summaryrefslogtreecommitdiff
path: root/libpod/kube.go
blob: 3cb0489b397306caeef30797d4c220cf2940bf99 (plain)
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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
package libpod

import (
	"context"
	"errors"
	"fmt"
	"math/rand"
	"os"
	"reflect"
	"sort"
	"strconv"
	"strings"
	"time"
	"unicode/utf8"

	"github.com/containers/common/libnetwork/types"
	"github.com/containers/common/pkg/config"
	cutil "github.com/containers/common/pkg/util"
	"github.com/containers/podman/v4/libpod/define"
	"github.com/containers/podman/v4/pkg/env"
	v1 "github.com/containers/podman/v4/pkg/k8s.io/api/core/v1"
	"github.com/containers/podman/v4/pkg/k8s.io/apimachinery/pkg/api/resource"
	v12 "github.com/containers/podman/v4/pkg/k8s.io/apimachinery/pkg/apis/meta/v1"
	"github.com/containers/podman/v4/pkg/k8s.io/apimachinery/pkg/util/intstr"
	"github.com/containers/podman/v4/pkg/lookup"
	"github.com/containers/podman/v4/pkg/namespaces"
	"github.com/containers/podman/v4/pkg/specgen"
	"github.com/containers/podman/v4/pkg/util"
	"github.com/opencontainers/runtime-spec/specs-go"
	"github.com/opencontainers/runtime-tools/generate"
	"github.com/sirupsen/logrus"
)

// GenerateForKube takes a slice of libpod containers and generates
// one v1.Pod description that includes just a single container.
func GenerateForKube(ctx context.Context, ctrs []*Container) (*v1.Pod, error) {
	// Generate the v1.Pod yaml description
	return simplePodWithV1Containers(ctx, ctrs)
}

// GenerateForKube takes a slice of libpod containers and generates
// one v1.Pod description
func (p *Pod) GenerateForKube(ctx context.Context) (*v1.Pod, []v1.ServicePort, error) {
	// Generate the v1.Pod yaml description
	var (
		ports        []v1.ContainerPort
		servicePorts []v1.ServicePort
	)

	allContainers, err := p.allContainers()
	if err != nil {
		return nil, servicePorts, err
	}
	// If the pod has no containers, no sense to generate YAML
	if len(allContainers) == 0 {
		return nil, servicePorts, fmt.Errorf("pod %s has no containers", p.ID())
	}
	// If only an infra container is present, makes no sense to generate YAML
	if len(allContainers) == 1 && p.HasInfraContainer() {
		return nil, servicePorts, fmt.Errorf("pod %s only has an infra container", p.ID())
	}

	extraHost := make([]v1.HostAlias, 0)
	hostNetwork := false
	if p.HasInfraContainer() {
		infraContainer, err := p.getInfraContainer()
		if err != nil {
			return nil, servicePorts, err
		}
		for _, host := range infraContainer.config.ContainerNetworkConfig.HostAdd {
			hostSli := strings.SplitN(host, ":", 2)
			if len(hostSli) != 2 {
				return nil, servicePorts, errors.New("invalid hostAdd")
			}
			extraHost = append(extraHost, v1.HostAlias{
				IP:        hostSli[1],
				Hostnames: []string{hostSli[0]},
			})
		}
		ports, err = portMappingToContainerPort(infraContainer.config.PortMappings)
		if err != nil {
			return nil, servicePorts, err
		}
		spState := newServicePortState()
		servicePorts, err = spState.containerPortsToServicePorts(ports)
		if err != nil {
			return nil, servicePorts, err
		}
		hostNetwork = infraContainer.NetworkMode() == string(namespaces.NetworkMode(specgen.Host))
	}
	pod, err := p.podWithContainers(ctx, allContainers, ports, hostNetwork)
	if err != nil {
		return nil, servicePorts, err
	}
	pod.Spec.HostAliases = extraHost

	// vendor/k8s.io/api/core/v1/types.go: v1.Container cannot save restartPolicy
	// so set it at here
	for _, ctr := range allContainers {
		if !ctr.IsInfra() {
			switch ctr.config.RestartPolicy {
			case define.RestartPolicyAlways:
				pod.Spec.RestartPolicy = v1.RestartPolicyAlways
			case define.RestartPolicyOnFailure:
				pod.Spec.RestartPolicy = v1.RestartPolicyOnFailure
			case define.RestartPolicyNo:
				pod.Spec.RestartPolicy = v1.RestartPolicyNever
			default: // some pod create from cmdline, such as "", so set it to Never
				pod.Spec.RestartPolicy = v1.RestartPolicyNever
			}
			break
		}
	}

	if p.SharesPID() {
		// unfortunately, go doesn't have a nice way to specify a pointer to a bool
		b := true
		pod.Spec.ShareProcessNamespace = &b
	}

	return pod, servicePorts, nil
}

func (p *Pod) getInfraContainer() (*Container, error) {
	infraID, err := p.InfraContainerID()
	if err != nil {
		return nil, err
	}
	return p.runtime.GetContainer(infraID)
}

