-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathcontainer.go
More file actions
1407 lines (1237 loc) · 37.5 KB
/
container.go
File metadata and controls
1407 lines (1237 loc) · 37.5 KB
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
package docker
import (
"bytes"
"encoding/json"
"errors"
"fmt"
"github.com/dotcloud/docker/archive"
"github.com/dotcloud/docker/graphdriver"
"github.com/dotcloud/docker/term"
"github.com/dotcloud/docker/utils"
"github.com/kr/pty"
"io"
"io/ioutil"
"log"
"net"
"os"
"os/exec"
"path"
"strconv"
"strings"
"sync"
"syscall"
"time"
)
type Container struct {
sync.Mutex
root string // Path to the "home" of the container, including metadata.
rootfs string // Path to the root filesystem of the container.
ID string
Created time.Time
Path string
Args []string
Config *Config
State State
Image string
network *NetworkInterface
NetworkSettings *NetworkSettings
SysInitPath string
ResolvConfPath string
HostnamePath string
HostsPath string
Name string
Driver string
cmd *exec.Cmd
stdout *utils.WriteBroadcaster
stderr *utils.WriteBroadcaster
stdin io.ReadCloser
stdinPipe io.WriteCloser
ptyMaster io.Closer
runtime *Runtime
waitLock chan struct{}
Volumes map[string]string
// Store rw/ro in a separate structure to preserve reverse-compatibility on-disk.
// Easier than migrating older container configs :)
VolumesRW map[string]bool
hostConfig *HostConfig
activeLinks map[string]*Link
}
// Note: the Config structure should hold only portable information about the container.
// Here, "portable" means "independent from the host we are running on".
// Non-portable information *should* appear in HostConfig.
type Config struct {
Hostname string
Domainname string
User string
Memory int64 // Memory limit (in bytes)
MemorySwap int64 // Total memory usage (memory + swap); set `-1' to disable swap
CpuShares int64 // CPU shares (relative weight vs. other containers)
AttachStdin bool
AttachStdout bool
AttachStderr bool
PortSpecs []string // Deprecated - Can be in the format of 8080/tcp
ExposedPorts map[Port]struct{}
Tty bool // Attach standard streams to a tty, including stdin if it is not closed.
OpenStdin bool // Open stdin
StdinOnce bool // If true, close stdin after the 1 attached client disconnects.
Env []string
Cmd []string
Dns []string
Image string // Name of the image as it was passed by the operator (eg. could be symbolic)
Volumes map[string]struct{}
VolumesFrom string
WorkingDir string
Entrypoint []string
NetworkDisabled bool
}
type HostConfig struct {
Binds []string
ContainerIDFile string
LxcConf []KeyValuePair
Privileged bool
PortBindings map[Port][]PortBinding
Links []string
PublishAllPorts bool
}
type BindMap struct {
SrcPath string
DstPath string
Mode string
}
var (
ErrContainerStart = errors.New("The container failed to start. Unkown error")
ErrContainerStartTimeout = errors.New("The container failed to start due to timed out.")
ErrInvalidWorikingDirectory = errors.New("The working directory is invalid. It needs to be an absolute path.")
