trainer/pkg/runtime/core/trainingruntime_test.go

481 lines
16 KiB
Go

/*
Copyright 2024 The Kubeflow Authors.
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 core
import (
"context"
"fmt"
"testing"
"github.com/google/go-cmp/cmp"
"github.com/google/go-cmp/cmp/cmpopts"
corev1 "k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/api/resource"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/runtime"
"k8s.io/apimachinery/pkg/util/intstr"
schedulerpluginsv1alpha1 "sigs.k8s.io/scheduler-plugins/apis/scheduling/v1alpha1"
trainer "github.com/kubeflow/trainer/pkg/apis/trainer/v1alpha1"
"github.com/kubeflow/trainer/pkg/constants"
jobsetplugin "github.com/kubeflow/trainer/pkg/runtime/framework/plugins/jobset"
testingutil "github.com/kubeflow/trainer/pkg/util/testing"
)
func TestTrainingRuntimeNewObjects(t *testing.T) {
resRequests := corev1.ResourceList{
corev1.ResourceCPU: resource.MustParse("1"),
}
// TODO (andreyvelich): Add more test cases.
cases := map[string]struct {
trainingRuntime *trainer.TrainingRuntime
trainJob *trainer.TrainJob
wantObjs []runtime.Object
wantError error
}{
// Test cases for the PlainML MLPolicy.
"succeeded to build PodGroup and JobSet with NumNodes from the TrainJob and container from the Runtime.": {
trainingRuntime: testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").
Label("conflictLabel", "overridden").
Annotation("conflictAnnotation", "overridden").
RuntimeSpec(
testingutil.MakeTrainingRuntimeSpecWrapper(testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").Spec).
ContainerDatasetModelInitializer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
WithMLPolicy(
testingutil.MakeMLPolicyWrapper().
WithNumNodes(100).
Obj(),
).
ContainerTrainer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
PodGroupPolicyCoschedulingSchedulingTimeout(120).
Obj(),
).Obj(),
trainJob: testingutil.MakeTrainJobWrapper(metav1.NamespaceDefault, "test-job").
Suspend(true).
UID("uid").
RuntimeRef(trainer.SchemeGroupVersion.WithKind(trainer.TrainingRuntimeKind), "test-runtime").
SpecLabel("conflictLabel", "override").
SpecAnnotation("conflictAnnotation", "override").
Trainer(
testingutil.MakeTrainJobTrainerWrapper().
NumNodes(30).
Obj(),
).
Obj(),
wantObjs: []runtime.Object{
testingutil.MakeJobSetWrapper(metav1.NamespaceDefault, "test-job").
ContainerDatasetModelInitializer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
NumNodes(30).
ContainerTrainer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
Suspend(true).
Label("conflictLabel", "override").
Annotation("conflictAnnotation", "override").
PodLabel(schedulerpluginsv1alpha1.PodGroupLabel, "test-job").
ControllerReference(trainer.SchemeGroupVersion.WithKind(trainer.TrainJobKind), "test-job", "uid").
Obj(),
testingutil.MakeSchedulerPluginsPodGroup(metav1.NamespaceDefault, "test-job").
ControllerReference(trainer.SchemeGroupVersion.WithKind(trainer.TrainJobKind), "test-job", "uid").
MinMember(31). // 31 replicas = 30 Trainer nodes + 1 Initializer.
MinResources(corev1.ResourceList{
// Trainer node has 30 CPUs + 2 CPUs from 2 initializer containers.
corev1.ResourceCPU: resource.MustParse("32"),
}).
SchedulingTimeout(120).
Obj(),
},
},
"succeeded to build JobSet with NumNodes from the Runtime and container from the TrainJob.": {
trainingRuntime: testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").RuntimeSpec(
testingutil.MakeTrainingRuntimeSpecWrapper(testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").Spec).
WithMLPolicy(
testingutil.MakeMLPolicyWrapper().
WithNumNodes(100).
Obj(),
).
