-
-
Notifications
You must be signed in to change notification settings - Fork 293
/
config.go
1150 lines (1026 loc) · 38.8 KB
/
config.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
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
// Copyright 2015 Matthew Holt
//
// 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 certmagic
import (
"bytes"
"context"
"crypto"
"crypto/rand"
"crypto/tls"
"crypto/x509"
"crypto/x509/pkix"
"encoding/asn1"
"encoding/json"
"errors"
"fmt"
"io/fs"
weakrand "math/rand"
"net"
"net/url"
"strings"
"time"
"github.com/mholt/acmez"
"github.com/mholt/acmez/acme"
"go.uber.org/zap"
"golang.org/x/crypto/ocsp"
"golang.org/x/net/idna"
)
// Config configures a certificate manager instance.
// An empty Config is not valid: use New() to obtain
// a valid Config.
type Config struct {
// How much of a certificate's lifetime becomes the
// renewal window, which is the span of time at the
// end of the certificate's validity period in which
// it should be renewed; for most certificates, the
// global default is good, but for extremely short-
// lived certs, you may want to raise this to ~0.5.
RenewalWindowRatio float64
// An optional event callback clients can set
// to subscribe to certain things happening
// internally by this config; invocations are
// synchronous, so make them return quickly!
// Functions should honor context cancellation.
//
// An error should only be returned to advise
// the emitter to abort or cancel an upcoming
// event. Some events, especially those that have
// already happened, cannot be aborted. For example,
// cert_obtaining can be canceled, but
// cert_obtained cannot. Emitters may choose to
// ignore returned errors.
OnEvent func(ctx context.Context, event string, data map[string]any) error
// DefaultServerName specifies a server name
// to use when choosing a certificate if the
// ClientHello's ServerName field is empty.
DefaultServerName string
// FallbackServerName specifies a server name
// to use when choosing a certificate if the
// ClientHello's ServerName field doesn't match
// any available certificate.
// EXPERIMENTAL: Subject to change or removal.
FallbackServerName string
// The state needed to operate on-demand TLS;
// if non-nil, on-demand TLS is enabled and
// certificate operations are deferred to
// TLS handshakes (or as-needed).
// TODO: Can we call this feature "Reactive/Lazy/Passive TLS" instead?
OnDemand *OnDemandConfig
// Adds the must staple TLS extension to the CSR.
MustStaple bool
// Sources for getting new, managed certificates;
// the default Issuer is ACMEIssuer. If multiple
// issuers are specified, they will be tried in
// turn until one succeeds.
Issuers []Issuer
// The source of new private keys for certificates;
// the default KeySource is StandardKeyGenerator.
KeySource KeyGenerator
// CertSelection chooses one of the certificates
// with which the ClientHello will be completed;
// if not set, DefaultCertificateSelector will
// be used.
CertSelection CertificateSelector
// OCSP configures how OCSP is handled. By default,
// OCSP responses are fetched for every certificate
// with a responder URL, and cached on disk. Changing
// these defaults is STRONGLY discouraged unless you
// have a compelling reason to put clients at greater
// risk and reduce their privacy.
OCSP OCSPConfig
// The storage to access when storing or loading
// TLS assets. Default is the local file system.
Storage Storage
// CertMagic will verify the storage configuration
// is acceptable before obtaining a certificate
// to avoid information loss after an expensive
// operation. If you are absolutely 100% sure your
// storage is properly configured and has sufficient
// space, you can disable this check to reduce I/O
// if that is expensive for you.
// EXPERIMENTAL: Option subject to change or removal.
DisableStorageCheck bool
// Set a logger to enable logging. If not set,
// a default logger will be created.
Logger *zap.Logger
// required pointer to the in-memory cert cache
certCache *Cache
}
// NewDefault makes a valid config based on the package
// Default config. Most users will call this function
// instead of New() since most use cases require only a
// single config for any and all certificates.
//
// If your requirements are more advanced (for example,
// multiple configs depending on the certificate), then use
// New() instead. (You will need to make your own Cache
// first.) If you only need a single Config to manage your
// certs (even if that config changes, as long as it is the
// only one), customize the Default package variable before
// calling NewDefault().
