K8s 镜像缓存管理 kube-fledged 认知

写在前面

  • 博文内容为K8s 镜像缓存管理 kube-fledged 认知
  • 内容涉及：
    • kube-fledged 简单介绍
    • 部署以及基本使用
  • 理解不足小伙伴帮忙指正

不必太纠结于当下，也不必太忧虑未来，当你经历过一些事情的时候，眼前的风景已经和从前不一样了。——村上春树

简单介绍

我们知道 k8s 上的容器调度需要在调度的节点行拉取当前容器的镜像，在一些特殊场景中，

  • 需要快速启动和/或扩展的应用程序。例如，由于数据量激增，执行实时数据处理的应用程序需要快速扩展。
  • 镜像比较庞大，涉及多个版本，节点存储有限，需要动态清理不需要的镜像
  • 无服务器函数通常需要在几分之一秒内立即对传入事件和启动容器做出反应。
  • 在边缘设备上运行的 IoT 应用程序，需要容忍边缘设备和镜像镜像仓库之间的间歇性网络连接。
  • 如果需要从专用仓库中拉取镜像，并且无法授予每个人从此镜像仓库拉取镜像的访问权限，则可以在群集的节点上提供镜像。
  • 如果集群管理员或操作员需要对应用程序进行升级，并希望事先验证是否可以成功拉取新镜像。

kube-fledged 是一个 kubernetes operator，用于直接在 Kubernetes 集群的 worker 节点上创建和管理容器镜像缓存。它允许用户定义镜像列表以及这些镜像应缓存到哪些工作节点上（即拉取）。因此，应用程序 Pod 几乎可以立即启动，因为不需要从镜像仓库中提取镜像。

kube-fledged 提供了 CRUD API 来管理镜像缓存的生命周期，并支持多个可配置的参数，可以根据自己的需要自定义功能。

Kubernetes 具有内置的镜像垃圾回收机制。节点中的 kubelet 会定期检查磁盘使用率是否达到特定阈值（可通过标志进行配置）。一旦达到这个阈值，kubelet 会自动删除节点中所有未使用的镜像。

需要在建议的解决方案中实现自动和定期刷新机制。如果镜像缓存中的镜像被 kubelet 的 gc 删除，下一个刷新周期会将已删除的镜像拉入镜像缓存中。这可确保镜像缓存是最新的。

设计流程

https://github.com/senthilrch/kube-fledged/blob/master/docs/kubefledged-architecture.png

部署 kube-fledged

Helm 方式部署

──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$mkdir  kube-fledged
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$cd kube-fledged
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$export KUBEFLEDGED_NAMESPACE=kube-fledged
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$kubectl create namespace ${KUBEFLEDGED_NAMESPACE}
namespace/kube-fledged created
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$helm repo add kubefledged-charts https://senthilrch.github.io/kubefledged-charts/
"kubefledged-charts" has been added to your repositories
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$helm repo update
Hang tight while we grab the latest from your chart repositories...
...Successfully got an update from the "kubefledged-charts" chart repository
...Successfully got an update from the "kubescape" chart repository
...Successfully got an update from the "rancher-stable" chart repository
...Successfully got an update from the "skm" chart repository
...Successfully got an update from the "openkruise" chart repository
...Successfully got an update from the "awx-operator" chart repository
...Successfully got an update from the "botkube" chart repository
Update Complete. ⎈Happy Helming!⎈

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$helm install --verify kube-fledged kubefledged-charts/kube-fledged -n ${KUBEFLEDGED_NAMESPACE} --wait

实际部署中发现，由于网络问题，chart 无法下载，所以通过 make deploy-using-yaml 使用 yaml 方式部署

Yaml 文件部署

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$git clone https://github.com/senthilrch/kube-fledged.git
正克隆到 'kube-fledged'...
remote: Enumerating objects: 10613, done.
remote: Counting objects: 100% (1501/1501), done.
remote: Compressing objects: 100% (629/629), done.
remote: Total 10613 (delta 845), reused 1357 (delta 766), pack-reused 9112
接收对象中: 100% (10613/10613), 34.58 MiB | 7.33 MiB/s, done.
处理 delta 中: 100% (4431/4431), done.
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$ls
kube-fledged
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$cd kube-fledged/
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged]
└─$make deploy-using-yaml
kubectl apply -f deploy/kubefledged-namespace.yaml

