-
Notifications
You must be signed in to change notification settings - Fork 173
/
ops_linux.go
1355 lines (1123 loc) · 38.1 KB
/
ops_linux.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 2016-2017 VMware, Inc. All Rights Reserved.
//
// 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 tether
import (
"bytes"
"context"
"errors"
"fmt"
"io/ioutil"
"net"
"net/url"
"os"
"path"
"path/filepath"
"strconv"
"strings"
"sync"
"syscall"
"time"
log "github.com/Sirupsen/logrus"
"github.com/d2g/dhcp4"
"github.com/docker/docker/pkg/archive"
"github.com/mdlayher/arp"
"github.com/mdlayher/ethernet"
"github.com/opencontainers/runc/libcontainer/user"
"github.com/vishvananda/netlink"
"github.com/vmware/vic/lib/dhcp"
"github.com/vmware/vic/lib/dhcp/client"
"github.com/vmware/vic/lib/etcconf"
"github.com/vmware/vic/pkg/ip"
"github.com/vmware/vic/pkg/trace"
"github.com/vmware/vmw-guestinfo/rpcout"
)
var (
defaultExecUser = &user.ExecUser{
Uid: syscall.Getuid(),
Gid: syscall.Getgid(),
Home: "/",
}
defaultUserSpecification = "0:0"
filesForMinOSLinux = map[string]os.FileMode{
"/etc/hostname": 0644,
"/etc/hosts": 0644,
"/etc/resolv.conf": 0644,
"/.tether/etc/hostname": 0644,
"/.tether/etc/hosts": 0644,
"/.tether/etc/resolv.conf": 0644,
}
)
const (
hostnameFile = "/etc/hostname"
hostnameFileBindSrc = "/.tether/etc/hostname"
hostsPathBindSrc = "/.tether/etc/hosts"
resolvConfPathBindSrc = "/.tether/etc/resolv.conf"
byLabelDir = "/dev/disk/by-label"
pciDevPath = "/sys/bus/pci/devices"
nfsFileSystemType = "nfs"
ext4FileSystemType = "ext4"
bridgeTableNumber = 201
// directory in which to perform the direct mount of disk for bind mount to actual
// target
diskBindBase = "/.filesystem-by-label/"
// used to isolate applications from the lost+found in the root of ext4
volumeDataDir = "/.vic.vol.data"
// temp directory to copy existing data to mounts
bindDir = "/.bind"
)
type BaseOperations struct {
dhcpLoops map[string]chan struct{}
dynEndpoints map[string][]*NetworkEndpoint
config Config
// Exclusive access to utilityPids table
utilityPidMutex sync.Mutex
// set of child PIDs for one-off non-persistent processes
utilityPids map[int]chan int
}
// NetLink gives us an interface to the netlink calls used so that
// we can test the calling code.
type Netlink interface {
LinkByName(string) (netlink.Link, error)
LinkSetName(netlink.Link, string) error
LinkSetDown(netlink.Link) error
LinkSetUp(netlink.Link) error
LinkSetAlias(netlink.Link, string) error
AddrList(netlink.Link, int) ([]netlink.Addr, error)
AddrAdd(netlink.Link, *netlink.Addr) error
AddrDel(netlink.Link, *netlink.Addr) error
RouteAdd(*netlink.Route) error
RouteDel(*netlink.Route) error
RuleList(family int) ([]netlink.Rule, error)
// Not quite netlink, but tightly associated
LinkBySlot(slot int32) (netlink.Link, error)
}
func (t *BaseOperations) LinkByName(name string) (netlink.Link, error) {
return netlink.LinkByName(name)
}
func (t *BaseOperations) LinkSetName(link netlink.Link, name string) error {
return netlink.LinkSetName(link, name)
}
func (t *BaseOperations) LinkSetDown(link netlink.Link) error {
return netlink.LinkSetDown(link)
}
func (t *BaseOperations) LinkSetUp(link netlink.Link) error {
return netlink.LinkSetUp(link)
}
func (t *BaseOperations) LinkSetAlias(link netlink.Link, alias string) error {
return netlink.LinkSetAlias(link, alias)
}
func (t *BaseOperations) AddrList(link netlink.Link, family int) ([]netlink.Addr, error) {
return netlink.AddrList(link, family)
}
func (t *BaseOperations) AddrAdd(link netlink.Link, addr *netlink.Addr) error {
return netlink.AddrAdd(link, addr)
}
func (t *BaseOperations) AddrDel(link netlink.Link, addr *netlink.Addr) error {
