forked from hashicorp/nomad
/
structs.go
1806 lines (1536 loc) · 48.9 KB
/
structs.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
package structs
import (
"bytes"
"crypto/sha1"
"errors"
"fmt"
"io"
"reflect"
"regexp"
"strings"
"time"
"github.com/hashicorp/go-msgpack/codec"
"github.com/hashicorp/go-multierror"
"github.com/hashicorp/go-version"
"github.com/hashicorp/nomad/helper/args"
)
var (
ErrNoLeader = fmt.Errorf("No cluster leader")
ErrNoRegionPath = fmt.Errorf("No path to region")
defaultServiceJobRestartPolicy = RestartPolicy{
Delay: 15 * time.Second,
Attempts: 2,
Interval: 1 * time.Minute,
}
defaultBatchJobRestartPolicy = RestartPolicy{
Delay: 15 * time.Second,
Attempts: 15,
}
)
type MessageType uint8
const (
NodeRegisterRequestType MessageType = iota
NodeDeregisterRequestType
NodeUpdateStatusRequestType
NodeUpdateDrainRequestType
JobRegisterRequestType
JobDeregisterRequestType
EvalUpdateRequestType
EvalDeleteRequestType
AllocUpdateRequestType
AllocClientUpdateRequestType
)
const (
// IgnoreUnknownTypeFlag is set along with a MessageType
// to indicate that the message type can be safely ignored
// if it is not recognized. This is for future proofing, so
// that new commands can be added in a way that won't cause
// old servers to crash when the FSM attempts to process them.
IgnoreUnknownTypeFlag MessageType = 128
)
// RPCInfo is used to describe common information about query
type RPCInfo interface {
RequestRegion() string
IsRead() bool
AllowStaleRead() bool
}
// QueryOptions is used to specify various flags for read queries
type QueryOptions struct {
// The target region for this query
Region string
// If set, wait until query exceeds given index. Must be provided
// with MaxQueryTime.
MinQueryIndex uint64
// Provided with MinQueryIndex to wait for change.
MaxQueryTime time.Duration
// If set, any follower can service the request. Results
// may be arbitrarily stale.
AllowStale bool
}
func (q QueryOptions) RequestRegion() string {
return q.Region
}
// QueryOption only applies to reads, so always true
func (q QueryOptions) IsRead() bool {
return true
}
func (q QueryOptions) AllowStaleRead() bool {
return q.AllowStale
}
type WriteRequest struct {
// The target region for this write
Region string
}
func (w WriteRequest) RequestRegion() string {
// The target region for this request
return w.Region
}
// WriteRequest only applies to writes, always false
func (w WriteRequest) IsRead() bool {
return false
}
func (w WriteRequest) AllowStaleRead() bool {
return false
}
// QueryMeta allows a query response to include potentially
// useful metadata about a query
type QueryMeta struct {
// This is the index associated with the read
Index uint64
// If AllowStale is used, this is time elapsed since
// last contact between the follower and leader. This
// can be used to gauge staleness.
LastContact time.Duration
// Used to indicate if there is a known leader node
KnownLeader bool
}
// WriteMeta allows a write response to includ e potentially
// useful metadata about the write
type WriteMeta struct {
// This is the index associated with the write
Index uint64
}
// NodeRegisterRequest is used for Node.Register endpoint
// to register a node as being a schedulable entity.
type NodeRegisterRequest struct {
Node *Node
WriteRequest
}
// NodeDeregisterRequest is used for Node.Deregister endpoint
// to deregister a node as being a schedulable entity.
type NodeDeregisterRequest struct {
NodeID string
WriteRequest
}
// NodeUpdateStatusRequest is used for Node.UpdateStatus endpoint
// to update the status of a node.
type NodeUpdateStatusRequest struct {
NodeID string
Status string
WriteRequest
}
// NodeUpdateDrainRequest is used for updatin the drain status
type NodeUpdateDrainRequest struct {
NodeID string
Drain bool
WriteRequest
}
// NodeEvaluateRequest is used to re-evaluate the ndoe
type NodeEvaluateRequest struct {
NodeID string
WriteRequest
}
// NodeSpecificRequest is used when we just need to specify a target node
type NodeSpecificRequest struct {
NodeID string
QueryOptions
}
// JobRegisterRequest is used for Job.Register endpoint
// to register a job as being a schedulable entity.
type JobRegisterRequest struct {
Job *Job
WriteRequest
}
// JobDeregisterRequest is used for Job.Deregister endpoint
// to deregister a job as being a schedulable entity.
