内容简介:作者简介Vyacheslav,拥有运维和项目管理经验的软件工程师这篇文章将承接我此前搭建的本地Docker开发环境,具体步骤已经放在在以下网址:
作者简介
Vyacheslav,拥有运维和项目管理经验的软件工程师
这篇文章将承接我此前搭建的本地 Docker 开发环境,具体步骤已经放在在以下网址:
https://github.com/Voronenko/t ... plate
除了经典的docker化的项目之外,我还有其他的Kubernetes项目。尽管Kubernetes已经成为容器编排的事实标准,但是不得不承认Kubernetes是一个既消耗资源又消耗金钱的平台。由于我并不经常需要外部集群,因此我使用轻量级K3s发行版来进行Kubernetes本地开发。
K3s是为IoT和边缘计算而构建的经过认证的Kubernetes发行版之一,还能够按产品规模部署到VM。
我使用K3s的方式是这样的:在我的工作笔记本上本地安装K3s,尽管有时我需要在本地部署较重的测试工作负载,为此,我准备了两个神器——两个运行ESXi的外部Intel NUCs。
默认情况下,K3s安装Traefik 1.x作为ingress,如果你对此十分满意,那么无需往下继续阅读了。
在我的场景中,我同时会牵涉到好几个项目,特别是经典的docker和docker swarm,因此我经常遇到在独立模式下部署Traefik的情况。
因此,本文其余部分将深入介绍如何将外部traefik2配置为K3s集群的ingress。
安装Kubernetes K3s系列集群
你可以按照常规方式使用命令 curl -sfL https://get.k3s.io | sh -
安装K3s,或者你可以使用轻量实用程序k3sup安装( https://github.com/alexellis/k3sup )。具体步骤在 之前的文章 介绍过。
与我们的设置不同的是,我们使用命令 --no-deploy traefik
专门安装了不带traefik组件的K3s。
export CLUSTER_MASTER=192.168.3.100 export CLUSTER_DEPLOY_USER=slavko k3sup install --ip $CLUSTER_MASTER --user $CLUSTER_DEPLOY_USER --k3s-extra-args '--no-deploy traefik'
执行后,你将获得使用kubectl所需的连接详细信息。安装K3s后,你可以快速检查是否可以看到节点。
Test your cluster with - export path to k3s cluster kubeconfig:
export KUBECONFIG=/home/slavko/kubeconfig
kubectl get node -o wide
注:这里没有固定的安装模式,你甚至可以使用docker-compose自行启动它。
server: image: rancher/k3s:v0.8.0 command: server --disable-agent --no-deploy traefik environment: - K3S_CLUSTER_SECRET=somethingtotallyrandom - K3S_KUBECONFIG_OUTPUT=/output/kubeconfig.yaml - K3S_KUBECONFIG_MODE=666 volumes: # k3s will generate a kubeconfig.yaml in this directory. This volume is mounted # on your host, so you can then 'export KUBECONFIG=/somewhere/on/your/host/out/kubeconfig.yaml', # in order for your kubectl commands to work. - /somewhere/on/your/host/out:/output # This directory is where you put all the (yaml) configuration files of # the Kubernetes resources. - /somewhere/on/your/host/in:/var/lib/rancher/k3s/server/manifests ports: - 6443:6443 node: image: rancher/k3s:v0.8.0 privileged: true links: - server environment: - K3S_URL=https://server:6443 - K3S_CLUSTER_SECRET=somethingtotallyrandom volumes: # this is where you would place a alternative traefik image (saved as a .tar file with # 'docker save'), if you want to use it, instead of the traefik:v2.0 image. - /sowewhere/on/your/host/custom-image:/var/lib/rancher/k3s/agent/images
配置Traefik 2,与Kubernetes一起使用
在文章开头提到的链接中,我已经在我的系统中安装了Traefik 2,并根据该链接内容,服务于一些需求。现在是时候配置Traefik 2 Kubernetes后端了。
Traefik 2使用CRD(自定义资源定义)来完成这一点。定义的最新示例可以在以下链接中找到,但这些示例仅适用于Traefik 2也作为Kubernetes工作负载的一部分执行的情况:
https://docs.traefik.io/refere ... -crd/
对于外部Traefik 2,我们仅需要以下描述的定义子集。
我们引入一系列自定义资源定义,以允许我们来描述我们的Kubernetes服务将会如何暴露到外部, traefik-crd.yaml
:
apiVersion: apiextensions.k8s.io/v1beta1 kind: CustomResourceDefinition metadata: name: ingressroutes.traefik.containo.us spec: group: traefik.containo.us version: v1alpha1 names: kind: IngressRoute plural: ingressroutes singular: ingressroute scope: Namespaced --- apiVersion: apiextensions.k8s.io/v1beta1 kind: CustomResourceDefinition metadata: name: ingressroutetcps.traefik.containo.us spec: group: traefik.containo.us version: v1alpha1 names: kind: IngressRouteTCP plural: ingressroutetcps singular: ingressroutetcp scope: Namespaced --- apiVersion: apiextensions.k8s.io/v1beta1 kind: CustomResourceDefinition metadata: name: middlewares.traefik.containo.us spec: group: traefik.containo.us version: v1alpha1 names: kind: Middleware plural: middlewares singular: middleware scope: Namespaced --- apiVersion: apiextensions.k8s.io/v1beta1 kind: CustomResourceDefinition metadata: name: tlsoptions.traefik.containo.us spec: group: traefik.containo.us version: v1alpha1 names: kind: TLSOption plural: tlsoptions singular: tlsoption scope: Namespaced --- apiVersion: apiextensions.k8s.io/v1beta1 kind: CustomResourceDefinition metadata: name: traefikservices.traefik.containo.us spec: group: traefik.containo.us version: v1alpha1 names: kind: TraefikService plural: traefikservices singular: traefikservice scope: Namespaced
同时,我们需要集群角色 traefik-ingress-controller
,以提供对服务、端点和secret的只读访问权限以及自定义的 traefik.containo.us
组, traefik-clusterrole.yaml
:
kind: ClusterRole apiVersion: rbac.authorization.k8s.io/v1beta1 metadata: name: traefik-ingress-controller rules: - apiGroups: - "" resources: - services - endpoints - secrets verbs: - get - list - watch - apiGroups: - extensions resources: - ingresses verbs: - get - list - watch - apiGroups: - extensions resources: - ingresses/status verbs: - update - apiGroups: - traefik.containo.us resources: - middlewares verbs: - get - list - watch - apiGroups: - traefik.containo.us resources: - ingressroutes verbs: - get - list - watch - apiGroups: - traefik.containo.us resources: - ingressroutetcps verbs: - get - list - watch - apiGroups: - traefik.containo.us resources: - tlsoptions verbs: - get - list - watch - apiGroups: - traefik.containo.us resources: - traefikservices verbs: - get - list - watch
最后,我们需要系统服务账号 traefik-ingress-controller
与之前创建的集群角色 traefik-ingress-controller
相关联。
{{{
kind: ServiceAccount
apiVersion: v1
metadata:
namespace: kube-system
name: traefik-ingress-controller
---
kind: ClusterRoleBinding
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
name: traefik-ingress-controller
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: ClusterRole
name: traefik-ingress-controller
subjects:
- kind: ServiceAccount
name: traefik-ingress-controller
namespace: kube-system
}}}
我们应用以上资源之后:
apply: kubectl apply -f traefik-crd.yaml kubectl apply -f traefik-clusterrole.yaml kubectl apply -f traefik-service-account.yaml
我们已经准备好开始调整Traefik 2
将Traefik 2指向K3s集群
根据Traefik文档的建议,当Traefik部署到Kubernetes中时,它将读取环境变量KUBERNETES_SERVICE_HOST和KUBERNETES_SERVICE_PORT或KUBECONFIG来构造端点。
在 /var/run/secrets/kubernetes.io/serviceaccount/token
中查找访问token,而SSL CA证书将在 /var/run/secrets/kubernetes.io/serviceaccount/ca.crt.
中查找。当部署到Kubernetes内部时,两者都会自动提供挂载。
当无法找到环境变量时,Traefik会尝试使用external-cluster客户端连接到Kubernetes API server。这一情况下,需要设置endpoint。具体来说,可以将其设置为kubectl代理使用的URL,以使用相关的kubeconfig授予的身份验证和授权连接到Kubernetes集群。
Traefik 2可以使用任何受支持的配置类型来静态配置-toml、yaml或命令行交换。
[providers.kubernetesCRD] endpoint = "http://localhost:8080" token = "mytoken"
providers: kubernetesCRD: endpoint = "http://localhost:8080" token = "mytoken" # ...