return netlink.AddrDel(link, addr)
}
func (t *BaseOperations) RouteAdd(route *netlink.Route) error {
return netlink.RouteAdd(route)
}
func (t *BaseOperations) RouteDel(route *netlink.Route) error {
return netlink.RouteDel(route)
}
func (t *BaseOperations) RuleList(family int) ([]netlink.Rule, error) {
return netlink.RuleList(family)
}
func (t *BaseOperations) LinkBySlot(slot int32) (netlink.Link, error) {
pciPath, err := slotToPCIPath(slot, 0)
if err != nil {
return nil, err
}
name, err := pciToLinkName(pciPath)
if err != nil {
return nil, err
}
log.Debugf("got link name: %#v", name)
return t.LinkByName(name)
}
// SetHostname sets both the kernel hostname and /etc/hostname to the specified string
func (t *BaseOperations) SetHostname(hostname string, aliases ...string) error {
defer trace.End(trace.Begin("setting hostname to " + hostname))
old, err := os.Hostname()
if err != nil {
log.Warnf("Unable to get current hostname - will not be able to revert on failure: %s", err)
}
err = Sys.Syscall.Sethostname([]byte(hostname))
if err != nil {
log.Errorf("Unable to set hostname: %s", err)
return err
}
log.Debugf("Updated kernel hostname")
// bind-mount /.tether/etc/hostname to /etc/hostname
if err = bindMount(hostnameFileBindSrc, hostnameFile); err != nil {
return err
}
// update /etc/hostname to match
err = ioutil.WriteFile(hostnameFile, []byte(hostname), 0644)
if err != nil {
log.Errorf("Failed to update hostname in %s", hostnameFile)
// revert the hostname
if old != "" {
log.Warnf("Reverting kernel hostname to %s", old)
err2 := Sys.Syscall.Sethostname([]byte(old))
if err2 != nil {
log.Errorf("Unable to revert kernel hostname - kernel and hostname file are out of sync! Error: %s", err2)
}
}
return err
}
if err := BindSys.Hosts.Load(); err != nil {
log.Errorf("Unable to load existing /etc/hosts file - modifications since last load will be overwritten: %s", err)
}
// add entry to hosts for resolution without nameservers
lo4 := net.IPv4(127, 0, 1, 1)
for _, a := range append(aliases, hostname) {
BindSys.Hosts.SetHost(a, lo4)
BindSys.Hosts.SetHost(a, net.IPv6loopback)
}
if err = bindMountAndSave(Sys.Hosts, BindSys.Hosts); err != nil {
return err
}
return nil
}
func slotToPCIPath(pciSlot int32, fun int32) (string, error) {
// see https://kb.vmware.com/kb/2047927
dev := pciSlot & 0x1f // DDDDD
bus := (pciSlot >> 5) & 0x1f // BBBBB
if fun == 0 {
fun = (pciSlot >> 10) & 0x7 // FFF
}
if bus == 0 {
return path.Join(pciDevPath, fmt.Sprintf("0000:%02x:%02x.%d", bus, dev, fun)), nil
}
// device on secondary bus, prepend pci bridge address, pciBridge0.pciSlotNumber is "17" aka "0x11"
bridgeSlot := 0x11 + (bus - 1)
bridgeAddr, err := slotToPCIPath(bridgeSlot, fun)
if err != nil {
return "", err
}
return path.Join(bridgeAddr, fmt.Sprintf("0000:*:%02x.0", dev)), nil
}
func pciToLinkName(pciPath string) (string, error) {
p := filepath.Join(pciPath, "net", "*")
matches, err := filepath.Glob(p)
if err != nil {
return "", err
}
if len(matches) != 1 {
return "", fmt.Errorf("%d eth interfaces match %s (%v)", len(matches), p, matches)
}
return path.Base(matches[0]), nil
}
func renameLink(t Netlink, link netlink.Link, slot int32, endpoint *NetworkEndpoint) (rlink netlink.Link, err error) {
rlink = link
defer func() {
// we still need to ensure that the link is up irrespective of path
err = t.LinkSetUp(link)
if err != nil {
err = fmt.Errorf("failed to bring link %s up: %s", link.Attrs().Name, err)
}
}()
if link.Attrs().Name == endpoint.Name || link.Attrs().Alias == endpoint.Name || endpoint.Name == "" {
// if the network is already identified, whether as primary name or alias it doesn't need repeating.