type JobDeregisterRequest struct {
JobID string
WriteRequest
}
// JobEvaluateRequest is used when we just need to re-evaluate a target job
type JobEvaluateRequest struct {
JobID string
WriteRequest
}
// JobSpecificRequest is used when we just need to specify a target job
type JobSpecificRequest struct {
JobID string
QueryOptions
}
// JobListRequest is used to parameterize a list request
type JobListRequest struct {
QueryOptions
}
// NodeListRequest is used to parameterize a list request
type NodeListRequest struct {
QueryOptions
}
// EvalUpdateRequest is used for upserting evaluations.
type EvalUpdateRequest struct {
Evals []*Evaluation
EvalToken string
WriteRequest
}
// EvalDeleteRequest is used for deleting an evaluation.
type EvalDeleteRequest struct {
Evals []string
Allocs []string
WriteRequest
}
// EvalSpecificRequest is used when we just need to specify a target evaluation
type EvalSpecificRequest struct {
EvalID string
QueryOptions
}
// EvalAckRequest is used to Ack/Nack a specific evaluation
type EvalAckRequest struct {
EvalID string
Token string
WriteRequest
}
// EvalDequeueRequest is used when we want to dequeue an evaluation
type EvalDequeueRequest struct {
Schedulers []string
Timeout time.Duration
WriteRequest
}
// EvalListRequest is used to list the evaluations
type EvalListRequest struct {
QueryOptions
}
// PlanRequest is used to submit an allocation plan to the leader
type PlanRequest struct {
Plan *Plan
WriteRequest
}
// AllocUpdateRequest is used to submit changes to allocations, either
// to cause evictions or to assign new allocaitons. Both can be done
// within a single transaction
type AllocUpdateRequest struct {
// Alloc is the list of new allocations to assign
Alloc []*Allocation
WriteRequest
}
// AllocListRequest is used to request a list of allocations
type AllocListRequest struct {
QueryOptions
}
// AllocSpecificRequest is used to query a specific allocation
type AllocSpecificRequest struct {
AllocID string
QueryOptions
}
// GenericRequest is used to request where no
// specific information is needed.
type GenericRequest struct {
QueryOptions
}
// GenericResponse is used to respond to a request where no
// specific response information is needed.
type GenericResponse struct {
WriteMeta
}
const (
ProtocolVersion = "protocol"
APIMajorVersion = "api.major"
APIMinorVersion = "api.minor"
)
// VersionResponse is used for the Status.Version reseponse
type VersionResponse struct {
Build string
Versions map[string]int
QueryMeta
}
// JobRegisterResponse is used to respond to a job registration
type JobRegisterResponse struct {
EvalID string
EvalCreateIndex uint64
JobModifyIndex uint64
QueryMeta
}
// JobDeregisterResponse is used to respond to a job deregistration
type JobDeregisterResponse struct {
EvalID string
EvalCreateIndex uint64
JobModifyIndex uint64
QueryMeta
}
// NodeUpdateResponse is used to respond to a node update
type NodeUpdateResponse struct {
HeartbeatTTL time.Duration
EvalIDs []string
EvalCreateIndex uint64
NodeModifyIndex uint64
QueryMeta
}
// NodeDrainUpdateResponse is used to respond to a node drain update
type NodeDrainUpdateResponse struct {
EvalIDs []string
EvalCreateIndex uint64
NodeModifyIndex uint64
QueryMeta
}
// NodeAllocsResponse is used to return allocs for a single node
type NodeAllocsResponse struct {
Allocs []*Allocation
QueryMeta
}
// SingleNodeResponse is used to return a single node
type SingleNodeResponse struct {
Node *Node
QueryMeta
}
// JobListResponse is used for a list request
type NodeListResponse struct {
Nodes []*NodeListStub
QueryMeta
}
// SingleJobResponse is used to return a single job
type SingleJobResponse struct {
Job *Job
QueryMeta
}
// JobListResponse is used for a list request
type JobListResponse struct {
Jobs []*JobListStub
QueryMeta
}
// SingleAllocResponse is used to return a single allocation
type SingleAllocResponse struct {
Alloc *Allocation
QueryMeta
}
// JobAllocationsResponse is used to return the allocations for a job
type JobAllocationsResponse struct {
Allocations []*AllocListStub
QueryMeta
}
// JobEvaluationsResponse is used to return the evaluations for a job
type JobEvaluationsResponse struct {
Evaluations []*Evaluation
QueryMeta
}
// SingleEvalResponse is used to return a single evaluation
type SingleEvalResponse struct {
Eval *Evaluation
QueryMeta
}
// EvalDequeueResponse is used to return from a dequeue
type EvalDequeueResponse struct {
Eval *Evaluation
Token string
QueryMeta
}
// PlanResponse is used to return from a PlanRequest
type PlanResponse struct {
Result *PlanResult
WriteMeta
}
// AllocListResponse is used for a list request
type AllocListResponse struct {
Allocations []*AllocListStub
QueryMeta
}
// EvalListResponse is used for a list request
type EvalListResponse struct {
Evaluations []*Evaluation
QueryMeta
}
// EvalAllocationsResponse is used to return the allocations for an evaluation
type EvalAllocationsResponse struct {
Allocations []*AllocListStub
QueryMeta
}
const (
NodeStatusInit = "initializing"
NodeStatusReady = "ready"
NodeStatusDown = "down"
)
// ShouldDrainNode checks if a given node status should trigger an
// evaluation. Some states don't require any further action.