--providers.kubernetescrd.endpoint=http://localhost:8080 --providers.kubernetescrd.token=mytoken
第一次运行时,如果你在外部有Traefik,很有可能没有 traefik-ingress-controller
访问token来指定mytoken。那么,你需要执行以下命令:
Check all possible clusters, as your .KUBECONFIG may have multiple contexts:
kubectl config view -o jsonpath='{"Cluster name\tServer\n"}{range .clusters[*]}{.name}{"\t"}{.cluster.server}{"\n"}{end}'
Output kind of
Alias tip: k config view -o jsonpath='{"Cluster name\tServer\n"}{range .clusters[*]}{.name}{"\t"}{.cluster.server}{"\n"}{end}'
Cluster name Server
default https://127.0.0.1:6443
You are interested in: "default", if you did not name it differently
Select name of cluster you want to interact with from above output:
export CLUSTER_NAME="default"
Point to the API server referring the cluster name
export APISERVER=$(kubectl config view -o jsonpath="{.clusters[?(@.name==\"$CLUSTER_NAME\")].cluster.server}")
usually https://127.0.0.1:6443
Gets the token value
export TOKEN=$(kubectl get secrets -o jsonpath="{.items[?(@.metadata.annotations['kubernetes\.io/service-account\.name']=='traefik-ingress-controller')].data.token}" --namespace kube-system|base64 --decode)
Explore the API with TOKEN
如果成功了,你应该收到以下响应:
{ "kind": "APIVersions", "versions": [ "v1" ], "serverAddressByClientCIDRs": [ { "clientCIDR": "0.0.0.0/0", "serverAddress": "192.168.3.100:6443" } ]
以及一些事实,如token:
eyJhbGciOiJSUzI1NiIsImtpZCI6IjBUeTQyNm5nakVWbW5PaTRRbDhucGlPeWhlTHhxTXZjUDJsRmNacURjVnMifQ.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.Mk8EBS4soO8uX-uSnV3o4qZKR6Iw6bgeSmPhHbJ2fjuqFgLnLh4ggxa-N9AqmCsEWiYjSi5oKAu986UEC-_kGQh3xaCYsUwlkM8147fsnwCbomSeGIct14JztVL9F8JwoDH6T0BOEjn-J9uY8-fUKYL_Y7uTrilhFapuILPsj_bFfgIeOOapRD0XshKBQV9Qzg8URxyQyfzl68ilm1Q13h3jLj8CFE2RlgEUFk8TqYH4T4fhfpvV-gNdmKJGODsJDI1hOuWUtBaH_ce9w6woC9K88O3FLKVi7fbvlDFrFoJ2iVZbrRALPjoFN92VA7a6R3pXUbKebTI3aUJiXyfXRQ根据上次响应的API server的外部地址:
https://192.168.3.100:6443
同样,提供的token中没有任何特殊之处:这是JWT的token,你可以使用 https://jwt.io/#debugger-io ,检查它的内容。
{ "alg": "RS256", "kid": "0Ty426ngjEVmnOi4Ql8npiOyheLxqMvcP2lFcZqDcVs" } { "iss": "kubernetes/serviceaccount", "kubernetes.io/serviceaccount/namespace": "kube-system", "kubernetes.io/serviceaccount/secret.name": "traefik-ingress-controller-token-vzc7v", "kubernetes.io/serviceaccount/service-account.name": "traefik-ingress-controller", "kubernetes.io/serviceaccount/service-account.uid": "d9577e91-7e64-4c06-bd82-cde949c81251", "sub": "system:serviceaccount:kube-system:traefik-ingress-controller" }
正确的配置非常重要,因此请确保对APISERVER的两个调用均返回合理的响应。
export APISERVER=YOURAPISERVER export TOKEN=YOURTOKEN curl -X GET $APISERVER/api --header "Authorization: Bearer $TOKEN" --insecure curl -X GET $APISERVER/api/v1/endpoints --header "Authorization: Bearer $TOKEN" --insecure
创建其他访问token
控制器循环确保每个服务账户都有一个带有API token的secret,可以像我们之前那样被发现。
此外,你还可以为一个服务账户创建额外的token,创建一个ServiceAccountToken类型的secret,并为服务账户添加一个注释,控制器会用生成的token来更新它。
{{{
apiVersion: v1
kind: Secret
namespace: kube-system
metadata:
name: traefik-manual-token
annotations:
kubernetes.io/service-account.name: traefik-ingress-controller
type: kubernetes.io/service-account-token
Any tokens for non-existent service accounts will be cleaned up by the token controller.