// if the name is empty then it should not be aliases or named directly. IPAM data should still be applied.
return link, nil
}
if strings.HasPrefix(link.Attrs().Name, "eth") {
log.Infof("Renaming link %s to %s", link.Attrs().Name, endpoint.Name)
err := t.LinkSetDown(link)
if err != nil {
detail := fmt.Sprintf("failed to set link %s down for rename: %s", link.Attrs().Name, err)
return nil, errors.New(detail)
}
err = t.LinkSetName(link, endpoint.Name)
if err != nil {
return nil, err
}
// reacquire link with updated attributes
link, err := t.LinkBySlot(slot)
if err != nil {
detail := fmt.Sprintf("unable to reacquire link %s after rename pass: %s", link.Attrs().Name, err)
return nil, errors.New(detail)
}
return link, nil
}
if link.Attrs().Alias == "" {
log.Infof("Aliasing link %s to %s", link.Attrs().Name, endpoint.Name)
err := t.LinkSetAlias(link, endpoint.Name)
if err != nil {
return nil, err
}
// reacquire link with updated attributes
link, err := t.LinkBySlot(slot)
if err != nil {
detail := fmt.Sprintf("unable to reacquire link %s after rename pass: %s", link.Attrs().Name, err)
return nil, errors.New(detail)
}
return link, nil
}
log.Warnf("Unable to add additional alias on link %s for %s", link.Attrs().Name, endpoint.Name)
return link, nil
}
func gratuitous(t Netlink, link netlink.Link, endpoint *NetworkEndpoint) error {
if ip.IsUnspecifiedIP(endpoint.Assigned.IP) {
// refusing to broadcast unspecified IP
return nil
}
intf, err := net.InterfaceByIndex(link.Attrs().Index)
if err != nil {
return err
}
client, err := arp.Dial(intf)
if err != nil {
return err
}
defer func() {
client.Close()
}()
srcHW := link.Attrs().HardwareAddr
srcPA := endpoint.Assigned.IP
gratuitousPacket, err := arp.NewPacket(arp.OperationRequest, srcHW, srcPA, ethernet.Broadcast, srcPA)
if err != nil {
return err
}
log.Debugf("Writing gratuitous arp on %s: %+v", endpoint.Name, gratuitousPacket)
return client.WriteTo(gratuitousPacket, ethernet.Broadcast)
}
func getDynamicIP(t Netlink, link netlink.Link, endpoint *NetworkEndpoint) (client.Client, error) {
var ack *dhcp.Packet
var err error
// use dhcp to acquire address
dc, err := client.NewClient(link.Attrs().Index, link.Attrs().HardwareAddr)
if err != nil {
return nil, err
}
params := []byte{byte(dhcp4.OptionSubnetMask)}
if ip.IsUnspecifiedIP(endpoint.Network.Gateway.IP) {
log.Debugf("No gateway IP. Asking DHCP.")
params = append(params, byte(dhcp4.OptionRouter))
}
if len(endpoint.Network.Nameservers) == 0 {
log.Debugf("No name servers configured. Asking DHCP.")
params = append(params, byte(dhcp4.OptionDomainNameServer))
}
dc.SetParameterRequestList(params...)