func ShouldDrainNode(status string) bool {
switch status {
case NodeStatusInit, NodeStatusReady:
return false
case NodeStatusDown:
return true
default:
panic(fmt.Sprintf("unhandled node status %s", status))
}
}
// ValidNodeStatus is used to check if a node status is valid
func ValidNodeStatus(status string) bool {
switch status {
case NodeStatusInit, NodeStatusReady, NodeStatusDown:
return true
default:
return false
}
}
// Node is a representation of a schedulable client node
type Node struct {
// ID is a unique identifier for the node. It can be constructed
// by doing a concatenation of the Name and Datacenter as a simple
// approach. Alternatively a UUID may be used.
ID string
// Datacenter for this node
Datacenter string
// Node name
Name string
// Attributes is an arbitrary set of key/value
// data that can be used for constraints. Examples
// include "kernel.name=linux", "arch=386", "driver.docker=1",
// "docker.runtime=1.8.3"
Attributes map[string]string
// Resources is the available resources on the client.
// For example 'cpu=2' 'memory=2048'
Resources *Resources
// Reserved is the set of resources that are reserved,
// and should be subtracted from the total resources for
// the purposes of scheduling. This may be provide certain
// high-watermark tolerances or because of external schedulers
// consuming resources.
Reserved *Resources
// Links are used to 'link' this client to external
// systems. For example 'consul=foo.dc1' 'aws=i-83212'
// 'ami=ami-123'
Links map[string]string
// Meta is used to associate arbitrary metadata with this
// client. This is opaque to Nomad.
Meta map[string]string
// NodeClass is an opaque identifier used to group nodes
// together for the purpose of determining scheduling pressure.
NodeClass string
// Drain is controlled by the servers, and not the client.
// If true, no jobs will be scheduled to this node, and existing
// allocations will be drained.
Drain bool
// Status of this node
Status string
// StatusDescription is meant to provide more human useful information
StatusDescription string
// Raft Indexes
CreateIndex uint64
ModifyIndex uint64
}
// TerminalStatus returns if the current status is terminal and
// will no longer transition.
func (n *Node) TerminalStatus() bool {
switch n.Status {
case NodeStatusDown:
return true
default:
return false
}
}
// Stub returns a summarized version of the node
func (n *Node) Stub() *NodeListStub {
return &NodeListStub{
ID: n.ID,
Datacenter: n.Datacenter,
Name: n.Name,
NodeClass: n.NodeClass,
Drain: n.Drain,
Status: n.Status,
StatusDescription: n.StatusDescription,
CreateIndex: n.CreateIndex,
ModifyIndex: n.ModifyIndex,
}
}
// NodeListStub is used to return a subset of job information
// for the job list
type NodeListStub struct {
ID string
Datacenter string
Name string
NodeClass string
Drain bool
Status string
StatusDescription string
CreateIndex uint64
ModifyIndex uint64
}
// Resources is used to define the resources available
// on a client
type Resources struct {
CPU int
MemoryMB int `mapstructure:"memory"`
DiskMB int `mapstructure:"disk"`
IOPS int
Networks []*NetworkResource
}
// Copy returns a deep copy of the resources
func (r *Resources) Copy() *Resources {
newR := new(Resources)
*newR = *r
n := len(r.Networks)
newR.Networks = make([]*NetworkResource, n)
for i := 0; i < n; i++ {
newR.Networks[i] = r.Networks[i].Copy()
}
return newR
}
// NetIndex finds the matching net index using device name
func (r *Resources) NetIndex(n *NetworkResource) int {
for idx, net := range r.Networks {
if net.Device == n.Device {
return idx
}
}
return -1
}
// Superset checks if one set of resources is a superset
// of another. This ignores network resources, and the NetworkIndex
// should be used for that.
