open-nomad/api/tasks.go

990 lines
26 KiB
Go

package api
import (
"fmt"
"path"
"path/filepath"
"strings"
"time"
)
const (
// RestartPolicyModeDelay causes an artificial delay till the next interval is
// reached when the specified attempts have been reached in the interval.
RestartPolicyModeDelay = "delay"
// RestartPolicyModeFail causes a job to fail if the specified number of
// attempts are reached within an interval.
RestartPolicyModeFail = "fail"
)
// MemoryStats holds memory usage related stats
type MemoryStats struct {
RSS uint64
Cache uint64
Swap uint64
Usage uint64
MaxUsage uint64
KernelUsage uint64
KernelMaxUsage uint64
Measured []string
}
// CpuStats holds cpu usage related stats
type CpuStats struct {
SystemMode float64
UserMode float64
TotalTicks float64
ThrottledPeriods uint64
ThrottledTime uint64
Percent float64
Measured []string
}
// ResourceUsage holds information related to cpu and memory stats
type ResourceUsage struct {
MemoryStats *MemoryStats
CpuStats *CpuStats
DeviceStats []*DeviceGroupStats
}
// TaskResourceUsage holds aggregated resource usage of all processes in a Task
// and the resource usage of the individual pids
type TaskResourceUsage struct {
ResourceUsage *ResourceUsage
Timestamp int64
Pids map[string]*ResourceUsage
}
// AllocResourceUsage holds the aggregated task resource usage of the
// allocation.
type AllocResourceUsage struct {
ResourceUsage *ResourceUsage
Tasks map[string]*TaskResourceUsage
Timestamp int64
}
// RestartPolicy defines how the Nomad client restarts
// tasks in a taskgroup when they fail
type RestartPolicy struct {
Interval *time.Duration
Attempts *int
Delay *time.Duration
Mode *string
}
func (r *RestartPolicy) Merge(rp *RestartPolicy) {
if rp.Interval != nil {
r.Interval = rp.Interval
}
if rp.Attempts != nil {
r.Attempts = rp.Attempts
}
if rp.Delay != nil {
r.Delay = rp.Delay
}
if rp.Mode != nil {
r.Mode = rp.Mode
}
}
// Reschedule configures how Tasks are rescheduled when they crash or fail.
type ReschedulePolicy struct {
// Attempts limits the number of rescheduling attempts that can occur in an interval.
Attempts *int `mapstructure:"attempts"`
// Interval is a duration in which we can limit the number of reschedule attempts.
Interval *time.Duration `mapstructure:"interval"`
// Delay is a minimum duration to wait between reschedule attempts.
// The delay function determines how much subsequent reschedule attempts are delayed by.
Delay *time.Duration `mapstructure:"delay"`
// DelayFunction determines how the delay progressively changes on subsequent reschedule
// attempts. Valid values are "exponential", "constant", and "fibonacci".
DelayFunction *string `mapstructure:"delay_function"`
// MaxDelay is an upper bound on the delay.
