Manual Scheduling

在 Kubernetes 中，手动调度框架允许您将 Pod 分配到特定节点，而无需依赖默认调度器。这对于测试、调试或处理特定工作负载非常有用。您可以通过在 Pod 的规范中设置 nodeName 字段来实现手动调度。以下是一个示例：

apiVersion: v1
kind: Pod
metadata:
  name: manual-scheduled-pod
spec:
  containers:
  - name: nginx
    image: nginx
  nodeName: your-node-name

将 your-node-name 替换为您希望 Pod 运行的节点名称。应用此配置后，Kubernetes 会直接将 Pod 放置在指定的节点上。

nodeName

nodeName 是比亲和性或者 nodeSelector 更为直接的形式。nodeName 是 Pod 规约中的一个字段。如果 nodeName 字段不为空，调度器会忽略该 Pod， 而指定节点上的 kubelet 会尝试将 Pod 放到该节点上。 使用 nodeName 规则的优先级会高于使用 nodeSelector 或亲和性与非亲和性的规则。

使用 nodeName 来选择节点的方式有一些局限性：

• 如果所指代的节点不存在，则 Pod 无法运行，而且在某些情况下可能会被自动删除。
• 如果所指代的节点无法提供用来运行 Pod 所需的资源，Pod 会失败， 而其失败原因中会给出是否因为内存或 CPU 不足而造成无法运行。
• 在云环境中的节点名称并不总是可预测的，也不总是稳定的。

警告：nodeName 旨在供自定义调度器或需要绕过任何已配置调度器的高级场景使用。 如果已分配的 Node 负载过重，绕过调度器可能会导致 Pod 失败。 你可以使用节点亲和性或 nodeselector 字段将 Pod 分配给特定 Node，而无需绕过调度器。

下面是一个使用 nodeName 字段的 Pod 规约示例：

apiVersion: v1
kind: Pod
metadata:
  name: nginx
spec:
  containers:
  - name: nginx
    image: nginx
  nodeName: kube-01

上面的 Pod 只能运行在节点 kube-01 之上。

Practice Detail

        Welcome to the KodeKloud Hands-On lab                                                                          
    __ ______  ____  ________ __ __    ____  __  ______ 
   / //_/ __ \/ __ \/ ____/ //_// /   / __ \/ / / / __ \
  / ,< / / / / / / / __/ / ,<  / /   / / / / / / / / / /
 / /| / /_/ / /_/ / /___/ /| |/ /___/ /_/ / /_/ / /_/ / 
/_/ |_\____/_____/_____/_/ |_/_____/\____/\____/_____/  
                                                        
            All rights reserved                                                                                        

controlplane ~ ➜  kubectl create -f nginx.yaml
pod/nginx created

controlplane ~ ➜  kubectl get pods
NAME    READY   STATUS    RESTARTS   AGE
nginx   0/1     Pending   0          8s

controlplane ~ ✖ kubectl describe pod nginx
Name:             nginx
Namespace:        default
Priority:         0
Service Account:  default
Node:             <none>
Labels:           <none>
Annotations:      <none>
Status:           Pending
IP:               
IPs:              <none>
Containers:
  nginx:
    Image:        nginx
    Port:         <none>
    Host Port:    <none>
    Environment:  <none>
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-bvfbc (ro)
Volumes:
  kube-api-access-bvfbc:
    Type:                    Projected (a volume that contains injected data from multiple sources)
    TokenExpirationSeconds:  3607
    ConfigMapName:           kube-root-ca.crt
    ConfigMapOptional:       <nil>
    DownwardAPI:             true
QoS Class:                   BestEffort
Node-Selectors:              <none>
Tolerations:                 node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
                             node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:                      <none>

controlplane ~ ➜  kubectl get pods -n kube-system
NAME                                   READY   STATUS    RESTARTS   AGE
coredns-7484cd47db-mhzfg               1/1     Running   0          10m
coredns-7484cd47db-pq6v7               1/1     Running   0          10m
etcd-controlplane                      1/1     Running   0          10m
kube-apiserver-controlplane            1/1     Running   0          10m
kube-controller-manager-controlplane   1/1     Running   0          10m
kube-proxy-dd6zp                       1/1     Running   0          9m48s
kube-proxy-mtmxb                       1/1     Running   0          10m

controlplane ~ ➜  echo "no scheduler present"
no scheduler present

controlplane ~ ➜  cat nginx.yaml 
---
apiVersion: v1
kind: Pod
metadata:
  name: nginx
spec:
  containers:
  -  image: nginx
     name: nginx


controlplane ~ ➜  vi nginx.yaml

controlplane ~ ➜  cat nginx.yaml 
---
apiVersion: v1
kind: Pod
metadata:
  name: nginx
spec:
  nodeName: node01
  containers:
  -  image: nginx
     name: nginx


controlplane ~ ➜  echo "Manually schedule the pod on node01"
Manually schedule the pod on node01

controlplane ~ ➜  cat nginx.yaml
---
apiVersion: v1
kind: Pod
metadata:
  name: nginx
spec:
  nodeName: node01
  containers:
  -  image: nginx
     name: nginx


controlplane ~ ➜  kubectl replace --force -f nginx.yaml
pod "nginx" deleted
pod/nginx replaced

controlplane ~ ➜  kubectl get pods -o wide
NAME    READY   STATUS    RESTARTS   AGE   IP           NODE     NOMINATED NODE   READINESS GATES
nginx   1/1     Running   0          43s   172.17.1.3   node01   <none>           <none>

controlplane ~ ➜  vi nginx.yaml

controlplane ~ ➜  cat nginx.yaml
---
apiVersion: v1
kind: Pod
metadata:
  name: nginx
spec:
  nodeName: controlplane
  containers:
  -  image: nginx
     name: nginx


controlplane ~ ➜  kubectl replace --force -f nginx.yaml
pod "nginx" deleted
pod/nginx replaced

controlplane ~ ➜  kubectl get pods -o wide
NAME    READY   STATUS    RESTARTS   AGE   IP           NODE           NOMINATED NODE   READINESS GATES
nginx   1/1     Running   0          8s    172.17.0.4   controlplane   <none>           <none>

controlplane ~ ➜  echo "Now schedule the same pod on the controlplane node."
Now schedule the same pod on the controlplane node.

Powered by Moshow@https://zhengkai.blog.csdn.net/