/
service.go
538 lines (470 loc) · 17.7 KB
/
service.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
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
// Copyright 2023 LY Corporation
//
// 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 token
import (
"context"
"fmt"
"net/http"
"path/filepath"
"runtime/metrics"
"sync"
"sync/atomic"
"time"
"github.com/AthenZ/athenz/clients/go/zts"
"github.com/cenkalti/backoff"
"github.com/prometheus/client_golang/prometheus"
"golang.org/x/sync/singleflight"
"github.com/AthenZ/k8s-athenz-sia/v3/pkg/config"
"github.com/AthenZ/k8s-athenz-sia/v3/pkg/daemon"
extutil "github.com/AthenZ/k8s-athenz-sia/v3/pkg/util"
"github.com/AthenZ/k8s-athenz-sia/v3/third_party/log"
"github.com/AthenZ/k8s-athenz-sia/v3/third_party/util"
)
type tokenService struct {
shutdownChan chan struct{}
shutdownWg sync.WaitGroup
group singleflight.Group
accessTokenCache TokenCache
roleTokenCache TokenCache
// keyFile string
// certFile string
// serverCACert string
// endpoint string
ztsClient *zts.ZTSClient
saService string
tokenRESTAPI bool
tokenType mode
tokenDir string
tokenRefresh time.Duration
tokenExpiryInSecond int
roleAuthHeader string
useTokenServer bool
tokenServer *http.Server
tokenServerRunning bool
shutdownDelay time.Duration
shutdownTimeout time.Duration
}
func New(ctx context.Context, idCfg *config.IdentityConfig) (daemon.Daemon, error) {
if ctx.Err() != nil {
log.Info("Skipped token provider initiation")
return nil, nil
}
// initialize token cache with placeholder
tt := newType(idCfg.TokenType)
tokenExpiryInSecond := int(idCfg.TokenExpiry.Seconds())
accessTokenCache := NewLockedTokenCache("accesstoken", idCfg.Namespace, idCfg.PodName)
roleTokenCache := NewLockedTokenCache("roletoken", idCfg.Namespace, idCfg.PodName)
for _, dr := range idCfg.TargetDomainRoles {
domain, role := dr.Domain, dr.Role
if tt&mACCESS_TOKEN != 0 {
accessTokenCache.Store(CacheKey{Domain: domain, Role: role, MaxExpiry: tokenExpiryInSecond}, &AccessToken{})
}
if tt&mROLE_TOKEN != 0 {
roleTokenCache.Store(CacheKey{Domain: domain, Role: role, MinExpiry: tokenExpiryInSecond}, &RoleToken{})
}
}
ztsClient, err := newZTSClient(idCfg.Reloader, idCfg.ServerCACert, idCfg.Endpoint)
if err != nil {
return nil, err
}
saService := extutil.ServiceAccountToService(idCfg.ServiceAccount)
if saService == "" {
// TODO: get service from svc cert
// https://github.com/AthenZ/athenz/blob/73b25572656f289cce501b4c2fe78f86656082e7/libs/go/athenzutils/principal.go
// func ExtractServicePrincipal(x509Cert x509.Certificate) (string, error)
}
// register prometheus metrics
if err := prometheus.Register(accessTokenCache); err != nil {
return nil, err
}
if err := prometheus.Register(roleTokenCache); err != nil {
return nil, err
}
// setup token service
ts := &tokenService{
shutdownChan: make(chan struct{}, 1),
accessTokenCache: accessTokenCache,
roleTokenCache: roleTokenCache,
ztsClient: ztsClient,
saService: saService,
tokenRESTAPI: idCfg.TokenServerRESTAPI,
tokenType: tt,
tokenDir: idCfg.TokenDir,
tokenRefresh: idCfg.TokenRefresh,
tokenExpiryInSecond: tokenExpiryInSecond,
roleAuthHeader: idCfg.RoleAuthHeader,
useTokenServer: idCfg.UseTokenServer,
shutdownDelay: idCfg.ShutdownDelay,
shutdownTimeout: idCfg.ShutdownTimeout,
}
// initialize tokens on mode=refresh or TOKEN_DIR is set
if !idCfg.Init || idCfg.TokenDir != "" {
errs := ts.updateTokenCaches(ctx, config.DEFAULT_MAX_ELAPSED_TIME_ON_INIT)
// TODO: if cap(errs) == len(errs), implies all token updates failed, should be fatal