MaxDelay *time.Duration `mapstructure:"max_delay"`
// Unlimited allows rescheduling attempts until they succeed
Unlimited *bool `mapstructure:"unlimited"`
}
func (r *ReschedulePolicy) Merge(rp *ReschedulePolicy) {
if rp == nil {
return
}
if rp.Interval != nil {
r.Interval = rp.Interval
}
if rp.Attempts != nil {
r.Attempts = rp.Attempts
}
if rp.Delay != nil {
r.Delay = rp.Delay
}
if rp.DelayFunction != nil {
r.DelayFunction = rp.DelayFunction
}
if rp.MaxDelay != nil {
r.MaxDelay = rp.MaxDelay
}
if rp.Unlimited != nil {
r.Unlimited = rp.Unlimited
}
}
func (r *ReschedulePolicy) Canonicalize(jobType string) {
dp := NewDefaultReschedulePolicy(jobType)
if r.Interval == nil {
r.Interval = dp.Interval
}
if r.Attempts == nil {
r.Attempts = dp.Attempts
}
if r.Delay == nil {
r.Delay = dp.Delay
}
if r.DelayFunction == nil {
r.DelayFunction = dp.DelayFunction
}
if r.MaxDelay == nil {
r.MaxDelay = dp.MaxDelay
}
if r.Unlimited == nil {
r.Unlimited = dp.Unlimited
}
}
// Affinity is used to serialize task group affinities
type Affinity struct {
LTarget string // Left-hand target
RTarget string // Right-hand target
Operand string // Constraint operand (<=, <, =, !=, >, >=), set_contains_all, set_contains_any
Weight *int8 // Weight applied to nodes that match the affinity. Can be negative
}
func NewAffinity(LTarget string, Operand string, RTarget string, Weight int8) *Affinity {
return &Affinity{
LTarget: LTarget,
RTarget: RTarget,
Operand: Operand,
Weight: int8ToPtr(Weight),
}
}
func (a *Affinity) Canonicalize() {
if a.Weight == nil {
a.Weight = int8ToPtr(50)
}
}
func NewDefaultReschedulePolicy(jobType string) *ReschedulePolicy {
var dp *ReschedulePolicy
switch jobType {
case "service":
// This needs to be in sync with DefaultServiceJobReschedulePolicy
// in nomad/structs/structs.go
dp = &ReschedulePolicy{
Delay: timeToPtr(30 * time.Second),
DelayFunction: stringToPtr("exponential"),
MaxDelay: timeToPtr(1 * time.Hour),
Unlimited: boolToPtr(true),
Attempts: intToPtr(0),
Interval: timeToPtr(0),
}
case "batch":
// This needs to be in sync with DefaultBatchJobReschedulePolicy
// in nomad/structs/structs.go
dp = &ReschedulePolicy{
Attempts: intToPtr(1),
Interval: timeToPtr(24 * time.Hour),
Delay: timeToPtr(5 * time.Second),
DelayFunction: stringToPtr("constant"),
MaxDelay: timeToPtr(0),
Unlimited: boolToPtr(false),
}
case "system":
dp = &ReschedulePolicy{
Attempts: intToPtr(0),
Interval: timeToPtr(0),
Delay: timeToPtr(0),
DelayFunction: stringToPtr(""),
MaxDelay: timeToPtr(0),
Unlimited: boolToPtr(false),
}
default:
// GH-7203: it is possible an unknown job type is passed to this
// function and we need to ensure a non-nil object is returned so that
// the canonicalization runs without panicking.
dp = &ReschedulePolicy{
Attempts: intToPtr(0),
Interval: timeToPtr(0),
Delay: timeToPtr(0),
DelayFunction: stringToPtr(""),
MaxDelay: timeToPtr(0),
Unlimited: boolToPtr(false),
}
}
return dp
}
func (r *ReschedulePolicy) Copy() *ReschedulePolicy {
if r == nil {
return nil
}
nrp := new(ReschedulePolicy)
*nrp = *r
return nrp
}
func (p *ReschedulePolicy) String() string {
if p == nil {
return ""
}
if *p.Unlimited {
return fmt.Sprintf("unlimited with %v delay, max_delay = %v", *p.DelayFunction, *p.MaxDelay)
}
return fmt.Sprintf("%v in %v with %v delay, max_delay = %v", *p.Attempts, *p.Interval, *p.DelayFunction, *p.MaxDelay)
}
// Spread is used to serialize task group allocation spread preferences
type Spread struct {
Attribute string
Weight *int8
SpreadTarget []*SpreadTarget
}
// SpreadTarget is used to serialize target allocation spread percentages
type SpreadTarget struct {
Value string
Percent uint8
}
func NewSpreadTarget(value string, percent uint8) *SpreadTarget {
return &SpreadTarget{
Value: value,
Percent: percent,
}
}
func NewSpread(attribute string, weight int8, spreadTargets []*SpreadTarget) *Spread {
return &Spread{
Attribute: attribute,
Weight: int8ToPtr(weight),
SpreadTarget: spreadTargets,
}
}
func (s *Spread) Canonicalize() {
if s.Weight == nil {
s.Weight = int8ToPtr(50)
}
}
// EphemeralDisk is an ephemeral disk object
type EphemeralDisk struct {
Sticky *bool
Migrate *bool
SizeMB *int `mapstructure:"size"`
}
func DefaultEphemeralDisk() *EphemeralDisk {
return &EphemeralDisk{
Sticky: boolToPtr(false),
Migrate: boolToPtr(false),
SizeMB: intToPtr(300),
}
}
func (e *EphemeralDisk) Canonicalize() {
if e.Sticky == nil {
e.Sticky = boolToPtr(false)
}
if e.Migrate == nil {
e.Migrate = boolToPtr(false)
}
if e.SizeMB == nil {
e.SizeMB = intToPtr(300)
}
}
// MigrateStrategy describes how allocations for a task group should be
// migrated between nodes (eg when draining).
