Skip to content

Commit b15fe21

Browse files
windsonseawindsonsea
committed
[zh] sync /workloads/pods/pod-qos.md
1 parent 190d50c commit b15fe21

File tree

1 file changed

+213
-0
lines changed
  • content/zh-cn/docs/concepts/workloads/pods

1 file changed

+213
-0
lines changed

‎content/zh-cn/docs/concepts/workloads/pods/pod-qos.md

Lines changed: 213 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,213 @@
1+
---
2+
title: Pod QoS 类
3+
content_type: concept
4+
weight: 85
5+
---
6+
<!--
7+
title: Pod Quality of Service Classes
8+
content_type: concept
9+
weight: 85
10+
-->
11+
12+
<!-- overview -->
13+
14+
<!--
15+
This page introduces _Quality of Service (QoS) classes_ in Kubernetes, and explains
16+
how Kubernetes assigns a QoS class to each Pods as a consequence of the resource
17+
constraints that you specify for the containers in that Pod. Kubernetes relies on this
18+
classification to make decisions about which Pods to evict when there are not enough
19+
available resources on a Node.
20+
-->
21+
本页介绍 Kubernetes 中的 **服务质量（Quality of Service，QoS）** 类，
22+
阐述 Kubernetes 如何根据为 Pod 中的容器指定的资源约束为每个 Pod 设置 QoS 类。
23+
Kubernetes 依赖这种分类来决定当 Node 上没有足够可用资源时要驱逐哪些 Pod。
24+
25+
<!-- body -->
26+
27+
<!--
28+
## Quality of Service classes
29+
-->
30+
## QoS 类 {#qos-class}
31+
32+
<!--
33+
Kubernetes classifies the Pods that you run and allocates each Pod into a specific
34+
_quality of service (QoS) class_. Kubernetes uses that classification to influence how different
35+
pods are handled. Kubernetes does this classification based on the
36+
[resource requests](/docs/concepts/configuration/manage-resources-containers/)
37+
of the {{< glossary_tooltip text="Containers" term_id="container" >}} in that Pod, along with
38+
how those requests relate to resource limits.
39+
This is known as {{< glossary_tooltip text="Quality of Service" term_id="qos-class" >}}
40+
(QoS) class. Kubernetes assigns every Pod a QoS class based on the resource requests
41+
and limits of its component Containers. QoS classes are used by Kubernetes to decide
42+
which Pods to evict from a Node experiencing
43+
[Node Pressure](/docs/concepts/scheduling-eviction/node-pressure-eviction/). The possible
44+
QoS classes are `Guaranteed`, `Burstable`, and `BestEffort`. When a Node runs out of resources,
45+
Kubernetes will first evict `BestEffort` Pods running on that Node, followed by `Burstable` and
46+
finally `Guaranteed` Pods. When this eviction is due to resource pressure, only Pods exceeding
47+
resource requests are candidates for eviction.
48+
-->
49+
Kubernetes 对你运行的 Pod 进行分类，并将每个 Pod 分配到特定的 **QoS 类**中。
50+
Kubernetes 使用这种分类来影响不同 Pod 被处理的方式。Kubernetes 基于 Pod
51+
中{{< glossary_tooltip text="容器" term_id="container" >}}的[资源请求](/zh-cn/docs/concepts/configuration/manage-resources-containers/)进行分类，
52+
同时确定这些请求如何与资源限制相关。
53+
这称为{{< glossary_tooltip text="服务质量" term_id="qos-class" >}} (QoS) 类。
54+
Kubernetes 基于每个 Pod 中容器的资源请求和限制为 Pod 设置 QoS 类。Kubernetes 使用 QoS
55+
类来决定从遇到[节点压力](/zh-cn/docs/concepts/scheduling-eviction/node-pressure-eviction/)的
56+
Node 中驱逐哪些 Pod。可选的 QoS 类有 `Guaranteed`、`Burstable` 和 `BestEffort`。
57+
当一个 Node 耗尽资源时，Kubernetes 将首先驱逐在该 Node 上运行的 `BestEffort` Pod，
58+
然后是 `Burstable` Pod，最后是 `Guaranteed` Pod。当这种驱逐是由于资源压力时，
59+
只有超出资源请求的 Pod 才是被驱逐的候选对象。
60+
61+
### Guaranteed
62+
63+
<!--
64+
Pods that are `Guaranteed` have the strictest resource limits and are least likely
65+
to face eviction. They are guaranteed not to be killed until they exceed their limits
66+
or there are no lower-priority Pods that can be preempted from the Node. They may
67+
not acquire resources beyond their specified limits. These Pods can also make
68+
