本文介绍: 默认情况下,Pod调度到哪个节点都是右Scheduler组件进行调度,对于在工作和学习中需要指定设备运行特定pod时就显得不实用,此时我们家就需要合理利用几个调度规则,主要包括全自动调度、定向调度、亲和性调度污点调度、容忍调度等主要功能自动部署一个容器的多个副本,以来持续保持使用者指定的副本数量(replicas)

目录

一.Pod调度简介

二.Deployment/RC全自动调度

1.简介

2.案例演示

(1)Deployment

(2)RC

三.nodeSelector/nodeName指定节点调度

1.原理简介

(1)nodeSelector原理

(2)nodeName原理

2.案例演示

3.kubernetes预定义pod标签

四.Affinity亲和性调度

1.简介

2.三类Affinity

(1)nodeAffinity,node亲和性

a.以node为主体,针对pod调度到node问题

b.两种表达方式和可配置项

c.案例演示

d.注意事项

(2)podAffinity,pod亲和性

a.以pod为主体,针对pod和pod部署在同一个拓扑的问题

b.两种表达方式和可配置项

c.案例演示

  (3)podAntiAffinity,pod反亲和性

a.以pod为主体,针对pod和pod不能部署在同一个拓扑的问题

b.案例演示

五.taint污点调度和Toleration容忍调度

1.污点调度(node方)

(1)三个污点策略

(2)添加污点/查看污点/去除污点

(3)案例演示

a.PreferNoSchedule

b.NoSchedule

c.NoExecute

2.容忍调度(pod方)

(1)容忍的配置参数

(2)案例演示


 

一.Pod调度简介

默认情况下,Pod调度到哪个节点都是由Scheduler组件进行调度,对于在工作和学习中需要指定设备运行特定pod时就显得不实用,此时我们就需要合理利用几个调度规则,主要包括全自动调度、定向调度、亲和性调度、污点调度、容忍调度等

二.Deployment/RC全自动调度

1.简介

主要功能自动部署一个容器的多个副本,以来持续保持使用者指定的副本数量(replicas)

2.案例演示

创建一个Deployment来管理pod,指定创建2个副本,自动调度pod创建完成后查看调度情况

(1)Deployment

[root@k8s-master pod]# kubectl create deployment my-nginx --image=nginx --replicas=2 --port=80 -n myns --dry-run=client -o yaml > mydeployment.yaml
[root@k8s-master pod]# cat mydeployment.yaml 
apiVersion: apps/v1
kind: Deployment
metadata:
  creationTimestamp: null
  labels:
    app: my-nginx
  name: my-nginx
  namespace: myns
spec:
  replicas: 2
  selector:
    matchLabels:
      app: my-nginx
  strategy: {}
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: my-nginx
    spec:
      containers:
      - image: nginx
        name: nginx
        ports:
        - containerPort: 80
        resources: {}
status: {}
​
[root@k8s-master pod]# kubectl apply -f mydeployment.yaml
​
[root@k8s-master pod]# kubectl get deploy -n myns   #创建了两个副本
NAME       READY   UP-TO-DATE   AVAILABLE   AGE
my-nginx   2/2     2            2           8m17s
[root@k8s-master pod]# kubectl get pods -n myns -o wide   #两个node分别分配到了各自的pod任务
NAME                        READY   STATUS    RESTARTS   AGE     IP               NODE        NOMINATED NODE   READINESS GATES
my-nginx-64f6999999-5j7nl   1/1     Running   0          8m40s   10.244.36.65     k8s-node1   <none>           <none>
my-nginx-64f6999999-hjbd4   1/1     Running   0          8m40s   10.244.169.130   k8s-node2   <none>           <none>

(2)RC

[root@k8s-master pod]# cat myrc.yaml 
apiVersion: v1
kind: ReplicationController
metadata:
  name: nginx-rc
  namespace: myns
spec:
  replicas: 2
  selector:
    app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx-container
        image: nginx
        ports:
        - containerPort: 80
[root@k8s-master pod]# kubectl apply -f myrc.yaml 
replicationcontroller/nginx-rc created
[root@k8s-master pod]# kubectl get rc -n myns
NAME       DESIRED   CURRENT   READY   AGE
nginx-rc   2         2         2       25s
[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME             READY   STATUS    RESTARTS   AGE   IP               NODE        NOMINATED NODE   READINESS GATES
nginx-rc-4qfm8   1/1     Running   0          36s   10.244.36.70     k8s-node1   <none>           <none>
nginx-rc-xmdbc   1/1     Running   0          36s   10.244.169.133   k8s-node2   <none>           <none>

