@@ -116,11 +116,11 @@ If none are enabled, you can skip creating Pod Disruption Budgets.
116
116
咨询集群管理员或联系云提供商,或者查询发布文档,以确定是否为集群启用了任何资源干扰源。
117
117
如果没有启用,可以不用创建 Pod Disruption Budgets(Pod 干扰预算)
118
118
119
+ {{< caution >}}
119
120
<!--
120
121
Not all voluntary disruptions are constrained by Pod Disruption Budgets. For example,
121
122
deleting deployments or pods bypasses Pod Disruption Budgets.
122
123
-->
123
- {{< caution >}}
124
124
并非所有的自愿干扰都会受到 Pod 干扰预算的限制。
125
125
例如,删除 Deployment 或 Pod 的删除操作就会跳过 Pod 干扰预算检查。
126
126
{{< /caution >}}
@@ -234,7 +234,7 @@ PDB 指定应用可以容忍的副本数量(相当于应该有多少副本)
234
234
The group of pods that comprise the application is specified using a label selector, the same
235
235
as the one used by the application's controller (deployment, stateful-set, etc).
236
236
-->
237
- 使用标签选择器来指定构成应用的一组 Pod,这与应用的控制器(Deployment, StatefulSet 等)
237
+ 使用标签选择器来指定构成应用的一组 Pod,这与应用的控制器(Deployment、 StatefulSet 等)
238
238
选择 Pod 的逻辑一样。
239
239
240
240
<!--
@@ -431,15 +431,23 @@ can happen, according to:
431
431
-->
432
432
## Pod 干扰状况 {#pod-disruption-conditions}
433
433
434
- {{< feature-state for_k8s_version="v1.25" state="alpha" >}}
434
+ {{< feature-state for_k8s_version="v1.26" state="beta" >}}
435
+
436
+ {{< note >}}
437
+ <!--
438
+ If you are using an older version of Kubernetes than {{< skew currentVersion >}}
439
+ please refer to the corresponding version of the documentation.
440
+ -->
441
+ 如果你正使用的 Kubernetes 版本早于 {{< skew currentVersion >}},请参阅对应版本的文档。
442
+ {{< /note >}}
435
443
436
444
{{< note >}}
437
445
<!--
438
- In order to use this behavior, you must enable the `PodDisruptionCondition `
446
+ In order to use this behavior, you must have the `PodDisruptionConditions `
439
447
[feature gate](/docs/reference/command-line-tools-reference/feature-gates/)
440
- in your cluster.
448
+ enabled in your cluster.
441
449
-->
442
- 要使用此行为,你必须在集群中启用 ` PodDisruptionCondition `
450
+ 要使用此行为,你必须在集群中启用 ` PodDisruptionConditions `
443
451
[ 特性门控] ( /zh-cn/docs/reference/command-line-tools-reference/feature-gates/ ) 。
444
452
{{< /note >}}
445
453
@@ -487,6 +495,15 @@ Taint Manager(`kube-controller-manager` 中节点生命周期控制器的一
487
495
: 绑定到一个不再存在的 Node 上的 Pod 将被
488
496
[ Pod 垃圾收集] ( /zh-cn/docs/concepts/workloads/pods/pod-lifecycle/#pod-garbage-collection ) 删除。
489
497
498
+ <!--
499
+ `TerminationByKubelet`
500
+ : Pod has been terminated by the kubelet, because of either {{<glossary_tooltip term_id="node-pressure-eviction" text="node pressure eviction">}} or the [graceful node shutdown](/docs/concepts/architecture/nodes/#graceful-node-shutdown).
501
+ -->
502
+ ` TerminationByKubelet `
503
+ : Pod
504
+ 由于{{<glossary_tooltip term_id="node-pressure-eviction" text="节点压力驱逐">}}或[ 节点体面关闭] ( /zh-cn/docs/concepts/architecture/nodes/#graceful-node-shutdown ) 而被
505
+ kubelet 终止。
506
+
490
507
{{< note >}}
491
508
<!--
492
509
A Pod disruption might be interrupted. The control plane might re-attempt to
@@ -501,6 +518,15 @@ Pod 的干扰可能会被中断。控制平面可能会重新尝试继续干扰
501
518
在这种情况下,一段时间后,Pod 干扰状况将被清除。
502
519
{{< /note >}}
503
520
521
+ <!--
522
+ When the `PodDisruptionConditions` feature gate is enabled,
523
+ along with cleaning up the pods, the Pod garbage collector (PodGC) will also mark them as failed if they are in a non-terminal
524
+ phase (see also [Pod garbage collection](/docs/concepts/workloads/pods/pod-lifecycle/#pod-garbage-collection)).
525
+ -->
526
+ 当 ` PodDisruptionConditions ` 特性门控被启用时,在清理 Pod 的同时,如果这些 Pod 处于非终止阶段,
527
+ 则 Pod 垃圾回收器 (PodGC) 也会将这些 Pod 标记为失效
528
+ (另见 [ Pod 垃圾回收] ( /zh-cn/docs/concepts/workloads/pods/pod-lifecycle/#pod-garbage-collection ) )。
529
+
504
530
<!--
505
531
When using a Job (or CronJob), you may want to use these Pod disruption conditions as part of your Job's
506
532
[Pod failure policy](/docs/concepts/workloads/controllers/job#pod-failure-policy).
@@ -563,16 +589,15 @@ the nodes in your cluster, such as a node or system software upgrade, here are s
563
589
disruptions largely overlaps with work to support autoscaling and tolerating
564
590
involuntary disruptions.
565
591
-->
566
-
567
592
- 接受升级期间的停机时间。
568
593
- 故障转移到另一个完整的副本集群。
569
- - 没有停机时间,但是对于重复的节点和人工协调成本可能是昂贵的。
594
+ - 没有停机时间,但是对于重复的节点和人工协调成本可能是昂贵的。
570
595
- 编写可容忍干扰的应用和使用 PDB。
571
- - 不停机。
572
- - 最小的资源重复。
573
- - 允许更多的集群管理自动化。
574
- - 编写可容忍干扰的应用是棘手的,但对于支持容忍自愿干扰所做的工作,和支持自动扩缩和容忍非
575
- 自愿干扰所做工作相比,有大量的重叠
596
+ - 不停机。
597
+ - 最小的资源重复。
598
+ - 允许更多的集群管理自动化。
599
+ - 编写可容忍干扰的应用是棘手的,但对于支持容忍自愿干扰所做的工作,和支持自动扩缩和容忍非
600
+ 自愿干扰所做工作相比,有大量的重叠
576
601
577
602
## {{% heading "whatsnext" %}}
578
603
0 commit comments