open-nomad/nomad/job_endpoint_hooks.go

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

321 lines
9.7 KiB
Go
Raw Normal View History

package nomad
import (
"fmt"
"github.com/hashicorp/go-multierror"
"github.com/hashicorp/nomad/helper"
"github.com/hashicorp/nomad/nomad/structs"
)
const (
attrVaultVersion = `${attr.vault.version}`
attrConsulVersion = `${attr.consul.version}`
attrNomadVersion = `${attr.nomad.version}`
attrNomadServiceDisco = `${attr.nomad.service_discovery}`
)
var (
// vaultConstraint is the implicit constraint added to jobs requesting a
// Vault token
vaultConstraint = &structs.Constraint{
LTarget: attrVaultVersion,
RTarget: ">= 0.6.1",
Operand: structs.ConstraintSemver,
}
// consulServiceDiscoveryConstraint is the implicit constraint added to
// task groups which include services utilising the Consul provider. The
// Consul version is pinned to a minimum of that which introduced the
// namespace feature.
consulServiceDiscoveryConstraint = &structs.Constraint{
LTarget: attrConsulVersion,
RTarget: ">= 1.7.0",
Operand: structs.ConstraintSemver,
}
// nativeServiceDiscoveryConstraint is the constraint injected into task
// groups that utilise Nomad's native service discovery feature. This is
// needed, as operators can disable the client functionality, and therefore
// we need to ensure task groups are placed where they can run
// successfully.
nativeServiceDiscoveryConstraint = &structs.Constraint{
LTarget: attrNomadServiceDisco,
RTarget: "true",
Operand: "=",
}
// nativeServiceDiscoveryChecksConstraint is the constraint injected into task
// groups that utilize Nomad's native service discovery checks feature. This
// is needed, as operators can have versions of Nomad pre-v1.4 mixed into a
// cluster with v1.4 servers, causing jobs to be placed on incompatible
// clients.
nativeServiceDiscoveryChecksConstraint = &structs.Constraint{
LTarget: attrNomadVersion,
RTarget: ">= 1.4.0",
Operand: structs.ConstraintSemver,
}
)
type admissionController interface {
Name() string
}
type jobMutator interface {
admissionController
Mutate(*structs.Job) (out *structs.Job, warnings []error, err error)
}
type jobValidator interface {
admissionController
Validate(*structs.Job) (warnings []error, err error)
}
func (j *Job) admissionControllers(job *structs.Job) (out *structs.Job, warnings []error, err error) {
// Mutators run first before validators, so validators view the final rendered job.
// So, mutators must handle invalid jobs.
out, warnings, err = j.admissionMutators(job)
if err != nil {
return nil, nil, err
}
validateWarnings, err := j.admissionValidators(job)
if err != nil {
return nil, nil, err
}
warnings = append(warnings, validateWarnings...)
return out, warnings, nil
}
// admissionMutator returns an updated job as well as warnings or an error.
func (j *Job) admissionMutators(job *structs.Job) (_ *structs.Job, warnings []error, err error) {
var w []error
for _, mutator := range j.mutators {
job, w, err = mutator.Mutate(job)
j.logger.Trace("job mutate results", "mutator", mutator.Name(), "warnings", w, "error", err)
if err != nil {
return nil, nil, fmt.Errorf("error in job mutator %s: %v", mutator.Name(), err)
}
warnings = append(warnings, w...)
}
return job, warnings, err
}
// admissionValidators returns a slice of validation warnings and a multierror
// of validation failures.
func (j *Job) admissionValidators(origJob *structs.Job) ([]error, error) {
// ensure job is not mutated
job := origJob.Copy()
var warnings []error
var errs error
for _, validator := range j.validators {
w, err := validator.Validate(job)
j.logger.Trace("job validate results", "validator", validator.Name(), "warnings", w, "error", err)
if err != nil {
errs = multierror.Append(errs, err)
}
warnings = append(warnings, w...)