三.nodeSelector/nodeName指定节点调度

1.原理简介

(1)nodeSelector原理

就是给具体的受管node打上标签,在部署pod时指定调度到特定标签的node,前提是这个集群中需要有指定的node,若没有这个特定node即使有其他正常运行的node,这个pod也无法完成调度任务。如果多个node拥有同一个标签,那么Scheduler会进一步选择一个可用的node进行配合调度任务。

(2)nodeName原理

强制指定node的名称,也就是“kubectl get nodes可以查出来的node名称

2.案例演示

(1)仍然使用mydeployment.yaml进行测试,nodeSelector为节点打上标签,,加上nodeSelector参数指定节点,运行完成后可以看到3个副本都运行在node1

[root@k8s-master pod]# kubectl label nodes k8s-node1 name=node1
​
[root@k8s-master pod]# kubectl get nodes --show-labels
NAME         STATUS   ROLES           AGE   VERSION   LABELS
k8s-master   Ready    control-plane   34m   v1.28.2   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-master,kubernetes.io/os=linux,node-role.kubernetes.io/control-plane=,node.kubernetes.io/exclude-from-external-load-balancers=
k8s-node1    Ready    <none>          34m   v1.28.2   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node1,kubernetes.io/os=linux,name=node1
k8s-node2    Ready    <none>          33m   v1.28.2   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node2,kubernetes.io/os=linux,name=node2
​
[root@k8s-master pod]# cat mydeployment.yaml 
apiVersion: apps/v1
kind: Deployment
metadata:
  creationTimestamp: null
  labels:
    app: my-nginx
  name: my-nginx
  namespace: myns
spec:
  replicas: 3
  selector:
    matchLabels:
      app: my-nginx
  strategy: {}
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: my-nginx
    spec:
      containers:
      - image: nginx
        name: nginx
        ports:
        - containerPort: 80
      nodeSelector:
        name: node1
        
[root@k8s-master pod]# kubectl apply -f mydeployment.yaml 
deployment.apps/my-nginx created
​
[root@k8s-master pod]# kubectl get deployments -n myns
NAME       READY   UP-TO-DATE   AVAILABLE   AGE
my-nginx   3/3     3            3           23s
​
[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME                        READY   STATUS    RESTARTS   AGE   IP             NODE        NOMINATED NODE   READINESS GATES
my-nginx-74db7ccb9b-gvgfk   1/1     Running   0          31s   10.244.36.67   k8s-node1   <none>           <none>
my-nginx-74db7ccb9b-kpsb2   1/1     Running   0          31s   10.244.36.66   k8s-node1   <none>           <none>
my-nginx-74db7ccb9b-xwpjt   1/1     Running   0          31s   10.244.36.68   k8s-node1   <none>           <none>

(2)使用nodeName同样可以实现

[root@k8s-master pod]# kubectl get nodes
NAME         STATUS   ROLES           AGE   VERSION
k8s-master   Ready    control-plane   70m   v1.28.2
k8s-node1    Ready    <none>          70m   v1.28.2
k8s-node2    Ready    <none>          70m   v1.28.2
​
[root@k8s-master pod]# cat mydeployment.yaml 
apiVersion: apps/v1
kind: Deployment
metadata:
  creationTimestamp: null
  labels:
    app: my-nginx
  name: my-nginx
  namespace: myns
spec:
  replicas: 3
  selector:
    matchLabels:
      app: my-nginx
  strategy: {}
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: my-nginx
    spec:
      containers:
      - image: nginx
        name: nginx
        ports:
        - containerPort: 80
      nodeName: k8s-node1
​
[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME                        READY   STATUS    RESTARTS   AGE   IP             NODE        NOMINATED NODE   READINESS GATES
my-nginx-7544fbf745-gwxv7   1/1     Running   0          85s   10.244.36.70   k8s-node1   <none>           <none>
my-nginx-7544fbf745-kn8lj   1/1     Running   0          85s   10.244.36.71   k8s-node1   <none>           <none>
my-nginx-7544fbf745-q6w6j   1/1     Running   0          85s   10.244.36.69   k8s-node1   <none>           <none>