ContainerTrainer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
ContainerTrainerEnv(
[]corev1.EnvVar{
{
Name: "TRAIN_JOB",
Value: "original",
},
{
Name: "RUNTIME",
Value: "test:runtime",
},
},
).
Obj(),
).Obj(),
trainJob: testingutil.MakeTrainJobWrapper(metav1.NamespaceDefault, "test-job").
UID("uid").
RuntimeRef(trainer.SchemeGroupVersion.WithKind(trainer.TrainingRuntimeKind), "test-runtime").
Trainer(
testingutil.MakeTrainJobTrainerWrapper().
Container("test:trainjob", []string{"trainjob"}, []string{"trainjob"}, resRequests).
ContainerEnv(
[]corev1.EnvVar{
{
Name: "TRAIN_JOB",
Value: "override",
},
{
Name: "TRAIN_JOB_CUSTOM",
Value: "test:trainjob",
},
}...,
).
Obj(),
).
Obj(),
wantObjs: []runtime.Object{
testingutil.MakeJobSetWrapper(metav1.NamespaceDefault, "test-job").
NumNodes(100).
ContainerTrainer("test:trainjob", []string{"trainjob"}, []string{"trainjob"}, resRequests).
ContainerTrainerEnv(
[]corev1.EnvVar{
{
Name: "TRAIN_JOB",
Value: "override",
},
{
Name: "TRAIN_JOB_CUSTOM",
Value: "test:trainjob",
},
{
Name: "RUNTIME",
Value: "test:runtime",
},
},
).
ControllerReference(trainer.SchemeGroupVersion.WithKind(trainer.TrainJobKind), "test-job", "uid").
Obj(),
},
},
"succeeded to build JobSet with dataset and model initializer from the TrainJob.": {
trainingRuntime: testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").RuntimeSpec(
testingutil.MakeTrainingRuntimeSpecWrapper(testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").Spec).
ContainerDatasetModelInitializer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
WithMLPolicy(
testingutil.MakeMLPolicyWrapper().
WithNumNodes(100).
Obj(),
).
ContainerTrainer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
Obj(),
).Obj(),
trainJob: testingutil.MakeTrainJobWrapper(metav1.NamespaceDefault, "test-job").
UID("uid").
RuntimeRef(trainer.SchemeGroupVersion.WithKind(trainer.TrainingRuntimeKind), "test-runtime").
Trainer(
testingutil.MakeTrainJobTrainerWrapper().
Obj(),
).
DatasetConfig(
testingutil.MakeTrainJobDatasetConfigWrapper().
StorageUri("hf://trainjob-dataset").
ContainerEnv(
[]corev1.EnvVar{
{
Name: "TRAIN_JOB",
Value: "test:trainjob:dataset",
},
},
).
SecretRef(corev1.LocalObjectReference{Name: "trainjob-secret-dataset"}).
Obj(),
).
ModelConfig(
testingutil.MakeTrainJobModelConfigWrapper().
StorageUri("hf://trainjob-model").
ContainerEnv(
[]corev1.EnvVar{
{
Name: "TRAIN_JOB",
Value: "test:trainjob:model",
},
},
).
SecretRef(corev1.LocalObjectReference{Name: "trainjob-secret-model"}).
Obj(),
).
Obj(),
wantObjs: []runtime.Object{
testingutil.MakeJobSetWrapper(metav1.NamespaceDefault, "test-job").
NumNodes(100).
ContainerTrainer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
ContainerDatasetModelInitializer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
ContainerDatasetInitializerEnv(
[]corev1.EnvVar{
{
Name: jobsetplugin.InitializerEnvStorageUri,
Value: "hf://trainjob-dataset",
},
{
Name: "TRAIN_JOB",
Value: "test:trainjob:dataset",
},
},
).
ContainerDatasetInitializerEnvFrom(
[]corev1.EnvFromSource{
{
SecretRef: &corev1.SecretEnvSource{
LocalObjectReference: corev1.LocalObjectReference{
Name: "trainjob-secret-dataset",
},
},
},
},
).