func (r *Resources) Superset(other *Resources) (bool, string) {
if r.CPU < other.CPU {
return false, "cpu exhausted"
}
if r.MemoryMB < other.MemoryMB {
return false, "memory exhausted"
}
if r.DiskMB < other.DiskMB {
return false, "disk exhausted"
}
if r.IOPS < other.IOPS {
return false, "iops exhausted"
}
return true, ""
}
// Add adds the resources of the delta to this, potentially
// returning an error if not possible.
func (r *Resources) Add(delta *Resources) error {
if delta == nil {
return nil
}
r.CPU += delta.CPU
r.MemoryMB += delta.MemoryMB
r.DiskMB += delta.DiskMB
r.IOPS += delta.IOPS
for _, n := range delta.Networks {
// Find the matching interface by IP or CIDR
idx := r.NetIndex(n)
if idx == -1 {
r.Networks = append(r.Networks, n.Copy())
} else {
r.Networks[idx].Add(n)
}
}
return nil
}
func (r *Resources) GoString() string {
return fmt.Sprintf("*%#v", *r)
}
type Port struct {
Label string
Value int `mapstructure:"static"`
}
// NetworkResource is used to represent available network
// resources
type NetworkResource struct {
Device string // Name of the device
CIDR string // CIDR block of addresses
IP string // IP address
MBits int // Throughput
ReservedPorts []Port // Reserved ports
DynamicPorts []Port // Dynamically assigned ports
}
// Copy returns a deep copy of the network resource
func (n *NetworkResource) Copy() *NetworkResource {
newR := new(NetworkResource)
*newR = *n
if n.ReservedPorts != nil {
newR.ReservedPorts = make([]Port, len(n.ReservedPorts))
copy(newR.ReservedPorts, n.ReservedPorts)
}
if n.DynamicPorts != nil {
newR.DynamicPorts = make([]Port, len(n.DynamicPorts))
copy(newR.DynamicPorts, n.DynamicPorts)
}
return newR
}
// Add adds the resources of the delta to this, potentially
// returning an error if not possible.
func (n *NetworkResource) Add(delta *NetworkResource) {
if len(delta.ReservedPorts) > 0 {
n.ReservedPorts = append(n.ReservedPorts, delta.ReservedPorts...)
}
n.MBits += delta.MBits
n.DynamicPorts = append(n.DynamicPorts, delta.DynamicPorts...)
}
func (n *NetworkResource) GoString() string {
return fmt.Sprintf("*%#v", *n)
}
func (n *NetworkResource) MapLabelToValues(port_map map[string]int) map[string]int {
labelValues := make(map[string]int)
ports := append(n.ReservedPorts, n.DynamicPorts...)
for _, port := range ports {
if mapping, ok := port_map[port.Label]; ok {
labelValues[port.Label] = mapping
} else {
labelValues[port.Label] = port.Value
}
}
return labelValues
}
const (
// JobTypeNomad is reserved for internal system tasks and is
// always handled by the CoreScheduler.
JobTypeCore = "_core"
JobTypeService = "service"
JobTypeBatch = "batch"
JobTypeSystem = "system"
)
const (
JobStatusPending = "pending" // Pending means the job is waiting on scheduling
JobStatusRunning = "running" // Running means the entire job is running
JobStatusComplete = "complete" // Complete means there was a clean termination
JobStatusDead = "dead" // Dead means there was abnormal termination
)
const (
// JobMinPriority is the minimum allowed priority
JobMinPriority = 1
// JobDefaultPriority is the default priority if not
// not specified.
JobDefaultPriority = 50
// JobMaxPriority is the maximum allowed priority
JobMaxPriority = 100
// Ensure CoreJobPriority is higher than any user
// specified job so that it gets priority. This is important
// for the system to remain healthy.