// GenerateForKube generates a v1.PersistentVolumeClaim from a libpod volume.
func (v *Volume) GenerateForKube() *v1.PersistentVolumeClaim {
	annotations := make(map[string]string)
	annotations[util.VolumeDriverAnnotation] = v.Driver()

	for k, v := range v.Options() {
		switch k {
		case "o":
			annotations[util.VolumeMountOptsAnnotation] = v
		case "device":
			annotations[util.VolumeDeviceAnnotation] = v
		case "type":
			annotations[util.VolumeTypeAnnotation] = v
		case "UID":
			annotations[util.VolumeUIDAnnotation] = v
		case "GID":
			annotations[util.VolumeGIDAnnotation] = v
		}
	}

	return &v1.PersistentVolumeClaim{
		TypeMeta: v12.TypeMeta{
			Kind:       "PersistentVolumeClaim",
			APIVersion: "v1",
		},
		ObjectMeta: v12.ObjectMeta{
			Name:              v.Name(),
			Labels:            v.Labels(),
			Annotations:       annotations,
			CreationTimestamp: v12.Now(),
		},
		Spec: v1.PersistentVolumeClaimSpec{
			Resources: v1.ResourceRequirements{
				Requests: map[v1.ResourceName]resource.Quantity{
					v1.ResourceStorage: resource.MustParse("1Gi"),
				},
			},
			AccessModes: []v1.PersistentVolumeAccessMode{
				v1.ReadWriteOnce,
			},
		},
	}
}

// YAMLPodSpec represents the same k8s API core PodSpec struct with a small
// change and that is having Containers as a pointer to YAMLContainer.
// Because Go doesn't omit empty struct and we want to omit Status in YAML
// if it's empty. Fixes: GH-11998
type YAMLPodSpec struct {
	v1.PodSpec
	Containers []*YAMLContainer `json:"containers"`
}

// YAMLPod represents the same k8s API core Pod struct with a small
// change and that is having Spec as a pointer to YAMLPodSpec and
// Status as a pointer to k8s API core PodStatus.
// Because Go doesn't omit empty struct and we want to omit Status in YAML
// if it's empty. Fixes: GH-11998
type YAMLPod struct {
	v1.Pod
	Spec   *YAMLPodSpec  `json:"spec,omitempty"`
	Status *v1.PodStatus `json:"status,omitempty"`
}

// YAMLService represents the same k8s API core Service struct with a small
// change and that is having Status as a pointer to k8s API core ServiceStatus.
// Because Go doesn't omit empty struct and we want to omit Status in YAML
// if it's empty. Fixes: GH-11998
type YAMLService struct {
	v1.Service
	Status *v1.ServiceStatus `json:"status,omitempty"`
}

// YAMLContainer represents the same k8s API core Container struct with a small
// change and that is having Resources as a pointer to k8s API core ResourceRequirements.
// Because Go doesn't omit empty struct and we want to omit Status in YAML
// if it's empty. Fixes: GH-11998
type YAMLContainer struct {
	v1.Container
	Resources *v1.ResourceRequirements `json:"resources,omitempty"`
}

// ConvertV1PodToYAMLPod takes k8s API core Pod and returns a pointer to YAMLPod
func ConvertV1PodToYAMLPod(pod *v1.Pod) *YAMLPod {
	cs := []*YAMLContainer{}
	for _, cc := range pod.Spec.Containers {
		var res *v1.ResourceRequirements
		if len(cc.Resources.Limits) > 0 || len(cc.Resources.Requests) > 0 {
			res = &cc.Resources
		}
		cs = append(cs, &YAMLContainer{Container: cc, Resources: res})
	}
	mpo := &YAMLPod{Pod: *pod}
	mpo.Spec = &YAMLPodSpec{PodSpec: pod.Spec, Containers: cs}
	for _, ctr := range pod.Spec.Containers {
		if ctr.SecurityContext == nil || ctr.SecurityContext.SELinuxOptions == nil {
			continue
		}
		selinuxOpts := ctr.SecurityContext.SELinuxOptions
		if selinuxOpts.User == "" && selinuxOpts.Role == "" && selinuxOpts.Type == "" && selinuxOpts.Level == "" {
			ctr.SecurityContext.SELinuxOptions = nil
		}
	}
	dnsCfg := pod.Spec.DNSConfig
	if dnsCfg != nil && (len(dnsCfg.Nameservers)+len(dnsCfg.Searches)+len(dnsCfg.Options) > 0) {
		mpo.Spec.DNSConfig = dnsCfg
	}
	status := pod.Status
	if status.Phase != "" || len(status.Conditions) > 0 ||
		status.Message != "" || status.Reason != "" ||
		status.NominatedNodeName != "" || status.HostIP != "" ||
		status.PodIP != "" || status.StartTime != nil ||
		len(status.InitContainerStatuses) > 0 || len(status.ContainerStatuses) > 0 || status.QOSClass != "" || len(status.EphemeralContainerStatuses) > 0 {
		mpo.Status = &status
	}
	return mpo
}