ContainerModelInitializerEnv(
[]corev1.EnvVar{
{
Name: jobsetplugin.InitializerEnvStorageUri,
Value: "hf://trainjob-model",
},
{
Name: "TRAIN_JOB",
Value: "test:trainjob:model",
},
},
).
ContainerModelInitializerEnvFrom(
[]corev1.EnvFromSource{
{
SecretRef: &corev1.SecretEnvSource{
LocalObjectReference: corev1.LocalObjectReference{
Name: "trainjob-secret-model",
},
},
},
},
).
ControllerReference(trainer.SchemeGroupVersion.WithKind(trainer.TrainJobKind), "test-job", "uid").
Obj(),
},
},
// Test cases for the Torch MLPolicy.
"succeeded to build JobSet with Torch values from the TrainJob": {
trainingRuntime: testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").RuntimeSpec(
testingutil.MakeTrainingRuntimeSpecWrapper(testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").Spec).
WithMLPolicy(
testingutil.MakeMLPolicyWrapper().
WithNumNodes(100).
TorchPolicy("auto", nil).
Obj(),
).
ContainerTrainer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
Obj(),
).Obj(),
trainJob: testingutil.MakeTrainJobWrapper(metav1.NamespaceDefault, "test-job").
UID("uid").
RuntimeRef(trainer.SchemeGroupVersion.WithKind(trainer.TrainingRuntimeKind), "test-runtime").
Trainer(
testingutil.MakeTrainJobTrainerWrapper().
NumNodes(30).
NumProcPerNode(intstr.FromInt32(3)).
Obj(),
).
Obj(),
wantObjs: []runtime.Object{
testingutil.MakeJobSetWrapper(metav1.NamespaceDefault, "test-job").
NumNodes(30).
ContainerTrainer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
ContainerTrainerPorts([]corev1.ContainerPort{{ContainerPort: constants.ContainerTrainerPort}}).
ContainerTrainerEnv(
[]corev1.EnvVar{
{
Name: constants.TorchEnvNumNodes,
Value: "30",
},
{
Name: constants.TorchEnvNumProcPerNode,
Value: "3",
},
{
Name: constants.TorchEnvNodeRank,
ValueFrom: &corev1.EnvVarSource{
FieldRef: &corev1.ObjectFieldSelector{
FieldPath: constants.JobCompletionIndexFieldPath,
},
},
},
{
Name: constants.TorchEnvMasterAddr,
Value: fmt.Sprintf("test-job-%s-0-0.test-job", constants.JobTrainerNode),
},
{
Name: constants.TorchEnvMasterPort,
Value: fmt.Sprintf("%d", constants.ContainerTrainerPort),
},
},
).
ControllerReference(trainer.SchemeGroupVersion.WithKind(trainer.TrainJobKind), "test-job", "uid").
Obj(),
},
},
"succeeded to build JobSet with Torch values from the Runtime and envs.": {
trainingRuntime: testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").RuntimeSpec(
testingutil.MakeTrainingRuntimeSpecWrapper(testingutil.MakeTrainingRuntimeWrapper(metav1.NamespaceDefault, "test-runtime").Spec).
WithMLPolicy(
testingutil.MakeMLPolicyWrapper().
WithNumNodes(100).
TorchPolicy("auto", nil).
Obj(),
).
ContainerTrainer("test:runtime", []string{"runtime"}, []string{"runtime"}, resRequests).
ContainerTrainerEnv(
[]corev1.EnvVar{
{
Name: "TRAIN_JOB",
Value: "original",
},
{
Name: "RUNTIME",
Value: "test:runtime",
},
},
).
Obj(),
).Obj(),
trainJob: testingutil.MakeTrainJobWrapper(metav1.NamespaceDefault, "test-job").
UID("uid").
RuntimeRef(trainer.SchemeGroupVersion.WithKind(trainer.TrainingRuntimeKind), "test-runtime").
Trainer(
testingutil.MakeTrainJobTrainerWrapper().
Container("test:trainjob", []string{"trainjob"}, []string{"trainjob"}, resRequests).
