单位网站建设管理情况,中国建筑公司排名一览表,大学信息化建设 网站群,市建设局领导名单在《研发工程师玩转Kubernetes——PVC通过storageClassName进行延迟绑定》一文中#xff0c;我们利用Node亲和性#xff0c;让Pod部署在节点ubuntud上。因为Pod使用的PVC可以部署在节点ubuntuc或者ubuntud上#xff0c;而系统为了让Pod可以部署成功#xff0c;则让PVC与Pod…在《研发工程师玩转Kubernetes——PVC通过storageClassName进行延迟绑定》一文中我们利用Node亲和性让Pod部署在节点ubuntud上。因为Pod使用的PVC可以部署在节点ubuntuc或者ubuntud上而系统为了让Pod可以部署成功则让PVC与Pod亲和的ubuntud上的PV绑定。这样Pod在自身节点亲和性和PVC上都满足了条件。 在一些业务场景下我们通过磁盘来保存数据而程序通过数据设置自身状态。如果一旦一个Pod崩溃我们希望新补充的Pod可以延续之前的状态。这个时候我们就可以使用PV的节点亲和性来完成上述调度。
PersistentVolume
下面PersistentVolume配置的节点亲和性要求资源只会在ubuntuc或者ubuntud上创建。
# default_storage_class_pv_ubuntucd.yaml
apiVersion: v1
kind: PersistentVolume
metadata:name: default-storage-class-pv-ubuntucdlabels:volume: lb-default-storage-class-pv
spec:capacity:storage: 1MivolumeMode: FilesystemaccessModes:- ReadWriteOncepersistentVolumeReclaimPolicy: Retainlocal:path: /tmpnodeAffinity:required:nodeSelectorTerms:- matchExpressions:- key: kubernetes.io/hostnameoperator: Invalues:- ubuntuc- ubuntudPersistentVolumeClaim
PVC直接和上述PV绑定。
# default_storage_class_pvc_600k.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:name: default-storage-class-pvc-600k
spec:resources:requests:storage: 600KiaccessModes:- ReadWriteOnceselector:matchLabels:volume: lb-default-storage-class-pvDeployment
下面这个Deployment会创建2个Pod。由于Pod的反亲和性它们被强制要求调度到不同Node上。
# default_deployment_one_on_node.yaml
apiVersion: apps/v1
kind: Deployment
metadata:name: default-pv-app-one-on-node-deployment
spec:replicas: 2selector:matchLabels:app: default-pv-app-one-on-nodetemplate:metadata:labels:app: default-pv-app-one-on-nodespec:containers:- name: default-pv-app-one-on-nodeimage: busyboxcommand: [/bin/sh, -c, if [ -f /tempdir/lockfile ] ! { set -C; 2/dev/null /tempdir/lockfile; }; then tail -f /tempdir/lockfile; else exec 3/tempdir/lockfile; if [ -n \$POD_NAME\ ]; then name$POD_NAME; else name\unknown\; fi; while true; do echo \this is $name.$name write something to lockfile\; echo \$name write something to lockfile\ 3; sleep 1; done; fi] volumeMounts:- name: default-pvc-volumemountPath: /tempdirenv:- name: POD_NAMEvalueFrom:fieldRef:fieldPath: metadata.namevolumes:- name: default-pvc-volumepersistentVolumeClaim:claimName: default-storage-class-pvc-600kaffinity:podAntiAffinity:requiredDuringSchedulingIgnoredDuringExecution:- labelSelector:matchExpressions:- key: appoperator: Invalues:- default-pv-app-one-on-nodetopologyKey: kubernetes.io/hostname实验
查看Pod状态
kubectl get pod -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
default-pv-app-one-on-node-deployment-76bf96bb5-7tzxd 1/1 Running 0 21s 10.1.202.210 ubuntud none none
default-pv-app-one-on-node-deployment-76bf96bb5-mtk67 1/1 Running 0 21s 10.1.43.207 ubuntuc none none可以看到两个Pod分别被调度到ubuntuc和ubuntud上。
调整Pod数量
kubectl scale deployment default-pv-app-one-on-node-deployment --replicas 3kubectl get pod -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
default-pv-app-one-on-node-deployment-76bf96bb5-7tzxd 1/1 Running 0 46s 10.1.202.210 ubuntud none none
default-pv-app-one-on-node-deployment-76bf96bb5-mtk67 1/1 Running 0 46s 10.1.43.207 ubuntuc none none
default-pv-app-one-on-node-deployment-76bf96bb5-9wvq5 0/1 Pending 0 3s none none none none可以看到新创建的Pod处于Pending状态。当前系统还有ubuntua、ubuntud、ubunutue三个节点它们由于不能满足Pod对PVC的要求实际是PVC绑定的PV的要求而没有被调度到。
kubectl get nodesNAME STATUS ROLES AGE VERSION
ubuntud Ready none 2d22h v1.27.4
ubuntuc Ready none 2d23h v1.27.4
ubuntue Ready none 2d22h v1.27.4
ubuntub Ready none 2d23h v1.27.4
ubuntua Ready none 2d23h v1.27.4恢复Pod
kubectl scale deployment default-pv-app-one-on-node-deployment --replicas 2
kubectl delete pod default-pv-app-one-on-node-deployment-76bf96bb5-7tzxd
kubectl get pod -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
default-pv-app-one-on-node-deployment-76bf96bb5-mtk67 1/1 Running 0 2m12s 10.1.43.207 ubuntuc none none
default-pv-app-one-on-node-deployment-76bf96bb5-qqtlm 1/1 Running 0 39s 10.1.202.211 ubuntud none none可以看到Pod被正确的调度到刚被删掉Pod的Node上进而可以继续使用它的PV从而实现服务状态恢复和衔接上的目的。