CoreJobPriority = JobMaxPriority * 2
)
// Job is the scope of a scheduling request to Nomad. It is the largest
// scoped object, and is a named collection of task groups. Each task group
// is further composed of tasks. A task group (TG) is the unit of scheduling
// however.
type Job struct {
// Region is the Nomad region that handles scheduling this job
Region string
// ID is a unique identifier for the job per region. It can be
// specified hierarchically like LineOfBiz/OrgName/Team/Project
ID string
// Name is the logical name of the job used to refer to it. This is unique
// per region, but not unique globally.
Name string
// Type is used to control various behaviors about the job. Most jobs
// are service jobs, meaning they are expected to be long lived.
// Some jobs are batch oriented meaning they run and then terminate.
// This can be extended in the future to support custom schedulers.
Type string
// Priority is used to control scheduling importance and if this job
// can preempt other jobs.
Priority int
// AllAtOnce is used to control if incremental scheduling of task groups
// is allowed or if we must do a gang scheduling of the entire job. This
// can slow down larger jobs if resources are not available.
AllAtOnce bool `mapstructure:"all_at_once"`
// Datacenters contains all the datacenters this job is allowed to span
Datacenters []string
// Constraints can be specified at a job level and apply to
// all the task groups and tasks.
Constraints []*Constraint
// TaskGroups are the collections of task groups that this job needs
// to run. Each task group is an atomic unit of scheduling and placement.
TaskGroups []*TaskGroup
// Update is used to control the update strategy
Update UpdateStrategy
// Meta is used to associate arbitrary metadata with this
// job. This is opaque to Nomad.
Meta map[string]string
// Job status
Status string
// StatusDescription is meant to provide more human useful information
StatusDescription string
// Raft Indexes
CreateIndex uint64
ModifyIndex uint64
}
// ExpandAllServiceNames traverses all Task Groups and makes them
// interpolate Job, Task group and Task names in all Service names
func (j *Job) ExpandAllServiceNames() {
for _, tg := range j.TaskGroups {
tg.ExpandAllServiceNames(j.Name)
}
}
// Validate is used to sanity check a job input
func (j *Job) Validate() error {
var mErr multierror.Error
if j.Region == "" {
mErr.Errors = append(mErr.Errors, errors.New("Missing job region"))
}
if j.ID == "" {
mErr.Errors = append(mErr.Errors, errors.New("Missing job ID"))
} else if strings.Contains(j.ID, " ") {
mErr.Errors = append(mErr.Errors, errors.New("Job ID contains a space"))
}
if j.Name == "" {
mErr.Errors = append(mErr.Errors, errors.New("Missing job name"))
}
if j.Type == "" {
mErr.Errors = append(mErr.Errors, errors.New("Missing job type"))
}
if j.Priority < JobMinPriority || j.Priority > JobMaxPriority {
mErr.Errors = append(mErr.Errors, fmt.Errorf("Job priority must be between [%d, %d]", JobMinPriority, JobMaxPriority))
}
if len(j.Datacenters) == 0 {
mErr.Errors = append(mErr.Errors, errors.New("Missing job datacenters"))
}
if len(j.TaskGroups) == 0 {
mErr.Errors = append(mErr.Errors, errors.New("Missing job task groups"))
}
for idx, constr := range j.Constraints {
if err := constr.Validate(); err != nil {
outer := fmt.Errorf("Constraint %d validation failed: %s", idx+1, err)
mErr.Errors = append(mErr.Errors, outer)
}
}
// Check for duplicate task groups
taskGroups := make(map[string]int)
for idx, tg := range j.TaskGroups {
if tg.Name == "" {
mErr.Errors = append(mErr.Errors, fmt.Errorf("Job task group %d missing name", idx+1))
} else if existing, ok := taskGroups[tg.Name]; ok {
mErr.Errors = append(mErr.Errors, fmt.Errorf("Job task group %d redefines '%s' from group %d", idx+1, tg.Name, existing+1))
} else {
taskGroups[tg.Name] = idx
}
if j.Type == "system" && tg.Count != 1 {
mErr.Errors = append(mErr.Errors,
fmt.Errorf("Job task group %d has count %d. Only count of 1 is supported with system scheduler",
idx+1, tg.Count))
}
}
// Validate the task group
for idx, tg := range j.TaskGroups {
if err := tg.Validate(); err != nil {
outer := fmt.Errorf("Task group %d validation failed: %s", idx+1, err)
mErr.Errors = append(mErr.Errors, outer)
}
}
return mErr.ErrorOrNil()
}
// LookupTaskGroup finds a task group by name
func (j *Job) LookupTaskGroup(name string) *TaskGroup {
for _, tg := range j.TaskGroups {
if tg.Name == name {
return tg
}
}
return nil
}
// Stub is used to return a summary of the job
func (j *Job) Stub() *JobListStub {
return &JobListStub{
ID: j.ID,
Name: j.Name,
Type: j.Type,
Priority: j.Priority,
Status: j.Status,
StatusDescription: j.StatusDescription,
CreateIndex: j.CreateIndex,
ModifyIndex: j.ModifyIndex,
}
}
// JobListStub is used to return a subset of job information
// for the job list
type JobListStub struct {
ID string
Name string
Type string
Priority int
Status string
StatusDescription string
CreateIndex uint64
ModifyIndex uint64
}
// UpdateStrategy is used to modify how updates are done
type UpdateStrategy struct {
// Stagger is the amount of time between the updates
Stagger time.Duration
// MaxParallel is how many updates can be done in parallel
MaxParallel int `mapstructure:"max_parallel"`
}
// Rolling returns if a rolling strategy should be used
func (u *UpdateStrategy) Rolling() bool {
return u.Stagger > 0 && u.MaxParallel > 0
}
// RestartPolicy influences how Nomad restarts Tasks when they
// crash or fail.