ErrConflictAttachDetach = errors.New("Conflicting options: -a and -d")
ErrConflictDetachAutoRemove = errors.New("Conflicting options: -rm and -d")
)
type KeyValuePair struct {
Key string
Value string
}
type PortBinding struct {
HostIp string
HostPort string
}
// 80/tcp
type Port string
func (p Port) Proto() string {
parts := strings.Split(string(p), "/")
if len(parts) == 1 {
return "tcp"
}
return parts[1]
}
func (p Port) Port() string {
return strings.Split(string(p), "/")[0]
}
func (p Port) Int() int {
i, err := parsePort(p.Port())
if err != nil {
panic(err)
}
return i
}
func NewPort(proto, port string) Port {
return Port(fmt.Sprintf("%s/%s", port, proto))
}
type PortMapping map[string]string // Deprecated
type NetworkSettings struct {
IPAddress string
IPPrefixLen int
Gateway string
Bridge string
PortMapping map[string]PortMapping // Deprecated
Ports map[Port][]PortBinding
}
func (settings *NetworkSettings) PortMappingAPI() []APIPort {
var mapping []APIPort
for port, bindings := range settings.Ports {
p, _ := parsePort(port.Port())
if len(bindings) == 0 {
mapping = append(mapping, APIPort{
PublicPort: int64(p),
Type: port.Proto(),
})
continue
}
for _, binding := range bindings {
p, _ := parsePort(port.Port())
h, _ := parsePort(binding.HostPort)
mapping = append(mapping, APIPort{
PrivatePort: int64(p),
PublicPort: int64(h),
Type: port.Proto(),
IP: binding.HostIp,
})
}
}
return mapping
}
// Inject the io.Reader at the given path. Note: do not close the reader
func (container *Container) Inject(file io.Reader, pth string) error {
if err := container.EnsureMounted(); err != nil {
return fmt.Errorf("inject: error mounting container %s: %s", container.ID, err)
}
// Return error if path exists
destPath := path.Join(container.RootfsPath(), pth)
if _, err := os.Stat(destPath); err == nil {
// Since err is nil, the path could be stat'd and it exists
return fmt.Errorf("%s exists", pth)
} else if !os.IsNotExist(err) {
// Expect err might be that the file doesn't exist, so
// if it's some other error, return that.
return err
}
// Make sure the directory exists
if err := os.MkdirAll(path.Join(container.RootfsPath(), path.Dir(pth)), 0755); err != nil {
return err
}
dest, err := os.Create(destPath)
if err != nil {
return err
}
defer dest.Close()
if _, err := io.Copy(dest, file); err != nil {
return err
}
return nil
}
func (container *Container) Cmd() *exec.Cmd {
return container.cmd
}
func (container *Container) When() time.Time {
return container.Created
}
func (container *Container) FromDisk() error {
data, err := ioutil.ReadFile(container.jsonPath())
if err != nil {
return err
}
// Load container settings
// udp broke compat of docker.PortMapping, but it's not used when loading a container, we can skip it
if err := json.Unmarshal(data, container); err != nil && !strings.Contains(err.Error(), "docker.PortMapping") {
return err
}
return container.readHostConfig()
}
func (container *Container) ToDisk() (err error) {
data, err := json.Marshal(container)
if err != nil {
return
}
err = ioutil.WriteFile(container.jsonPath(), data, 0666)
if err != nil {
return
}
return container.writeHostConfig()
}
func (container *Container) readHostConfig() error {
container.hostConfig = &HostConfig{}
// If the hostconfig file does not exist, do not read it.
// (We still have to initialize container.hostConfig,
// but that's OK, since we just did that above.)