type MigrateStrategy struct {
MaxParallel *int `mapstructure:"max_parallel"`
HealthCheck *string `mapstructure:"health_check"`
MinHealthyTime *time.Duration `mapstructure:"min_healthy_time"`
HealthyDeadline *time.Duration `mapstructure:"healthy_deadline"`
}
func DefaultMigrateStrategy() *MigrateStrategy {
return &MigrateStrategy{
MaxParallel: intToPtr(1),
HealthCheck: stringToPtr("checks"),
MinHealthyTime: timeToPtr(10 * time.Second),
HealthyDeadline: timeToPtr(5 * time.Minute),
}
}
func (m *MigrateStrategy) Canonicalize() {
if m == nil {
return
}
defaults := DefaultMigrateStrategy()
if m.MaxParallel == nil {
m.MaxParallel = defaults.MaxParallel
}
if m.HealthCheck == nil {
m.HealthCheck = defaults.HealthCheck
}
if m.MinHealthyTime == nil {
m.MinHealthyTime = defaults.MinHealthyTime
}
if m.HealthyDeadline == nil {
m.HealthyDeadline = defaults.HealthyDeadline
}
}
func (m *MigrateStrategy) Merge(o *MigrateStrategy) {
if o.MaxParallel != nil {
m.MaxParallel = o.MaxParallel
}
if o.HealthCheck != nil {
m.HealthCheck = o.HealthCheck
}
if o.MinHealthyTime != nil {
m.MinHealthyTime = o.MinHealthyTime
}
if o.HealthyDeadline != nil {
m.HealthyDeadline = o.HealthyDeadline
}
}
func (m *MigrateStrategy) Copy() *MigrateStrategy {
if m == nil {
return nil
}
nm := new(MigrateStrategy)
*nm = *m
return nm
}
// VolumeRequest is a representation of a storage volume that a TaskGroup wishes to use.
type VolumeRequest struct {
Name string
Type string
Source string
ReadOnly bool `hcl:"read_only"`
MountOptions *CSIMountOptions `hcl:"mount_options"`
ExtraKeysHCL []string `hcl:",unusedKeys" json:"-"`
}
const (
VolumeMountPropagationPrivate = "private"
VolumeMountPropagationHostToTask = "host-to-task"
VolumeMountPropagationBidirectional = "bidirectional"
)
// VolumeMount represents the relationship between a destination path in a task
// and the task group volume that should be mounted there.
type VolumeMount struct {
Volume *string
Destination *string
ReadOnly *bool `mapstructure:"read_only"`
PropagationMode *string `mapstructure:"propagation_mode"`
}
func (vm *VolumeMount) Canonicalize() {
if vm.PropagationMode == nil {
vm.PropagationMode = stringToPtr(VolumeMountPropagationPrivate)
}
if vm.ReadOnly == nil {
vm.ReadOnly = boolToPtr(false)
}
}
// TaskGroup is the unit of scheduling.
type TaskGroup struct {
Name *string
Count *int
Constraints []*Constraint
Affinities []*Affinity
Tasks []*Task
Spreads []*Spread
Volumes map[string]*VolumeRequest
RestartPolicy *RestartPolicy
ReschedulePolicy *ReschedulePolicy
EphemeralDisk *EphemeralDisk
Update *UpdateStrategy
Migrate *MigrateStrategy
Networks []*NetworkResource
Meta map[string]string
Services []*Service
ShutdownDelay *time.Duration `mapstructure:"shutdown_delay"`
StopAfterClientDisconnect *time.Duration `mapstructure:"stop_after_client_disconnect"`
Scaling *ScalingPolicy
}
// NewTaskGroup creates a new TaskGroup.