3.kubernetes预定义pod标签

可以看到查看node的labels后,有一长串现有标签,这些就是k8s预定义的pod标签,当然也可以只指定标签进一步指定node进行pod调度。其中基本上都是指定的基于设备架构类型和操作系统类型

[root@k8s-master pod]# kubectl get nodes --show-labels
NAME         STATUS   ROLES           AGE   VERSION   LABELS
k8s-master   Ready    control-plane   41m   v1.28.2   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-master,kubernetes.io/os=linux,node-role.kubernetes.io/control-plane=,node.kubernetes.io/exclude-from-external-load-balancers=
k8s-node1    Ready    <none>          40m   v1.28.2   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node1,kubernetes.io/os=linux,name=node1
k8s-node2    Ready    <none>          40m   v1.28.2   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node2,kubernetes.io/os=linux,name=node2

四.Affinity亲和性调度

1.简介

对pod配置亲和性调度,可以解决上面介绍的定向调度存在“没有满足匹配条件的node时,即使其他有可用node,pod也无法成功调度”的问题,亲和性调度会先选择满足条件的node,若没有,也可以调度到其他可用node上

2.三类Affinity

(1)nodeAffinity,node亲和性

a.以node为主体,针对pod调度到node问题

b.两种表达方式和可配置

[root@k8s-master pod]# kubectl explain pod.spec.affinity.nodeAffinity
KIND:       Pod
VERSION:    v1
​
  preferredDuringSchedulingIgnoredDuringExecution   <[]PreferredSchedulingTerm>
    #强调优先满足指定规则,可以理解为尽量满足该规则,实在是不满足再为其分配其他可用node,调度器尝试但不强求调度pod到node上,若有多个优先级规则那么可以设置权值来调整调度顺序
  - weight: 1  #权值,越小优先级越高,1-100
    preference:   #节点选择器,关联weight
      matchExpressions:   #按照节点的标签来匹配选择目标
      - key: kubernetes.io/arch  #同上
        operator: In
        values:
        - amd64
  requiredDuringSchedulingIgnoredDuringExecution    <NodeSelector>
    #必须满足指定的规则才能调度pod至node上,硬性条件
    nodeSelectorTerms:   #节点选择列表
    - matchExpressions:   #或matchFields,当然比较推荐matchExpressions,表示按照节点的标签来匹配选择调度目标
      - key: kubernetes.io/hostname   #标签的键
        operator: In  
        #operator关系符,可以是In(属于)、NotIn(不属于)、Exists(存在)、DoseNotExists(不存在)、Gt(大于)、Lt(小于)
        values:   #标签的值
        - k8s-node1

c.案例演示

下例要求必须运行在hostname为k8s-node1的节点,尽量运行在架构为amd64的设备

[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME        READY   STATUS    RESTARTS   AGE   IP             NODE        NOMINATED NODE   READINESS GATES
nginx-pod   1/1     Running   0          9s    10.244.36.82   k8s-node1   <none>           <none>
[root@k8s-master pod]# cat myaffinity.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: nginx-pod
  namespace: myns
spec:
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: kubernetes.io/hostname
            operator: In
            values:
            - k8s-node1
      preferredDuringSchedulingIgnoredDuringExecution:
      - weight: 1
        preference:
          matchExpressions:
          - key: kubernetes.io/arch
            operator: In
            values:
            - amd64
  containers:
    - name: nginx-container
      image: nginx:latest
      ports:
      - name: nginx-port
        containerPort: 80