_, err := os.Stat(container.hostConfigPath())
if os.IsNotExist(err) {
return nil
}
data, err := ioutil.ReadFile(container.hostConfigPath())
if err != nil {
return err
}
return json.Unmarshal(data, container.hostConfig)
}
func (container *Container) writeHostConfig() (err error) {
data, err := json.Marshal(container.hostConfig)
if err != nil {
return
}
return ioutil.WriteFile(container.hostConfigPath(), data, 0666)
}
func (container *Container) generateEnvConfig(env []string) error {
data, err := json.Marshal(env)
if err != nil {
return err
}
ioutil.WriteFile(container.EnvConfigPath(), data, 0600)
return nil
}
func (container *Container) generateLXCConfig() error {
fo, err := os.Create(container.lxcConfigPath())
if err != nil {
return err
}
defer fo.Close()
return LxcTemplateCompiled.Execute(fo, container)
}
func (container *Container) startPty() error {
ptyMaster, ptySlave, err := pty.Open()
if err != nil {
return err
}
container.ptyMaster = ptyMaster
container.cmd.Stdout = ptySlave
container.cmd.Stderr = ptySlave
// Copy the PTYs to our broadcasters
go func() {
defer container.stdout.CloseWriters()
utils.Debugf("startPty: begin of stdout pipe")
io.Copy(container.stdout, ptyMaster)
utils.Debugf("startPty: end of stdout pipe")
}()
// stdin
if container.Config.OpenStdin {
container.cmd.Stdin = ptySlave
container.cmd.SysProcAttr.Setctty = true
go func() {
defer container.stdin.Close()
utils.Debugf("startPty: begin of stdin pipe")
io.Copy(ptyMaster, container.stdin)
utils.Debugf("startPty: end of stdin pipe")
}()
}
if err := container.cmd.Start(); err != nil {
return err
}
ptySlave.Close()
return nil
}
func (container *Container) start() error {
container.cmd.Stdout = container.stdout
container.cmd.Stderr = container.stderr
if container.Config.OpenStdin {
stdin, err := container.cmd.StdinPipe()
if err != nil {
return err
}
go func() {
defer stdin.Close()
utils.Debugf("start: begin of stdin pipe")
io.Copy(stdin, container.stdin)
utils.Debugf("start: end of stdin pipe")
}()
}
return container.cmd.Start()
}
func (container *Container) Attach(stdin io.ReadCloser, stdinCloser io.Closer, stdout io.Writer, stderr io.Writer) chan error {
var cStdout, cStderr io.ReadCloser
var nJobs int
errors := make(chan error, 3)
if stdin != nil && container.Config.OpenStdin {
nJobs += 1
if cStdin, err := container.StdinPipe(); err != nil {
errors <- err
} else {
go func() {
utils.Debugf("attach: stdin: begin")
defer utils.Debugf("attach: stdin: end")
// No matter what, when stdin is closed (io.Copy unblock), close stdout and stderr
if container.Config.StdinOnce && !container.Config.Tty {
defer cStdin.Close()
} else {
if cStdout != nil {
defer cStdout.Close()
}
if cStderr != nil {
defer cStderr.Close()
}
}
if container.Config.Tty {
_, err = utils.CopyEscapable(cStdin, stdin)
} else {
_, err = io.Copy(cStdin, stdin)
}
if err == io.ErrClosedPipe {
err = nil
}
if err != nil {
utils.Errorf("attach: stdin: %s", err)
}
errors <- err
}()
}
}
if stdout != nil {
nJobs += 1
if p, err := container.StdoutPipe(); err != nil {
errors <- err
} else {
cStdout = p
go func() {
utils.Debugf("attach: stdout: begin")
defer utils.Debugf("attach: stdout: end")
// If we are in StdinOnce mode, then close stdin
if container.Config.StdinOnce && stdin != nil {
defer stdin.Close()
}
if stdinCloser != nil {
defer stdinCloser.Close()
}
_, err := io.Copy(stdout, cStdout)
if err == io.ErrClosedPipe {
err = nil
}
if err != nil {
utils.Errorf("attach: stdout: %s", err)
}
errors <- err
}()
}
} else {
go func() {
if stdinCloser != nil {
defer stdinCloser.Close()
}
if cStdout, err := container.StdoutPipe(); err != nil {
utils.Errorf("attach: stdout pipe: %s", err)
} else {
io.Copy(&utils.NopWriter{}, cStdout)
}
}()
}
if stderr != nil {
nJobs += 1
if p, err := container.StderrPipe(); err != nil {
errors <- err
} else {
cStderr = p
go func() {
utils.Debugf("attach: stderr: begin")
defer utils.Debugf("attach: stderr: end")
// If we are in StdinOnce mode, then close stdin
if container.Config.StdinOnce && stdin != nil {
defer stdin.Close()
}
if stdinCloser != nil {
defer stdinCloser.Close()
}
_, err := io.Copy(stderr, cStderr)
if err == io.ErrClosedPipe {
err = nil
}
if err != nil {
utils.Errorf("attach: stderr: %s", err)
}
errors <- err
}()
}
} else {
go func() {
if stdinCloser != nil {
defer stdinCloser.Close()
}
if cStderr, err := container.StderrPipe(); err != nil {
utils.Errorf("attach: stdout pipe: %s", err)
} else {
io.Copy(&utils.NopWriter{}, cStderr)
}
}()
}
return utils.Go(func() error {
if cStdout != nil {
defer cStdout.Close()
}
if cStderr != nil {
defer cStderr.Close()
}
// FIXME: how to clean up the stdin goroutine without the unwanted side effect
// of closing the passed stdin? Add an intermediary io.Pipe?