//
// All calls to NewDefault() will return configs that use the
// same, default certificate cache. All configs returned
// by NewDefault() are based on the values of the fields of
// Default at the time it is called.
//
// This is the only way to get a config that uses the
// default certificate cache.
func NewDefault() *Config {
defaultCacheMu.Lock()
if defaultCache == nil {
defaultCache = NewCache(CacheOptions{
// the cache will likely need to renew certificates,
// so it will need to know how to do that, which
// depends on the certificate being managed and which
// can change during the lifetime of the cache; this
// callback makes it possible to get the latest and
// correct config with which to manage the cert,
// but if the user does not provide one, we can only
// assume that we are to use the default config
GetConfigForCert: func(Certificate) (*Config, error) {
return NewDefault(), nil
},
Logger: Default.Logger,
})
}
certCache := defaultCache
defaultCacheMu.Unlock()
return newWithCache(certCache, Default)
}
// New makes a new, valid config based on cfg and
// uses the provided certificate cache. certCache
// MUST NOT be nil or this function will panic.
//
// Use this method when you have an advanced use case
// that requires a custom certificate cache and config
// that may differ from the Default. For example, if
// not all certificates are managed/renewed the same
// way, you need to make your own Cache value with a
// GetConfigForCert callback that returns the correct
// configuration for each certificate. However, for
// the vast majority of cases, there will be only a
// single Config, thus the default cache (which always
// uses the default Config) and default config will
// suffice, and you should use NewDefault() instead.
func New(certCache *Cache, cfg Config) *Config {
if certCache == nil {
panic("a certificate cache is required")
}
if certCache.options.GetConfigForCert == nil {
panic("cache must have GetConfigForCert set in its options")
}
return newWithCache(certCache, cfg)
}
// newWithCache ensures that cfg is a valid config by populating
// zero-value fields from the Default Config. If certCache is
// nil, this function panics.
func newWithCache(certCache *Cache, cfg Config) *Config {
if certCache == nil {
panic("cannot make a valid config without a pointer to a certificate cache")
}
if cfg.OnDemand == nil {
cfg.OnDemand = Default.OnDemand
}
if !cfg.MustStaple {
cfg.MustStaple = Default.MustStaple
}
if cfg.Issuers == nil {
cfg.Issuers = Default.Issuers
if cfg.Issuers == nil {
// at least one issuer is absolutely required if not nil
cfg.Issuers = []Issuer{NewACMEIssuer(&cfg, DefaultACME)}
}
}
if cfg.RenewalWindowRatio == 0 {
cfg.RenewalWindowRatio = Default.RenewalWindowRatio
}
if cfg.OnEvent == nil {
cfg.OnEvent = Default.OnEvent
}
if cfg.KeySource == nil {
cfg.KeySource = Default.KeySource
}
if cfg.DefaultServerName == "" {
cfg.DefaultServerName = Default.DefaultServerName
}
if cfg.FallbackServerName == "" {
cfg.FallbackServerName = Default.FallbackServerName
}
if cfg.Storage == nil {
cfg.Storage = Default.Storage
}
if cfg.Logger == nil {
cfg.Logger = Default.Logger
}
// absolutely don't allow a nil storage,
// because that would make almost anything
// a config can do pointless
if cfg.Storage == nil {
cfg.Storage = defaultFileStorage
}
// absolutely don't allow a nil logger either,
// because that would result in panics
if cfg.Logger == nil {
cfg.Logger = defaultLogger
}
cfg.certCache = certCache
return &cfg
}
// ManageSync causes the certificates for domainNames to be managed
// according to cfg. If cfg.OnDemand is not nil, then this simply
// whitelists the domain names and defers the certificate operations
// to when they are needed. Otherwise, the certificates for each
// name are loaded from storage or obtained from the CA. If loaded
// from storage, they are renewed if they are expiring or expired.
// It then caches the certificate in memory and is prepared to serve
// them up during TLS handshakes.
//
// Note that name whitelisting for on-demand management only takes
// effect if cfg.OnDemand.DecisionFunc is not set (is nil); it will
// not overwrite an existing DecisionFunc, nor will it overwrite
// its decision; i.e. the implicit whitelist is only used if no
// DecisionFunc is set.