第一次部署，发现镜像拉不下来

┌──[root@vms100.liruilongs.github.io]-[~]
└─$kubectl get all -n kube-fledged
NAME                                               READY   STATUS                  RESTARTS         AGE
pod/kube-fledged-controller-df69f6565-drrqg        0/1     CrashLoopBackOff        35 (5h59m ago)   21h
pod/kube-fledged-webhook-server-7bcd589bc4-b7kg2   0/1     Init:CrashLoopBackOff   35 (5h58m ago)   21h
pod/kubefledged-controller-55f848cc67-7f4rl        1/1     Running                 0                21h
pod/kubefledged-webhook-server-597dbf4ff5-l8fbh    0/1     Init:CrashLoopBackOff   34 (6h ago)      21hNAME                                  TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)    AGE
service/kube-fledged-webhook-server   ClusterIP   10.100.194.199   <none>        3443/TCP   21h
service/kubefledged-webhook-server    ClusterIP   10.101.191.206   <none>        3443/TCP   21hNAME                                          READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/kube-fledged-controller       0/1     1            0           21h
deployment.apps/kube-fledged-webhook-server   0/1     1            0           21h
deployment.apps/kubefledged-controller        0/1     1            0           21h
deployment.apps/kubefledged-webhook-server    0/1     1            0           21hNAME                                                     DESIRED   CURRENT   READY   AGE
replicaset.apps/kube-fledged-controller-df69f6565        1         1         0       21h
replicaset.apps/kube-fledged-webhook-server-7bcd589bc4   1         1         0       21h
replicaset.apps/kubefledged-controller-55f848cc67        1         1         0       21h
replicaset.apps/kubefledged-webhook-server-597dbf4ff5    1         1         0       21h
┌──[root@vms100.liruilongs.github.io]-[~]
└─$

这里我们找一下要拉取的镜像

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$cat *.yaml | grep image:- image: senthilrch/kubefledged-controller:v0.10.0- image: senthilrch/kubefledged-webhook-server:v0.10.0- image: senthilrch/kubefledged-webhook-server:v0.10.0

单独拉取一些，当前使用 ansible 在所有工作节点批量操作

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$ansible k8s_node -m shell -a "docker pull docker.io/senthilrch/kubefledged-cri-client:v0.10.0" -i host.yaml

其他相关的镜像都拉取一下

操作完成之后容器状态全部正常

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$kubectl -n kube-fledged get all
NAME                                               READY   STATUS    RESTARTS   AGE
pod/kube-fledged-controller-df69f6565-wdb4g        1/1     Running   0          13h
pod/kube-fledged-webhook-server-7bcd589bc4-j8xxp   1/1     Running   0          13h
pod/kubefledged-controller-55f848cc67-klxlm        1/1     Running   0          13h
pod/kubefledged-webhook-server-597dbf4ff5-ktbsh    1/1     Running   0          13hNAME                                  TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)    AGE
service/kube-fledged-webhook-server   ClusterIP   10.100.194.199   <none>        3443/TCP   36h
service/kubefledged-webhook-server    ClusterIP   10.101.191.206   <none>        3443/TCP   36hNAME                                          READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/kube-fledged-controller       1/1     1            1           36h
deployment.apps/kube-fledged-webhook-server   1/1     1            1           36h
deployment.apps/kubefledged-controller        1/1     1            1           36h
deployment.apps/kubefledged-webhook-server    1/1     1            1           36hNAME                                                     DESIRED   CURRENT   READY   AGE
replicaset.apps/kube-fledged-controller-df69f6565        1         1         1       36h
replicaset.apps/kube-fledged-webhook-server-7bcd589bc4   1         1         1       36h
replicaset.apps/kubefledged-controller-55f848cc67        1         1         1       36h
replicaset.apps/kubefledged-webhook-server-597dbf4ff5    1         1         1       36h

验证是否安装成功

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged]
└─$kubectl get pods -n kube-fledged -l app=kubefledged
NAME                                          READY   STATUS    RESTARTS   AGE
kubefledged-controller-55f848cc67-klxlm       1/1     Running   0          16h
kubefledged-webhook-server-597dbf4ff5-ktbsh   1/1     Running   0          16h
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged]
└─$kubectl get imagecaches -n kube-fledged
No resources found in kube-fledged namespace.

