Files
kubernetes/pkg/api/v1/resource/helpers.go
Eric Ernst 80ee072b85 pod-overhead: utilize pod overhead for cgroup sizing, eviction handling
Pod and burstable QoS cgroups should take overhead of running a sandbox
into account if the PodOverhead feature is enabled. These helper
functions are utilized by Kubelet for sizing the pod and burstable QoS
cgroups.

Pod overhead is added to resource requests, regardless of the initial
request values. A particular resource pod overhead is only added to a
resource limit if a non-zero limit already existed.

This commit updates eviction handling to also take Pod Overhead into
account (if the feature is enabled).

Signed-off-by: Eric Ernst <eric.ernst@intel.com>
2019-08-19 17:40:47 -07:00

251 lines
8.7 KiB
Go

/*
Copyright 2014 The Kubernetes 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 resource
import (
"fmt"
"math"
"strconv"
"k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/api/resource"
utilfeature "k8s.io/apiserver/pkg/util/feature"
"k8s.io/kubernetes/pkg/features"
)
// addResourceList adds the resources in newList to list
func addResourceList(list, newList v1.ResourceList) {
for name, quantity := range newList {
if value, ok := list[name]; !ok {
list[name] = *quantity.Copy()
} else {
value.Add(quantity)
list[name] = value
}
}
}
// maxResourceList sets list to the greater of list/newList for every resource
// either list
func maxResourceList(list, new v1.ResourceList) {
for name, quantity := range new {
if value, ok := list[name]; !ok {
list[name] = *quantity.Copy()
continue
} else {
if quantity.Cmp(value) > 0 {
list[name] = *quantity.Copy()
}
}
}
}
// PodRequestsAndLimits returns a dictionary of all defined resources summed up for all
// containers of the pod. If PodOverhead feature is enabled, pod overhead is added to the
// total container resource requests and to the total container limits which have a
// non-zero quantity.
func PodRequestsAndLimits(pod *v1.Pod) (reqs, limits v1.ResourceList) {
reqs, limits = v1.ResourceList{}, v1.ResourceList{}
for _, container := range pod.Spec.Containers {
addResourceList(reqs, container.Resources.Requests)
addResourceList(limits, container.Resources.Limits)
}
// init containers define the minimum of any resource
for _, container := range pod.Spec.InitContainers {
maxResourceList(reqs, container.Resources.Requests)
maxResourceList(limits, container.Resources.Limits)
}
// if PodOverhead feature is supported, add overhead for running a pod
// to the sum of reqeuests and to non-zero limits:
if pod.Spec.Overhead != nil && utilfeature.DefaultFeatureGate.Enabled(features.PodOverhead) {
addResourceList(reqs, pod.Spec.Overhead)
for name, quantity := range pod.Spec.Overhead {
if value, ok := limits[name]; ok && !value.IsZero() {
value.Add(quantity)
limits[name] = value
}
}
}
return
}
// GetResourceRequestQuantity finds and returns the request quantity for a specific resource.
func GetResourceRequestQuantity(pod *v1.Pod, resourceName v1.ResourceName) resource.Quantity {
requestQuantity := resource.Quantity{}
switch resourceName {
case v1.ResourceCPU:
requestQuantity = resource.Quantity{Format: resource.DecimalSI}
case v1.ResourceMemory, v1.ResourceStorage, v1.ResourceEphemeralStorage:
requestQuantity = resource.Quantity{Format: resource.BinarySI}
default:
requestQuantity = resource.Quantity{Format: resource.DecimalSI}
}
if resourceName == v1.ResourceEphemeralStorage && !utilfeature.DefaultFeatureGate.Enabled(features.LocalStorageCapacityIsolation) {
// if the local storage capacity isolation feature gate is disabled, pods request 0 disk
return requestQuantity
}
for _, container := range pod.Spec.Containers {
if rQuantity, ok := container.Resources.Requests[resourceName]; ok {
requestQuantity.Add(rQuantity)
}
}
for _, container := range pod.Spec.InitContainers {
if rQuantity, ok := container.Resources.Requests[resourceName]; ok {
if requestQuantity.Cmp(rQuantity) < 0 {
requestQuantity = rQuantity.DeepCopy()
}
}
}
// if PodOverhead feature is supported, add overhead for running a pod
// to the total requests if the resource total is non-zero
if pod.Spec.Overhead != nil && utilfeature.DefaultFeatureGate.Enabled(features.PodOverhead) {
if podOverhead, ok := pod.Spec.Overhead[resourceName]; ok && !requestQuantity.IsZero() {
requestQuantity.Add(podOverhead)
}
}
return requestQuantity
}
// GetResourceRequest finds and returns the request value for a specific resource.