for _, err := range errs {
log.Errorf("Failed to refresh tokens after multiple retries: %s", err.Error())
}
if err := ts.writeFilesWithRetry(ctx, config.DEFAULT_MAX_ELAPSED_TIME_ON_INIT); err != nil {
log.Errorf("Failed to write token files after multiple retries: %s", err.Error())
}
}
// create token server
if idCfg.Init {
log.Infof("Token server is disabled for init mode: address[%s]", idCfg.TokenServerAddr)
return ts, nil
}
if idCfg.TokenServerAddr == "" || tt == 0 {
log.Infof("Token server is disabled due to insufficient options: address[%s], token-type[%s]", idCfg.TokenServerAddr, idCfg.TokenType)
return ts, nil
}
tokenServer := &http.Server{
Addr: idCfg.TokenServerAddr,
Handler: newHandlerFunc(ts, idCfg.TokenServerTimeout),
TLSConfig: nil,
}
if idCfg.TokenServerTLSCertPath != "" && idCfg.TokenServerTLSKeyPath != "" {
tokenServer.TLSConfig, err = NewTLSConfig(idCfg.TokenServerTLSCAPath, idCfg.TokenServerTLSCertPath, idCfg.TokenServerTLSKeyPath)
if err != nil {
return nil, err
}
}
ts.tokenServer = tokenServer
return ts, nil
}
// Start starts the token server, refreshes tokens periodically and reports memory usage periodically
func (ts *tokenService) Start(ctx context.Context) error {
if ctx.Err() != nil {
log.Info("Skipped token provider start")
return nil
}
// starts the token server
if ts.tokenServer != nil {
log.Infof("Starting token provider server[%s]", ts.tokenServer.Addr)
ts.shutdownWg.Add(1)
go func() {
defer ts.shutdownWg.Done()
listenAndServe := func() error {
if ts.tokenServer.TLSConfig != nil {
return ts.tokenServer.ListenAndServeTLS("", "")
}
return ts.tokenServer.ListenAndServe()
}
if err := listenAndServe(); err != nil && err != http.ErrServerClosed {
log.Fatalf("Failed to start token provider server: %s", err.Error())
}
log.Info("Stopped token provider server")
}()
if err := daemon.WaitForServerReady(ts.tokenServer.Addr, ts.tokenServer.TLSConfig != nil); err != nil {
log.Errorf("Failed to confirm token provider server ready: %s", err.Error())
return err
}
ts.tokenServerRunning = true
}
// refreshes tokens periodically
if ts.tokenRefresh > 0 {
t := time.NewTicker(ts.tokenRefresh)
ts.shutdownWg.Add(1)
go func() {
defer t.Stop()
defer ts.shutdownWg.Done()
for {
log.Infof("Will refresh cached tokens within %s", ts.tokenRefresh.String())
select {
case <-ts.shutdownChan:
log.Info("Stopped token provider daemon")
return
case <-t.C:
// skip refresh if context is done but Shutdown() is not called
if ctx.Err() != nil {
log.Info("Skipped to refresh cached tokens")
continue
}
// backoff retry until TOKEN_REFRESH_INTERVAL / 4 OR context is done
for _, err := range ts.updateTokenCaches(ctx, ts.tokenRefresh/4) {
log.Errorf("Failed to refresh tokens after multiple retries: %s", err.Error())
}
// backoff retry until TOKEN_REFRESH_INTERVAL / 4 OR context is done
if err := ts.writeFilesWithRetry(ctx, ts.tokenRefresh/4); err != nil {
log.Errorf("Failed to write token files after multiple retries: %s", err.Error())
}
}
}
}()
}
// reports memory usage periodically
reportTicker := time.NewTicker(time.Minute)
ts.shutdownWg.Add(1)
go func() {
defer reportTicker.Stop()
defer ts.shutdownWg.Done()
for {
select {
case <-ts.shutdownChan:
log.Info("Stopped memory reporter daemon")
return
case <-reportTicker.C:
// skip report if context is done but Shutdown() is not called
if ctx.Err() != nil {
continue
}
ts.reportMemory()
}
}
}()
return nil
}
func (ts *tokenService) Shutdown() {
log.Info("Initiating shutdown of token provider daemon ...")
