详解 K8s 镜像缓存管理 kube-fledged
- 2024-04-15 广东
本文字数:14505 字
阅读完需:约 48 分钟
本文分享自华为云社区《K8s 镜像缓存管理 kube-fledged 认知》,作者: 山河已无恙。
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) 21h
NAME 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 21h
NAME 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 21h
NAME 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]
└─$kubectl -n kube-fledged get all
NAME READY STATUS RESTARTS AGE
pod/kube-fledged-controller-df69f6565-wdb4g 1/1 Running 0 13h
pod/kube-fledged-webhook-server-7bcd589bc4-j8xxp 1/1 Running 0 13h
pod/kubefledged-controller-55f848cc67-klxlm 1/1 Running 0 13h
pod/kubefledged-webhook-server-597dbf4ff5-ktbsh 1/1 Running 0 13h
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/kube-fledged-webhook-server ClusterIP 10.100.194.199 <none> 3443/TCP 36h
service/kubefledged-webhook-server ClusterIP 10.101.191.206 <none> 3443/TCP 36h
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/kube-fledged-controller 1/1 1 1 36h
deployment.apps/kube-fledged-webhook-server 1/1 1 1 36h
deployment.apps/kubefledged-controller 1/1 1 1 36h
deployment.apps/kubefledged-webhook-server 1/1 1 1 36h
NAME DESIRED CURRENT READY AGE
replicaset.apps/kube-fledged-controller-df69f6565 1 1 1 36h
replicaset.apps/kube-fledged-webhook-server-7bcd589bc4 1 1 1 36h
replicaset.apps/kubefledged-controller-55f848cc67 1 1 1 36h
replicaset.apps/kubefledged-webhook-server-597dbf4ff5 1 1 1 36h
验证是否安装成功
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged]
└─$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]
└─$cat kubefledged-imagecache.yaml
---
apiVersion: kubefledged.io/v1alpha2
kind: ImageCache
metadata:
# Name of the image cache. A cluster can have multiple image cache objects
name: imagecache1
namespace: kube-fledged
# The kubernetes namespace to be used for this image cache. You can choose a different namepace as per your preference
labels:
app: kubefledged
kubefledged: imagecache
spec:
# The "cacheSpec" field allows a user to define a list of images and onto which worker nodes those images should be cached (i.e. pre-pulled).
cacheSpec:
# Specifies a list of images (nginx:1.23.1) with no node selector, hence these images will be cached in all the nodes in the cluster
- images:
- ghcr.io/jitesoft/nginx:1.23.1
# Specifies a list of images (cassandra:v7 and etcd:3.5.4-0) with a node selector, hence these images will be cached only on the nodes selected by the node selector
- images:
- us.gcr.io/k8s-artifacts-prod/cassandra:v7
- us.gcr.io/k8s-artifacts-prod/etcd:3.5.4-0
nodeSelector:
tier: backend
# Specifies a list of image pull secrets to pull images from private repositories into the cache
imagePullSecrets:
- name: myregistrykey
官方的 Demo 中对应的 镜像拉取不下来,所以换一下
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$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]
└─$cat kubefledged-imagecache.yaml
---
apiVersion: kubefledged.io/v1alpha2
kind: ImageCache
metadata:
# Name of the image cache. A cluster can have multiple image cache objects
name: imagecache1
namespace: kube-fledged
# The kubernetes namespace to be used for this image cache. You can choose a different namepace as per your preference
labels:
app: kubefledged
kubefledged: imagecache
spec:
# The "cacheSpec" field allows a user to define a list of images and onto which worker nodes those images should be cached (i.e. pre-pulled).
cacheSpec:
# Specifies a list of images (nginx:1.23.1) with no node selector, hence these images will be cached in all the nodes in the cluster
- images:
- liruilong/my-busybox:latest
# Specifies a list of images (cassandra:v7 and etcd:3.5.4-0) with a node selector, hence these images will be cached only on the nodes selected by the node selector
- images:
- liruilong/hikvision-sdk-config-ftp:latest
nodeSelector:
kubernetes.io/hostname: vms105.liruilongs.github.io
# Specifies a list of image pull secrets to pull images from private repositories into the cache
#imagePullSecrets:
#- name: myregistrykey
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$
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]
└─$cat kubefledged-imagecache.yaml
---
apiVersion: kubefledged.io/v1alpha2
kind: ImageCache
metadata:
# Name of the image cache. A cluster can have multiple image cache objects
name: imagecache1
namespace: kube-fledged
# The kubernetes namespace to be used for this image cache. You can choose a different namepace as per your preference
labels:
app: kubefledged
kubefledged: imagecache
spec:
# The "cacheSpec" field allows a user to define a list of images and onto which worker nodes those images should be cached (i.e. pre-pulled).
cacheSpec:
# Specifies a list of images (nginx:1.23.1) with no node selector, hence these images will be cached in all the nodes in the cluster
#- images:
#- liruilong/my-busybox:latest
# Specifies a list of images (cassandra:v7 and etcd:3.5.4-0) with a node selector, hence these images will be cached only on the nodes selected by the node selector
- images:
- liruilong/hikvision-sdk-config-ftp:latest
nodeSelector:
kubernetes.io/hostname: vms105.liruilongs.github.io
# Specifies a list of image pull secrets to pull images from private repositories into the cache
#imagePullSecrets:
#- name: myregistrykey
┌──[root@vms100.liruilongs.github.io]-[~/ansible/kube-fledged/kube-fledged/deploy]
└─$
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
版权声明: 本文为 InfoQ 作者【华为云开发者联盟】的原创文章。
原文链接:【http://xie.infoq.cn/article/290edacf41823a2de450e21d3】。文章转载请联系作者。
华为云开发者联盟
提供全面深入的云计算技术干货 2020-07-14 加入
生于云,长于云,让开发者成为决定性力量
评论