type RestartPolicy struct {
Attempts int
Interval time.Duration
Delay time.Duration
}
func (r *RestartPolicy) Validate() error {
if r.Interval == 0 {
return nil
}
if time.Duration(r.Attempts)*r.Delay > r.Interval {
return fmt.Errorf("Nomad can't restart the TaskGroup %v times in an interval of %v with a delay of %v", r.Attempts, r.Interval, r.Delay)
}
return nil
}
func NewRestartPolicy(jobType string) *RestartPolicy {
switch jobType {
case JobTypeService, JobTypeSystem:
rp := defaultServiceJobRestartPolicy
return &rp
case JobTypeBatch:
rp := defaultBatchJobRestartPolicy
return &rp
}
return nil
}
// TaskGroup is an atomic unit of placement. Each task group belongs to
// a job and may contain any number of tasks. A task group support running
// in many replicas using the same configuration..
type TaskGroup struct {
// Name of the task group
Name string
// Count is the number of replicas of this task group that should
// be scheduled.
Count int
// Constraints can be specified at a task group level and apply to
// all the tasks contained.
Constraints []*Constraint
//RestartPolicy of a TaskGroup
RestartPolicy *RestartPolicy
// Tasks are the collection of tasks that this task group needs to run
Tasks []*Task
// Meta is used to associate arbitrary metadata with this
// task group. This is opaque to Nomad.
Meta map[string]string
}
// ExpandAllServiceNames traverses over all Tasks and makes them to interpolate
// values of Job, Task Group and Task names in all Service Names
func (tg *TaskGroup) ExpandAllServiceNames(job string) {
for _, task := range tg.Tasks {
task.ExpandAllServiceNames(job, tg.Name)
}
}
// Validate is used to sanity check a task group
func (tg *TaskGroup) Validate() error {
var mErr multierror.Error
if tg.Name == "" {
mErr.Errors = append(mErr.Errors, errors.New("Missing task group name"))
}
if tg.Count <= 0 {
mErr.Errors = append(mErr.Errors, errors.New("Task group count must be positive"))
}
if len(tg.Tasks) == 0 {
mErr.Errors = append(mErr.Errors, errors.New("Missing tasks for task group"))
}
for idx, constr := range tg.Constraints {
if err := constr.Validate(); err != nil {
outer := fmt.Errorf("Constraint %d validation failed: %s", idx+1, err)
mErr.Errors = append(mErr.Errors, outer)
}
}
if tg.RestartPolicy != nil {
if err := tg.RestartPolicy.Validate(); err != nil {
mErr.Errors = append(mErr.Errors, err)
}
} else {
mErr.Errors = append(mErr.Errors, fmt.Errorf("Task Group %v should have a restart policy", tg.Name))
}
// Check for duplicate tasks
tasks := make(map[string]int)
for idx, task := range tg.Tasks {
if task.Name == "" {
mErr.Errors = append(mErr.Errors, fmt.Errorf("Task %d missing name", idx+1))
} else if existing, ok := tasks[task.Name]; ok {
mErr.Errors = append(mErr.Errors, fmt.Errorf("Task %d redefines '%s' from task %d", idx+1, task.Name, existing+1))
} else {
tasks[task.Name] = idx
}
}