若将硬限制处改为不存在的node,那么会调度错误

apiVersion: v1
kind: Pod
metadata:
  name: nginx-pod
  namespace: myns
spec:
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: kubernetes.io/hostname
            operator: In
            values:
            - k8s-node3
      #preferredDuringSchedulingIgnoredDuringExecution:
      #- weight: 1
        #preference:
          #matchExpressions:
          #- key: kubernetes.io/arch
            #operator: In
            #values:
            #- amd64
  containers:
    - name: nginx-container
      image: nginx:latest
      ports:
      - name: nginx-port
        containerPort: 80
​
[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME        READY   STATUS    RESTARTS   AGE   IP       NODE     NOMINATED NODE   READINESS GATES
nginx-pod   0/1     Pending   0          6s    <none>   <none>   <none>           <none>
[root@k8s-master pod]# kubectl describe pod nginx-pod -n myns | tail -5
                             node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
  Type     Reason            Age   From               Message
  ----     ------            ----  ----               -------
  Warning  FailedScheduling  25s   default-scheduler  0/3 nodes are available: 1 node(s) had untolerated taint {node-role.kubernetes.io/control-plane: }, 2 node(s) didn't match Pod's node affinity/selector. preemption: 0/3 nodes are available: 3 Preemption is not helpful for scheduling..

若将软限制更改为一个不存在的node,那么会自动挑选一个可用node接受调度任务

[root@k8s-master pod]# cat myaffinity.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: nginx-pod
  namespace: myns
spec:
  affinity:
    nodeAffinity:
      #requiredDuringSchedulingIgnoredDuringExecution:
        #nodeSelectorTerms:
        #- matchExpressions:
          #- key: kubernetes.io/hostname
            #operator: In
            #values:
            #- k8s-node3
      preferredDuringSchedulingIgnoredDuringExecution:
      - weight: 1
        preference:
          matchExpressions:
          - key: kubernetes.io/hostname
            operator: In
            values:
            - k8s-node3
  containers:
    - name: nginx-container
      image: nginx:latest
      ports:
      - name: nginx-port
        containerPort: 80
​
[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME        READY   STATUS    RESTARTS   AGE   IP             NODE        NOMINATED NODE   READINESS GATES
nginx-pod   1/1     Running   0          8s    10.244.36.83   k8s-node1   <none>           <none>

d.注意事项

定向调度nodeSelector和nodeAffinity可以同时配置,但需要两者的条件都满足才能正常运行调度

pod被调度到node后node的标签发生了改变不符合亲和性调度匹配规则时,系统将忽略此变化(requiredDuringSchedulingIgnoredDuringExecution表达方式中的IgnoredDuringExecution就表示这个意思)

定义了多个nodeSelectorTerms时满足一个即可

定义了多个matchExpressions时需要都满足才成功

(2)podAffinity,pod亲和性

a.以pod为主体,针对pod和pod部署在同一个拓扑的问题

为减少网络通信性能损耗,多用于两个pod应用频繁交互互动的情况

b.两种表达方式和可配置

[root@k8s-master pod]# kubectl explain  pod.spec.affinity.podAffinity
KIND:       Pod
VERSION:    v1
​
FIELDS:
  preferredDuringSchedulingIgnoredDuringExecution   <[]WeightedPodAffinityTerm>
    #软限制,尽量满足
  - weight: 3        
      podAffinityTerm:   #选项
        labelSelector:
          matchExpressions:
          - key: kubernetes.io/arch
            operator: In
            values:
            - amd64
        topologyKey: kubernetes.io/arch 
​
  requiredDuringSchedulingIgnoredDuringExecution    <[]PodAffinityTerm>
    #硬限制,必须满足
  - labelSelector:   #标签选择器
    matchExpressions:   #同nodeAffinity
    - key: name
      operator: In
      values:
      - su1
    namespaces: xxx     #指定参照pod的名称空间
    topologyKey: kubernetes.io/hostname  #指定调度作用域,kubernetes.io/hostname(以Node节点为区分范围),kubernetes.io/os(以Node节点的操作系统类型来区分)等
    #matchLabels    可指定多个matchExpressions内容