kubectl describe secrets/traefik-manual-token
}}}
用以下命令创建:
kubectl create -f ./traefik-service-account-secret.yaml kubectl describe secret traefik-manual-token
删除/无效:
kubectl delete secret traefik-manual-token
对外部traefik 2的更改构成定义
我们需要在文章开头给出的链接中获得的traefik2配置进行哪些更改?
https://github.com/Voronenko/t ... plate
a) 我们在新文件夹kubernetes_data中存储ca.crt文件,该文件用于验证对Kubernetes授权的调用。这是可以在kubeconfig文件的clusters-> cluster-> certificate-authority-data下找到的证书。
该volume将映射在 /var/run/secrets/kubernetes.io/serviceaccount
下以获取官方Traefik 2镜像
volumes: ... - ./kubernetes_data:/var/run/secrets/kubernetes.io/serviceaccount
b) 调整Traefik 2 kubernetescrd后端以提供3个参数:endpoint、证书路径和token。请注意,作为外部Traefik作为docker容器,你需要指定正确的endpoint地址,并确保以安全的方式进行。
- "--providers.kubernetescrd=true" - "--providers.kubernetescrd.endpoint=https://192.168.3.100:6443" - "--providers.kubernetescrd.certauthfilepath=/var/run/secrets/kubernetes.io/serviceaccount/ca.crt" - "--providers.kubernetescrd.token=YOURTOKENWITHOUTANYQUOTES
如果你都执行正确了,那么你现在应该在Traefik UI上看到了一些希望。如果你没有看到traefik,或者在运行Traefik时有问题,你可以查看之后的故障排除部分。
现在是时候通过Trafik 2暴露一些Kubernetes服务了,以确保Traefik 2能够作为ingress工作。让我们来看经典案例whoami服务,whoami-service.yaml
apiVersion: v1 kind: Service metadata: name: whoami spec: ports: - protocol: TCP name: web port: 80 selector: app: whoami --- kind: Deployment apiVersion: apps/v1 metadata: namespace: default name: whoami labels: app: whoami spec: replicas: 2 selector: matchLabels: app: whoami template: metadata: labels: app: whoami spec: containers: - name: whoami image: containous/whoami ports: - name: web containerPort: 80
并且以http或https的方式暴露它, whoami.k.voronenko.net
全限定域名下的 whoami-ingress-route.yaml
。
apiVersion: traefik.containo.us/v1alpha1 kind: IngressRoute metadata: name: ingressroute-notls namespace: default spec: entryPoints: - web routes: - match: Host(`whoami.k.voronenko.net`) kind: Rule services: - name: whoami port: 80 --- apiVersion: traefik.containo.us/v1alpha1 kind: IngressRoute metadata: name: ingressroute-tls namespace: default spec: entryPoints: - websecure routes: - match: Host(`whoami.k.voronenko.net`) kind: Rule services: - name: whoami port: 80 tls: certResolver: default
然后应用它:
kubectl apply -f whoami-service.yaml kubectl apply -f whoami-ingress-route.yaml
应用后,你应该会在Traefik dashboard上看到一些希望,即KubernetesCRD后端。
正如你所看到的,Traefik已经检测到我们的K3s Kubernetes集群上运行的新工作负载,而且它与我们在同一个盒子上的经典Docker工作负载(如portainer)很好地共存。
让我们检查一下Traefik 2是否将Traefik路由到了我们的Kubernetes工作负载:如你所见,你可以在http和https endpoint上成功地接触到whoami工作负载,浏览器接受你的证书为可信任的“绿标签”。
我们的目标达到了!我们在本地笔记本上配置了Traefik 2。Traefik 2将你的docker或Kubernetes工作流暴露在http或https endpoint上。带可选的 letsencrypt 的 Traefik 2 将负责 https。
故障排查
正如你所知,在配置过程可能存在多个问题,你可以考虑使用一些分析工具,如:
https://github.com/Voronenko/d ... 3L185
我特别建议:
a) VMWare octant:这是一个基于Web的功能强大的Kubernetes dashboard,你可以在上面使用你的kubeconfig
b) Rakess:这是一个独立 工具 也是一个kubectl插件,用于显示Kubernetes服务器资源的访问矩阵( https://github.com/corneliusweig/rakkess )
检查系统账户的凭据
rakkess --sa kube-system:traefik-ingress-controller
c) kubectl
检查哪些角色与服务账户相关联
kubectl get clusterrolebindings -o json | jq -r ' .items[] | select( .subjects // [] | .[] | [.kind,.namespace,.name] == ["ServiceAccount","kube-system","traefik-ingress-controller"] ) | .metadata.name'
d) Traefik 文档:例如kubernetescrd后端提供了更多配置开关的方式。