err = dc.Request()
if err != nil {
log.Errorf("error sending dhcp request: %s", err)
return nil, err
}
ack = dc.LastAck()
if ack.YourIP() == nil || ack.SubnetMask() == nil {
err = fmt.Errorf("dhcp assigned nil ip or subnet mask")
log.Error(err)
return nil, err
}
log.Infof("DHCP response: IP=%s, SubnetMask=%s, Gateway=%s, DNS=%s, Lease Time=%s", ack.YourIP(), ack.SubnetMask(), ack.Gateway(), ack.DNS(), ack.LeaseTime())
defer func() {
if err != nil && ack != nil {
dc.Release()
}
}()
return dc, nil
}
func updateEndpoint(newIP *net.IPNet, endpoint *NetworkEndpoint) {
log.Debugf("updateEndpoint(%s, %+v)", newIP, endpoint)
dhcp := endpoint.DHCP
if dhcp == nil {
endpoint.Assigned = *newIP
endpoint.Network.Assigned.Gateway = endpoint.Network.Gateway
endpoint.Network.Assigned.Nameservers = endpoint.Network.Nameservers
return
}
log.Debugf("DHCP data: %#v", dhcp)
log.Debugf("DHCP DNS Servers %s: ", dhcp.Nameservers)
endpoint.Assigned = dhcp.Assigned
endpoint.Network.Assigned.Gateway = dhcp.Gateway
if len(dhcp.Nameservers) > 0 {
endpoint.Network.Assigned.Nameservers = dhcp.Nameservers
}
}
func linkAddrUpdate(old, new *net.IPNet, t Netlink, link netlink.Link) error {
log.Infof("setting ip address %s for link %s", new, link.Attrs().Name)
if old != nil && !old.IP.Equal(new.IP) {
log.Debugf("removing old address %s", old)
if err := t.AddrDel(link, &netlink.Addr{IPNet: old}); err != nil {
if errno, ok := err.(syscall.Errno); !ok || errno != syscall.EADDRNOTAVAIL {
log.Errorf("failed to remove existing address %s: %s", old, err)
return err
}
}
log.Debugf("removed old address %s for link %s", old, link.Attrs().Name)
}
// assign IP to NIC
if err := t.AddrAdd(link, &netlink.Addr{IPNet: new}); err != nil {
if errno, ok := err.(syscall.Errno); !ok || errno != syscall.EEXIST {
log.Errorf("failed to assign ip %s for link %s", new, link.Attrs().Name)
return err
}
log.Warnf("address %s already set on interface %s", new, link.Attrs().Name)
}
log.Debugf("added address %s to link %s", new, link.Attrs().Name)
return nil
}
func updateRoutes(newIP *net.IPNet, t Netlink, link netlink.Link, endpoint *NetworkEndpoint) error {
gw := endpoint.Network.Assigned.Gateway
if ip.IsUnspecifiedIP(gw.IP) {
return nil
}
if endpoint.Network.Default {
return updateDefaultRoute(newIP, t, link, endpoint)
}
for _, d := range endpoint.Network.Destinations {
r := &netlink.Route{
LinkIndex: link.Attrs().Index,
Dst: &d,
Gw: gw.IP,
}
if err := t.RouteAdd(r); err != nil && !os.IsNotExist(err) {
log.Errorf("failed to add route for destination %s via gateway %s", d, gw.IP)
}
}
return nil
}
func bridgeTableExists(t Netlink) bool {
rules, err := t.RuleList(syscall.AF_INET)
if err != nil {
return false
}
for _, r := range rules {
if r.Table == bridgeTableNumber {
return true
}
}
return false
}
func updateDefaultRoute(newIP *net.IPNet, t Netlink, link netlink.Link, endpoint *NetworkEndpoint) error {
gw := endpoint.Network.Assigned.Gateway
// Add routes
if !endpoint.Network.Default || ip.IsUnspecifiedIP(gw.IP) {
log.Debugf("not setting route for network: default=%v gateway=%s", endpoint.Network.Default, gw.IP)
return nil
}
// #nosec: Errors unhandled.