//
// This method is synchronous, meaning that certificates for all
// domainNames must be successfully obtained (or renewed) before
// it returns. It returns immediately on the first error for any
// of the given domainNames. This behavior is recommended for
// interactive use (i.e. when an administrator is present) so
// that errors can be reported and fixed immediately.
func (cfg *Config) ManageSync(ctx context.Context, domainNames []string) error {
return cfg.manageAll(ctx, domainNames, false)
}
// ManageAsync is the same as ManageSync, except that ACME
// operations are performed asynchronously (in the background).
// This method returns before certificates are ready. It is
// crucial that the administrator monitors the logs and is
// notified of any errors so that corrective action can be
// taken as soon as possible. Any errors returned from this
// method occurred before ACME transactions started.
//
// As long as logs are monitored, this method is typically
// recommended for non-interactive environments.
//
// If there are failures loading, obtaining, or renewing a
// certificate, it will be retried with exponential backoff
// for up to about 30 days, with a maximum interval of about
// 24 hours. Cancelling ctx will cancel retries and shut down
// any goroutines spawned by ManageAsync.
func (cfg *Config) ManageAsync(ctx context.Context, domainNames []string) error {
return cfg.manageAll(ctx, domainNames, true)
}
// ClientCredentials returns a list of TLS client certificate chains for the given identifiers.
// The return value can be used in a tls.Config to enable client authentication using managed certificates.
// Any certificates that need to be obtained or renewed for these identifiers will be managed accordingly.
func (cfg *Config) ClientCredentials(ctx context.Context, identifiers []string) ([]tls.Certificate, error) {
err := cfg.manageAll(ctx, identifiers, false)
if err != nil {
return nil, err
}
var chains []tls.Certificate
for _, id := range identifiers {
certRes, err := cfg.loadCertResourceAnyIssuer(ctx, id)
if err != nil {
return chains, err
}
chain, err := tls.X509KeyPair(certRes.CertificatePEM, certRes.PrivateKeyPEM)
if err != nil {
return chains, err
}
chains = append(chains, chain)
}
return chains, nil
}
func (cfg *Config) manageAll(ctx context.Context, domainNames []string, async bool) error {
if ctx == nil {
ctx = context.Background()
}
for _, domainName := range domainNames {
// if on-demand is configured, defer obtain and renew operations
if cfg.OnDemand != nil {
if !cfg.OnDemand.allowlistContains(domainName) {
cfg.OnDemand.hostAllowlist = append(cfg.OnDemand.hostAllowlist, domainName)
}
continue
}
// TODO: consider doing this in a goroutine if async, to utilize multiple cores while loading certs
// otherwise, begin management immediately
err := cfg.manageOne(ctx, domainName, async)
if err != nil {
return err
}
}
return nil
}
func (cfg *Config) manageOne(ctx context.Context, domainName string, async bool) error {
// first try loading existing certificate from storage
cert, err := cfg.CacheManagedCertificate(ctx, domainName)
if err != nil {
if !errors.Is(err, fs.ErrNotExist) {
return fmt.Errorf("%s: caching certificate: %v", domainName, err)
}
// if we don't have one in storage, obtain one
obtain := func() error {
var err error
if async {
err = cfg.ObtainCertAsync(ctx, domainName)
} else {
err = cfg.ObtainCertSync(ctx, domainName)
}
if err != nil {
return fmt.Errorf("%s: obtaining certificate: %w", domainName, err)
}
cert, err = cfg.CacheManagedCertificate(ctx, domainName)
if err != nil {
return fmt.Errorf("%s: caching certificate after obtaining it: %v", domainName, err)
}
return nil
}
if async {