func GetResourceRequest(pod *v1.Pod, resource v1.ResourceName) int64 {
if resource == v1.ResourcePods {
return 1
}
requestQuantity := GetResourceRequestQuantity(pod, resource)
if resource == v1.ResourceCPU {
return requestQuantity.MilliValue()
}
return requestQuantity.Value()
}
// ExtractResourceValueByContainerName extracts the value of a resource
// by providing container name
func ExtractResourceValueByContainerName(fs *v1.ResourceFieldSelector, pod *v1.Pod, containerName string) (string, error) {
container, err := findContainerInPod(pod, containerName)
if err != nil {
return "", err
}
return ExtractContainerResourceValue(fs, container)
}
// ExtractResourceValueByContainerNameAndNodeAllocatable extracts the value of a resource
// by providing container name and node allocatable
func ExtractResourceValueByContainerNameAndNodeAllocatable(fs *v1.ResourceFieldSelector, pod *v1.Pod, containerName string, nodeAllocatable v1.ResourceList) (string, error) {
realContainer, err := findContainerInPod(pod, containerName)
if err != nil {
return "", err
}
container := realContainer.DeepCopy()
MergeContainerResourceLimits(container, nodeAllocatable)
return ExtractContainerResourceValue(fs, container)
}
// ExtractContainerResourceValue extracts the value of a resource
// in an already known container
func ExtractContainerResourceValue(fs *v1.ResourceFieldSelector, container *v1.Container) (string, error) {
divisor := resource.Quantity{}
if divisor.Cmp(fs.Divisor) == 0 {
divisor = resource.MustParse("1")
} else {
divisor = fs.Divisor
}
switch fs.Resource {
case "limits.cpu":
return convertResourceCPUToString(container.Resources.Limits.Cpu(), divisor)
case "limits.memory":
return convertResourceMemoryToString(container.Resources.Limits.Memory(), divisor)
case "limits.ephemeral-storage":
return convertResourceEphemeralStorageToString(container.Resources.Limits.StorageEphemeral(), divisor)
case "requests.cpu":
return convertResourceCPUToString(container.Resources.Requests.Cpu(), divisor)
case "requests.memory":
return convertResourceMemoryToString(container.Resources.Requests.Memory(), divisor)
case "requests.ephemeral-storage":
return convertResourceEphemeralStorageToString(container.Resources.Requests.StorageEphemeral(), divisor)
}
return "", fmt.Errorf("unsupported container resource : %v", fs.Resource)
}
// convertResourceCPUToString converts cpu value to the format of divisor and returns
// ceiling of the value.
func convertResourceCPUToString(cpu *resource.Quantity, divisor resource.Quantity) (string, error) {
c := int64(math.Ceil(float64(cpu.MilliValue()) / float64(divisor.MilliValue())))
return strconv.FormatInt(c, 10), nil
}
// convertResourceMemoryToString converts memory value to the format of divisor and returns
// ceiling of the value.
func convertResourceMemoryToString(memory *resource.Quantity, divisor resource.Quantity) (string, error) {
m := int64(math.Ceil(float64(memory.Value()) / float64(divisor.Value())))
return strconv.FormatInt(m, 10), nil
}
// convertResourceEphemeralStorageToString converts ephemeral storage value to the format of divisor and returns
// ceiling of the value.
func convertResourceEphemeralStorageToString(ephemeralStorage *resource.Quantity, divisor resource.Quantity) (string, error) {
m := int64(math.Ceil(float64(ephemeralStorage.Value()) / float64(divisor.Value())))
return strconv.FormatInt(m, 10), nil
}
// findContainerInPod finds a container by its name in the provided pod
func findContainerInPod(pod *v1.Pod, containerName string) (*v1.Container, error) {
for _, container := range pod.Spec.Containers {
if container.Name == containerName {
return &container, nil
}
}
for _, container := range pod.Spec.InitContainers {
if container.Name == containerName {
return &container, nil
}
}
return nil, fmt.Errorf("container %s not found", containerName)
}
// MergeContainerResourceLimits checks if a limit is applied for
// the container, and if not, it sets the limit to the passed resource list.
func MergeContainerResourceLimits(container *v1.Container,
allocatable v1.ResourceList) {
if container.Resources.Limits == nil {
container.Resources.Limits = make(v1.ResourceList)
}
for _, resource := range []v1.ResourceName{v1.ResourceCPU, v1.ResourceMemory, v1.ResourceEphemeralStorage} {
if quantity, exists := container.Resources.Limits[resource]; !exists || quantity.IsZero() {
if cap, exists := allocatable[resource]; exists {
container.Resources.Limits[resource] = *cap.Copy()
}
}
}
}