}
return warnings, errs
}
// jobCanonicalizer calls job.Canonicalize (sets defaults and initializes
// fields) and returns any errors as warnings.
type jobCanonicalizer struct{}
func (jobCanonicalizer) Name() string {
return "canonicalize"
}
func (jobCanonicalizer) Mutate(job *structs.Job) (*structs.Job, []error, error) {
job.Canonicalize()
return job, nil, nil
}
// jobImpliedConstraints adds constraints to a job implied by other job fields
// and blocks.
type jobImpliedConstraints struct{}
func (jobImpliedConstraints) Name() string {
return "constraints"
}
func (jobImpliedConstraints) Mutate(j *structs.Job) (*structs.Job, []error, error) {
// Get the Vault blocks in the job
vaultBlocks := j.Vault()
// Get the required signals
signals := j.RequiredSignals()
// Identify which task groups are utilising Nomad native service discovery.
nativeServiceDisco := j.RequiredNativeServiceDiscovery()
// Identify which task groups are utilising Consul service discovery.
consulServiceDisco := j.RequiredConsulServiceDiscovery()
// Hot path
if len(signals) == 0 && len(vaultBlocks) == 0 &&
nativeServiceDisco.Empty() && len(consulServiceDisco) == 0 {
return j, nil, nil
}
// Iterate through all the task groups within the job and add any required
// constraints. When adding new implicit constraints, they should go inside
// this single loop, with a new constraintMatcher if needed.
for _, tg := range j.TaskGroups {
// If the task group utilises Vault, run the mutator.
if _, ok := vaultBlocks[tg.Name]; ok {
mutateConstraint(constraintMatcherLeft, tg, vaultConstraint)
}
// Check whether the task group is using signals. In the case that it
// is, we flatten the signals and build a constraint, then run the
// mutator.
if tgSignals, ok := signals[tg.Name]; ok {
required := helper.UniqueMapSliceValues(tgSignals)
sigConstraint := getSignalConstraint(required)
mutateConstraint(constraintMatcherFull, tg, sigConstraint)
}
// If the task group utilises Nomad service discovery, run the mutator.
if nativeServiceDisco.Basic.Contains(tg.Name) {
mutateConstraint(constraintMatcherFull, tg, nativeServiceDiscoveryConstraint)
}
// If the task group utilizes NSD checks, run the mutator.
if nativeServiceDisco.Checks.Contains(tg.Name) {
mutateConstraint(constraintMatcherFull, tg, nativeServiceDiscoveryChecksConstraint)
}
// If the task group utilises Consul service discovery, run the mutator.
if ok := consulServiceDisco[tg.Name]; ok {
mutateConstraint(constraintMatcherLeft, tg, consulServiceDiscoveryConstraint)
}
}
return j, nil, nil
}
// constraintMatcher is a custom type which helps control how constraints are
// identified as being present within a task group.
type constraintMatcher uint
const (
// constraintMatcherFull ensures that a constraint is only considered found
// when they match totally. This check is performed using the
// structs.Constraint Equal function.
constraintMatcherFull constraintMatcher = iota
// constraintMatcherLeft ensure that a constraint is considered found if
// the constraints LTarget is matched only. This allows an existing
// constraint to override the proposed implicit one.
constraintMatcherLeft
)
// mutateConstraint is a generic mutator used to set implicit constraints
// within the task group if they are needed.
func mutateConstraint(matcher constraintMatcher, taskGroup *structs.TaskGroup, constraint *structs.Constraint) {
var found bool
// It's possible to switch on the matcher within the constraint loop to
// reduce repetition. This, however, means switching per constraint,
// therefore we do it here.
switch matcher {
case constraintMatcherFull:
for _, c := range taskGroup.Constraints {
if c.Equal(constraint) {
found = true
break
}
}
case constraintMatcherLeft:
for _, c := range taskGroup.Constraints {
if c.LTarget == constraint.LTarget {
found = true
break
}
}
}
// If we didn't find a suitable constraint match, add one.