// GenerateKubeServiceFromV1Pod creates a v1 service object from a v1 pod object
func GenerateKubeServiceFromV1Pod(pod *v1.Pod, servicePorts []v1.ServicePort) (YAMLService, error) {
	service := YAMLService{}
	selector := make(map[string]string)
	selector["app"] = pod.Labels["app"]
	ports := servicePorts
	if len(ports) == 0 {
		p, err := containersToServicePorts(pod.Spec.Containers)
		if err != nil {
			return service, err
		}
		ports = p
	}
	serviceSpec := v1.ServiceSpec{
		Ports:    ports,
		Selector: selector,
		Type:     v1.ServiceTypeNodePort,
	}
	service.Spec = serviceSpec
	service.ObjectMeta = pod.ObjectMeta
	tm := v12.TypeMeta{
		Kind:       "Service",
		APIVersion: pod.TypeMeta.APIVersion,
	}
	service.TypeMeta = tm
	return service, nil
}

// servicePortState allows calling containerPortsToServicePorts for a single service
type servicePortState struct {
	// A program using the shared math/rand state with the default seed will produce the same sequence of pseudo-random numbers
	// for each execution. Use a private RNG state not to interfere with other users.
	rng       *rand.Rand
	usedPorts map[int]struct{}
}

func newServicePortState() servicePortState {
	return servicePortState{
		rng:       rand.New(rand.NewSource(time.Now().UnixNano())),
		usedPorts: map[int]struct{}{},
	}
}

func TruncateKubeAnnotation(str string) string {
	str = strings.TrimSpace(str)
	if utf8.RuneCountInString(str) < define.MaxKubeAnnotation {
		return str
	}
	trunc := string([]rune(str)[:define.MaxKubeAnnotation])
	logrus.Warnf("Truncation Annotation: %q to %q: Kubernetes only allows %d characters", str, trunc, define.MaxKubeAnnotation)
	return trunc
}

// containerPortsToServicePorts takes a slice of containerports and generates a
// slice of service ports
func (state *servicePortState) containerPortsToServicePorts(containerPorts []v1.ContainerPort) ([]v1.ServicePort, error) {
	sps := make([]v1.ServicePort, 0, len(containerPorts))
	for _, cp := range containerPorts {
		var nodePort int
		attempt := 0
		for {
			// Legal nodeport range is 30000-32767
			nodePort = 30000 + state.rng.Intn(32767-30000+1)
			if _, found := state.usedPorts[nodePort]; !found {
				state.usedPorts[nodePort] = struct{}{}
				break
			}
			attempt++
			if attempt >= 100 {
				return nil, fmt.Errorf("too many attempts trying to generate a unique NodePort number")
			}
		}
		servicePort := v1.ServicePort{
			Protocol:   cp.Protocol,
			Port:       cp.ContainerPort,
			NodePort:   int32(nodePort),
			Name:       strconv.Itoa(int(cp.ContainerPort)),
			TargetPort: intstr.Parse(strconv.Itoa(int(cp.ContainerPort))),
		}
		sps = append(sps, servicePort)
	}
	return sps, nil
}

// containersToServicePorts takes a slice of v1.Containers and generates an
// inclusive list of serviceports to expose
func containersToServicePorts(containers []v1.Container) ([]v1.ServicePort, error) {
	state := newServicePortState()
	sps := make([]v1.ServicePort, 0, len(containers))
	for _, ctr := range containers {
		ports, err := state.containerPortsToServicePorts(ctr.Ports)
		if err != nil {
			return nil, err
		}
		sps = append(sps, ports...)
	}
	return sps, nil
}

func (p *Pod) podWithContainers(ctx context.Context, containers []*Container, ports []v1.ContainerPort, hostNetwork bool) (*v1.Pod, error) {
	deDupPodVolumes := make(map[string]*v1.Volume)
	first := true
	podContainers := make([]v1.Container, 0, len(containers))
	podInitCtrs := []v1.Container{}
	podAnnotations := make(map[string]string)
	dnsInfo := v1.PodDNSConfig{}

	// Let's sort the containers in order of created time
	// This will ensure that the init containers are defined in the correct order in the kube yaml
	sort.Slice(containers, func(i, j int) bool { return containers[i].CreatedTime().Before(containers[j].CreatedTime()) })

	for _, ctr := range containers {
		if !ctr.IsInfra() {
			for k, v := range ctr.config.Spec.Annotations {
				podAnnotations[fmt.Sprintf("%s/%s", k, removeUnderscores(ctr.Name()))] = TruncateKubeAnnotation(v)
			}
			// Convert auto-update labels into kube annotations
			for k, v := range getAutoUpdateAnnotations(ctr.Name(), ctr.Labels()) {
				podAnnotations[k] = TruncateKubeAnnotation(v)
			}
			isInit := ctr.IsInitCtr()

			ctr, volumes, _, annotations, err := containerToV1Container(ctx, ctr)
			if err != nil {
				return nil, err
			}
			for k, v := range annotations {
				podAnnotations[define.BindMountPrefix+k] = TruncateKubeAnnotation(v)
			}
			// Since port bindings for the pod are handled by the
			// infra container, wipe them here.
			ctr.Ports = nil