for i := 0; i < nJobs; i += 1 {
utils.Debugf("attach: waiting for job %d/%d", i+1, nJobs)
if err := <-errors; err != nil {
utils.Errorf("attach: job %d returned error %s, aborting all jobs", i+1, err)
return err
}
utils.Debugf("attach: job %d completed successfully", i+1)
}
utils.Debugf("attach: all jobs completed successfully")
return nil
})
}
func (container *Container) Start() (err error) {
container.Lock()
defer container.Unlock()
if container.State.IsRunning() {
return fmt.Errorf("The container %s is already running.", container.ID)
}
defer func() {
if err != nil {
container.cleanup()
}
}()
if err := container.EnsureMounted(); err != nil {
return err
}
if container.runtime.networkManager.disabled {
container.Config.NetworkDisabled = true
container.buildHostnameAndHostsFiles("127.0.1.1")
} else {
if err := container.allocateNetwork(); err != nil {
return err
}
container.buildHostnameAndHostsFiles(container.NetworkSettings.IPAddress)
}
// Make sure the config is compatible with the current kernel
if container.Config.Memory > 0 && !container.runtime.capabilities.MemoryLimit {
log.Printf("WARNING: Your kernel does not support memory limit capabilities. Limitation discarded.\n")
container.Config.Memory = 0
}
if container.Config.Memory > 0 && !container.runtime.capabilities.SwapLimit {
log.Printf("WARNING: Your kernel does not support swap limit capabilities. Limitation discarded.\n")
container.Config.MemorySwap = -1
}
if container.runtime.capabilities.IPv4ForwardingDisabled {
log.Printf("WARNING: IPv4 forwarding is disabled. Networking will not work")
}
// Create the requested bind mounts
binds := make(map[string]BindMap)
// Define illegal container destinations
illegalDsts := []string{"/", "."}
for _, bind := range container.hostConfig.Binds {
// FIXME: factorize bind parsing in parseBind
var src, dst, mode string
arr := strings.Split(bind, ":")
if len(arr) == 2 {
src = arr[0]
dst = arr[1]
mode = "rw"
} else if len(arr) == 3 {
src = arr[0]
dst = arr[1]
mode = arr[2]
} else {
return fmt.Errorf("Invalid bind specification: %s", bind)
}
// Bail if trying to mount to an illegal destination
for _, illegal := range illegalDsts {
if dst == illegal {
return fmt.Errorf("Illegal bind destination: %s", dst)
}
}
bindMap := BindMap{
SrcPath: src,
DstPath: dst,
Mode: mode,
}
binds[path.Clean(dst)] = bindMap
}
if container.Volumes == nil || len(container.Volumes) == 0 {
container.Volumes = make(map[string]string)
container.VolumesRW = make(map[string]bool)
}
// Apply volumes from another container if requested
if container.Config.VolumesFrom != "" {
containerSpecs := strings.Split(container.Config.VolumesFrom, ",")
for _, containerSpec := range containerSpecs {
mountRW := true
specParts := strings.SplitN(containerSpec, ":", 2)
switch len(specParts) {
case 0:
return fmt.Errorf("Malformed volumes-from specification: %s", container.Config.VolumesFrom)
case 2:
switch specParts[1] {
case "ro":
mountRW = false
case "rw": // mountRW is already true
default:
return fmt.Errorf("Malformed volumes-from speficication: %s", containerSpec)
}
}
c := container.runtime.Get(specParts[0])
if c == nil {
return fmt.Errorf("Container %s not found. Impossible to mount its volumes", container.ID)
}
for volPath, id := range c.Volumes {
if _, exists := container.Volumes[volPath]; exists {
continue
}
if err := os.MkdirAll(path.Join(container.RootfsPath(), volPath), 0755); err != nil {
return err
}
container.Volumes[volPath] = id
if isRW, exists := c.VolumesRW[volPath]; exists {
container.VolumesRW[volPath] = isRW && mountRW
}
}
}
}
volumesDriver := container.runtime.volumes.driver