c.案例演示

下例要求将两个nginx的pod运行在同一个node,并且是以hostname为调度作用域

[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME        READY   STATUS    RESTARTS   AGE   IP             NODE        NOMINATED NODE   READINESS GATES
my-nginx1   1/1     Running   0          9s    10.244.36.80   k8s-node1   <none>           <none>
my-nginx2   1/1     Running   0          9s    10.244.36.81   k8s-node1   <none>           <none>
[root@k8s-master pod]# cat twopod.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: my-nginx1
  labels:
    name: su1
  namespace: myns
spec:
  containers:
  - name: my-nginx1
    image: nginx
​
---
​
apiVersion: v1
kind: Pod
metadata: 
  name: my-nginx2
  labels:
    name: su2
  namespace: myns
spec:
  containers:
  - name: my-nginx2
    image: nginx
  affinity:
    podAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
      - labelSelector:
          matchExpressions:
          - key: name
            operator: In
            values:
            - su1
        topologyKey: kubernetes.io/hostname
      preferredDuringSchedulingIgnoredDuringExecution:
      - weight: 3        
        podAffinityTerm:
          labelSelector:
            matchExpressions:
            - key: kubernetes.io/arch
              operator: In
              values:
              - amd64
          topologyKey: kubernetes.io/arch 

  (3)podAntiAffinity,pod反亲和性

a.以pod为主体,针对pod和pod不能部署在同一个拓扑的问题

为提高服务可用性,多用于pod应用有多个副本,让应用分布到各个node上的情况,用法和podAffinity一致

b.案例演示

简单测试下效果就将上面podAffinity例子改为podAntiAffinity即可

[root@k8s-master pod]# cat twopod.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: my-nginx1
  labels:
    name: su1
  namespace: myns
spec:
  containers:
  - name: my-nginx1
    image: nginx
​
---
​
apiVersion: v1
kind: Pod
metadata: 
  name: my-nginx2
  labels:
    name: su2
  namespace: myns
spec:
  containers:
  - name: my-nginx2
    image: nginx
  affinity:
    podAntiAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
      - labelSelector:
          matchExpressions:
          - key: name
            operator: In
            values:
            - su1
        topologyKey: kubernetes.io/hostname
      preferredDuringSchedulingIgnoredDuringExecution:
      - weight: 3        
        podAffinityTerm:
          labelSelector:
            matchExpressions:
            - key: kubernetes.io/arch
              operator: In
              values:
              - amd64
          topologyKey: kubernetes.io/arch 
          
[root@k8s-master pod]# kubectl get pods -n myns -o wide   #已将其运行在其他node
NAME        READY   STATUS    RESTARTS   AGE   IP               NODE        NOMINATED NODE   READINESS GATES
my-nginx1   1/1     Running   0          9s    10.244.36.84     k8s-node1   <none>           <none>
my-nginx2   1/1     Running   0          9s    10.244.169.132   k8s-node2   <none>           <none>

五.taint污点调度和Toleration容忍调度

1.污点调度(node方)

也就是在node上添加一个taint属性,来标识此node是否允许pod调度

(1)三个污点策略

PreferNoSchedule

能不来你就别来了,但是要是没有其他node可以调度了可以过来

NoSchedule

新pod就别来了,我现在只负责我的老pod了

NoExecute

老pod我也不负责了,新的我也调度不了

(2)添加污点/查看污点/去除污点

#添加
kubectl taint nodes nodename 键=值:污点策略
#查看
kubectl describe nodes nodename | grep Taints
#删除
kubectl taint nodes nodename 键:污点策略-
#删除所有
kubectl taint nodes nodename 键-

(3)案例演示

模拟目前集群中只有k8s-node1一个node可用,在该node上演示污点调度的三种模式效果

[root@k8s-master pod]# kubectl get nodes -o wide
NAME         STATUS   ROLES           AGE     VERSION   INTERNAL-IP     EXTERNAL-IP   OS-IMAGE                KERNEL-VERSION           CONTAINER-RUNTIME
k8s-master   Ready    control-plane   8m32s   v1.28.2   192.168.2.150   <none>        CentOS Linux 7 (Core)   3.10.0-1160.el7.x86_64   containerd://1.6.24
k8s-node1    Ready    <none>          7m42s   v1.28.2   192.168.2.151   <none>        CentOS Linux 7 (Core)   3.10.0-1160.el7.x86_64   containerd://1.6.24