// Leave the job name empty so as to allow duplicate 'obtain'
// jobs; this is because Caddy calls ManageAsync() before the
// previous config is stopped (and before its context is
// canceled), which means that if an obtain job is still
// running for the same domain, Submit() would not queue the
// new one because it is still running, even though it is
// (probably) about to be canceled (it might not if the new
// config fails to finish loading, however). In any case, we
// presume it is safe to enqueue a duplicate obtain job because
// either the old one (or sometimes the new one) is about to be
// canceled. This seems like reasonable logic for any consumer
// of this lib. See https://github.com/caddyserver/caddy/issues/3202
jm.Submit(cfg.Logger, "", obtain)
return nil
}
return obtain()
}
// for an existing certificate, make sure it is renewed; or if it is revoked,
// force a renewal even if it's not expiring
renew := func() error {
// first, ensure status is not revoked (it was just refreshed in CacheManagedCertificate above)
if !cert.Expired() && cert.ocsp != nil && cert.ocsp.Status == ocsp.Revoked {
_, err = cfg.forceRenew(ctx, cfg.Logger, cert)
return err
}
// otherwise, simply renew the certificate if needed
if cert.NeedsRenewal(cfg) {
var err error
if async {
err = cfg.RenewCertAsync(ctx, domainName, false)
} else {
err = cfg.RenewCertSync(ctx, domainName, false)
}
if err != nil {
return fmt.Errorf("%s: renewing certificate: %w", domainName, err)
}
// successful renewal, so update in-memory cache
_, err = cfg.reloadManagedCertificate(ctx, cert)
if err != nil {
return fmt.Errorf("%s: reloading renewed certificate into memory: %v", domainName, err)
}
}
return nil
}
if async {
jm.Submit(cfg.Logger, "renew_"+domainName, renew)
return nil
}
return renew()
}
// Unmanage causes the certificates for domainNames to stop being managed.
// If there are certificates for the supplied domain names in the cache, they
// are evicted from the cache.
func (cfg *Config) Unmanage(domainNames []string) {
var deleteQueue []Certificate
for _, domainName := range domainNames {
certs := cfg.certCache.AllMatchingCertificates(domainName)
for _, cert := range certs {
if !cert.managed {
continue
}
deleteQueue = append(deleteQueue, cert)
}
}
cfg.certCache.mu.Lock()
for _, cert := range deleteQueue {
cfg.certCache.removeCertificate(cert)
}
cfg.certCache.mu.Unlock()
}
// ObtainCertSync generates a new private key and obtains a certificate for
// name using cfg in the foreground; i.e. interactively and without retries.
// It stows the renewed certificate and its assets in storage if successful.
// It DOES NOT load the certificate into the in-memory cache. This method
// is a no-op if storage already has a certificate for name.
func (cfg *Config) ObtainCertSync(ctx context.Context, name string) error {
return cfg.obtainCert(ctx, name, true)
}
// ObtainCertAsync is the same as ObtainCertSync(), except it runs in the
// background; i.e. non-interactively, and with retries if it fails.
func (cfg *Config) ObtainCertAsync(ctx context.Context, name string) error {
return cfg.obtainCert(ctx, name, false)
}
func (cfg *Config) obtainCert(ctx context.Context, name string, interactive bool) error {
if len(cfg.Issuers) == 0 {
return fmt.Errorf("no issuers configured; impossible to obtain or check for existing certificate in storage")
}
// if storage has all resources for this certificate, obtain is a no-op
if cfg.storageHasCertResourcesAnyIssuer(ctx, name) {
return nil
}
// ensure storage is writeable and readable
// TODO: this is not necessary every time; should only perform check once every so often for each storage, which may require some global state...