_, defaultNet, _ := net.ParseCIDR("0.0.0.0/0")
// delete default route first
if err := t.RouteDel(&netlink.Route{LinkIndex: link.Attrs().Index, Dst: defaultNet}); err != nil {
if errno, ok := err.(syscall.Errno); !ok || errno != syscall.ESRCH {
return fmt.Errorf("could not update default route: %s", err)
}
}
// delete the default route for the bridge.out table, if it exists
bTablePresent := bridgeTableExists(t)
if bTablePresent {
if err := t.RouteDel(&netlink.Route{LinkIndex: link.Attrs().Index, Dst: defaultNet, Table: bridgeTableNumber}); err != nil {
if errno, ok := err.(syscall.Errno); !ok || errno != syscall.ESRCH {
return fmt.Errorf("could not update default route for bridge.out table: %s", err)
}
}
}
log.Infof("Setting default gateway to %s", gw.IP)
route := &netlink.Route{LinkIndex: link.Attrs().Index, Dst: defaultNet, Gw: gw.IP}
if err := t.RouteAdd(route); err != nil {
return fmt.Errorf("failed to add gateway route for endpoint %s: %s", endpoint.Network.Name, err)
}
if bTablePresent {
// Gateway IP may not contain a network mask, so it is taken from the assigned interface configuration
// where network mask has to be defined.
gwNet := &net.IPNet{
IP: gw.IP.Mask(newIP.Mask),
Mask: newIP.Mask,
}
log.Debugf("Adding route to default gateway network: %s/%s", gwNet.IP, gwNet.Mask)
route = &netlink.Route{LinkIndex: link.Attrs().Index, Dst: gwNet, Table: bridgeTableNumber}
if err := t.RouteAdd(route); err != nil {
// if IP address has changed and it stays within the same subnet it will cause already exists error,
// so we can safely ignore it.
errno, ok := err.(syscall.Errno)
if !ok || errno != syscall.EEXIST {
return fmt.Errorf(
"failed to add gateway network route for table bridge.out for endpoint %s: %s",
endpoint.Network.Name, err)
}
}
route = &netlink.Route{LinkIndex: link.Attrs().Index, Dst: defaultNet, Gw: gw.IP, Table: bridgeTableNumber}
if err := t.RouteAdd(route); err != nil {
return fmt.Errorf("failed to add gateway route for table bridge.out for endpoint %s: %s", endpoint.Network.Name, err)
}
}
log.Infof("updated default route to %s interface, gateway: %s", endpoint.Network.Name, gw.IP)
return nil
}
func (t *BaseOperations) updateHosts(endpoint *NetworkEndpoint) error {
log.Debugf("%+v", endpoint)
// Add /etc/hosts entry
if endpoint.Network.Name == "" {
return nil
}
if err := BindSys.Hosts.Load(); err != nil {
log.Errorf("Unable to load existing /etc/hosts file - modifications since last load will be overwritten: %s", err)
}
BindSys.Hosts.SetHost(fmt.Sprintf("%s.localhost", endpoint.Network.Name), endpoint.Assigned.IP)
if err := bindMountAndSave(Sys.Hosts, BindSys.Hosts); err != nil {
return err
}
return nil
}
func (t *BaseOperations) updateNameservers(endpoint *NetworkEndpoint) error {
gw := endpoint.Network.Assigned.Gateway
ns := endpoint.Network.Assigned.Nameservers
if len(ns) > 0 && len(endpoint.Network.Nameservers) > 0 {
log.Debugf("DHCP server returned DNS server configuration, it will be ignored")
}
// Manually set DNS servers should always be DNS servers that are being in use.
if len(endpoint.Network.Nameservers) > 0 {
ns = endpoint.Network.Nameservers
}
// Add nameservers
// This is incredibly trivial for now - should be updated to a less messy approach
if len(ns) > 0 {
BindSys.ResolvConf.AddNameservers(ns...)