使用 kubefledged

创建镜像缓存对象

根据 Demo 文件，创建镜像缓存对象

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged]
└─$cd deploy/
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$cat kubefledged-imagecache.yaml
---
apiVersion: kubefledged.io/v1alpha2
kind: ImageCache
metadata:# Name of the image cache. A cluster can have multiple image cache objectsname: imagecache1namespace: kube-fledged# The kubernetes namespace to be used for this image cache. You can choose a different namepace as per your preferencelabels:app: kubefledgedkubefledged: imagecache
spec:# The "cacheSpec" field allows a user to define a list of images and onto which worker nodes those images should be cached (i.e. pre-pulled).cacheSpec:# Specifies a list of images (nginx:1.23.1) with no node selector, hence these images will be cached in all the nodes in the cluster- images:- ghcr.io/jitesoft/nginx:1.23.1# Specifies a list of images (cassandra:v7 and etcd:3.5.4-0) with a node selector, hence these images will be cached only on the nodes selected by the node selector- images:- us.gcr.io/k8s-artifacts-prod/cassandra:v7- us.gcr.io/k8s-artifacts-prod/etcd:3.5.4-0nodeSelector:tier: backend# Specifies a list of image pull secrets to pull images from private repositories into the cacheimagePullSecrets:- name: myregistrykey

官方的 Demo 中对应的 镜像拉取不下来，所以换一下

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$docker pull us.gcr.io/k8s-artifacts-prod/cassandra:v7
Error response from daemon: Get "https://us.gcr.io/v2/": net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$

为了测试选择器标签的使用,我们找一个节点的标签单独做镜像缓存

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$kubectl get nodes  --show-labels

同时我们直接从公有仓库拉取镜像，所以不需要 imagePullSecrets 对象

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$vim kubefledged-imagecache.yaml

修改后的 yaml 文件

  • 添加了一个所有节点的 liruilong/my-busybox:latest 镜像缓存
  • 添加了一个 kubernetes.io/hostname: vms105.liruilongs.github.io 对应标签选择器的 liruilong/hikvision-sdk-config-ftp:latest 镜像缓存
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$cat kubefledged-imagecache.yaml
---
apiVersion: kubefledged.io/v1alpha2
kind: ImageCache
metadata:# Name of the image cache. A cluster can have multiple image cache objectsname: imagecache1namespace: kube-fledged# The kubernetes namespace to be used for this image cache. You can choose a different namepace as per your preferencelabels:app: kubefledgedkubefledged: imagecache
spec:# The "cacheSpec" field allows a user to define a list of images and onto which worker nodes those images should be cached (i.e. pre-pulled).cacheSpec:# Specifies a list of images (nginx:1.23.1) with no node selector, hence these images will be cached in all the nodes in the cluster- images:- liruilong/my-busybox:latest# Specifies a list of images (cassandra:v7 and etcd:3.5.4-0) with a node selector, hence these images will be cached only on the nodes selected by the node selector- images:- liruilong/hikvision-sdk-config-ftp:latestnodeSelector:kubernetes.io/hostname: vms105.liruilongs.github.io# Specifies a list of image pull secrets to pull images from private repositories into the cache#imagePullSecrets:#- name: myregistrykey
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$

直接创建报错了

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$kubectl create -f kubefledged-imagecache.yaml
Error from server (InternalError): error when creating "kubefledged-imagecache.yaml": Internal error occurred: failed calling webhook "validate-image-cache.kubefledged.io": failed to call webhook: Post "https://kubefledged-webhook-server.kube-fledged.svc:3443/validate-image-cache?timeout=1s": x509: certificate signed by unknown authority (possibly because of "crypto/rsa: verification error" while trying to verify candidate authority certificate "kubefledged.io")
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$kubectl get imagecaches -n kube-fledged
No resources found in kube-fledged namespace.
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$

解决办法，删除对应的对象，重新创建

我在当前项目的一个 issues 下面找到了解决办法 https://github.com/senthilrch/kube-fledged/issues/76

看起来这是因为 Webhook CA 是硬编码的，但是当 webhook 服务器启动时，会生成一个新的 CA 捆绑包并更新 webhook 配置。当发生另一个部署时，将重新应用原始 CA 捆绑包，并且 Webhook 请求开始失败，直到再次重新启动 Webhook 组件以修补捆绑包init-server

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged]
└─$make remove-kubefledged-and-operator
# Remove kubefledged
kubectl delete -f deploy/kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_v1alpha2_kubefledged_cr.yaml
error: resource mapping not found for name: "kube-fledged" namespace: "kube-fledged" from "deploy/kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_v1alpha2_kubefledged_cr.yaml": no matches for kind "KubeFledged" in version "charts.helm.kubefledged.io/v1alpha2"
ensure CRDs are installed first