// Create the requested volumes if they don't exist
for volPath := range container.Config.Volumes {
volPath = path.Clean(volPath)
// Skip existing volumes
if _, exists := container.Volumes[volPath]; exists {
continue
}
var srcPath string
var isBindMount bool
srcRW := false
// If an external bind is defined for this volume, use that as a source
if bindMap, exists := binds[volPath]; exists {
isBindMount = true
srcPath = bindMap.SrcPath
if strings.ToLower(bindMap.Mode) == "rw" {
srcRW = true
}
// Otherwise create an directory in $ROOT/volumes/ and use that
} else {
// Do not pass a container as the parameter for the volume creation.
// The graph driver using the container's information ( Image ) to
// create the parent.
c, err := container.runtime.volumes.Create(nil, nil, "", "", nil)
if err != nil {
return err
}
srcPath, err = volumesDriver.Get(c.ID)
if err != nil {
return fmt.Errorf("Driver %s failed to get volume rootfs %s: %s", volumesDriver, c.ID, err)
}
srcRW = true // RW by default
}
container.Volumes[volPath] = srcPath
container.VolumesRW[volPath] = srcRW
// Create the mountpoint
rootVolPath := path.Join(container.RootfsPath(), volPath)
if err := os.MkdirAll(rootVolPath, 0755); err != nil {
return err
}
// Do not copy or change permissions if we are mounting from the host
if srcRW && !isBindMount {
volList, err := ioutil.ReadDir(rootVolPath)
if err != nil {
return err
}
if len(volList) > 0 {
srcList, err := ioutil.ReadDir(srcPath)
if err != nil {
return err
}
if len(srcList) == 0 {
// If the source volume is empty copy files from the root into the volume
if err := archive.CopyWithTar(rootVolPath, srcPath); err != nil {
return err
}
var stat syscall.Stat_t
if err := syscall.Stat(rootVolPath, &stat); err != nil {
return err
}
var srcStat syscall.Stat_t
if err := syscall.Stat(srcPath, &srcStat); err != nil {
return err
}
// Change the source volume's ownership if it differs from the root
// files that where just copied
if stat.Uid != srcStat.Uid || stat.Gid != srcStat.Gid {
if err := os.Chown(srcPath, int(stat.Uid), int(stat.Gid)); err != nil {
return err
}
}
}
}
}
}
if err := container.generateLXCConfig(); err != nil {
return err
}
var lxcStart string = "lxc-start"
if container.hostConfig.Privileged && container.runtime.capabilities.AppArmor {
lxcStart = path.Join(container.runtime.config.Root, "lxc-start-unconfined")
}
params := []string{
lxcStart,
"-n", container.ID,
"-f", container.lxcConfigPath(),
"--",
"/.dockerinit",
}
// Networking
if !container.Config.NetworkDisabled {
params = append(params, "-g", container.network.Gateway.String())
}
// User
if container.Config.User != "" {
params = append(params, "-u", container.Config.User)
}
// Setup environment
env := []string{
"HOME=/",
"PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin",
"container=lxc",
"HOSTNAME=" + container.Config.Hostname,
}
if container.Config.Tty {
env = append(env, "TERM=xterm")
}
// Init any links between the parent and children
runtime := container.runtime
children, err := runtime.Children(container.Name)
if err != nil {
return err
}
if len(children) > 0 {
container.activeLinks = make(map[string]*Link, len(children))
// If we encounter an error make sure that we rollback any network
// config and ip table changes
rollback := func() {
for _, link := range container.activeLinks {
link.Disable()
}
container.activeLinks = nil
}
for p, child := range children {
link, err := NewLink(container, child, p, runtime.networkManager.bridgeIface)
if err != nil {
rollback()
return err
}
container.activeLinks[link.Alias()] = link
if err := link.Enable(); err != nil {
rollback()
