This repository has been archived by the owner on Nov 1, 2022. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 262
/
operator.go
310 lines (273 loc) · 10.2 KB
/
operator.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
package operator
import (
"fmt"
"github.com/fluxcd/helm-operator/pkg/chartsync"
"os"
"path"
"sync"
"time"
"github.com/go-kit/kit/log"
"github.com/google/go-cmp/cmp"
corev1 "k8s.io/api/core/v1"
k8serrors "k8s.io/apimachinery/pkg/api/errors"
"k8s.io/apimachinery/pkg/util/runtime"
"k8s.io/apimachinery/pkg/util/wait"
"k8s.io/client-go/kubernetes"
"k8s.io/client-go/kubernetes/scheme"
typedcorev1 "k8s.io/client-go/kubernetes/typed/core/v1"
"k8s.io/client-go/tools/cache"
"k8s.io/client-go/tools/record"
"k8s.io/client-go/util/workqueue"
"github.com/fluxcd/helm-operator/internal/lockedfile"
helmfluxv1 "github.com/fluxcd/helm-operator/pkg/apis/helm.fluxcd.io/v1"
ifscheme "github.com/fluxcd/helm-operator/pkg/client/clientset/versioned/scheme"
hrv1 "github.com/fluxcd/helm-operator/pkg/client/informers/externalversions/helm.fluxcd.io/v1"
iflister "github.com/fluxcd/helm-operator/pkg/client/listers/helm.fluxcd.io/v1"
"github.com/fluxcd/helm-operator/pkg/release"
"github.com/fluxcd/helm-operator/pkg/status"
)
const (
controllerAgentName = "helm-operator"
ReleaseSynced = "ReleaseSynced"
FailedReleaseSync = "FailedReleaseSync"
)
// Controller is the operator implementation for HelmRelease resources
type Controller struct {
logger log.Logger
logDiffs bool
hrLister iflister.HelmReleaseLister
hrSynced cache.InformerSynced
release *release.Release
gitChartSync *chartsync.GitChartSync
// workqueue is a rate limited work queue. This is used to queue work to be
// processed instead of performing it as soon as a change happens. This
// means we can ensure we only process a fixed amount of resources at a
// time, and makes it easy to ensure we are never processing the same item
// simultaneously in two different workers.
releaseWorkqueue workqueue.RateLimitingInterface
// recorder is an event recorder for recording Event resources to the
// Kubernetes API.
recorder record.EventRecorder
}
// New returns a new helm-operator
func New(
logger log.Logger,
logReleaseDiffs bool,
kubeclientset kubernetes.Interface,
hrInformer hrv1.HelmReleaseInformer,
releaseWorkqueue workqueue.RateLimitingInterface,
release *release.Release,
gitChartSync *chartsync.GitChartSync) *Controller {
// Add helm-operator types to the default Kubernetes Scheme so Events can be
// logged for helm-operator types.
ifscheme.AddToScheme(scheme.Scheme)
eventBroadcaster := record.NewBroadcaster()
eventBroadcaster.StartRecordingToSink(&typedcorev1.EventSinkImpl{Interface: kubeclientset.CoreV1().Events("")})
recorder := eventBroadcaster.NewRecorder(scheme.Scheme, corev1.EventSource{Component: controllerAgentName})
controller := &Controller{
logger: logger,
logDiffs: logReleaseDiffs,
hrLister: hrInformer.Lister(),
hrSynced: hrInformer.Informer().HasSynced,
releaseWorkqueue: releaseWorkqueue,
recorder: recorder,
release: release,
gitChartSync: gitChartSync,
}
controller.logger.Log("info", "setting up event handlers")
hrInformer.Informer().AddEventHandler(cache.ResourceEventHandlerFuncs{
AddFunc: func(new interface{}) {
if _, ok := checkCustomResourceType(controller.logger, new); ok {
releaseCount.Add(1)
controller.enqueueJob(new)
}
},
UpdateFunc: func(old, new interface{}) {
controller.enqueueUpdateJob(old, new)
},
DeleteFunc: func(old interface{}) {
if hr, ok := checkCustomResourceType(controller.logger, old); ok {
releaseCount.Add(-1)
if err := controller.release.Uninstall(hr.DeepCopy()); err != nil {
controller.logger.Log("error", err)
}
status.ObserveReleaseConditions(&hr, nil)
}
},
})
controller.logger.Log("info", "event handlers set up")
return controller
}
// Run starts workers handling the enqueued events. It will block until
// stopCh is closed, at which point it will shutdown the workqueue and
// wait for workers to finish processing their current work items.
func (c *Controller) Run(threadiness int, stopCh <-chan struct{}, wg *sync.WaitGroup) {
defer runtime.HandleCrash()
defer c.releaseWorkqueue.ShutDown()
c.logger.Log("info", "starting operator")
c.logger.Log("info", "starting workers")
for i := 0; i < threadiness; i++ {
wg.Add(1)
go wait.Until(c.runWorker, time.Second, stopCh)
}
<-stopCh
for i := 0; i < threadiness; i++ {
wg.Done()
}
c.logger.Log("info", "stopping workers")
}
// runWorker is a long-running function calling the
// processNextWorkItem function to read and process a message
// on a workqueue.
func (c *Controller) runWorker() {
for c.processNextWorkItem() {
}
}
// processNextWorkItem will read a single work item off the workqueue and
// attempt to process it, by calling the syncHandler.
func (c *Controller) processNextWorkItem() bool {
releaseQueueLength.Set(float64(c.releaseWorkqueue.Len()))
obj, shutdown := c.releaseWorkqueue.Get()
if shutdown {
return false
}
// wrapping block in a func to defer c.workqueue.Done
err := func(obj interface{}) error {
// We call Done here so the workqueue knows we have finished
// processing this item. We must call Forget if we do not want
// this work item being re-queued. If a transient error
// occurs, we do not call Forget. Instead the item is put back
// on the workqueue and attempted again after a back-off
// period.