close(ts.shutdownChan)
if ts.tokenServer != nil && ts.tokenServerRunning {
log.Infof("Delaying token provider server shutdown for %s to shutdown gracefully ...", ts.shutdownDelay.String())
time.Sleep(ts.shutdownDelay)
ctx, cancel := context.WithTimeout(context.Background(), ts.shutdownTimeout)
defer cancel()
ts.tokenServer.SetKeepAlivesEnabled(false)
if err := ts.tokenServer.Shutdown(ctx); err != nil {
// graceful shutdown error or timeout should be fatal
log.Errorf("Failed to shutdown token provider server: %s", err.Error())
}
}
// wait for graceful shutdown
ts.shutdownWg.Wait()
}
// GroupDoResult contains token and its requestID after singleFlight.group.Do()
type GroupDoResult struct {
requestID string
token Token
}
// requestTokenToZts sends a request to ZTS server to fetch either role token or access token.
// for mode, it only accepts mROLE_TOKEN or mACCESS_TOKEN
// it also stores in cache for you after successful fetch.
func (ts *tokenService) requestTokenToZts(k CacheKey, m mode, requestID string) (GroupDoResult, error) {
tokenName := "" // tokenName is used for logger (role token or access token only)
isRoleTokenRequested := m == mROLE_TOKEN
isAccessTokenRequested := m == mACCESS_TOKEN
if isRoleTokenRequested {
tokenName = "role token"
} else if isAccessTokenRequested {
tokenName = "access token"
} else {
return GroupDoResult{requestID: requestID, token: nil}, fmt.Errorf("Invalid mode: %d", m)
}
log.Debugf("Attempting to fetch %s from Athenz ZTS server: target[%s], requestID[%s]", tokenName, k.String(), requestID)
r, err, shared := ts.group.Do(k.UniqueId(tokenName), func() (interface{}, error) {
// define variables before request to ZTS
var fetchedToken Token
var err error
if isRoleTokenRequested {
fetchedToken, err = fetchRoleToken(ts.ztsClient, k)
} else { // isAccessTokenRequested
fetchedToken, err = fetchAccessToken(ts.ztsClient, k, ts.saService)
}
if err != nil {
log.Debugf("Failed to fetch %s from Athenz ZTS server: target[%s], requestID[%s]", tokenName, k.String(), requestID)
return GroupDoResult{requestID: requestID, token: nil}, err
}
if isRoleTokenRequested {
ts.roleTokenCache.Store(k, fetchedToken)
} else { // isAccessTokenRequested
ts.accessTokenCache.Store(k, fetchedToken)
}
log.Infof("Successfully updated %s cache: target[%s], requestID[%s]", tokenName, k.String(), requestID)
return GroupDoResult{requestID: requestID, token: fetchedToken}, nil
})
result := r.(GroupDoResult)
log.Debugf("requestID: [%s] handledRequestId: [%s] target: [%s]", requestID, result.requestID, k.String())
if shared && result.requestID != requestID { // if it is shared and not the actual performer:
if err == nil {
log.Infof("Successfully updated %s cache by coalescing requests to a leader request: target[%s], leaderRequestID[%s], requestID[%s]", tokenName, k.String(), result.requestID, requestID)