			// We add the original port declarations from the libpod infra container
			// to the first kubernetes container description because otherwise we loose
			// the original container/port bindings.
			// Add the port configuration to the first regular container or the first
			// init container if only init containers have been created in the pod.
			if first && len(ports) > 0 && (!isInit || len(containers) == 2) {
				ctr.Ports = ports
				first = false
			}
			if isInit {
				podInitCtrs = append(podInitCtrs, ctr)
				continue
			}
			podContainers = append(podContainers, ctr)
			// Deduplicate volumes, so if containers in the pod share a volume, it's only
			// listed in the volumes section once
			for _, vol := range volumes {
				vol := vol
				deDupPodVolumes[vol.Name] = &vol
			}
		} else {
			_, _, infraDNS, _, err := containerToV1Container(ctx, ctr)
			if err != nil {
				return nil, err
			}
			if infraDNS != nil {
				if servers := infraDNS.Nameservers; len(servers) > 0 {
					dnsInfo.Nameservers = servers
				}
				if searches := infraDNS.Searches; len(searches) > 0 {
					dnsInfo.Searches = searches
				}
				if options := infraDNS.Options; len(options) > 0 {
					dnsInfo.Options = options
				}
			}
		}
	}
	podVolumes := make([]v1.Volume, 0, len(deDupPodVolumes))
	for _, vol := range deDupPodVolumes {
		podVolumes = append(podVolumes, *vol)
	}

	return newPodObject(
		p.Name(),
		podAnnotations,
		podInitCtrs,
		podContainers,
		podVolumes,
		&dnsInfo,
		hostNetwork), nil
}

func newPodObject(podName string, annotations map[string]string, initCtrs, containers []v1.Container, volumes []v1.Volume, dnsOptions *v1.PodDNSConfig, hostNetwork bool) *v1.Pod {
	tm := v12.TypeMeta{
		Kind:       "Pod",
		APIVersion: "v1",
	}

	// Add a label called "app" with the containers name as a value
	labels := make(map[string]string)
	labels["app"] = removeUnderscores(podName)
	om := v12.ObjectMeta{
		// The name of the pod is container_name-libpod
		Name:   podName,
		Labels: labels,
		// CreationTimestamp seems to be required, so adding it; in doing so, the timestamp
		// will reflect time this is run (not container create time) because the conversion
		// of the container create time to v1 Time is probably not warranted nor worthwhile.
		CreationTimestamp: v12.Now(),
		Annotations:       annotations,
	}
	ps := v1.PodSpec{
		Containers:     containers,
		HostNetwork:    hostNetwork,
		InitContainers: initCtrs,
		Volumes:        volumes,
	}
	if dnsOptions != nil && (len(dnsOptions.Nameservers)+len(dnsOptions.Searches)+len(dnsOptions.Options) > 0) {
		ps.DNSConfig = dnsOptions
	}
	p := v1.Pod{
		TypeMeta:   tm,
		ObjectMeta: om,
		Spec:       ps,
	}
	return &p
}

// simplePodWithV1Containers is a function used by inspect when kube yaml needs to be generated
// for a single container.  we "insert" that container description in a pod.
func simplePodWithV1Containers(ctx context.Context, ctrs []*Container) (*v1.Pod, error) {
	kubeCtrs := make([]v1.Container, 0, len(ctrs))
	kubeInitCtrs := []v1.Container{}
	kubeVolumes := make([]v1.Volume, 0)
	hostNetwork := true
	podDNS := v1.PodDNSConfig{}
	kubeAnnotations := make(map[string]string)
	ctrNames := make([]string, 0, len(ctrs))
	for _, ctr := range ctrs {
		ctrNames = append(ctrNames, removeUnderscores(ctr.Name()))
		for k, v := range ctr.config.Spec.Annotations {
			kubeAnnotations[fmt.Sprintf("%s/%s", k, removeUnderscores(ctr.Name()))] = TruncateKubeAnnotation(v)
		}

		// Convert auto-update labels into kube annotations
		for k, v := range getAutoUpdateAnnotations(ctr.Name(), ctr.Labels()) {
			kubeAnnotations[k] = TruncateKubeAnnotation(v)
		}

		isInit := ctr.IsInitCtr()

		if !ctr.HostNetwork() {
			hostNetwork = false
		}
		kubeCtr, kubeVols, ctrDNS, annotations, err := containerToV1Container(ctx, ctr)
		if err != nil {
			return nil, err
		}
		for k, v := range annotations {
			kubeAnnotations[define.BindMountPrefix+k] = TruncateKubeAnnotation(v)
		}
		if isInit {
			kubeInitCtrs = append(kubeInitCtrs, kubeCtr)
		} else {
			kubeCtrs = append(kubeCtrs, kubeCtr)
		}
		kubeVolumes = append(kubeVolumes, kubeVols...)
		// Combine DNS information in sum'd structure
		if ctrDNS != nil {
			// nameservers
			if servers := ctrDNS.Nameservers; servers != nil {
				if podDNS.Nameservers == nil {
					podDNS.Nameservers = make([]string, 0)
				}
				for _, s := range servers {
					if !cutil.StringInSlice(s, podDNS.Nameservers) { // only append if it does not exist
						podDNS.Nameservers = append(podDNS.Nameservers, s)
					}
				}
			}
			// search domains
			if domains := ctrDNS.Searches; domains != nil {
				if podDNS.Searches == nil {
					podDNS.Searches = make([]string, 0)
				}
				for _, d := range domains {
					if !cutil.StringInSlice(d, podDNS.Searches) { // only append if it does not exist
						podDNS.Searches = append(podDNS.Searches, d)
					}
				}
			}
			// dns options
			if options := ctrDNS.Options; options != nil {
				if podDNS.Options == nil {
					podDNS.Options = make([]v1.PodDNSConfigOption, 0)
				}
				podDNS.Options = append(podDNS.Options, options...)
			}
		} // end if ctrDNS
	}
	podName := removeUnderscores(ctrs[0].Name())
	// Check if the pod name and container name will end up conflicting
	// Append -pod if so
	if cutil.StringInSlice(podName, ctrNames) {
		podName += "-pod"
	}

	return newPodObject(
		podName,
		kubeAnnotations,
		kubeInitCtrs,
		kubeCtrs,
		kubeVolumes,
		&podDNS,
		hostNetwork), nil
}