defer c.releaseWorkqueue.Done(obj)
var key string
var ok bool
// We expect strings to come off the workqueue. These are of
// the form "namespace/hr(custom resource) name". We do this
// as the delayed nature of the workqueue means the items in
// the informer cache may actually be more up to date than
// when the item was initially put onto the workqueue.
if key, ok = obj.(string); !ok {
// As the item in the workqueue is actually invalid, we call
// Forget not to get into a loop of attempting to
// process a work item that is invalid.
c.releaseWorkqueue.Forget(obj)
runtime.HandleError(fmt.Errorf("expected string in workqueue but got %#v", obj))
return nil
}
// Run the syncHandler, passing it the namespace/name string of the
// HelmRelease resource to sync the corresponding Chart release.
// If the sync failed, then we return while the item will get requeued
if err := c.syncHandler(key); err != nil {
return fmt.Errorf("errored syncing HelmRelease '%s': %s", key, err.Error())
}
// If no error occurs we Forget this item so it does not
// get queued again until another change happens.
c.releaseWorkqueue.Forget(obj)
return nil
}(obj)
if err != nil {
runtime.HandleError(err)
return true
}
return true
}
// syncHandler acts according to the action
// Deletes/creates or updates a Chart release
func (c *Controller) syncHandler(key string) error {
// Retrieve namespace and Custom Resource name from the key
namespace, name, err := cache.SplitMetaNamespaceKey(key)
if err != nil {
c.logger.Log("error", fmt.Sprintf("key '%s' is invalid: %v", key, err))
runtime.HandleError(fmt.Errorf("key '%s' is invalid", key))
return nil
}
// acquire lock
unlock, err := c.lock(fmt.Sprintf("%s-%s", namespace, name))
if err != nil {
c.logger.Log("info", fmt.Sprintf("could not obtain lock: %s", err))
return nil
}
defer unlock()
// Custom Resource hr contains all information we need to know about the Chart release
hr, err := c.hrLister.HelmReleases(namespace).Get(name)
if err != nil {
if k8serrors.IsNotFound(err) {
c.logger.Log("info", fmt.Sprintf("HelmRelease '%s' referred to in work queue no longer exists", key))
runtime.HandleError(fmt.Errorf("HelmRelease '%s' referred to in work queue no longer exists", key))
return nil
}
c.logger.Log("error", err.Error())
return err
}
err = c.release.Sync(hr.DeepCopy())
if err != nil {
c.recorder.Event(hr, corev1.EventTypeWarning, FailedReleaseSync,
fmt.Sprintf("synchronization of release '%s' in namespace '%s' failed: %s", hr.GetReleaseName(), hr.GetTargetNamespace(), err.Error()))
} else {
c.recorder.Event(hr, corev1.EventTypeNormal, ReleaseSynced,
fmt.Sprintf("managed release '%s' in namespace '%s' synchronized", hr.GetReleaseName(), hr.GetTargetNamespace()))
}
return nil
}
func (c *Controller) lock(name string) (unlock func(), err error) {
lockFile := path.Join(os.TempDir(), name+".lock")
mutex := lockedfile.MutexAt(lockFile)
return mutex.Lock()
}
// enqueueJob takes a HelmRelease resource and converts it into a namespace/name
// string which is then put onto the work queue. This method should not be
// passed resources of any type other than HelmRelease.
func (c *Controller) enqueueJob(obj interface{}) {
key, err := getCacheKey(obj)
if err != nil {
return
}
c.releaseWorkqueue.AddRateLimited(key)
releaseQueueLength.Set(float64(c.releaseWorkqueue.Len()))
}
// enqueueUpdateJob decides if there is a genuine resource update
func (c *Controller) enqueueUpdateJob(old, new interface{}) {
oldHr, ok := checkCustomResourceType(c.logger, old)
if !ok {
return
}
newHr, ok := checkCustomResourceType(c.logger, new)
if !ok {
return
}
diff := cmp.Diff(oldHr.Spec, newHr.Spec)
// Filter out any update notifications that are due to status
// updates, as the dry-run that determines if we should upgrade
// is expensive, but _without_ filtering out updates that are
// from the periodic refresh, as we still want to detect (and
// undo) mutations to Helm charts.
if sDiff := cmp.Diff(oldHr.Status, newHr.Status); diff == "" && sDiff != "" {
return
}
// if there is a change in the chartsource (ref change, eg),
// its possible that the mirror's ref-sha be obsolete w.r.t. upstream
// repo's ref-sha. To avoid spurious deploy, mirror is synced before
// doing helm reconciliation
if csDiff := cmp.Diff(oldHr.Spec.ChartSource, newHr.Spec.ChartSource); csDiff != "" {
c.gitChartSync.SyncMirror(&newHr)
}
c.enqueueJob(new)
}
func checkCustomResourceType(logger log.Logger, obj interface{}) (helmfluxv1.HelmRelease, bool) {
var hr *helmfluxv1.HelmRelease
var ok bool
if hr, ok = obj.(*helmfluxv1.HelmRelease); !ok {
logger.Log("error", fmt.Sprintf("HelmRelease Event Watch received an invalid object: %#v", obj))
return helmfluxv1.HelmRelease{}, false
}
return *hr, true
}
func getCacheKey(obj interface{}) (string, error) {
key, err := cache.MetaNamespaceKeyFunc(obj)
if err != nil {
runtime.HandleError(err)
return "", err
}
return key, nil
}