if !found {
taskGroup.Constraints = append(taskGroup.Constraints, constraint)
}
}
// jobValidate validates a Job and task drivers and returns an error if there is
// a validation problem or if the Job is of a type a user is not allowed to
// submit.
type jobValidate struct{}
func (jobValidate) Name() string {
return "validate"
}
func (jobValidate) Validate(job *structs.Job) (warnings []error, err error) {
validationErrors := new(multierror.Error)
if err := job.Validate(); err != nil {
multierror.Append(validationErrors, err)
}
// Get any warnings
jobWarnings := job.Warnings()
if jobWarnings != nil {
if multi, ok := jobWarnings.(*multierror.Error); ok {
// Unpack multiple warnings
warnings = append(warnings, multi.Errors...)
} else {
warnings = append(warnings, jobWarnings)
}
}
// TODO: Validate the driver configurations. These had to be removed in 0.9
// to support driver plugins, but see issue: #XXXX for more info.
if job.Type == structs.JobTypeCore {
multierror.Append(validationErrors, fmt.Errorf("job type cannot be core"))
}
if len(job.Payload) != 0 {
multierror.Append(validationErrors, fmt.Errorf("job can't be submitted with a payload, only dispatched"))
}
return warnings, validationErrors.ErrorOrNil()
}
Allow configuring memory oversubscription (#10466) Cluster operators want to have better control over memory oversubscription and may want to enable/disable it based on their experience. This PR adds a scheduler configuration field to control memory oversubscription. It's additional field that can be set in the [API via Scheduler Config](https://www.nomadproject.io/api-docs/operator/scheduler), or [the agent server config](https://www.nomadproject.io/docs/configuration/server#configuring-scheduler-config). I opted to have the memory oversubscription be an opt-in, but happy to change it. To enable it, operators should call the API with: ```json { "MemoryOversubscriptionEnabled": true } ``` If memory oversubscription is disabled, submitting jobs specifying `memory_max` will get a "Memory oversubscription is not enabled" warnings, but the jobs will be accepted without them accessing the additional memory. The warning message is like: ``` $ nomad job run /tmp/j Job Warnings: 1 warning(s): * Memory oversubscription is not enabled; Task cache.redis memory_max value will be ignored ==> Monitoring evaluation "7c444157" Evaluation triggered by job "example" ==> Monitoring evaluation "7c444157" Evaluation within deployment: "9d826f13" Allocation "aa5c3cad" created: node "9272088e", group "cache" Evaluation status changed: "pending" -> "complete" ==> Evaluation "7c444157" finished with status "complete" # then you can examine the Alloc AllocatedResources to validate whether the task is allowed to exceed memory: $ nomad alloc status -json aa5c3cad | jq '.AllocatedResources.Tasks["redis"].Memory' { "MemoryMB": 256, "MemoryMaxMB": 0 } ```
2021-04-30 02:09:56 +00:00
type memoryOversubscriptionValidate struct {
srv *Server
}
func (*memoryOversubscriptionValidate) Name() string {
return "memory_oversubscription"
}
func (v *memoryOversubscriptionValidate) Validate(job *structs.Job) (warnings []error, err error) {
_, c, err := v.srv.State().SchedulerConfig()
if err != nil {
return nil, err
}
if c != nil && c.MemoryOversubscriptionEnabled {
return nil, nil
}
for _, tg := range job.TaskGroups {
for _, t := range tg.Tasks {
if t.Resources != nil && t.Resources.MemoryMaxMB != 0 {
warnings = append(warnings, fmt.Errorf("Memory oversubscription is not enabled; Task \"%v.%v\" memory_max value will be ignored. Update the Scheduler Configuration to allow oversubscription.", tg.Name, t.Name))
}
}
}
return warnings, err
}