// containerToV1Container converts information we know about a libpod container
// to a V1.Container specification.
func containerToV1Container(ctx context.Context, c *Container) (v1.Container, []v1.Volume, *v1.PodDNSConfig, map[string]string, error) {
	kubeContainer := v1.Container{}
	kubeVolumes := []v1.Volume{}
	annotations := make(map[string]string)
	kubeSec, err := generateKubeSecurityContext(c)
	if err != nil {
		return kubeContainer, kubeVolumes, nil, annotations, err
	}

	// NOTE: a privileged container mounts all of /dev/*.
	if !c.Privileged() && len(c.config.Spec.Linux.Devices) > 0 {
		// TODO Enable when we can support devices and their names
		kubeContainer.VolumeDevices = generateKubeVolumeDeviceFromLinuxDevice(c.config.Spec.Linux.Devices)
		return kubeContainer, kubeVolumes, nil, annotations, fmt.Errorf("linux devices: %w", define.ErrNotImplemented)
	}

	if len(c.config.UserVolumes) > 0 {
		volumeMounts, volumes, localAnnotations, err := libpodMountsToKubeVolumeMounts(c)
		if err != nil {
			return kubeContainer, kubeVolumes, nil, nil, err
		}
		annotations = localAnnotations
		kubeContainer.VolumeMounts = volumeMounts
		kubeVolumes = append(kubeVolumes, volumes...)
	}

	portmappings, err := c.PortMappings()
	if err != nil {
		return kubeContainer, kubeVolumes, nil, annotations, err
	}
	ports, err := portMappingToContainerPort(portmappings)
	if err != nil {
		return kubeContainer, kubeVolumes, nil, annotations, err
	}

	// Handle command and arguments.
	if ep := c.Entrypoint(); len(ep) > 0 {
		// If we have an entrypoint, set the container's command as
		// arguments.
		kubeContainer.Command = ep
		kubeContainer.Args = c.Command()
	} else {
		kubeContainer.Command = c.Command()
	}

	kubeContainer.Name = removeUnderscores(c.Name())
	_, image := c.Image()

	// The infra container may have been created with an overlay root FS
	// instead of an infra image.  If so, set the imageto the default K8s
	// pause one and make sure it's in the storage by pulling it down if
	// missing.
	if image == "" && c.IsInfra() {
		image = c.runtime.config.Engine.InfraImage
		if _, err := c.runtime.libimageRuntime.Pull(ctx, image, config.PullPolicyMissing, nil); err != nil {
			return kubeContainer, nil, nil, nil, err
		}
	}

	kubeContainer.Image = image
	kubeContainer.Stdin = c.Stdin()
	img, _, err := c.runtime.libimageRuntime.LookupImage(image, nil)
	if err != nil {
		return kubeContainer, kubeVolumes, nil, annotations, fmt.Errorf("looking up image %q of container %q: %w", image, c.ID(), err)
	}
	imgData, err := img.Inspect(ctx, nil)
	if err != nil {
		return kubeContainer, kubeVolumes, nil, annotations, err
	}
	// If the user doesn't set a command/entrypoint when creating the container with podman and
	// is using the image command or entrypoint from the image, don't add it to the generated kube yaml
	if reflect.DeepEqual(imgData.Config.Cmd, kubeContainer.Command) || reflect.DeepEqual(imgData.Config.Entrypoint, kubeContainer.Command) {
		kubeContainer.Command = nil
	}

	if c.WorkingDir() != "/" && imgData.Config.WorkingDir != c.WorkingDir() {
		kubeContainer.WorkingDir = c.WorkingDir()
	}

	if imgData.User == c.User() {
		kubeSec.RunAsGroup, kubeSec.RunAsUser = nil, nil
	}

	envVariables, err := libpodEnvVarsToKubeEnvVars(c.config.Spec.Process.Env, imgData.Config.Env)
	if err != nil {
		return kubeContainer, kubeVolumes, nil, annotations, err
	}
	kubeContainer.Env = envVariables

	kubeContainer.Ports = ports
	// This should not be applicable
	// container.EnvFromSource =
	kubeContainer.SecurityContext = kubeSec
	kubeContainer.StdinOnce = false
	kubeContainer.TTY = c.config.Spec.Process.Terminal

	if c.config.Spec.Linux != nil &&
		c.config.Spec.Linux.Resources != nil {
		if c.config.Spec.Linux.Resources.Memory != nil &&
			c.config.Spec.Linux.Resources.Memory.Limit != nil {
			if kubeContainer.Resources.Limits == nil {
				kubeContainer.Resources.Limits = v1.ResourceList{}
			}