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged]
└─$make deploy-using-yaml
kubectl apply -f deploy/kubefledged-namespace.yaml
namespace/kube-fledged created
kubectl apply -f deploy/kubefledged-crd.yaml
customresourcedefinition.apiextensions.k8s.io/imagecaches.kubefledged.io unchanged
....................
kubectl rollout status deployment kubefledged-webhook-server -n kube-fledged --watch
Waiting for deployment "kubefledged-webhook-server" rollout to finish: 0 of 1 updated replicas are available...
deployment "kubefledged-webhook-server" successfully rolled out
kubectl get pods -n kube-fledged
NAME                                          READY   STATUS    RESTARTS   AGE
kubefledged-controller-55f848cc67-76c4v       1/1     Running   0          112s
kubefledged-webhook-server-597dbf4ff5-56h6z   1/1     Running   0          66s

重新创建缓存对象，创建成功

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$kubectl create -f kubefledged-imagecache.yaml
imagecache.kubefledged.io/imagecache1 created
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$kubectl get imagecaches -n kube-fledged
NAME          AGE
imagecache1   10s
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$

查看当前被纳管的镜像缓存

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$kubectl get imagecaches imagecache1 -n kube-fledged -o json
{"apiVersion": "kubefledged.io/v1alpha2","kind": "ImageCache","metadata": {"creationTimestamp": "2024-03-01T15:08:42Z","generation": 83,"labels": {"app": "kubefledged","kubefledged": "imagecache"},"name": "imagecache1","namespace": "kube-fledged","resourceVersion": "20169836","uid": "3a680a57-d8ab-444f-b9c9-4382459c5c72"},"spec": {"cacheSpec": [{"images": ["liruilong/my-busybox:latest"]},{"images": ["liruilong/hikvision-sdk-config-ftp:latest"],"nodeSelector": {"kubernetes.io/hostname": "vms105.liruilongs.github.io"}}]},"status": {"completionTime": "2024-03-02T01:06:47Z","message": "All requested images pulled succesfully to respective nodes","reason": "ImageCacheRefresh","startTime": "2024-03-02T01:05:33Z","status": "Succeeded"}
}
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged]
└─$

通过 ansible 来验证

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$ansible all -m shell -a "docker images | grep liruilong/my-busybox" -i host.yaml
192.168.26.102 | CHANGED | rc=0 >>
liruilong/my-busybox                                                        latest    497b83a63aad   11 months ago   1.24MB
192.168.26.101 | CHANGED | rc=0 >>
liruilong/my-busybox                                                        latest    497b83a63aad   11 months ago   1.24MB
192.168.26.103 | CHANGED | rc=0 >>
liruilong/my-busybox                                                        latest    497b83a63aad   11 months ago   1.24MB
192.168.26.105 | CHANGED | rc=0 >>
liruilong/my-busybox                                                        latest    497b83a63aad   11 months ago   1.24MB
192.168.26.100 | CHANGED | rc=0 >>
liruilong/my-busybox                                                        latest    497b83a63aad   11 months ago   1.24MB
192.168.26.106 | CHANGED | rc=0 >>
liruilong/my-busybox                                                        latest    497b83a63aad   11 months ago   1.24MB
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$ansible all -m shell -a "docker images | grep liruilong/hikvision-sdk-config-ftp" -i host.yaml
192.168.26.102 | FAILED | rc=1 >>
non-zero return code
192.168.26.100 | FAILED | rc=1 >>
non-zero return code
192.168.26.103 | FAILED | rc=1 >>
non-zero return code
192.168.26.105 | CHANGED | rc=0 >>
liruilong/hikvision-sdk-config-ftp                                          latest            a02cd03b4342   4 months ago    830MB
192.168.26.101 | FAILED | rc=1 >>
non-zero return code
192.168.26.106 | FAILED | rc=1 >>
non-zero return code
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$

开启自动刷新

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$kubectl annotate imagecaches imagecache1 -n kube-fledged kubefledged.io/refresh-imagecache=
imagecache.kubefledged.io/imagecache1 annotated
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$

添加镜像缓存

添加一个新的镜像缓存

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$kubectl get imagecaches.kubefledged.io  -n kube-fledged  imagecache1 -o json
{"apiVersion": "kubefledged.io/v1alpha2","kind": "ImageCache","metadata": {"creationTimestamp": "2024-03-01T15:08:42Z","generation": 92,"labels": {"app": "kubefledged","kubefledged": "imagecache"},"name": "imagecache1","namespace": "kube-fledged","resourceVersion": "20175233","uid": "3a680a57-d8ab-444f-b9c9-4382459c5c72"},"spec": {"cacheSpec": [{"images": ["liruilong/my-busybox:latest","liruilong/jdk1.8_191:latest"]},{"images": ["liruilong/hikvision-sdk-config-ftp:latest"],"nodeSelector": {"kubernetes.io/hostname": "vms105.liruilongs.github.io"}}]},"status": {"completionTime": "2024-03-02T01:43:32Z","message": "All requested images pulled succesfully to respective nodes","reason": "ImageCacheUpdate","startTime": "2024-03-02T01:40:34Z","status": "Succeeded"}
}
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$