} else {
log.Debugf("Failed to fetch %s while coalescing requests to a leader request: target[%s], leaderRequestID[%s], requestID[%s], err[%s]", tokenName, k.String(), result.requestID, requestID, err)
}
}
return result, err
}
func (ts *tokenService) updateTokenCaches(ctx context.Context, maxElapsedTime time.Duration) []error {
var atErrorCount, rtErrorCount atomic.Int64
atTargets := ts.accessTokenCache.Keys()
rtTargets := ts.roleTokenCache.Keys()
log.Infof("Attempting to fetch tokens from Athenz ZTS server: access token targets[%v], role token targets[%v]...", atTargets, rtTargets)
var wg sync.WaitGroup
echan := make(chan error, len(atTargets)+len(rtTargets))
for _, t := range atTargets {
wg.Add(1)
go func(key CacheKey) {
defer wg.Done()
err := ts.updateTokenWithRetry(ctx, maxElapsedTime, key, mACCESS_TOKEN)
if err != nil {
echan <- err
atErrorCount.Add(1)
}
}(t)
}
for _, t := range rtTargets {
wg.Add(1)
go func(key CacheKey) {
defer wg.Done()
err := ts.updateTokenWithRetry(ctx, maxElapsedTime, key, mROLE_TOKEN)
if err != nil {
echan <- err
rtErrorCount.Add(1)
}
}(t)
}
// wait for ALL token updates to complete
wg.Wait()
log.Infof("Token cache updated. accesstoken:success[%d],error[%d]; roletoken:success[%d],error[%d]", int64(len(atTargets))-atErrorCount.Load(), atErrorCount.Load(), int64(len(rtTargets))-rtErrorCount.Load(), rtErrorCount.Load())
// collect errors
close(echan)
errs := make([]error, 0, len(atTargets)+len(rtTargets))
for err := range echan {
errs = append(errs, err)
}
return errs
}
func (ts *tokenService) updateTokenWithRetry(ctx context.Context, maxElapsedTime time.Duration, key CacheKey, tt mode) error {
// backoff config with first retry delay of 5s, and then 10s, 20s, ...
b := backoff.NewExponentialBackOff()
b.InitialInterval = 5 * time.Second
b.Multiplier = 2
b.MaxElapsedTime = maxElapsedTime
operation := func() error {
return ts.updateToken(key, tt)
}
notifyOnErr := func(err error, backoffDelay time.Duration) {
log.Errorf("Failed to refresh tokens: %s. Retrying in %s", err.Error(), backoffDelay)
}
return backoff.RetryNotify(operation, backoff.WithContext(b, ctx), notifyOnErr)
}
func (d *tokenService) updateToken(key CacheKey, tt mode) error {
updateAccessToken := func(key CacheKey) error {
_, err := d.requestTokenToZts(key, mACCESS_TOKEN, "daemon_access_token_update")
return err
}
updateRoleToken := func(key CacheKey) error {
_, err := d.requestTokenToZts(key, mROLE_TOKEN, "daemon_role_token_update")
return err
}
switch tt {
case mACCESS_TOKEN:
return updateAccessToken(key)
case mROLE_TOKEN:
return updateRoleToken(key)
default:
return fmt.Errorf("Invalid token type: %d", tt)
}
}
func (ts *tokenService) writeFilesWithRetry(ctx context.Context, maxElapsedTime time.Duration) error {
// backoff config with first retry delay of 5s, and then 10s, 20s, ...
