mirror of
https://github.com/cozystack/cozystack.git
synced 2026-11-09 14:02:56 +09:00
Aleksei Sviridkin
02df4af713
A tenant picking a GPU class needs to know whether a VM of it is likely to start, and a hoster needs to see a class that can never serve a VM before a tenant's start is refused. With no GPUNode and no binder in step 1, both figures come from the Node and Pod objects alone: free units are a node's allocatable count of the class's resource name minus what its pods request, over nodes that are Ready, not cordoned and carry no taint the class does not tolerate. One pass covers every class, because a class's ResourceNameConflict depends on the classes created before it. The KubeVirt CR is read on each pass and on a one-minute recheck instead of watched, so the controller needs no KubeVirt kinds served to start. The class status is readable by every authenticated user, so the inventory and load behind it, the allocatable and requested totals per class, go to metrics instead. Assisted-by: LLM Signed-off-by: Aleksei Sviridkin <f@lex.la>
122 lines
3.7 KiB
Go
122 lines
3.7 KiB
Go
/*
|
|
Copyright 2026 The Cozystack 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 controller
|
|
|
|
import (
|
|
"context"
|
|
"testing"
|
|
"time"
|
|
|
|
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/types"
|
|
clientgoscheme "k8s.io/client-go/kubernetes/scheme"
|
|
"sigs.k8s.io/controller-runtime/pkg/client"
|
|
"sigs.k8s.io/controller-runtime/pkg/client/fake"
|
|
"sigs.k8s.io/controller-runtime/pkg/client/interceptor"
|
|
|
|
cozyv1alpha1 "github.com/cozystack/cozystack/api/v1alpha1"
|
|
)
|
|
|
|
const (
|
|
rnL40S = "nvidia.com/NVIDIA_L40S-24C"
|
|
rnH100 = "nvidia.com/GH100_H100"
|
|
)
|
|
|
|
var epoch = time.Date(2026, 1, 1, 0, 0, 0, 0, time.UTC)
|
|
|
|
func testScheme(t *testing.T) *runtime.Scheme {
|
|
t.Helper()
|
|
s := runtime.NewScheme()
|
|
if err := clientgoscheme.AddToScheme(s); err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if err := cozyv1alpha1.AddToScheme(s); err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
return s
|
|
}
|
|
|
|
func newClient(t *testing.T, funcs *interceptor.Funcs, objs ...client.Object) client.WithWatch {
|
|
t.Helper()
|
|
b := fake.NewClientBuilder().WithScheme(testScheme(t)).WithObjects(objs...).
|
|
WithStatusSubresource(&cozyv1alpha1.GPUClass{})
|
|
if funcs != nil {
|
|
b = b.WithInterceptorFuncs(*funcs)
|
|
}
|
|
return b.Build()
|
|
}
|
|
|
|
// gpuClass is created age minutes after epoch, so tests order classes by age.
|
|
func gpuClass(name, resourceName string, age int, mutate ...func(*cozyv1alpha1.GPUClass)) *cozyv1alpha1.GPUClass {
|
|
c := &cozyv1alpha1.GPUClass{
|
|
ObjectMeta: metav1.ObjectMeta{Name: name, Generation: 1, CreationTimestamp: metav1.NewTime(epoch.Add(time.Duration(age) * time.Minute))},
|
|
Spec: cozyv1alpha1.GPUClassSpec{
|
|
Vendor: "nvidia", Model: "L40S", Mode: cozyv1alpha1.GPUModePassthrough,
|
|
Source: cozyv1alpha1.GPUClassSource{DevicePlugin: &cozyv1alpha1.DevicePluginSource{ResourceName: resourceName}},
|
|
},
|
|
}
|
|
for _, m := range mutate {
|
|
m(c)
|
|
}
|
|
return c
|
|
}
|
|
|
|
func gpuNode(name string, alloc map[string]int64, mutate ...func(*corev1.Node)) *corev1.Node {
|
|
n := &corev1.Node{
|
|
ObjectMeta: metav1.ObjectMeta{Name: name},
|
|
Status: corev1.NodeStatus{
|
|
Allocatable: corev1.ResourceList{},
|
|
Conditions: []corev1.NodeCondition{{Type: corev1.NodeReady, Status: corev1.ConditionTrue}},
|
|
},
|
|
}
|
|
for rn, v := range alloc {
|
|
n.Status.Allocatable[corev1.ResourceName(rn)] = *resource.NewQuantity(v, resource.DecimalSI)
|
|
}
|
|
for _, m := range mutate {
|
|
m(n)
|
|
}
|
|
return n
|
|
}
|
|
|
|
func gpuPod(name, node, resourceName string, units int64, mutate ...func(*corev1.Pod)) *corev1.Pod {
|
|
q := *resource.NewQuantity(units, resource.DecimalSI)
|
|
p := &corev1.Pod{
|
|
ObjectMeta: metav1.ObjectMeta{Namespace: "tenant-a", Name: name},
|
|
Spec: corev1.PodSpec{NodeName: node, Containers: []corev1.Container{{
|
|
Name: "compute",
|
|
Resources: corev1.ResourceRequirements{Limits: corev1.ResourceList{corev1.ResourceName(resourceName): q}},
|
|
}}},
|
|
Status: corev1.PodStatus{Phase: corev1.PodRunning},
|
|
}
|
|
for _, m := range mutate {
|
|
m(p)
|
|
}
|
|
return p
|
|
}
|
|
|
|
func getClass(t *testing.T, c client.Client, name string) *cozyv1alpha1.GPUClass {
|
|
t.Helper()
|
|
out := &cozyv1alpha1.GPUClass{}
|
|
if err := c.Get(context.Background(), types.NamespacedName{Name: name}, out); err != nil {
|
|
t.Fatalf("get GPUClass %s: %v", name, err)
|
|
}
|
|
return out
|
|
}
|