log.Infof("Added nameservers: %+v", ns)
} else if !ip.IsUnspecifiedIP(gw.IP) {
BindSys.ResolvConf.AddNameservers(gw.IP)
log.Infof("Added nameserver: %s", gw.IP)
}
if err := bindMountAndSave(Sys.ResolvConf, BindSys.ResolvConf); err != nil {
return err
}
return nil
}
func (t *BaseOperations) Apply(endpoint *NetworkEndpoint) error {
defer trace.End(trace.Begin("applying endpoint configuration for " + endpoint.Network.Name))
return ApplyEndpoint(t, t, endpoint)
}
func ApplyEndpoint(nl Netlink, t *BaseOperations, endpoint *NetworkEndpoint) error {
if endpoint.configured {
log.Infof("skipping applying config for network %s as it has been applied already", endpoint.Network.Name)
return nil // already applied
}
log.Debugf("Static name servers: %s", endpoint.Network.Nameservers)
// Locate interface
slot, err := strconv.Atoi(endpoint.ID)
if err != nil {
return fmt.Errorf("endpoint ID must be a base10 numeric pci slot identifier: %s", err)
}
defer func() {
if err == nil {
log.Infof("successfully applied config for network %s", endpoint.Network.Name)
endpoint.configured = true
}
}()
var link netlink.Link
link, err = nl.LinkBySlot(int32(slot))
if err != nil {
return fmt.Errorf("unable to acquire reference to link %s: %s", endpoint.ID, err)
}
// rename the link if needed
link, err = renameLink(nl, link, int32(slot), endpoint)
if err != nil {
return fmt.Errorf("unable to reacquire link %s after rename pass: %s", endpoint.ID, err)
}
var dc client.Client
defer func() {
if err != nil && dc != nil {
dc.Release()
}
}()
var newIP *net.IPNet
if endpoint.IsDynamic() && endpoint.DHCP == nil {
if e, ok := t.dynEndpoints[endpoint.ID]; ok {
// endpoint shares NIC, copy over DHCP
endpoint.DHCP = e[0].DHCP
}
}
log.Debugf("%+v", endpoint)
if endpoint.IsDynamic() {
if endpoint.DHCP == nil {
dc, err = getDynamicIP(nl, link, endpoint)
if err != nil {
return err
}
ack := dc.LastAck()
endpoint.DHCP = &DHCPInfo{
Assigned: net.IPNet{IP: ack.YourIP(), Mask: ack.SubnetMask()},
Nameservers: ack.DNS(),
Gateway: net.IPNet{IP: ack.Gateway(), Mask: ack.SubnetMask()},
}
}
newIP = &endpoint.DHCP.Assigned
} else {
newIP = endpoint.IP
if newIP.IP.Equal(net.IPv4zero) {
// managed externally
return nil
}
}
var old *net.IPNet
if !ip.IsUnspecifiedIP(endpoint.Assigned.IP) {
old = &endpoint.Assigned
}
if err = linkAddrUpdate(old, newIP, nl, link); err != nil {
return err
}
updateEndpoint(newIP, endpoint)
// notify our peers of the assigned address
err = gratuitous(t, link, endpoint)
if err != nil {
log.Errorf("Failed to issue gratuitous ARP broadcast for %s: %s", endpoint.Name, err)
// continue regardless as this will eventually self-correct
// if the error is pernacious then we'll likely hit the other early error returns
}
if err = updateRoutes(newIP, nl, link, endpoint); err != nil {
return err
}
if err = t.updateHosts(endpoint); err != nil {
return err
}
Sys.ResolvConf.RemoveNameservers(endpoint.Network.Assigned.Nameservers...)
if err = t.updateNameservers(endpoint); err != nil {
return err
}
if endpoint.IsDynamic() {
eps := t.dynEndpoints[endpoint.ID]
found := false
for _, e := range eps {
if e == endpoint {
found = true
break
}
}
if !found {
eps = append(eps, endpoint)
t.dynEndpoints[endpoint.ID] = eps
}
}
// add renew/release loop if necessary
if dc != nil {
if _, ok := t.dhcpLoops[endpoint.ID]; !ok {
stop := make(chan struct{})
if err != nil {
log.Errorf("could not make DHCP client id for link %s: %s", link.Attrs().Name, err)
} else {
t.dhcpLoops[endpoint.ID] = stop
go t.dhcpLoop(stop, endpoint, dc)
}
}
}
return nil
}
func (t *BaseOperations) dhcpLoop(stop chan struct{}, e *NetworkEndpoint, dc client.Client) {
// wait half of the remaining lease time before trying again if we fail to renew
exp := time.After(dc.LastAck().LeaseTime() / 2)
// the divisor starts set to 4 so that we can update it AFTER the duration is calculated. This allows combining the check for integer overflow with
// the minimum wait time.