通过 ansible 确认

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$ansible all -m shell -a "docker images | grep liruilong/jdk1.8_191" -i host.yaml
192.168.26.101 | FAILED | rc=1 >>
non-zero return code
192.168.26.100 | FAILED | rc=1 >>
non-zero return code
192.168.26.102 | FAILED | rc=1 >>
non-zero return code
192.168.26.103 | FAILED | rc=1 >>
non-zero return code
192.168.26.105 | FAILED | rc=1 >>
non-zero return code
192.168.26.106 | FAILED | rc=1 >>
non-zero return code
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$ansible all -m shell -a "docker images | grep liruilong/jdk1.8_191" -i host.yaml
192.168.26.101 | CHANGED | rc=0 >>
liruilong/jdk1.8_191                                                        latest    17dbd4002a8c   5 years ago     170MB
192.168.26.102 | CHANGED | rc=0 >>
liruilong/jdk1.8_191                                                        latest    17dbd4002a8c   5 years ago     170MB
192.168.26.100 | CHANGED | rc=0 >>
liruilong/jdk1.8_191                                                        latest    17dbd4002a8c   5 years ago     170MB
192.168.26.103 | CHANGED | rc=0 >>
liruilong/jdk1.8_191                                                        latest                                      17dbd4002a8c   5 years ago     170MB
192.168.26.105 | CHANGED | rc=0 >>
liruilong/jdk1.8_191                                                        latest            17dbd4002a8c   5 years ago     170MB
192.168.26.106 | CHANGED | rc=0 >>
liruilong/jdk1.8_191                                                        latest            17dbd4002a8c   5 years ago     170MB
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$

删除镜像缓存

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$kubectl edit imagecaches imagecache1 -n kube-fledged
imagecache.kubefledged.io/imagecache1 edited
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$kubectl get imagecaches.kubefledged.io  -n kube-fledged  imagecache1 -o json
{"apiVersion": "kubefledged.io/v1alpha2","kind": "ImageCache","metadata": {"creationTimestamp": "2024-03-01T15:08:42Z","generation": 94,"labels": {"app": "kubefledged","kubefledged": "imagecache"},"name": "imagecache1","namespace": "kube-fledged","resourceVersion": "20175766","uid": "3a680a57-d8ab-444f-b9c9-4382459c5c72"},"spec": {"cacheSpec": [{"images": ["liruilong/jdk1.8_191:latest"]},{"images": ["liruilong/hikvision-sdk-config-ftp:latest"],"nodeSelector": {"kubernetes.io/hostname": "vms105.liruilongs.github.io"}}]},"status": {"message": "Image cache is being updated. Please view the status after some time","reason": "ImageCacheUpdate","startTime": "2024-03-02T01:48:03Z","status": "Processing"}
}

通过 Ansible 确认，可以看到无论是 mastere 上的节点还是 work 的节点，对应的镜像缓存都被清理

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$ansible all -m shell -a "docker images | grep liruilong/my-busybox" -i host.yaml
192.168.26.102 | CHANGED | rc=0 >>
liruilong/my-busybox                                                        latest    497b83a63aad   11 months ago   1.24MB
192.168.26.101 | CHANGED | rc=0 >>
liruilong/my-busybox                                                        latest    497b83a63aad   11 months ago   1.24MB
192.168.26.105 | FAILED | rc=1 >>
non-zero return code
192.168.26.100 | CHANGED | rc=0 >>
liruilong/my-busybox                                                        latest    497b83a63aad   11 months ago   1.24MB
192.168.26.103 | FAILED | rc=1 >>
non-zero return code
192.168.26.106 | FAILED | rc=1 >>
non-zero return code
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$ansible all -m shell -a "docker images | grep liruilong/my-busybox" -i host.yaml
192.168.26.105 | FAILED | rc=1 >>
non-zero return code
192.168.26.102 | FAILED | rc=1 >>
non-zero return code
192.168.26.103 | FAILED | rc=1 >>
non-zero return code
192.168.26.101 | FAILED | rc=1 >>
non-zero return code
192.168.26.100 | FAILED | rc=1 >>
non-zero return code
192.168.26.106 | FAILED | rc=1 >>
non-zero return code
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$

这里需要注意如果清除所有的镜像缓存，那么需要把 images 下的数组 写成 “”.

┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$kubectl edit imagecaches imagecache1 -n kube-fledged
imagecache.kubefledged.io/imagecache1 edited
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$ansible all -m shell -a "docker images | grep liruilong/jdk1.8_191" -i host.yaml
192.168.26.102 | FAILED | rc=1 >>
non-zero return code
192.168.26.101 | FAILED | rc=1 >>
non-zero return code
192.168.26.100 | FAILED | rc=1 >>
non-zero return code
192.168.26.105 | FAILED | rc=1 >>
non-zero return code
192.168.26.103 | FAILED | rc=1 >>
non-zero return code
192.168.26.106 | FAILED | rc=1 >>
non-zero return code
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$kubectl get imagecaches.kubefledged.io  -n kube-fledged  imagecache1 -o json
{"apiVersion": "kubefledged.io/v1alpha2","kind": "ImageCache","metadata": {"creationTimestamp": "2024-03-01T15:08:42Z","generation": 98,"labels": {"app": "kubefledged","kubefledged": "imagecache"},"name": "imagecache1","namespace": "kube-fledged","resourceVersion": "20176849","uid": "3a680a57-d8ab-444f-b9c9-4382459c5c72"},"spec": {"cacheSpec": [{"images": [""]},{"images": ["liruilong/hikvision-sdk-config-ftp:latest"],"nodeSelector": {"kubernetes.io/hostname": "vms105.liruilongs.github.io"}}]},"status": {"completionTime": "2024-03-02T01:52:16Z","message": "All cached images succesfully deleted from respective nodes","reason": "ImageCacheUpdate","startTime": "2024-03-02T01:51:47Z","status": "Succeeded"}
}
┌──[root@vms100.liruilongs.github.io]-[~/ansible]
└─$

如果通过下面的方式删除，直接注释调对应的标签

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$cat kubefledged-imagecache.yaml
---
apiVersion: kubefledged.io/v1alpha2
kind: ImageCache
metadata:# Name of the image cache. A cluster can have multiple image cache objectsname: imagecache1namespace: kube-fledged# The kubernetes namespace to be used for this image cache. You can choose a different namepace as per your preferencelabels:app: kubefledgedkubefledged: imagecache
spec:# The "cacheSpec" field allows a user to define a list of images and onto which worker nodes those images should be cached (i.e. pre-pulled).cacheSpec:# Specifies a list of images (nginx:1.23.1) with no node selector, hence these images will be cached in all the nodes in the cluster#- images:#- liruilong/my-busybox:latest# Specifies a list of images (cassandra:v7 and etcd:3.5.4-0) with a node selector, hence these images will be cached only on the nodes selected by the node selector- images:- liruilong/hikvision-sdk-config-ftp:latestnodeSelector:kubernetes.io/hostname: vms105.liruilongs.github.io# Specifies a list of image pull secrets to pull images from private repositories into the cache#imagePullSecrets:#- name: myregistrykey
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$

那么会报下面的错

┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$kubectl edit imagecaches imagecache1 -n kube-fledged
error: imagecaches.kubefledged.io "imagecache1" could not be patched: admission webhook "validate-image-cache.kubefledged.io" denied the request: Mismatch in no. of image lists
You can run `kubectl replace -f /tmp/kubectl-edit-4113815075.yaml` to try this update again.

博文部分内容参考

© 文中涉及参考链接内容版权归原作者所有，如有侵权请告知,如果你认可它不要吝啬星星哦 😃

https://github.com/senthilrch/kube-fledged

© 2018-2024 liruilonger@gmail.com, All rights reserved. 保持署名-非商用-相同方式共享(CC BY-NC-SA 4.0)

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.mzph.cn/news/716417.shtml

如若内容造成侵权/违法违规/事实不符，请联系多彩编程网进行投诉反馈email:809451989@qq.com，一经查实，立即删除！

相关文章

springboot236基于springboot在线课程管理系统的设计与实现

springboot236基于springboot在线课程管理系统的设计与实现

基于SpringBoot在线课程管理系统的设计与实现 摘要 本文首先介绍了在线课程管理系统的现状及开发背景&#xff0c;然后论述了系统的设计目标、系统需求、总体设计方案以及系统的详细设计和实现&#xff0c;最后对在线课程管理系统进行了系统检测并提出了还需要改进的问题。本系…
阅读更多...
Spring Data Redis 使用方式