err := cfg.checkStorage(ctx)
if err != nil {
return fmt.Errorf("failed storage check: %v - storage is probably misconfigured", err)
}
log := cfg.Logger.Named("obtain")
log.Info("acquiring lock", zap.String("identifier", name))
// ensure idempotency of the obtain operation for this name
lockKey := cfg.lockKey(certIssueLockOp, name)
err = acquireLock(ctx, cfg.Storage, lockKey)
if err != nil {
return fmt.Errorf("unable to acquire lock '%s': %v", lockKey, err)
}
defer func() {
log.Info("releasing lock", zap.String("identifier", name))
if err := releaseLock(ctx, cfg.Storage, lockKey); err != nil {
log.Error("unable to unlock",
zap.String("identifier", name),
zap.String("lock_key", lockKey),
zap.Error(err))
}
}()
log.Info("lock acquired", zap.String("identifier", name))
f := func(ctx context.Context) error {
// check if obtain is still needed -- might have been obtained during lock
if cfg.storageHasCertResourcesAnyIssuer(ctx, name) {
log.Info("certificate already exists in storage", zap.String("identifier", name))
return nil
}
log.Info("obtaining certificate", zap.String("identifier", name))
if err := cfg.emit(ctx, "cert_obtaining", map[string]any{"identifier": name}); err != nil {
return fmt.Errorf("obtaining certificate aborted by event handler: %w", err)
}
// if storage has a private key already, use it; otherwise we'll generate our own
privKey, privKeyPEM, issuers, err := cfg.reusePrivateKey(ctx, name)
if err != nil {
return err
}
if privKey == nil {
privKey, err = cfg.KeySource.GenerateKey()
if err != nil {
return err
}
privKeyPEM, err = PEMEncodePrivateKey(privKey)
if err != nil {
return err
}
}
csr, err := cfg.generateCSR(privKey, []string{name})
if err != nil {
return err
}
// try to obtain from each issuer until we succeed
var issuedCert *IssuedCertificate
var issuerUsed Issuer
var issuerKeys []string
for i, issuer := range issuers {
issuerKeys = append(issuerKeys, issuer.IssuerKey())
log.Debug(fmt.Sprintf("trying issuer %d/%d", i+1, len(cfg.Issuers)),
zap.String("issuer", issuer.IssuerKey()))
if prechecker, ok := issuer.(PreChecker); ok {
err = prechecker.PreCheck(ctx, []string{name}, interactive)
if err != nil {
continue
}
}
issuedCert, err = issuer.Issue(ctx, csr)
if err == nil {
issuerUsed = issuer
break
}
// err is usually wrapped, which is nice for simply printing it, but
// with our structured error logs we only need the problem string
errToLog := err
var problem acme.Problem
if errors.As(err, &problem) {
errToLog = problem
}
log.Error("could not get certificate from issuer",
zap.String("identifier", name),
zap.String("issuer", issuer.IssuerKey()),
zap.Error(errToLog))
}
if err != nil {
cfg.emit(ctx, "cert_failed", map[string]any{
"renewal": false,
"identifier": name,
"issuers": issuerKeys,
"error": err,
})
// only the error from the last issuer will be returned, but we logged the others
return fmt.Errorf("[%s] Obtain: %w", name, err)
}
// success - immediately save the certificate resource
certRes := CertificateResource{
SANs: namesFromCSR(csr),
CertificatePEM: issuedCert.Certificate,
PrivateKeyPEM: privKeyPEM,
IssuerData: issuedCert.Metadata,
issuerKey: issuerUsed.IssuerKey(),
}
err = cfg.saveCertResource(ctx, issuerUsed, certRes)
if err != nil {
return fmt.Errorf("[%s] Obtain: saving assets: %v", name, err)
}
log.Info("certificate obtained successfully", zap.String("identifier", name))
cfg.emit(ctx, "cert_obtained", map[string]any{
"renewal": false,
"identifier": name,
"issuers": issuerUsed.IssuerKey(),
"storage_key": certRes.NamesKey(),
})
return nil
}
if interactive {
err = f(ctx)
} else {
err = doWithRetry(ctx, log, f)
}
return err
}
// reusePrivateKey looks for a private key for domain in storage in the configured issuers
// paths. For the first private key it finds, it returns that key both decoded and PEM-encoded,
// as well as the reordered list of issuers to use instead of cfg.Issuers (because if a key
// is found, that issuer should be tried first, so it is moved to the front in a copy of
// cfg.Issuers).