b := backoff.NewExponentialBackOff()
b.InitialInterval = 5 * time.Second
b.Multiplier = 2
b.MaxElapsedTime = maxElapsedTime
return backoff.RetryNotify(ts.writeFiles, backoff.WithContext(b, ctx), func(err error, backoffDelay time.Duration) {
log.Errorf("Failed to write token files: %s. Retrying in %s", err.Error(), backoffDelay)
})
}
func (d *tokenService) writeFiles() error {
if d.tokenDir == "" {
log.Debugf("Skipping to write token files to directory[%s]", d.tokenDir)
return nil
}
w := util.NewWriter()
err := d.accessTokenCache.Range(func(k CacheKey, t Token) error {
domain := t.Domain()
role := t.Role()
at := t.Raw()
if at == "" {
// skip placeholder token added during daemon creation
return nil
}
log.Infof("[New Access Token] Domain: %s, Role: %s", domain, role)
outPath := filepath.Join(d.tokenDir, domain+":role."+role+".accesstoken")
log.Debugf("Saving Access Token[%d bytes] at %s", len(at), outPath)
if err := w.AddBytes(outPath, 0644, []byte(at)); err != nil {
return fmt.Errorf("unable to save access token: %w", err)
}
return nil
})
if err != nil {
return err
}
err = d.roleTokenCache.Range(func(k CacheKey, t Token) error {
domain := t.Domain()
role := t.Role()
rt := t.Raw()
if rt == "" {
// skip placeholder token added during daemon creation
return nil
}
log.Infof("[New Role Token] Domain: %s, Role: %s", domain, role)
outPath := filepath.Join(d.tokenDir, domain+":role."+role+".roletoken")
log.Debugf("Saving Role Token[%d bytes] at %s", len(rt), outPath)
if err := w.AddBytes(outPath, 0644, []byte(rt)); err != nil {
return fmt.Errorf("unable to save role token: %w", err)
}
return nil
})
if err != nil {
return err
}
return w.Save()
}
func (ts *tokenService) reportMemory() {
// gather golang metrics
const sysMemMetric = "/memory/classes/total:bytes" // go_memstats_sys_bytes
const heapMemMetric = "/memory/classes/heap/objects:bytes" // go_memstats_heap_alloc_bytes
const releasedHeapMemMetric = "/memory/classes/heap/released:bytes" // go_memstats_heap_released_bytes
// https://pkg.go.dev/runtime/metrics#pkg-examples
// https://github.com/prometheus/client_golang/blob/3f8bd73e9b6d1e20e8e1536622bd0fda8bb3cb50/prometheus/go_collector_latest.go#L32
samples := make([]metrics.Sample, 3)
samples[0].Name = sysMemMetric
samples[1].Name = heapMemMetric
samples[2].Name = releasedHeapMemMetric
metrics.Read(samples)
validSample := func(s metrics.Sample) float64 {
name, value := s.Name, s.Value
switch value.Kind() {
case metrics.KindUint64:
return float64(value.Uint64())
case metrics.KindFloat64:
return value.Float64()
case metrics.KindBad:
// Check if the metric is actually supported. If it's not, the resulting value will always have kind KindBad.
panic(fmt.Sprintf("%q: metric is no longer supported", name))
default:
// Check if the metrics specification has changed.
panic(fmt.Sprintf("%q: unexpected metric Kind: %v\n", name, value.Kind()))
}
}
sysMemValue := validSample(samples[0])
heapMemValue := validSample(samples[1])
releasedHeapMemValue := validSample(samples[2])
sysMemInUse := sysMemValue - releasedHeapMemValue
// gather token cache metrics
atcSize := ts.accessTokenCache.Size()
atcLen := ts.accessTokenCache.Len()
rtcSize := ts.roleTokenCache.Size()
rtcLen := ts.roleTokenCache.Len()
totalSize := atcSize + rtcSize
totalLen := atcLen + rtcLen
// report as log message
toMB := func(f float64) float64 {
return f / 1024 / 1024
}
log.Infof("system_memory_inuse[%.1fMB]; go_memstats_heap_alloc_bytes[%.1fMB]; accesstoken:cached_token_bytes[%.1fMB],entries[%d]; roletoken:cached_token_bytes[%.1fMB],entries[%d]; total:cached_token_bytes[%.1fMB],entries[%d]; cache_token_ratio:sys[%.1f%%],heap[%.1f%%]", toMB(sysMemInUse), toMB(heapMemValue), toMB(float64(atcSize)), atcLen, toMB(float64(rtcSize)), rtcLen, toMB(float64(totalSize)), totalLen, float64(totalSize)/sysMemInUse*100, float64(totalSize)/heapMemValue*100)
// TODO: memory triggers
// if mem > warn threshold, warning log
// if mem > error threshold, binary heap dump, i.e. debug.WriteHeapDump(os.Stdout.Fd())
}