Spring Data Redis 使用方式

在Java中操作Redis 1. 在Java中操作Redis1.1 Redis的Java客户端1.2 Spring Data Redis使用方式1.2.1 介绍1.2.1 环境搭建1.2.3 操作常见类型数据 1. 在Java中操作Redis 1.1 Redis的Java客户端 Redis 的 Java 客户端很多&#xff0c;常用的几种&#xff1a; JedisLettuceSpri…
阅读更多...
关于Axios接口请求超时处理与重试的方法教程

关于Axios接口请求超时处理与重试的方法教程

在前端开发中&#xff0c;使用Axios作为HTTP客户端库进行接口请求是非常常见的做法。然而&#xff0c;在实际开发中&#xff0c;我们经常会遇到网络不稳定或服务器响应缓慢导致接口请求超时的情况。为了提高用户体验和程序的稳定性&#xff0c;我们需要实现接口请求超时的处理与…
阅读更多...
UnityAPI的学习——Transform类

UnityAPI的学习——Transform类

Transform类继承自Component类&#xff0c;并实现了IEnumberable接口。Transform是GameObject必须拥有得一个组件&#xff0c;用来管理所在GameObject对象的坐标位置、选择角度、和大小缩放。 Transform实现了IEnumberable接口&#xff0c;因此可以在程序中使用foreach()方法快…
阅读更多...
echarts vue 动画效果的水球图、波浪图教程

echarts vue 动画效果的水球图、波浪图教程

1、安装插件 前提是已经安装了echarts&#xff08;我的版本是4.2.1&#xff09; npm install echarts-liquidfill --save 我安装了3.1.0版本的&#xff0c;结果运行时报错"TypeError: wave.ensureState is not a function" 原因&#xff1a;echarts版本和echarts-l…
阅读更多...
miniconda3彻底删除虚拟环境

miniconda3彻底删除虚拟环境

退出虚拟环境&#xff1a;确保您不在要删除的虚拟环境中。如果在&#xff0c;使用命令 conda deactivate 来退出当前激活的虚拟环境。查看虚拟环境列表&#xff1a;运行命令 conda env list 或 conda info -e 来查看所有存在的虚拟环境及其路径。删除虚拟环境&#xff1a;使用命…
阅读更多...
在VMware中安装CentOS 7并配置Docker

在VMware中安装CentOS 7并配置Docker

VMware安装CentOS 7 一、介绍 该文章介绍如何使用启动U盘在虚拟机里面安装系统&#xff0c;虚拟机版本为VMware Workstation 16 pro&#xff0c;Linux版本为CentOS Linux release 7.9.2009 (Core)。 二、安装 1、创建虚拟机 点击创建新的虚拟机 选择典型就可以了&#xf…
阅读更多...
前缀和算法题（区间次方和、小蓝平衡和、大石头的搬运工、最大数组和）

前缀和算法题（区间次方和、小蓝平衡和、大石头的搬运工、最大数组和）

一、前缀和的原理和特点 prefix表示前缀和&#xff0c;前缀和由一个用户输入的数组生成。对于一个数组a[]&#xff08;下标从1开始&#xff09;&#xff0c;我们定义一个前缀和数组prefix[]&#xff0c;满足&#xff1a; prefix有一个重要的特性&#xff0c;可以用于快速生成p…
阅读更多...
WordPress建站入门教程：如何安装本地WordPress网站运行环境？

WordPress建站入门教程：如何安装本地WordPress网站运行环境？

有些站长想要搭建WordPress网站&#xff0c;又担心自己玩不转&#xff0c;白白浪费购买域名和主机空间的费用。像这种情况&#xff0c;最好的做法就是在自己电脑上安装一个WordPress网站运行环境&#xff0c;然后在本地电脑搭建WordPress&#xff0c;等熟悉掌握后再考虑购买域名…
阅读更多...
设计模式学习笔记——抽象工厂模式

设计模式学习笔记——抽象工厂模式

设计模式&#xff08;创建型&#xff09;—— 抽象工厂模式 在工厂模式中&#xff0c;我们为每一个类都设计了一个工厂&#xff0c;以此来获取该类的对象&#xff0c;但缺点就是一旦类多了&#xff0c;工厂就多了&#xff0c;这时候我们可以考虑这些类间是否有关联&#xff0c…
阅读更多...
多输入多输出 | MATLAB实现GWO-Elman灰狼优化循环神经网络多输入多输出预测