divisor := time.Duration(4)
for {
select {
case <-stop:
// release the ip
log.Infof("releasing IP address for network %s", e.Name)
dc.Release()
return
case <-exp:
log.Infof("renewing IP address for network %s", e.Name)
err := dc.Renew()
if err != nil {
log.Errorf("failed to renew ip address for network %s: %s", e.Name, err)
duration := dc.LastAck().LeaseTime() / divisor
// for now go with a minimum retry of 1min
if duration < time.Minute {
duration = time.Minute
} else if divisor*2 > divisor {
divisor *= 2
}
exp = time.After(duration)
continue
}
ack := dc.LastAck()
log.Infof("successfully renewed ip address: IP=%s, SubnetMask=%s, Gateway=%s, DNS=%s, Lease Time=%s", ack.YourIP(), ack.SubnetMask(), ack.Gateway(), ack.DNS(), ack.LeaseTime())
e.DHCP = &DHCPInfo{
Assigned: net.IPNet{IP: ack.YourIP(), Mask: ack.SubnetMask()},
Gateway: net.IPNet{IP: ack.Gateway(), Mask: ack.SubnetMask()},
Nameservers: ack.DNS(),
}
// TODO: determine if there are actually any changes to apply before performing updates
e.configured = false
t.Apply(e)
if err = t.config.UpdateNetworkEndpoint(e); err != nil {
log.Error(err)
}
// update any endpoints that share this NIC
for _, d := range t.dynEndpoints[e.ID] {
if e == d {
// already applied above
continue
}
d.DHCP = e.DHCP
d.configured = false
t.Apply(d)
if err = t.config.UpdateNetworkEndpoint(d); err != nil {
log.Error(err)
}
}
t.config.Flush()
exp = time.After(ack.LeaseTime() / 2)
}
}
}
// MountLabel performs a mount with the label and target being absolute paths
func (t *BaseOperations) MountLabel(ctx context.Context, label, target string) error {
defer trace.End(trace.Begin(fmt.Sprintf("Mounting %s on %s", label, target)))
bindTarget := path.Join(BindSys.Root, diskBindBase, label)
fi, err := os.Stat(target)
if err != nil {
// #nosec
if err := os.MkdirAll(target, 0755); err != nil {
// same as MountFileSystem error for consistency
return fmt.Errorf("unable to create mount point %s: %s", target, err)
}
}
// convert the label to a filesystem path
label = "/dev/disk/by-label/" + label
var e1, e2 error
_, e1 = os.Stat(bindTarget)
if e1 == nil {
// bindTarget exists, check whether or not bindTarget is a mount point
e2 = os.Remove(bindTarget)
}
if (e1 == nil && e2 == nil) || os.IsNotExist(e1) {
// #nosec
if err := os.MkdirAll(bindTarget, 0755); err != nil {
return fmt.Errorf("unable to create mount point %s: %s", bindTarget, err)
}
if err := mountDeviceLabel(ctx, label, bindTarget); err != nil {
return err
}
}
// at this point bindTarget should be mounted successfully
mntsrc := path.Join(bindTarget, volumeDataDir)
mnttype := "bind"
mntflags := uintptr(syscall.MS_BIND)
// if the volume contains a volumeDataDir directory then mount that instead of the root of the filesystem
// if we cannot read it we go with the root of the filesystem
_, err = os.Stat(mntsrc)
if err != nil {
if os.IsNotExist(err) {
// if there's not such directory then revert to using the device directly
log.Info("No %s data directory in volume, mounting filesystem directly", volumeDataDir)
mntsrc = label
mnttype = ext4FileSystemType
mntflags = syscall.MS_NOATIME
} else {
return fmt.Errorf("unable to determine whether lost+found masking is required: %s", err)
}
}
if err := Sys.Syscall.Mount(mntsrc, target, mnttype, mntflags, ""); err != nil {
// consistent with MountFileSystem
detail := fmt.Sprintf("mounting %s on %s failed: %s", label, target, err)
return errors.New(detail)
}
// change the ownership of the target directory to the original uid/gid
if fi != nil {
sys, ok := fi.Sys().(*syscall.Stat_t)
if !ok {
return fmt.Errorf("unable to get uid/gid from existing target directory %s", target)
}
uid := int(sys.Uid)
gid := int(sys.Gid)
log.Debugf("Setting target uid/gid to the mount source as %d/%d", uid, gid)
if err := os.Chown(target, uid, gid); err != nil {
return fmt.Errorf("unable to change the owner of the mount point %s: %s", target, err)
}
log.Debugf("Setting target %s permissions to the mount source as: %#o",
target, fi.Mode())
if err := os.Chmod(target, fi.Mode()); err != nil {
return fmt.Errorf("failed to set target %s mount permissions as %#o: %s",
target, fi.Mode(), err)
}
}
return nil
}
// mountDeviceLabel mounts the device to target
func mountDeviceLabel(ctx context.Context, label string, target string) error {
// wait for device label to show up until the ctx deadline exceeds.