func (cfg *Config) reusePrivateKey(ctx context.Context, domain string) (privKey crypto.PrivateKey, privKeyPEM []byte, issuers []Issuer, err error) {
// make a copy of cfg.Issuers so that if we have to reorder elements, we don't
// inadvertently mutate the configured issuers (see append calls below)
issuers = make([]Issuer, len(cfg.Issuers))
copy(issuers, cfg.Issuers)
for i, issuer := range issuers {
// see if this issuer location in storage has a private key for the domain
privateKeyStorageKey := StorageKeys.SitePrivateKey(issuer.IssuerKey(), domain)
privKeyPEM, err = cfg.Storage.Load(ctx, privateKeyStorageKey)
if errors.Is(err, fs.ErrNotExist) {
err = nil // obviously, it's OK to not have a private key; so don't prevent obtaining a cert
continue
}
if err != nil {
return nil, nil, nil, fmt.Errorf("loading existing private key for reuse with issuer %s: %v", issuer.IssuerKey(), err)
}
// we loaded a private key; try decoding it so we can use it
privKey, err = PEMDecodePrivateKey(privKeyPEM)
if err != nil {
return nil, nil, nil, err
}
// since the private key was found in storage for this issuer, move it
// to the front of the list so we prefer this issuer first
issuers = append([]Issuer{issuer}, append(issuers[:i], issuers[i+1:]...)...)
break
}
return
}
// storageHasCertResourcesAnyIssuer returns true if storage has all the
// certificate resources in storage from any configured issuer. It checks
// all configured issuers in order.
func (cfg *Config) storageHasCertResourcesAnyIssuer(ctx context.Context, name string) bool {
for _, iss := range cfg.Issuers {
if cfg.storageHasCertResources(ctx, iss, name) {
return true
}
}
return false
}
// RenewCertSync renews the certificate for name using cfg in the foreground;
// i.e. interactively and without retries. It stows the renewed certificate
// and its assets in storage if successful. It DOES NOT update the in-memory
// cache with the new certificate. The certificate will not be renewed if it
// is not close to expiring unless force is true.
//
// Renewing a certificate is the same as obtaining a certificate, except that
// the existing private key already in storage is reused.
func (cfg *Config) RenewCertSync(ctx context.Context, name string, force bool) error {
return cfg.renewCert(ctx, name, force, true)
}
// RenewCertAsync is the same as RenewCertSync(), except it runs in the
// background; i.e. non-interactively, and with retries if it fails.
func (cfg *Config) RenewCertAsync(ctx context.Context, name string, force bool) error {
return cfg.renewCert(ctx, name, force, false)
}
func (cfg *Config) renewCert(ctx context.Context, name string, force, interactive bool) error {
if len(cfg.Issuers) == 0 {
return fmt.Errorf("no issuers configured; impossible to renew or check existing certificate in storage")
}
// ensure storage is writeable and readable
// TODO: this is not necessary every time; should only perform check once every so often for each storage, which may require some global state...
err := cfg.checkStorage(ctx)
if err != nil {
return fmt.Errorf("failed storage check: %v - storage is probably misconfigured", err)
}
log := cfg.Logger.Named("renew")
log.Info("acquiring lock", zap.String("identifier", name))
// ensure idempotency of the renew operation for this name
lockKey := cfg.lockKey(certIssueLockOp, name)
err = acquireLock(ctx, cfg.Storage, lockKey)
if err != nil {
return fmt.Errorf("unable to acquire lock '%s': %v", lockKey, err)
}
defer func() {
log.Info("releasing lock", zap.String("identifier", name))
if err := releaseLock(ctx, cfg.Storage, lockKey); err != nil {
log.Error("unable to unlock",
zap.String("identifier", name),
zap.String("lock_key", lockKey),
zap.Error(err))
}
}()
log.Info("lock acquired", zap.String("identifier", name))
f := func(ctx context.Context) error {
// prepare for renewal (load PEM cert, key, and meta)
certRes, err := cfg.loadCertResourceAnyIssuer(ctx, name)
if err != nil {
return err
}
// check if renew is still needed - might have been renewed while waiting for lock
timeLeft, needsRenew := cfg.managedCertNeedsRenewal(certRes)
