k8s（kubernetes）拉取本地镜像部署节点

2023-07-16 04:22| 来源: 网络整理| 查看: 265

本人在虚拟机上部署了1master、2node的集群，想要使用本地打包的镜像hyperledger/fabric-orderer：1.4.3，一开始不清楚原因，只在master上安装了：

docker load -i fabric-orderer.tar

随后查看：

[root@k8s-master my-test]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE hyperledger/fabric-orderer 1.4.3 b666a6ebbe09 17 months ago 173 MB

然后创建测试fabric-orderer.yaml文件：

--- apiVersion: extensions/v1beta1 kind: Deployment metadata: name: my-test spec: replicas: 2 template: metadata: labels: k8s-app: fabric-orderer spec: containers: - name: fabric-orderer-select-node1 image: hyperledger/fabric-orderer:1.4.3 imagePullPolicy: Never ports: - containerPort: 80

随后创建该deployment：

[root@k8s-master my-test]# kubectl create -f fabric-orderer.yaml deployment "my-test" created

但是查看dashboard发现一片飘红，于是用命令查看：

[root@k8s-master my-test]# kubectl get pods --all-namespaces NAMESPACE NAME READY STATUS RESTARTS AGE default my-test-66cf46b4d-cn7qr 0/1 ErrImageNeverPull 0 19s default my-test-66cf46b4d-xgkqd 0/1 ErrImageNeverPull 0 19s

再具体一点：

[root@k8s-master my-test]# kubectl describe pod my-test-66cf46b4d-cn7qr Name: my-test-66cf46b4d-cn7qr Namespace: default Node: k8s-node-2/192.168.203.203 Start Time: Tue, 26 Jan 2021 15:42:28 +0800 Labels: k8s-app=fabric-orderer pod-template-hash=227902608 Annotations: Status: Pending IP: Controlled By: ReplicaSet/my-test-66cf46b4d Containers: fabric-orderer-demo: Container ID: Image: fabric-orderer Image ID: Port: State: Waiting Reason: ErrImageNeverPull Ready: False Restart Count: 0 Environment: Mounts: /var/run/secrets/kubernetes.io/serviceaccount from default-token-bzrrv (ro) Conditions: Type Status Initialized True Ready False PodScheduled True Volumes: default-token-bzrrv: Type: Secret (a volume populated by a Secret) SecretName: default-token-bzrrv Optional: false QoS Class: BestEffort Node-Selectors: Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s node.kubernetes.io/unreachable:NoExecute for 300s Events: Type Reason Age From Message ---- ------ ---- ---- ------- Normal Scheduled 52s default-scheduler Successfully assigned my-test-66cf46b4d-cn7qr to k8s-node-2 Normal SuccessfulMountVolume 51s kubelet, k8s-node-2 MountVolume.SetUp succeeded for volume "default-token-bzrrv" Warning ErrImageNeverPull 32s (x8 over 49s) kubelet, k8s-node-2 Container image "fabric-orderer" is not present with pull policy of Never Warning Failed 32s (x8 over 49s) kubelet, k8s-node-2 Error: ErrImageNeverPull Normal SandboxChanged 31s (x8 over 49s) kubelet, k8s-node-2 Pod sandbox changed, it will be killed and re-created.

可以从最下面的Events看到两个warning，尤其是第一个，发现我的fabric-orderer没有部署上，理所应当的拉取不了，经过一顿查询：参考，每一个要使用该镜像的节点上都要事先docker load上该镜像，于是在k8s-node-1和k8s-node-2上：

docker load -i fabric-orderer.tar

随后再次：

[root@k8s-master my-test]# kubectl create -f fabric-orderer.yaml deployment "my-test" created

这次成功了，当然这只是临时的用法，因为这样意味着你要把所有用到的镜像都打包好，部署到每个节点上。

不怕麻烦而且条件允许的话，还是创建一个本地镜像仓库为好（一劳永逸）。

【本文地址】

公司简介

联系我们