func NewTaskGroup(name string, count int) *TaskGroup {
return &TaskGroup{
Name: stringToPtr(name),
Count: intToPtr(count),
}
}
// Canonicalize sets defaults and merges settings that should be inherited from the job
func (g *TaskGroup) Canonicalize(job *Job) {
if g.Name == nil {
g.Name = stringToPtr("")
}
if g.Count == nil {
if g.Scaling != nil && g.Scaling.Min != nil {
g.Count = intToPtr(int(*g.Scaling.Min))
} else {
g.Count = intToPtr(1)
}
}
if g.Scaling != nil {
g.Scaling.Canonicalize(*g.Count)
}
if g.EphemeralDisk == nil {
g.EphemeralDisk = DefaultEphemeralDisk()
} else {
g.EphemeralDisk.Canonicalize()
}
// Merge the update policy from the job
if ju, tu := job.Update != nil, g.Update != nil; ju && tu {
// Merge the jobs and task groups definition of the update strategy
jc := job.Update.Copy()
jc.Merge(g.Update)
g.Update = jc
} else if ju && !job.Update.Empty() {
// Inherit the jobs as long as it is non-empty.
jc := job.Update.Copy()
g.Update = jc
}
if g.Update != nil {
g.Update.Canonicalize()
}
// Merge the reschedule policy from the job
if jr, tr := job.Reschedule != nil, g.ReschedulePolicy != nil; jr && tr {
jobReschedule := job.Reschedule.Copy()
jobReschedule.Merge(g.ReschedulePolicy)
g.ReschedulePolicy = jobReschedule
} else if jr {
jobReschedule := job.Reschedule.Copy()
g.ReschedulePolicy = jobReschedule
}
// Only use default reschedule policy for non system jobs
if g.ReschedulePolicy == nil && *job.Type != "system" {
g.ReschedulePolicy = NewDefaultReschedulePolicy(*job.Type)
}
if g.ReschedulePolicy != nil {
g.ReschedulePolicy.Canonicalize(*job.Type)
}
// Merge the migrate strategy from the job
if jm, tm := job.Migrate != nil, g.Migrate != nil; jm && tm {
jobMigrate := job.Migrate.Copy()
jobMigrate.Merge(g.Migrate)
g.Migrate = jobMigrate
} else if jm {
jobMigrate := job.Migrate.Copy()
g.Migrate = jobMigrate
}
// Merge with default reschedule policy
if g.Migrate == nil && *job.Type == "service" {
g.Migrate = &MigrateStrategy{}
}
if g.Migrate != nil {
g.Migrate.Canonicalize()
}
var defaultRestartPolicy *RestartPolicy
switch *job.Type {
case "service", "system":
defaultRestartPolicy = defaultServiceJobRestartPolicy()
default:
defaultRestartPolicy = defaultBatchJobRestartPolicy()
}
if g.RestartPolicy != nil {
defaultRestartPolicy.Merge(g.RestartPolicy)
}
g.RestartPolicy = defaultRestartPolicy
for _, t := range g.Tasks {
t.Canonicalize(g, job)
}
for _, spread := range g.Spreads {
spread.Canonicalize()
}
for _, a := range g.Affinities {
a.Canonicalize()
}
for _, n := range g.Networks {
n.Canonicalize()
}
for _, s := range g.Services {
s.Canonicalize(nil, g, job)
}
}
// These needs to be in sync with DefaultServiceJobRestartPolicy in
// in nomad/structs/structs.go
func defaultServiceJobRestartPolicy() *RestartPolicy {
return &RestartPolicy{
Delay: timeToPtr(15 * time.Second),
Attempts: intToPtr(2),
Interval: timeToPtr(30 * time.Minute),
Mode: stringToPtr(RestartPolicyModeFail),
}
}
// These needs to be in sync with DefaultBatchJobRestartPolicy in
// in nomad/structs/structs.go
func defaultBatchJobRestartPolicy() *RestartPolicy {
return &RestartPolicy{
Delay: timeToPtr(15 * time.Second),
Attempts: intToPtr(3),
Interval: timeToPtr(24 * time.Hour),
Mode: stringToPtr(RestartPolicyModeFail),
}
}
// Constrain is used to add a constraint to a task group.