			qty := kubeContainer.Resources.Limits.Memory()
			qty.Set(*c.config.Spec.Linux.Resources.Memory.Limit)
			kubeContainer.Resources.Limits[v1.ResourceMemory] = *qty
		}

		if c.config.Spec.Linux.Resources.CPU != nil &&
			c.config.Spec.Linux.Resources.CPU.Quota != nil &&
			c.config.Spec.Linux.Resources.CPU.Period != nil {
			quota := *c.config.Spec.Linux.Resources.CPU.Quota
			period := *c.config.Spec.Linux.Resources.CPU.Period

			if quota > 0 && period > 0 {
				cpuLimitMilli := int64(1000 * util.PeriodAndQuotaToCores(period, quota))

				// Kubernetes: precision finer than 1m is not allowed
				if cpuLimitMilli >= 1 {
					if kubeContainer.Resources.Limits == nil {
						kubeContainer.Resources.Limits = v1.ResourceList{}
					}

					qty := kubeContainer.Resources.Limits.Cpu()
					qty.SetMilli(cpuLimitMilli)
					kubeContainer.Resources.Limits[v1.ResourceCPU] = *qty
				}
			}
		}
	}

	// Obtain the DNS entries from the container
	dns := v1.PodDNSConfig{}

	// DNS servers
	if servers := c.config.DNSServer; len(servers) > 0 {
		dnsServers := make([]string, 0)
		for _, server := range servers {
			dnsServers = append(dnsServers, server.String())
		}
		dns.Nameservers = dnsServers
	}

	// DNS search domains
	if searches := c.config.DNSSearch; len(searches) > 0 {
		dns.Searches = searches
	}

	// DNS options
	if options := c.config.DNSOption; len(options) > 0 {
		dnsOptions := make([]v1.PodDNSConfigOption, 0)
		for _, option := range options {
			// the option can be "k:v" or just "k", no delimiter is required
			opts := strings.SplitN(option, ":", 2)
			dnsOpt := v1.PodDNSConfigOption{
				Name:  opts[0],
				Value: &opts[1],
			}
			dnsOptions = append(dnsOptions, dnsOpt)
		}
		dns.Options = dnsOptions
	}
	return kubeContainer, kubeVolumes, &dns, annotations, nil
}

// portMappingToContainerPort takes an portmapping and converts
// it to a v1.ContainerPort format for kube output
func portMappingToContainerPort(portMappings []types.PortMapping) ([]v1.ContainerPort, error) {
	containerPorts := make([]v1.ContainerPort, 0, len(portMappings))
	for _, p := range portMappings {
		protocols := strings.Split(p.Protocol, ",")
		for _, proto := range protocols {
			var protocol v1.Protocol
			switch strings.ToUpper(proto) {
			case "TCP":
				// do nothing as it is the default protocol in k8s, there is no need to explicitly
				// add it to the generated yaml
			case "UDP":
				protocol = v1.ProtocolUDP
			case "SCTP":
				protocol = v1.ProtocolSCTP
			default:
				return containerPorts, fmt.Errorf("unknown network protocol %s", p.Protocol)
			}
			for i := uint16(0); i < p.Range; i++ {
				cp := v1.ContainerPort{
					// Name will not be supported
					HostPort:      int32(p.HostPort + i),
					HostIP:        p.HostIP,
					ContainerPort: int32(p.ContainerPort + i),
					Protocol:      protocol,
				}
				containerPorts = append(containerPorts, cp)
			}
		}
	}
	return containerPorts, nil
}

// libpodEnvVarsToKubeEnvVars converts a key=value string slice to []v1.EnvVar
func libpodEnvVarsToKubeEnvVars(envs []string, imageEnvs []string) ([]v1.EnvVar, error) {
	defaultEnv := env.DefaultEnvVariables()
	envVars := make([]v1.EnvVar, 0, len(envs))
	imageMap := make(map[string]string, len(imageEnvs))
	for _, ie := range imageEnvs {
		split := strings.SplitN(ie, "=", 2)
		imageMap[split[0]] = split[1]
	}
	for _, e := range envs {
		split := strings.SplitN(e, "=", 2)
		if len(split) != 2 {
			return envVars, fmt.Errorf("environment variable %s is malformed; should be key=value", e)
		}
		if defaultEnv[split[0]] == split[1] {
			continue
		}
		if imageMap[split[0]] == split[1] {
			continue
		}
		ev := v1.EnvVar{
			Name:  split[0],
			Value: split[1],
		}
		envVars = append(envVars, ev)
	}
	return envVars, nil
}