WaitForDevice:
for {
select {
case <-ctx.Done():
detail := fmt.Sprintf("timed out waiting for %s to appear", label)
return errors.New(detail)
default:
_, err := os.Stat(label)
if err == nil || !os.IsNotExist(err) {
break WaitForDevice
}
// sleep for 1 ms to reduce pressure on cpu
time.Sleep(time.Millisecond)
}
}
if err := Sys.Syscall.Mount(label, target, ext4FileSystemType, syscall.MS_NOATIME, ""); err != nil {
// consistent with MountFileSystem
detail := fmt.Sprintf("Actual mount: mounting %s on %s failed: %s", label, target, err)
return errors.New(detail)
}
return nil
}
// MountTarget performs a mount based on the target path from the source url
// This assumes that the source url is valid and available.
func (t *BaseOperations) MountTarget(ctx context.Context, source url.URL, target string, mountOptions string) error {
defer trace.End(trace.Begin(fmt.Sprintf("Mounting %s on %s", source.String(), target)))
// #nosec
if err := os.MkdirAll(target, 0755); err != nil {
// same as MountLabel error for consistency
return fmt.Errorf("unable to create mount point %s: %s", target, err)
}
var rawSource bytes.Buffer
rawSource.WriteString(source.Hostname())
rawSource.WriteByte(':')
// ensure the path is absolute - not using path.Clean to allow arbitrary content
// so as not to bias what can be used for a share identifier.
if len(source.Path) == 0 || source.Path[0] != '/' {
rawSource.WriteByte('/')
}
rawSource.WriteString(source.Path)
ips, err := net.LookupIP(source.Hostname())
if err != nil {
log.Errorf("failing to resolve %s. mounting %s on %s failed: %s", source.Hostname(), source.String(), target, err)
return err
}
for _, ip := range ips {
//NOTE: the mountOptions of syscall mount only accept addr=ip. addr=FQDN doesn't work
//We resolve the ip address nearest the mounting action.
mountOptionsIP := strings.Replace(mountOptions, "addr="+source.Hostname(), "addr="+ip.String(), -1)
// NOTE: by default we are supporting "NOATIME" and it can be configurable later. this must be specfied as a flag.
// Additionally, we must parse out the "ro" option and supply it as a flag as well for this flavor of the mount call.
if err = Sys.Syscall.Mount(rawSource.String(), target, nfsFileSystemType, syscall.MS_NOATIME, mountOptionsIP); err != nil {
log.Debugf("mounting %s with resolved ip: %s on %s failed: %s", source.String(), ip.String(), target, err)
} else {
return nil
}
}
log.Errorf("mounting %s on %s failed: %s", source.String(), target, err)
return err
}
// CopyExistingContent copies the underlying files shadowed by a mount on a directory
// to the volume mounted on the directory
// see bug https://github.com/vmware/vic/issues/3482
func (t *BaseOperations) CopyExistingContent(source string) error {
defer trace.End(trace.Begin(fmt.Sprintf("copyExistingContent from %s", source)))
source = filepath.Clean(source)
bind := path.Join(BindSys.Root, bindDir)
// if mounted volume is not empty skip the copy task
if empty, err := isEmpty(source); err != nil || !empty {
if err != nil {
log.Errorf("error checking directory for contents %s: %+v", source, err)
return err
}
log.Debugf("Skipping copy as volume %s is not empty", source)