func (g *TaskGroup) Constrain(c *Constraint) *TaskGroup {
g.Constraints = append(g.Constraints, c)
return g
}
// AddMeta is used to add a meta k/v pair to a task group
func (g *TaskGroup) SetMeta(key, val string) *TaskGroup {
if g.Meta == nil {
g.Meta = make(map[string]string)
}
g.Meta[key] = val
return g
}
// AddTask is used to add a new task to a task group.
func (g *TaskGroup) AddTask(t *Task) *TaskGroup {
g.Tasks = append(g.Tasks, t)
return g
}
// AddAffinity is used to add a new affinity to a task group.
func (g *TaskGroup) AddAffinity(a *Affinity) *TaskGroup {
g.Affinities = append(g.Affinities, a)
return g
}
// RequireDisk adds a ephemeral disk to the task group
func (g *TaskGroup) RequireDisk(disk *EphemeralDisk) *TaskGroup {
g.EphemeralDisk = disk
return g
}
// AddSpread is used to add a new spread preference to a task group.
func (g *TaskGroup) AddSpread(s *Spread) *TaskGroup {
g.Spreads = append(g.Spreads, s)
return g
}
// LogConfig provides configuration for log rotation
type LogConfig struct {
MaxFiles *int `mapstructure:"max_files"`
MaxFileSizeMB *int `mapstructure:"max_file_size"`
}
func DefaultLogConfig() *LogConfig {
return &LogConfig{
MaxFiles: intToPtr(10),
MaxFileSizeMB: intToPtr(10),
}
}
func (l *LogConfig) Canonicalize() {
if l.MaxFiles == nil {
l.MaxFiles = intToPtr(10)
}
if l.MaxFileSizeMB == nil {
l.MaxFileSizeMB = intToPtr(10)
}
}
// DispatchPayloadConfig configures how a task gets its input from a job dispatch
type DispatchPayloadConfig struct {
File string
}
const (
TaskLifecycleHookPrestart = "prestart"
)
type TaskLifecycle struct {
Hook string `mapstructure:"hook"`
Sidecar bool `mapstructure:"sidecar"`
}
// Determine if lifecycle has user-input values
func (l *TaskLifecycle) Empty() bool {
return l == nil || (l.Hook == "")
}
// Task is a single process in a task group.
type Task struct {
Name string
Driver string
User string
Lifecycle *TaskLifecycle
Config map[string]interface{}
Constraints []*Constraint
Affinities []*Affinity
Env map[string]string
Services []*Service
Resources *Resources
RestartPolicy *RestartPolicy
Meta map[string]string
KillTimeout *time.Duration `mapstructure:"kill_timeout"`
LogConfig *LogConfig `mapstructure:"logs"`
Artifacts []*TaskArtifact
Vault *Vault
Templates []*Template
DispatchPayload *DispatchPayloadConfig
VolumeMounts []*VolumeMount
CSIPluginConfig *TaskCSIPluginConfig `mapstructure:"csi_plugin" json:",omitempty"`
Leader bool
ShutdownDelay time.Duration `mapstructure:"shutdown_delay"`
KillSignal string `mapstructure:"kill_signal"`
Kind string
}
func (t *Task) Canonicalize(tg *TaskGroup, job *Job) {
if t.Resources == nil {
t.Resources = &Resources{}
}
t.Resources.Canonicalize()
if t.KillTimeout == nil {
t.KillTimeout = timeToPtr(5 * time.Second)
}
if t.LogConfig == nil {
t.LogConfig = DefaultLogConfig()
} else {
t.LogConfig.Canonicalize()
}
for _, artifact := range t.Artifacts {
artifact.Canonicalize()
}
if t.Vault != nil {
t.Vault.Canonicalize()
}
for _, tmpl := range t.Templates {
tmpl.Canonicalize()
}
for _, s := range t.Services {
s.Canonicalize(t, tg, job)
}
for _, a := range t.Affinities {
a.Canonicalize()
}
for _, vm := range t.VolumeMounts {
vm.Canonicalize()
}
if t.Lifecycle.Empty() {
t.Lifecycle = nil
}
if t.CSIPluginConfig != nil {
t.CSIPluginConfig.Canonicalize()
}
if t.RestartPolicy == nil {
t.RestartPolicy = tg.RestartPolicy
} else {
tgrp := &RestartPolicy{}
*tgrp = *tg.RestartPolicy
tgrp.Merge(t.RestartPolicy)
t.RestartPolicy = tgrp
}
}
// TaskArtifact is used to download artifacts before running a task.