// libpodMountsToKubeVolumeMounts converts the containers mounts to a struct kube understands
func libpodMountsToKubeVolumeMounts(c *Container) ([]v1.VolumeMount, []v1.Volume, map[string]string, error) {
	namedVolumes, mounts := c.SortUserVolumes(c.config.Spec)
	vms := make([]v1.VolumeMount, 0, len(mounts))
	vos := make([]v1.Volume, 0, len(mounts))
	annotations := make(map[string]string)

	var suffix string
	for index, m := range mounts {
		for _, opt := range m.Options {
			if opt == "Z" || opt == "z" {
				annotations[m.Source] = opt
				break
			}
		}
		vm, vo, err := generateKubeVolumeMount(m)
		if err != nil {
			return vms, vos, annotations, err
		}
		// Name will be the same, so use the index as suffix
		suffix = fmt.Sprintf("-%d", index)
		vm.Name += suffix
		vo.Name += suffix
		vms = append(vms, vm)
		vos = append(vos, vo)
	}
	for _, v := range namedVolumes {
		vm, vo := generateKubePersistentVolumeClaim(v)
		vms = append(vms, vm)
		vos = append(vos, vo)
	}
	return vms, vos, annotations, nil
}

// generateKubePersistentVolumeClaim converts a ContainerNamedVolume to a Kubernetes PersistentVolumeClaim
func generateKubePersistentVolumeClaim(v *ContainerNamedVolume) (v1.VolumeMount, v1.Volume) {
	ro := cutil.StringInSlice("ro", v.Options)

	// To avoid naming conflicts with any host path mounts, add a unique suffix to the volume's name.
	name := v.Name + "-pvc"

	vm := v1.VolumeMount{}
	vm.Name = name
	vm.MountPath = v.Dest
	vm.ReadOnly = ro

	pvc := v1.PersistentVolumeClaimVolumeSource{ClaimName: v.Name, ReadOnly: ro}
	vs := v1.VolumeSource{}
	vs.PersistentVolumeClaim = &pvc
	vo := v1.Volume{Name: name, VolumeSource: vs}

	return vm, vo
}

// generateKubeVolumeMount takes a user specified mount and returns
// a kubernetes VolumeMount (to be added to the container) and a kubernetes Volume
// (to be added to the pod)
func generateKubeVolumeMount(m specs.Mount) (v1.VolumeMount, v1.Volume, error) {
	vm := v1.VolumeMount{}
	vo := v1.Volume{}

	name, err := convertVolumePathToName(m.Source)
	if err != nil {
		return vm, vo, err
	}
	// To avoid naming conflicts with any persistent volume mounts, add a unique suffix to the volume's name.
	name += "-host"
	vm.Name = name
	vm.MountPath = m.Destination
	if cutil.StringInSlice("ro", m.Options) {
		vm.ReadOnly = true
	}

	vo.Name = name
	vo.HostPath = &v1.HostPathVolumeSource{}
	vo.HostPath.Path = m.Source
	isDir, err := isHostPathDirectory(m.Source)
	// neither a directory or a file lives here, default to creating a directory
	// TODO should this be an error instead?
	var hostPathType v1.HostPathType
	switch {
	case err != nil:
		hostPathType = v1.HostPathDirectoryOrCreate
	case isDir:
		hostPathType = v1.HostPathDirectory
	default:
		hostPathType = v1.HostPathFile
	}
	vo.HostPath.Type = &hostPathType

	return vm, vo, nil
}

func isHostPathDirectory(hostPathSource string) (bool, error) {
	info, err := os.Stat(hostPathSource)
	if err != nil {
		return false, err
	}
	return info.Mode().IsDir(), nil
}

func convertVolumePathToName(hostSourcePath string) (string, error) {
	if len(hostSourcePath) == 0 {
		return "", errors.New("hostSourcePath must be specified to generate volume name")
	}
	if len(hostSourcePath) == 1 {
		if hostSourcePath != "/" {
			return "", fmt.Errorf("hostSourcePath malformatted: %s", hostSourcePath)
		}
		// add special case name
		return "root", nil
	}
	// First, trim trailing slashes, then replace slashes with dashes.
	// Thus, /mnt/data/ will become mnt-data
	return strings.ReplaceAll(strings.Trim(hostSourcePath, "/"), "/", "-"), nil
}

func determineCapAddDropFromCapabilities(defaultCaps, containerCaps []string) *v1.Capabilities {
	var (
		drop = []v1.Capability{}
		add  = []v1.Capability{}
	)
	dedupDrop := make(map[string]bool)
	dedupAdd := make(map[string]bool)
	// Find caps in the defaultCaps but not in the container's
	// those indicate a dropped cap
	for _, capability := range defaultCaps {
		if !cutil.StringInSlice(capability, containerCaps) {
			if _, ok := dedupDrop[capability]; !ok {
				drop = append(drop, v1.Capability(capability))
				dedupDrop[capability] = true
			}
		}
	}
	// Find caps in the container but not in the defaults; those indicate
	// an added cap
	for _, capability := range containerCaps {
		if !cutil.StringInSlice(capability, defaultCaps) {
			if _, ok := dedupAdd[capability]; !ok {
				add = append(add, v1.Capability(capability))
				dedupAdd[capability] = true
			}
		}
	}

	return &v1.Capabilities{
		Add:  add,
		Drop: drop,
	}
}

func capAddDrop(caps *specs.LinuxCapabilities) (*v1.Capabilities, error) {
	g, err := generate.New("linux")
	if err != nil {
		return nil, err
	}

	defCaps := g.Config.Process.Capabilities
	// Combine all the default capabilities into a slice
	defaultCaps := make([]string, 0, len(defCaps.Ambient)+len(defCaps.Bounding)+len(defCaps.Effective)+len(defCaps.Inheritable)+len(defCaps.Permitted))
	defaultCaps = append(defaultCaps, defCaps.Ambient...)
	defaultCaps = append(defaultCaps, defCaps.Bounding...)
	defaultCaps = append(defaultCaps, defCaps.Effective...)
	defaultCaps = append(defaultCaps, defCaps.Inheritable...)
	defaultCaps = append(defaultCaps, defCaps.Permitted...)