return err
}
for _, envVar := range link.ToEnv() {
env = append(env, envVar)
}
}
}
for _, elem := range container.Config.Env {
env = append(env, elem)
}
if err := container.generateEnvConfig(env); err != nil {
return err
}
if container.Config.WorkingDir != "" {
workingDir := path.Clean(container.Config.WorkingDir)
utils.Debugf("[working dir] working dir is %s", workingDir)
if err := os.MkdirAll(path.Join(container.RootfsPath(), workingDir), 0755); err != nil {
return nil
}
params = append(params,
"-w", workingDir,
)
}
// Program
params = append(params, "--", container.Path)
params = append(params, container.Args...)
if RootIsShared() {
// lxc-start really needs / to be non-shared, or all kinds of stuff break
// when lxc-start unmount things and those unmounts propagate to the main
// mount namespace.
// What we really want is to clone into a new namespace and then
// mount / MS_REC|MS_SLAVE, but since we can't really clone or fork
// without exec in go we have to do this horrible shell hack...
shellString :=
"mount --make-rslave /; exec " +
utils.ShellQuoteArguments(params)
params = []string{
"unshare", "-m", "--", "/bin/sh", "-c", shellString,
}
}
container.cmd = exec.Command(params[0], params[1:]...)
// Setup logging of stdout and stderr to disk
if err := container.runtime.LogToDisk(container.stdout, container.logPath("json"), "stdout"); err != nil {
return err
}
if err := container.runtime.LogToDisk(container.stderr, container.logPath("json"), "stderr"); err != nil {
return err
}
container.cmd.SysProcAttr = &syscall.SysProcAttr{Setsid: true}
if container.Config.Tty {
err = container.startPty()
} else {
err = container.start()
}
if err != nil {
return err
}
// FIXME: save state on disk *first*, then converge
// this way disk state is used as a journal, eg. we can restore after crash etc.
container.State.SetRunning(container.cmd.Process.Pid)
// Init the lock
container.waitLock = make(chan struct{})
container.ToDisk()
go container.monitor()
defer utils.Debugf("Container running: %v", container.State.IsRunning())
// We wait for the container to be fully running.
// Timeout after 5 seconds. In case of broken pipe, just retry.
// Note: The container can run and finish correctly before
// the end of this loop
for now := time.Now(); time.Since(now) < 5*time.Second; {
// If the container dies while waiting for it, just return
if !container.State.IsRunning() {
return nil
}
output, err := exec.Command("lxc-info", "-s", "-n", container.ID).CombinedOutput()
if err != nil {
utils.Debugf("Error with lxc-info: %s (%s)", err, output)
output, err = exec.Command("lxc-info", "-s", "-n", container.ID).CombinedOutput()
if err != nil {
utils.Debugf("Second Error with lxc-info: %s (%s)", err, output)
return err
}
}
if strings.Contains(string(output), "RUNNING") {
return nil
}
utils.Debugf("Waiting for the container to start (running: %v): %s", container.State.IsRunning(), bytes.TrimSpace(output))
time.Sleep(50 * time.Millisecond)
}
if container.State.IsRunning() {
return ErrContainerStartTimeout
}
return ErrContainerStart
}
func (container *Container) Run() error {
if err := container.Start(); err != nil {
return err
}
container.Wait()
return nil
}
func (container *Container) Output() (output []byte, err error) {
pipe, err := container.StdoutPipe()
if err != nil {
return nil, err
}
defer pipe.Close()
if err := container.Start(); err != nil {
return nil, err
}
output, err = ioutil.ReadAll(pipe)
container.Wait()
return output, err
}
// Container.StdinPipe returns a WriteCloser which can be used to feed data
// to the standard input of the container's active process.