ContainerEnv(
[]corev1.EnvVar{
{
Name: "TRAIN_JOB",
Value: "override",
},
{
Name: "TRAIN_JOB_CUSTOM",
Value: "test:trainjob",
},
}...,
).
Obj(),
).
Obj(),
wantObjs: []runtime.Object{
testingutil.MakeJobSetWrapper(metav1.NamespaceDefault, "test-job").
NumNodes(100).
ContainerTrainer("test:trainjob", []string{"trainjob"}, []string{"trainjob"}, resRequests).
ContainerTrainerPorts([]corev1.ContainerPort{{ContainerPort: constants.ContainerTrainerPort}}).
ContainerTrainerEnv(
[]corev1.EnvVar{
{
Name: "TRAIN_JOB",
Value: "override",
},
{
Name: "TRAIN_JOB_CUSTOM",
Value: "test:trainjob",
},
{
Name: constants.TorchEnvNumNodes,
Value: "100",
},
{
Name: constants.TorchEnvNumProcPerNode,
Value: "auto",
},
{
Name: constants.TorchEnvNodeRank,
ValueFrom: &corev1.EnvVarSource{
FieldRef: &corev1.ObjectFieldSelector{
FieldPath: constants.JobCompletionIndexFieldPath,
},
},
},
{
Name: constants.TorchEnvMasterAddr,
Value: fmt.Sprintf("test-job-%s-0-0.test-job", constants.JobTrainerNode),
},
{
Name: constants.TorchEnvMasterPort,
Value: fmt.Sprintf("%d", constants.ContainerTrainerPort),
},
{
Name: "RUNTIME",
Value: "test:runtime",
},
},
).
ControllerReference(trainer.SchemeGroupVersion.WithKind(trainer.TrainJobKind), "test-job", "uid").
Obj(),
},
},
// Failed test cases.
"missing trainingRuntime resource": {
trainJob: testingutil.MakeTrainJobWrapper(metav1.NamespaceDefault, "test-job-3").
UID("uid").
RuntimeRef(trainer.SchemeGroupVersion.WithKind(trainer.TrainingRuntimeKind), "test-runtime-3").
Trainer(
testingutil.MakeTrainJobTrainerWrapper().
Obj(),
).
Obj(),
wantError: errorNotFoundSpecifiedTrainingRuntime,
},
}
cmpOpts := []cmp.Option{
cmpopts.SortSlices(func(a, b runtime.Object) bool {
return a.GetObjectKind().GroupVersionKind().String() < b.GetObjectKind().GroupVersionKind().String()
}),
cmpopts.SortSlices(func(a, b corev1.EnvVar) bool {
return a.Name < b.Name
}),
cmpopts.EquateEmpty(),
cmpopts.SortMaps(func(a, b string) bool { return a < b }),
}
for name, tc := range cases {
t.Run(name, func(t *testing.T) {
ctx, cancel := context.WithCancel(context.Background())
t.Cleanup(cancel)
clientBuilder := testingutil.NewClientBuilder()
if tc.trainingRuntime != nil {
clientBuilder.WithObjects(tc.trainingRuntime)
}
c := clientBuilder.Build()
trainingRuntime, err := NewTrainingRuntime(ctx, c, testingutil.AsIndex(clientBuilder))
if err != nil {
t.Fatal(err)
}
objs, err := trainingRuntime.NewObjects(ctx, tc.trainJob)
if diff := cmp.Diff(tc.wantError, err, cmpopts.EquateErrors()); len(diff) != 0 {
t.Errorf("Unexpected error (-want,+got):\n%s", diff)
}
resultObjs, err := testingutil.ToObject(c.Scheme(), objs...)
if err != nil {
t.Errorf("Pipeline built unrecognizable objects: %v", err)
}
if diff := cmp.Diff(tc.wantObjs, resultObjs, cmpOpts...); len(diff) != 0 {
t.Errorf("Unexpected objects (-want,+got):\n%s", diff)
}
})
}
}