	// Combine all the container's capabilities into a slice
	containerCaps := make([]string, 0, len(caps.Ambient)+len(caps.Bounding)+len(caps.Effective)+len(caps.Inheritable)+len(caps.Permitted))
	containerCaps = append(containerCaps, caps.Ambient...)
	containerCaps = append(containerCaps, caps.Bounding...)
	containerCaps = append(containerCaps, caps.Effective...)
	containerCaps = append(containerCaps, caps.Inheritable...)
	containerCaps = append(containerCaps, caps.Permitted...)

	calculatedCaps := determineCapAddDropFromCapabilities(defaultCaps, containerCaps)
	return calculatedCaps, nil
}

// generateKubeSecurityContext generates a securityContext based on the existing container
func generateKubeSecurityContext(c *Container) (*v1.SecurityContext, error) {
	privileged := c.Privileged()
	ro := c.IsReadOnly()
	allowPrivEscalation := !c.config.Spec.Process.NoNewPrivileges

	var capabilities *v1.Capabilities
	if !privileged {
		// Running privileged adds all caps.
		newCaps, err := capAddDrop(c.config.Spec.Process.Capabilities)
		if err != nil {
			return nil, err
		}
		capabilities = newCaps
	}

	sc := v1.SecurityContext{
		// RunAsNonRoot is an optional parameter; our first implementations should be root only; however
		// I'm leaving this as a bread-crumb for later
		//RunAsNonRoot:             &nonRoot,
	}
	if capabilities != nil {
		sc.Capabilities = capabilities
	}
	var selinuxOpts v1.SELinuxOptions
	opts := strings.SplitN(c.config.Spec.Annotations[define.InspectAnnotationLabel], ":", 2)
	switch len(opts) {
	case 2:
		switch opts[0] {
		case "type":
			selinuxOpts.Type = opts[1]
			sc.SELinuxOptions = &selinuxOpts
		case "level":
			selinuxOpts.Level = opts[1]
			sc.SELinuxOptions = &selinuxOpts
		}
	case 1:
		if opts[0] == "disable" {
			selinuxOpts.Type = "spc_t"
			sc.SELinuxOptions = &selinuxOpts
		}
	}

	if !allowPrivEscalation {
		sc.AllowPrivilegeEscalation = &allowPrivEscalation
	}
	if privileged {
		sc.Privileged = &privileged
	}
	if ro {
		sc.ReadOnlyRootFilesystem = &ro
	}
	if c.User() != "" {
		if !c.batched {
			c.lock.Lock()
			defer c.lock.Unlock()
		}
		if err := c.syncContainer(); err != nil {
			return nil, fmt.Errorf("unable to sync container during YAML generation: %w", err)
		}

		mountpoint := c.state.Mountpoint
		if mountpoint == "" {
			var err error
			mountpoint, err = c.mount()
			if err != nil {
				return nil, fmt.Errorf("failed to mount %s mountpoint: %w", c.ID(), err)
			}
			defer func() {
				if err := c.unmount(false); err != nil {
					logrus.Errorf("Failed to unmount container: %v", err)
				}
			}()
		}
		logrus.Debugf("Looking in container for user: %s", c.User())

		execUser, err := lookup.GetUserGroupInfo(mountpoint, c.User(), nil)
		if err != nil {
			return nil, err
		}
		uid := int64(execUser.Uid)
		gid := int64(execUser.Gid)
		sc.RunAsUser = &uid
		sc.RunAsGroup = &gid
	}
	return &sc, nil
}

// generateKubeVolumeDeviceFromLinuxDevice takes a list of devices and makes a VolumeDevice struct for kube
func generateKubeVolumeDeviceFromLinuxDevice(devices []specs.LinuxDevice) []v1.VolumeDevice {
	volumeDevices := make([]v1.VolumeDevice, 0, len(devices))
	for _, d := range devices {
		vd := v1.VolumeDevice{
			// TBD How are we going to sync up these names
			//Name:
			DevicePath: d.Path,
		}
		volumeDevices = append(volumeDevices, vd)
	}
	return volumeDevices
}

func removeUnderscores(s string) string {
	return strings.ReplaceAll(s, "_", "")
}

// getAutoUpdateAnnotations searches for auto-update container labels
// and returns them as kube annotations
func getAutoUpdateAnnotations(ctrName string, ctrLabels map[string]string) map[string]string {
	autoUpdateLabel := "io.containers.autoupdate"
	annotations := make(map[string]string)

	ctrName = removeUnderscores(ctrName)
	for k, v := range ctrLabels {
		if strings.Contains(k, autoUpdateLabel) {
			// since labels can variate between containers within a pod, they will be
			// identified with the container name when converted into kube annotations
			kc := fmt.Sprintf("%s/%s", k, ctrName)
			annotations[kc] = TruncateKubeAnnotation(v)
		}
	}

	return annotations
}