use of exclusive CPUs using the
69+
[`static`](/docs/tasks/administer-cluster/cpu-management-policies/#static-policy) CPU management policy.
70+
-->
71+
`Guaranteed` Pod 具有最严格的资源限制，并且最不可能面临驱逐。
72+
在这些 Pod 超过其自身的限制或者从 Node 上没有可以抢占的低优先级 Pod 之前，
73+
这些 Pod 保证不会被杀死。这些 Pod 不可以获得超出其指定 limit 的资源。这些 Pod 也可以使用
74+
[`static`](/zh-cn/docs/tasks/administer-cluster/cpu-management-policies/#static-policy)
75+
CPU 管理策略来使用独占的 CPU。
76+
77+
<!--
78+
#### Criteria
79+
80+
For a Pod to be given a QoS class of `Guaranteed`:
81+
-->
82+
#### 判据
83+
84+
Pod 被赋予 `Guaranteed` QoS 类的几个判据：
85+
86+
<!--
87+
* Every Container in the Pod must have a memory limit and a memory request.
88+
* For every Container in the Pod, the memory limit must equal the memory request.
89+
* Every Container in the Pod must have a CPU limit and a CPU request.
90+
* For every Container in the Pod, the CPU limit must equal the CPU request.
91+
-->
92+
* Pod 中的每个容器必须有内存 limit 和内存 request。
93+
* 对于 Pod 中的每个容器，内存 limit 必须等于内存 request。
94+
* Pod 中的每个容器必须有 CPU limit 和 CPU request。
95+
* 对于 Pod 中的每个容器，CPU limit 必须等于 CPU request。
96+
97+
### Burstable
98+
99+
<!--
100+
Pods that are `Burstable` have some lower-bound resource guarantees based on the request, but
101+
do not require a specific limit. If a limit is not specified, it defaults to a
102+
limit equivalent to the capacity of the Node, which allows the Pods to flexibly increase
103+
their resources if resources are available. In the event of Pod eviction due to Node
104+
resource pressure, these Pods are evicted only after all `BestEffort` Pods are evicted.
105+
Because a `Burstable` Pod can include a Container that has no resource limits or requests, a Pod
106+
that is `Burstable` can try to use any amount of node resources.
107+
-->
108+
`Burstable` Pod 有一些基于 request 的资源下限保证，但不需要特定的 limit。
109+
如果未指定 limit，则默认为其 limit 等于 Node 容量，这允许 Pod 在资源可用时灵活地增加其资源。
110+
在由于 Node 资源压力导致 Pod 被驱逐的情况下，只有在所有 `BestEffort` Pod 被驱逐后
111+
这些 Pod 才会被驱逐。因为 `Burstable` Pod 可以包括没有资源 limit 或资源 request 的容器，
112+
所以 `Burstable` Pod 可以尝试使用任意数量的节点资源。
113+
114+
<!--
115+
#### Criteria
116+
117+
A Pod is given a QoS class of `Burstable` if:
118+
119+
* The Pod does not meet the criteria for QoS class `Guaranteed`.
120+
* At least one Container in the Pod has a memory or CPU request or limit.
121+
-->
122+
#### 判据
123+
124+
Pod 被赋予 `Burstable` QoS 类的几个判据：
125+
126+
* Pod 不满足针对 QoS 类 `Guaranteed` 的判据。
127+
* Pod 中至少一个容器有内存或 CPU request 或 limit。
128+
129+
### BestEffort
130+
131+
<!--
132+
Pods in the `BestEffort` QoS class can use node resources that aren't specifically assigned
133+
to Pods in other QoS classes. For example, if you have a node with 16 CPU cores available to the
134+
kubelet, and you assign 4 CPU cores to a `Guaranteed` Pod, then a Pod in the `BestEffort`
135+
QoS class can try to use any amount of the remaining 12 CPU cores.
136+
137+
The kubelet prefers to evict `BestEffort` Pods if the node comes under resource pressure.
138+
-->
139+
`BestEffort` QoS 类中的 Pod 可以使用未专门分配给其他 QoS 类中的 Pod 的节点资源。
140+
例如若你有一个节点有 16 核 CPU 可供 kubelet 使用，并且你将 4 核 CPU 分配给一个 `Guaranteed` Pod，
141+
那么 `BestEffort` QoS 类中的 Pod 可以尝试任意使用剩余的 12 核 CPU。
142+
143+
如果节点遇到资源压力，kubelet 将优先驱逐 `BestEffort` Pod。
144+
145+
<!--
146+
#### Criteria
147+
148+
A Pod has a QoS class of `BestEffort` if it doesn't meet the criteria for either `Guaranteed`
149+
or `Burstable`. In other words, a Pod is `BestEffort` only if none of the Containers in the Pod have a
150+
memory limit or a memory request, and none of the Containers in the Pod have a
151+
CPU limit or a CPU request.
152+
Containers in a Pod can request other resources (not CPU or memory) and still be classified as