if !needsRenew {
if force {
log.Info("certificate does not need to be renewed, but renewal is being forced",
zap.String("identifier", name),
zap.Duration("remaining", timeLeft))
} else {
log.Info("certificate appears to have been renewed already",
zap.String("identifier", name),
zap.Duration("remaining", timeLeft))
return nil
}
}
log.Info("renewing certificate",
zap.String("identifier", name),
zap.Duration("remaining", timeLeft))
if err := cfg.emit(ctx, "cert_obtaining", map[string]any{
"renewal": true,
"identifier": name,
"forced": force,
"remaining": timeLeft,
"issuer": certRes.issuerKey, // previous/current issuer
}); err != nil {
return fmt.Errorf("renewing certificate aborted by event handler: %w", err)
}
privateKey, err := PEMDecodePrivateKey(certRes.PrivateKeyPEM)
if err != nil {
return err
}
csr, err := cfg.generateCSR(privateKey, []string{name})
if err != nil {
return err
}
// try to obtain from each issuer until we succeed
var issuedCert *IssuedCertificate
var issuerUsed Issuer
var issuerKeys []string
for _, issuer := range cfg.Issuers {
issuerKeys = append(issuerKeys, issuer.IssuerKey())
if prechecker, ok := issuer.(PreChecker); ok {
err = prechecker.PreCheck(ctx, []string{name}, interactive)
if err != nil {
continue
}
}
issuedCert, err = issuer.Issue(ctx, csr)
if err == nil {
issuerUsed = issuer
break
}
// err is usually wrapped, which is nice for simply printing it, but
// with our structured error logs we only need the problem string
errToLog := err
var problem acme.Problem
if errors.As(err, &problem) {
errToLog = problem
}
log.Error("could not get certificate from issuer",
zap.String("identifier", name),
zap.String("issuer", issuer.IssuerKey()),
zap.Error(errToLog))
}
if err != nil {
cfg.emit(ctx, "cert_failed", map[string]any{
"renewal": true,
"identifier": name,
"remaining": timeLeft,
"issuers": issuerKeys,
"storage_key": certRes.NamesKey(),
"error": err,
})
// only the error from the last issuer will be returned, but we logged the others
return fmt.Errorf("[%s] Renew: %w", name, err)
}
// success - immediately save the renewed certificate resource
newCertRes := CertificateResource{
SANs: namesFromCSR(csr),
CertificatePEM: issuedCert.Certificate,
PrivateKeyPEM: certRes.PrivateKeyPEM,
IssuerData: issuedCert.Metadata,
issuerKey: issuerUsed.IssuerKey(),
}
err = cfg.saveCertResource(ctx, issuerUsed, newCertRes)
if err != nil {
return fmt.Errorf("[%s] Renew: saving assets: %v", name, err)
}
log.Info("certificate renewed successfully", zap.String("identifier", name))
cfg.emit(ctx, "cert_obtained", map[string]any{
"renewal": true,
"remaining": timeLeft,
"identifier": name,
"issuer": issuerUsed.IssuerKey(),
"storage_key": certRes.NamesKey(),
})
return nil
}
if interactive {
err = f(ctx)
} else {
err = doWithRetry(ctx, log, f)
}
return err
}
func (cfg *Config) generateCSR(privateKey crypto.PrivateKey, sans []string) (*x509.CertificateRequest, error) {
csrTemplate := new(x509.CertificateRequest)
for _, name := range sans {
if ip := net.ParseIP(name); ip != nil {
csrTemplate.IPAddresses = append(csrTemplate.IPAddresses, ip)
} else if strings.Contains(name, "@") {
csrTemplate.EmailAddresses = append(csrTemplate.EmailAddresses, name)
} else if u, err := url.Parse(name); err == nil && strings.Contains(name, "/") {
csrTemplate.URIs = append(csrTemplate.URIs, u)
} else {
// convert IDNs to ASCII according to RFC 5280 section 7
normalizedName, err := idna.ToASCII(name)
if err != nil {
return nil, fmt.Errorf("converting identifier '%s' to ASCII: %v", name, err)
}
csrTemplate.DNSNames = append(csrTemplate.DNSNames, normalizedName)
}
}
if cfg.MustStaple {
csrTemplate.ExtraExtensions = append(csrTemplate.ExtraExtensions, mustStapleExtension)
}
csrDER, err := x509.CreateCertificateRequest(rand.Reader, csrTemplate, privateKey)
if err != nil {
return nil, err
}
return x509.ParseCertificateRequest(csrDER)
}
// RevokeCert revokes the certificate for domain via ACME protocol. It requires
// that cfg.Issuers is properly configured with the same issuer that issued the
// certificate being revoked. See RFC 5280 §5.3.1 for reason codes.