多输入多输出 | MATLAB实现GWO-Elman灰狼优化循环神经网络多输入多输出预测

多输入多输出 | MATLAB实现GWO-Elman灰狼优化循环神经网络多输入多输出预测 目录 多输入多输出 | MATLAB实现GWO-Elman灰狼优化循环神经网络多输入多输出预测预测效果基本介绍程序设计往期精彩参考资料 预测效果 基本介绍 Matlab实现GWO-Elman灰狼优化循环神经网络多输入多输出…
阅读更多...
kernel bypass 是什么？

kernel bypass 是什么？

文章目录 一、kernel bypass 是什么二、Kernel Bypass技术优缺点三、Kernel Bypass技术应用领域四、Kernel Bypass的实现方式 一、kernel bypass 是什么 Kernel Bypass是一种技术&#xff0c;旨在通过绕过操作系统核来提高网络数据包处理的性能和降低延迟。它的主要优点是高性能…
阅读更多...
[LeetBook]【学习日记】有序链表合并

[LeetBook]【学习日记】有序链表合并

21. 合并两个有序链表 将两个升序链表合并为一个新的 升序 链表并返回。新链表是通过拼接给定的两个链表的所有节点组成的。 示例 1&#xff1a; 输入&#xff1a;l1 [1,2,4], l2 [1,3,4] 输出&#xff1a;[1,1,2,3,4,4] 示例 2&#xff1a; 输入&#xff1a;l1 [], l2 [] …
阅读更多...
如何在电脑上中恢复已删除的视频

如何在电脑上中恢复已删除的视频

您可以在电脑中恢复已删除的视频&#xff0c;无需任何繁琐的工作。您所需要做的就是阅读本文&#xff0c;了解恢复已删除视频的最佳方法。 一次错误的点击可能会夺走您以视频形式存储的宝贵记忆。嗯&#xff0c;有些视频不适合删除&#xff0c;您希望永远保留它们。失去这些宝…
阅读更多...
如何使用Docker搭建StackEdit编辑器并结合内网穿透实现远程办公

如何使用Docker搭建StackEdit编辑器并结合内网穿透实现远程办公

文章目录 前言1. ubuntu安装VNC2. 设置vnc开机启动3. windows 安装VNC viewer连接工具4. 内网穿透4.1 安装cpolar【支持使用一键脚本命令安装】4.2 创建隧道映射4.3 测试公网远程访问 5. 配置固定TCP地址5.1 保留一个固定的公网TCP端口地址5.2 配置固定公网TCP端口地址5.3 测试…
阅读更多...
优选算法|【双指针】|1089.复写零

优选算法|【双指针】|1089.复写零

目录 题目描述 题目解析 算法原理讲解 代码 题目描述 1089. 复写零 给你一个长度固定的整数数组 arr &#xff0c;请你将该数组中出现的每个零都复写一遍&#xff0c;并将其余的元素向右平移。 注意&#xff1a;请不要在超过该数组长度的位置写入元素。请对输入的数组 就…
阅读更多...
LeetCode受限条件下可到达节点的数目

LeetCode受限条件下可到达节点的数目

题目描述 现有一棵由 n 个节点组成的无向树&#xff0c;节点编号从 0 到 n - 1 &#xff0c;共有 n - 1 条边。 给你一个二维整数数组 edges &#xff0c;长度为 n - 1 &#xff0c;其中 edges[i] [ai, bi] 表示树中节点 ai 和 bi 之间存在一条边。另给你一个整数数组 restr…
阅读更多...
OJ：移除链表元素

OJ：移除链表元素

203. 移除链表元素 - 力扣&#xff08;LeetCode&#xff09; 思路&#xff1a;这个题可以直接在原链表上进行修改&#xff0c;但是修改链表的指向是有点麻烦的&#xff0c;所以我们给两个指针&#xff0c;phead和ptail,这是新链表的两个指针&#xff0c;再给一个指针pcur来遍历…
阅读更多...
Java和JavaScript区别

Java和JavaScript区别

1. Java和javaScript都是面向对象语言 2. 他两除了名字相似之外没有任何关系 3. Java是一种真正的面向对象语言&#xff0c;不管开发什么程序都要设计对象&#xff1b;而JavaScript是种脚本语言&#xff0c;主要实现前端页面的交互&#xff0c;比如验证表单&#xff0c;弹窗提…
阅读更多...
Sqli-labs靶场第12关详解[Sqli-labs-less-12]

Sqli-labs靶场第12关详解[Sqli-labs-less-12]

Sqli-labs-Less-12 #手工注入 post传参了 根据题目看&#xff0c;像一个登录页面&#xff0c;尝试使用布尔型盲注测试能否登录网站 1. Username输入a a" 测试是否会有报错&#xff0c;burp抓包 报错&#xff1a;syntax to use near "a"") and passw…
阅读更多...
最新文章

Copyright @ 2022~2023