type TaskArtifact struct {
GetterSource *string `mapstructure:"source"`
GetterOptions map[string]string `mapstructure:"options"`
GetterMode *string `mapstructure:"mode"`
RelativeDest *string `mapstructure:"destination"`
}
func (a *TaskArtifact) Canonicalize() {
if a.GetterMode == nil {
a.GetterMode = stringToPtr("any")
}
if a.GetterSource == nil {
// Shouldn't be possible, but we don't want to panic
a.GetterSource = stringToPtr("")
}
if a.RelativeDest == nil {
switch *a.GetterMode {
case "file":
// File mode should default to local/filename
dest := *a.GetterSource
dest = path.Base(dest)
dest = filepath.Join("local", dest)
a.RelativeDest = &dest
default:
// Default to a directory
a.RelativeDest = stringToPtr("local/")
}
}
}
type Template struct {
SourcePath *string `mapstructure:"source"`
DestPath *string `mapstructure:"destination"`
EmbeddedTmpl *string `mapstructure:"data"`
ChangeMode *string `mapstructure:"change_mode"`
ChangeSignal *string `mapstructure:"change_signal"`
Splay *time.Duration `mapstructure:"splay"`
Perms *string `mapstructure:"perms"`
LeftDelim *string `mapstructure:"left_delimiter"`
RightDelim *string `mapstructure:"right_delimiter"`
Envvars *bool `mapstructure:"env"`
VaultGrace *time.Duration `mapstructure:"vault_grace"`
}
func (tmpl *Template) Canonicalize() {
if tmpl.SourcePath == nil {
tmpl.SourcePath = stringToPtr("")
}
if tmpl.DestPath == nil {
tmpl.DestPath = stringToPtr("")
}
if tmpl.EmbeddedTmpl == nil {
tmpl.EmbeddedTmpl = stringToPtr("")
}
if tmpl.ChangeMode == nil {
tmpl.ChangeMode = stringToPtr("restart")
}
if tmpl.ChangeSignal == nil {
if *tmpl.ChangeMode == "signal" {
tmpl.ChangeSignal = stringToPtr("SIGHUP")
} else {
tmpl.ChangeSignal = stringToPtr("")
}
} else {
sig := *tmpl.ChangeSignal
tmpl.ChangeSignal = stringToPtr(strings.ToUpper(sig))
}
if tmpl.Splay == nil {
tmpl.Splay = timeToPtr(5 * time.Second)
}
if tmpl.Perms == nil {
tmpl.Perms = stringToPtr("0644")
}
if tmpl.LeftDelim == nil {
tmpl.LeftDelim = stringToPtr("{{")
}
if tmpl.RightDelim == nil {
tmpl.RightDelim = stringToPtr("}}")
}
if tmpl.Envvars == nil {
tmpl.Envvars = boolToPtr(false)
}
//COMPAT(0.12) VaultGrace is deprecated and unused as of Vault 0.5
if tmpl.VaultGrace == nil {
tmpl.VaultGrace = timeToPtr(0)
}
}
type Vault struct {
Policies []string
Env *bool
ChangeMode *string `mapstructure:"change_mode"`
ChangeSignal *string `mapstructure:"change_signal"`
}
func (v *Vault) Canonicalize() {
if v.Env == nil {
v.Env = boolToPtr(true)
}
if v.ChangeMode == nil {
v.ChangeMode = stringToPtr("restart")
}
if v.ChangeSignal == nil {
v.ChangeSignal = stringToPtr("SIGHUP")
}
}
// NewTask creates and initializes a new Task.
func NewTask(name, driver string) *Task {
return &Task{
Name: name,
Driver: driver,
}
}
// Configure is used to configure a single k/v pair on
// the task.
func (t *Task) SetConfig(key string, val interface{}) *Task {
if t.Config == nil {
t.Config = make(map[string]interface{})
}
t.Config[key] = val
return t
}
// SetMeta is used to add metadata k/v pairs to the task.
