()
glog.Infof("Starting AvailableConditionController")
defer glog.Infof("Shutting down AvailableConditionController")
if !controllers.WaitForCacheSync("AvailableConditionController", stopCh, c.apiServiceSynced, c.servicesSynced, c.endpointsSynced) {
return
}
for i := 0; i < threadiness; i++ {
go wait.Until(c.runWorker, time.Second, stopCh)
}
<-stopCh
}
func (c *AvailableConditionController) runWorker() {
for c.processNextWorkItem() {
}
}
// processNextWorkItem deals with one key off the queue. It returns false when it's time to quit.
func (c *AvailableConditionController) processNextWorkItem() bool {
key, quit := c.queue.Get()
if quit {
return false
}
defer c.queue.Done(key)
err := c.syncFn(key.(string))
if err == nil {
c.queue.Forget(key)
return true
}
utilruntime.HandleError(fmt.Errorf("%v failed with: %v", key, err))
c.queue.AddRateLimited(key)
return true
}
func (c *AvailableConditionController) enqueue(obj *apiregistration.APIService) {
key, err := cache.DeletionHandlingMetaNamespaceKeyFunc(obj)
if err != nil {
glog.Errorf("Couldn't get key for object %#v: %v", obj, err)
return
}
c.queue.Add(key)
}
func (c *AvailableConditionController) addAPIService(obj interface{}) {
castObj := obj.(*apiregistration.APIService)
glog.V(4).Infof("Adding %s", castObj.Name)
c.enqueue(castObj)
}
func (c *AvailableConditionController) updateAPIService(obj, _ interface{}) {
castObj := obj.(*apiregistration.APIService)
glog.V(4).Infof("Updating %s", castObj.Name)
c.enqueue(castObj)
}
func (c *AvailableConditionController) deleteAPIService(obj interface{}) {
castObj, ok := obj.(*apiregistration.APIService)
if !ok {
tombstone, ok := obj.(cache.DeletedFinalStateUnknown)
if !ok {
glog.Errorf("Couldn't get object from tombstone %#v", obj)
return
}
castObj, ok = tombstone.Obj.(*apiregistration.APIService)
if !ok {
glog.Errorf("Tombstone contained object that is not expected %#v", obj)
return
}
}
glog.V(4).Infof("Deleting %q", castObj.Name)
c.enqueue(castObj)
}
// there aren't very many apiservices, just check them all.
func (c *AvailableConditionController) getAPIServicesFor(obj runtime.Object) []*apiregistration.APIService {
metadata, err := meta.Accessor(obj)
if err != nil {
utilruntime.HandleError(err)
return nil
}
var ret []*apiregistration.APIService
apiServiceList, _ := c.apiServiceLister.List(labels.Everything())
for _, apiService := range apiServiceList {
if apiService.Spec.Service == nil {
continue
}
if apiService.Spec.Service.Namespace == metadata.GetNamespace() && apiService.Spec.Service.Name == metadata.GetName() {
ret = append(ret, apiService)
}
}
return ret
}
// TODO, think of a way to avoid checking on every service manipulation
func (c *AvailableConditionController) addService(obj interface{}) {
for _, apiService := range c.getAPIServicesFor(obj.(*v1.Service)) {
c.enqueue(apiService)
}
}
func (c *AvailableConditionController) updateService(obj, _ interface{}) {
for _, apiService := range c.getAPIServicesFor(obj.(*v1.Service)) {
c.enqueue(apiService)
}
}
func (c *AvailableConditionController) deleteService(obj interface{}) {
castObj, ok := obj.(*v1.Service)
if !ok {
tombstone, ok := obj.(cache.DeletedFinalStateUnknown)
if !ok {
glog.Errorf("Couldn't get object from tombstone %#v", obj)
return
}
castObj, ok = tombstone.Obj.(*v1.Service)
if !ok {
glog.Errorf("Tombstone contained object that is not expected %#v", obj)
return
}
}
for _, apiService := range c.getAPIServicesFor(castObj) {
c.enqueue(apiService)
}
}
func (c *AvailableConditionController) addEndpoints(obj interface{}) {
for _, apiService := range c.getAPIServicesFor(obj.(*v1.Endpoints)) {
c.enqueue(apiService)
}
}
func (c *AvailableConditionController) updateEndpoints(obj, _ interface{}) {
for _, apiService := range c.getAPIServicesFor(obj.(*v1.Endpoints)) {
c.enqueue(apiService)
}
}
func (c *AvailableConditionController) deleteEndpoints(obj interface{}) {
castObj, ok := obj.(*v1.Endpoints)
if !ok {
tombstone, ok := obj.(cache.DeletedFinalStateUnknown)
if !ok {
glog.Errorf("Couldn't get object from tombstone %#v", obj)
return
}
castObj, ok = tombstone.Obj.(*v1.Endpoints)
if !ok {
glog.Errorf("Tombstone contained object that is not expected %#v", obj)
return
}
}
for _, apiService := range c.getAPIServicesFor(castObj) {
c.enqueue(apiService)
}
}