//
// The certificate assets are deleted from storage after successful revocation
// to prevent reuse.
func (cfg *Config) RevokeCert(ctx context.Context, domain string, reason int, interactive bool) error {
for i, issuer := range cfg.Issuers {
issuerKey := issuer.IssuerKey()
rev, ok := issuer.(Revoker)
if !ok {
return fmt.Errorf("issuer %d (%s) is not a Revoker", i, issuerKey)
}
certRes, err := cfg.loadCertResource(ctx, issuer, domain)
if err != nil {
return err
}
if !cfg.Storage.Exists(ctx, StorageKeys.SitePrivateKey(issuerKey, domain)) {
return fmt.Errorf("private key not found for %s", certRes.SANs)
}
err = rev.Revoke(ctx, certRes, reason)
if err != nil {
return fmt.Errorf("issuer %d (%s): %v", i, issuerKey, err)
}
err = cfg.deleteSiteAssets(ctx, issuerKey, domain)
if err != nil {
return fmt.Errorf("certificate revoked, but unable to fully clean up assets from issuer %s: %v", issuerKey, err)
}
}
return nil
}
// TLSConfig is an opinionated method that returns a recommended, modern
// TLS configuration that can be used to configure TLS listeners. Aside
// from safe, modern defaults, this method sets two critical fields on the
// TLS config which are required to enable automatic certificate
// management: GetCertificate and NextProtos.
//
// The GetCertificate field is necessary to get certificates from memory
// or storage, including both manual and automated certificates. You
// should only change this field if you know what you are doing.
//
// The NextProtos field is pre-populated with a special value to enable
// solving the TLS-ALPN ACME challenge. Because this method does not
// assume any particular protocols after the TLS handshake is completed,
// you will likely need to customize the NextProtos field by prepending
// your application's protocols to the slice. For example, to serve
// HTTP, you will need to prepend "h2" and "http/1.1" values. Be sure to
// leave the acmez.ACMETLS1Protocol value intact, however, or TLS-ALPN
// challenges will fail (which may be acceptable if you are not using
// ACME, or specifically, the TLS-ALPN challenge).
//
// Unlike the package TLS() function, this method does not, by itself,
// enable certificate management for any domain names.
func (cfg *Config) TLSConfig() *tls.Config {
return &tls.Config{
// these two fields necessary for TLS-ALPN challenge
GetCertificate: cfg.GetCertificate,
NextProtos: []string{acmez.ACMETLS1Protocol},
// the rest recommended for modern TLS servers
MinVersion: tls.VersionTLS12,
CurvePreferences: []tls.CurveID{
tls.X25519,
tls.CurveP256,
},
CipherSuites: preferredDefaultCipherSuites(),
PreferServerCipherSuites: true,
}
}
// getChallengeInfo loads the challenge info from either the internal challenge memory
// or the external storage (implying distributed solving). The second return value
// indicates whether challenge info was loaded from external storage. If true, the
// challenge is being solved in a distributed fashion; if false, from internal memory.
// If no matching challenge information can be found, an error is returned.
func (cfg *Config) getChallengeInfo(ctx context.Context, identifier string) (Challenge, bool, error) {
// first, check if our process initiated this challenge; if so, just return it
chalData, ok := GetACMEChallenge(identifier)
if ok {
return chalData, false, nil
}
// otherwise, perhaps another instance in the cluster initiated it; check
// the configured storage to retrieve challenge data
var chalInfo acme.Challenge
var chalInfoBytes []byte
var tokenKey string
for _, issuer := range cfg.Issuers {
ds := distributedSolver{
storage: cfg.Storage,
storageKeyIssuerPrefix: storageKeyACMECAPrefix(issuer.IssuerKey()),
}
tokenKey = ds.challengeTokensKey(identifier)
var err error
chalInfoBytes, err = cfg.Storage.Load(ctx, tokenKey)
if err == nil {
break
}
if errors.Is(err, fs.ErrNotExist) {
continue
}
return Challenge{}, false, fmt.Errorf("opening distributed challenge token file %s: %v", tokenKey, err)
}