153+
`BestEffort`.
154+
-->
155+
#### 判据
156+
157+
如果 Pod 不满足 `Guaranteed` 或 `Burstable` 的判据，则它的 QoS 类为 `BestEffort`。
158+
换言之，只有当 Pod 中的所有容器没有内存 limit 或内存 request，也没有 CPU limit 或
159+
CPU request 时，Pod 才是 `BestEffort`。Pod 中的容器可以请求（除 CPU 或内存之外的）
160+
其他资源并且仍然被归类为 `BestEffort`。
161+
162+
<!--
163+
## Some behavior is independent of QoS class {#class-independent-behavior}
164+
165+
Certain behavior is independent of the QoS class assigned by Kubernetes. For example:
166+
-->
167+
## 某些行为独立于 QoS 类 {#class-independent-behavior}
168+
169+
某些行为独立于 Kubernetes 分配的 QoS 类。例如：
170+
171+
<!--
172+
* Any Container exceeding a resource limit will be killed and restarted by the kubelet without
173+
affecting other Containers in that Pod.
174+
* If a Container exceeds its resource request and the node it runs on faces
175+
resource pressure, the Pod it is in becomes a candidate for [eviction](/docs/concepts/scheduling-eviction/node-pressure-eviction/).
176+
If this occurs, all Containers in the Pod will be terminated. Kubernetes may create a
177+
replacement Pod, usually on a different node.
178+
-->
179+
* 所有超过资源 limit 的容器都将被 kubelet 杀死并重启，而不会影响该 Pod 中的其他容器。
180+
* 如果一个容器超出了自身的资源 request，且该容器运行的节点面临资源压力，则该容器所在的 Pod
181+
就会成为被[驱逐](/zh-cn/docs/concepts/scheduling-eviction/node-pressure-eviction/)的候选对象。
182+
如果出现这种情况，Pod 中的所有容器都将被终止。Kubernetes 通常会在不同的节点上创建一个替代的 Pod。
183+
<!--
184+
* The resource request of a Pod is equal to the sum of the resource requests of
185+
its component Containers, and the resource limit of a Pod is equal to the sum of
186+
the resource limits of its component Containers.
187+
* The kube-scheduler does not consider QoS class when selecting which Pods to
188+
[preempt](/docs/concepts/scheduling-eviction/pod-priority-preemption/#preemption).
189+
Preemption can occur when a cluster does not have enough resources to run all the Pods
190+
you defined.
191+
-->
192+
* Pod 的资源 request 等于其成员容器的资源 request 之和，Pod 的资源 limit 等于其组成容器的资源 limit 之和。
193+
* kube-scheduler 在选择要[抢占](/zh-cn/docs/concepts/scheduling-eviction/pod-priority-preemption/#preemption)的
194+
Pod 时不考虑 QoS 类。当集群没有足够的资源来运行你所定义的所有 Pod 时，就会发生抢占。
195+
196+
## {{% heading "whatsnext" %}}
197+
198+
<!--
199+
* Learn about [resource management for Pods and Containers](/docs/concepts/configuration/manage-resources-containers/).
200+
* Learn about [Node-pressure eviction](/docs/concepts/scheduling-eviction/node-pressure-eviction/).
201+
* Learn about [Pod priority and preemption](/docs/concepts/scheduling-eviction/pod-priority-preemption/).
202+
* Learn about [Pod disruptions](/docs/concepts/workload/pods/disruptions/).
203+
* Learn how to [assign memory resources to containers and pods](/docs/tasks/configure-pod-container/assign-memory-resource/).
204+
* Learn how to [assign CPU resources to containers and pods](/docs/tasks/configure-pod-container/assign-cpu-resource/).
205+
* Learn how to [configure Quality of Service for Pods](/docs/tasks/configure-pod-container/quality-service-pod/).
206+
-->
207+
* 进一步了解[为 Pod 和容器管理资源](/zh-cn/docs/concepts/configuration/manage-resources-containers/)。
208+
* 进一步了解[节点压力驱逐](/zh-cn/docs/concepts/scheduling-eviction/node-pressure-eviction/)。
209+
* 进一步了解 [Pod 优先级和抢占](/zh-cn/docs/concepts/scheduling-eviction/pod-priority-preemption/)。
210+
* 进一步了解 [Pod 干扰](/zh-cn/docs/concepts/workloads/pods/disruptions/)。
211+
* 进一步了解如何[为容器和 Pod 分配内存资源](/zh-cn/docs/tasks/configure-pod-container/assign-memory-resource/)。
212+
* 进一步了解如何[为容器和 Pod 分配 CPU 资源](/zh-cn/docs/tasks/configure-pod-container/assign-cpu-resource/)。
213+
* 进一步了解如何[配置 Pod 的服务质量](/zh-cn/docs/tasks/configure-pod-container/quality-service-pod/)。

0 commit comments

Comments
 (0)