// Container.StdoutPipe and Container.StderrPipe each return a ReadCloser
// which can be used to retrieve the standard output (and error) generated
// by the container's active process. The output (and error) are actually
// copied and delivered to all StdoutPipe and StderrPipe consumers, using
// a kind of "broadcaster".
func (container *Container) StdinPipe() (io.WriteCloser, error) {
return container.stdinPipe, nil
}
func (container *Container) StdoutPipe() (io.ReadCloser, error) {
reader, writer := io.Pipe()
container.stdout.AddWriter(writer, "")
return utils.NewBufReader(reader), nil
}
func (container *Container) StderrPipe() (io.ReadCloser, error) {
reader, writer := io.Pipe()
container.stderr.AddWriter(writer, "")
return utils.NewBufReader(reader), nil
}
func (container *Container) buildHostnameAndHostsFiles(IP string) {
container.HostnamePath = path.Join(container.root, "hostname")
ioutil.WriteFile(container.HostnamePath, []byte(container.Config.Hostname+"\n"), 0644)
hostsContent := []byte(`
127.0.0.1 localhost
::1 localhost ip6-localhost ip6-loopback
fe00::0 ip6-localnet
ff00::0 ip6-mcastprefix
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
`)
container.HostsPath = path.Join(container.root, "hosts")
if container.Config.Domainname != "" {
hostsContent = append([]byte(fmt.Sprintf("%s\t%s.%s %s\n", IP, container.Config.Hostname, container.Config.Domainname, container.Config.Hostname)), hostsContent...)
} else {
hostsContent = append([]byte(fmt.Sprintf("%s\t%s\n", IP, container.Config.Hostname)), hostsContent...)
}
ioutil.WriteFile(container.HostsPath, hostsContent, 0644)
}
func (container *Container) allocateNetwork() error {
if container.Config.NetworkDisabled {
return nil
}
var (
iface *NetworkInterface
err error
)
if container.State.IsGhost() {
if manager := container.runtime.networkManager; manager.disabled {
iface = &NetworkInterface{disabled: true}
} else {
iface = &NetworkInterface{
IPNet: net.IPNet{IP: net.ParseIP(container.NetworkSettings.IPAddress), Mask: manager.bridgeNetwork.Mask},
Gateway: manager.bridgeNetwork.IP,
manager: manager,
}
if iface != nil && iface.IPNet.IP != nil {
ipNum := ipToInt(iface.IPNet.IP)
manager.ipAllocator.inUse[ipNum] = struct{}{}
} else {
iface, err = container.runtime.networkManager.Allocate()
if err != nil {
return err
}
}
}
} else {
iface, err = container.runtime.networkManager.Allocate()
if err != nil {
return err
}
}
if container.Config.PortSpecs != nil {
utils.Debugf("Migrating port mappings for container: %s", strings.Join(container.Config.PortSpecs, ", "))
if err := migratePortMappings(container.Config, container.hostConfig); err != nil {
return err
}
container.Config.PortSpecs = nil
if err := container.writeHostConfig(); err != nil {
return err
}
}
var (
portSpecs = make(map[Port]struct{})
bindings = make(map[Port][]PortBinding)
)