func (t *Task) SetMeta(key, val string) *Task {
if t.Meta == nil {
t.Meta = make(map[string]string)
}
t.Meta[key] = val
return t
}
// Require is used to add resource requirements to a task.
func (t *Task) Require(r *Resources) *Task {
t.Resources = r
return t
}
// Constraint adds a new constraints to a single task.
func (t *Task) Constrain(c *Constraint) *Task {
t.Constraints = append(t.Constraints, c)
return t
}
// AddAffinity adds a new affinity to a single task.
func (t *Task) AddAffinity(a *Affinity) *Task {
t.Affinities = append(t.Affinities, a)
return t
}
// SetLogConfig sets a log config to a task
func (t *Task) SetLogConfig(l *LogConfig) *Task {
t.LogConfig = l
return t
}
// TaskState tracks the current state of a task and events that caused state
// transitions.
type TaskState struct {
State string
Failed bool
Restarts uint64
LastRestart time.Time
StartedAt time.Time
FinishedAt time.Time
Events []*TaskEvent
}
const (
TaskSetup = "Task Setup"
TaskSetupFailure = "Setup Failure"
TaskDriverFailure = "Driver Failure"
TaskDriverMessage = "Driver"
TaskReceived = "Received"
TaskFailedValidation = "Failed Validation"
TaskStarted = "Started"
TaskTerminated = "Terminated"
TaskKilling = "Killing"
TaskKilled = "Killed"
TaskRestarting = "Restarting"
TaskNotRestarting = "Not Restarting"
TaskDownloadingArtifacts = "Downloading Artifacts"
TaskArtifactDownloadFailed = "Failed Artifact Download"
TaskSiblingFailed = "Sibling Task Failed"
TaskSignaling = "Signaling"
TaskRestartSignal = "Restart Signaled"
TaskLeaderDead = "Leader Task Dead"
TaskBuildingTaskDir = "Building Task Directory"
)
// TaskEvent is an event that effects the state of a task and contains meta-data
// appropriate to the events type.
type TaskEvent struct {
Type string
Time int64
DisplayMessage string
Details map[string]string
// DEPRECATION NOTICE: The following fields are all deprecated. see TaskEvent struct in structs.go for details.
FailsTask bool
RestartReason string
SetupError string
DriverError string
DriverMessage string
ExitCode int
Signal int
Message string
KillReason string
KillTimeout time.Duration
KillError string
StartDelay int64
DownloadError string
ValidationError string
DiskLimit int64
DiskSize int64
FailedSibling string
VaultError string
TaskSignalReason string
TaskSignal string
GenericSource string
}
// CSIPluginType is an enum string that encapsulates the valid options for a
// CSIPlugin stanza's Type. These modes will allow the plugin to be used in
// different ways by the client.
type CSIPluginType string
const (
// CSIPluginTypeNode indicates that Nomad should only use the plugin for
// performing Node RPCs against the provided plugin.
CSIPluginTypeNode CSIPluginType = "node"
// CSIPluginTypeController indicates that Nomad should only use the plugin for
// performing Controller RPCs against the provided plugin.
CSIPluginTypeController CSIPluginType = "controller"
// CSIPluginTypeMonolith indicates that Nomad can use the provided plugin for
// both controller and node rpcs.
CSIPluginTypeMonolith CSIPluginType = "monolith"
)
// TaskCSIPluginConfig contains the data that is required to setup a task as a
// CSI plugin. This will be used by the csi_plugin_supervisor_hook to configure
// mounts for the plugin and initiate the connection to the plugin catalog.
type TaskCSIPluginConfig struct {
// ID is the identifier of the plugin.
// Ideally this should be the FQDN of the plugin.
ID string `mapstructure:"id"`
// CSIPluginType instructs Nomad on how to handle processing a plugin
Type CSIPluginType `mapstructure:"type"`
// MountDir is the destination that nomad should mount in its CSI
// directory for the plugin. It will then expect a file called CSISocketName
// to be created by the plugin, and will provide references into
// "MountDir/CSIIntermediaryDirname/VolumeName/AllocID for mounts.
//
// Default is /csi.
MountDir string `mapstructure:"mount_dir"`
}
func (t *TaskCSIPluginConfig) Canonicalize() {
if t.MountDir == "" {
t.MountDir = "/csi"
}
}