a.PreferNoSchedule

可以看到设置PreferNoSchedule后在无其他可用节点时仍然mytaint1可以被调度上来

[root@k8s-master pod]# kubectl taint nodes k8s-node1 status=taint:PreferNoSchedule
node/k8s-node1 tainted
[root@k8s-master pod]# kubectl run mytaint1 --image=nginx -n myns
pod/mytaint1 created
[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME       READY   STATUS    RESTARTS   AGE   IP             NODE        NOMINATED NODE   READINESS GATES
mytaint1   1/1     Running   0          9s    10.244.36.66   k8s-node1   <none>           <none>
[root@k8s-master pod]# kubectl taint nodes k8s-node1 status:PreferNoSchedule-
node/k8s-node1 untainted

b.NoSchedule

可以看到设置NoSchedule后,新pod无法被调度,但不会影响原有的mytaint1

[root@k8s-master pod]# kubectl taint nodes k8s-node1 status=taint:NoSchedule
node/k8s-node1 tainted
[root@k8s-master pod]# kubectl run mytaint2 --image=nginx -n myns
pod/mytaint2 created
[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME       READY   STATUS    RESTARTS   AGE     IP             NODE        NOMINATED NODE   READINESS GATES
mytaint1   1/1     Running   0          2m54s   10.244.36.66   k8s-node1   <none>           <none>
mytaint2   0/1     Pending   0          9s      <none>         <none>      <none>           <none>
[root@k8s-master pod]# kubectl taint nodes k8s-node1 status:NoSchedule-
node/k8s-node1 untainted

c.NoExecute

可以看到在设置NoExecute后,node上原有的pod都被剔除,只剩mytaint3,他同样也无法被调度

[root@k8s-master pod]# kubectl taint nodes k8s-node1 status=taint:NoExecute
node/k8s-node1 tainted
[root@k8s-master pod]# kubectl run mytaint3 --image=nginx -n myns
pod/mytaint3 created
[root@k8s-master pod]# kkubec get pods -n myns -o wide
-bash: kkubec: command not found
[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME       READY   STATUS    RESTARTS   AGE   IP       NODE     NOMINATED NODE   READINESS GATES
mytaint3   0/1     Pending   0          16s   <none>   <none>   <none>           <none>

2.容忍调度(pod方)

就是实在需要将pod调度到污点node上,可以利用容忍调度

(1)容忍的配置参数

[root@k8s-master pod]# kubectl explain pod.spec.tolerations
KIND:       Pod
VERSION:    v1
​
FIELDS:
  effect    <string>
    #指定对应想要调度node的污点策略,若是空则匹配所有策略,必须和目标node的策略一致
  key   <string>
    #指定对应想要调度node的污点键,必须和目标node的污点键一致
  operator  <string>
    #运算符,只支持Equal和Exists,默认Exists
  tolerationSeconds <integer>
    #容忍时间设置pod在node上的存活时间,只有在污点策略为NoExecute才生效
  value <string>
    #指定对应想要调度node的污点键对应的值,必须和目标node的污点键值一致

(2)案例演示

在污点为“status=taint:NoExecute”的k8s-node1上成功调度my-to

[root@k8s-master pod]# kubectl describe node k8s-node1 | grep Taints
Taints:             status=taint:NoExecute
​
[root@k8s-master pod]# cat mytoleration.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: my-to
  namespace: myns
spec:
  containers:
  - name: my-nginx
    image: nginx
  tolerations:
  - key: status
    operator: Equal
    value: taint
    effect: NoExecute
    
[root@k8s-master pod]# kubectl get pods -n myns -o wide
NAME    READY   STATUS    RESTARTS   AGE     IP             NODE        NOMINATED NODE   READINESS GATES
my-to   1/1     Running   0          6m48s   10.244.36.68   k8s-node1   <none>           <none>

 

 

 

 

原文地址:https://blog.csdn.net/weixin_64334766/article/details/134596464

本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。

如若转载,请注明出处:http://www.7code.cn/show_1186.html

如若内容造成侵权/违法违规/事实不符,请联系代码007邮箱:suwngjj01@126.com进行投诉反馈,一经查实,立即删除

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注