--providers.kubernetescrd (Default: "false") Enable Kubernetes backend with default settings. --providers.kubernetescrd.certauthfilepath (Default: "") Kubernetes certificate authority file path (not needed for in-cluster client). --providers.kubernetescrd.disablepasshostheaders (Default: "false") Kubernetes disable PassHost Headers. --providers.kubernetescrd.endpoint (Default: "") Kubernetes server endpoint (required for external cluster client). --providers.kubernetescrd.ingressclass (Default: "") Value of kubernetes.io/ingress.class annotation to watch for. --providers.kubernetescrd.labelselector (Default: "") Kubernetes label selector to use. --providers.kubernetescrd.namespaces (Default: "") Kubernetes namespaces. --providers.kubernetescrd.throttleduration (Default: "0") Ingress refresh throttle duration --providers.kubernetescrd.token (Default: "") Kubernetes bearer token (not needed for in-cluster client). --providers.kubernetesingress (Default: "false") Enable Kubernetes backend with default settings. --providers.kubernetesingress.certauthfilepath (Default: "") Kubernetes certificate authority file path (not needed for in-cluster client). --providers.kubernetesingress.disablepasshostheaders (Default: "false") Kubernetes disable PassHost Headers. --providers.kubernetesingress.endpoint (Default: "") Kubernetes server endpoint (required for external cluster client). --providers.kubernetesingress.ingressclass (Default: "") Value of kubernetes.io/ingress.class annotation to watch for. --providers.kubernetesingress.ingressendpoint.hostname (Default: "") Hostname used for Kubernetes Ingress endpoints. --providers.kubernetesingress.ingressendpoint.ip (Default: "") IP used for Kubernetes Ingress endpoints. --providers.kubernetesingress.ingressendpoint.publishedservice (Default: "") Published Kubernetes Service to copy status from. --providers.kubernetesingress.labelselector (Default: "") Kubernetes Ingress label selector to use. --providers.kubernetesingress.namespaces (Default: "") Kubernetes namespaces. --providers.kubernetesingress.throttleduration (Default: "0") Ingress refresh throttle duration --providers.kubernetesingress.token (Default: "") Kubernetes bearer token (not needed for in-cluster client).
e) 确保Traefik有足够的权限可以访问apiserver endpoint
如果你希望Traefik为你查询信息:通过在配置中放置一些错误的apiserver地址,可以查看访问的endpoint和查询顺序。有了这些知识和你的Traefik Kubernetes token,你就可以使用Traefik凭证检查这些endpoint是否可以访问。
traefik_1 | E0421 12:30:12.624877 1 reflector.go:125] pkg/mod/k8s.io/client-go@v0.0.0-20190718183610-8e956561bbf5/tools/cache/reflector.go:98: Failed to list *v1.Endpoints: Get https://192.168.3.101:6443/api/v1/endpoints?limit=500&resourceVersion=0: traefik_1 | E0421 12:30:12.625341 1 reflector.go:125] pkg/mod/k8s.io/client-go@v0.0.0-20190718183610-8e956561bbf5/tools/cache/reflector.go:98: Failed to list *v1.Service: Get https://192.168.3.101:6443/api/v1/services?limit=500&resourceVersion=0: traefik_1 | E0421 12:30:12.625395 1 reflector.go:125] pkg/mod/k8s.io/client-go@v0.0.0-20190718183610-8e956561bbf5/tools/cache/reflector.go:98: Failed to list *v1beta1.Ingress: Get https://192.168.3.101:6443/apis/extensions/v1beta1/ingresses?limit=500&resourceVersion=0: traefik_1 | E0421 12:30:12.625449 1 reflector.go:125] pkg/mod/k8s.io/client-go@v0.0.0-20190718183610-8e956561bbf5/tools/cache/reflector.go:98: Failed to list *v1alpha1.Middleware: Get https://192.168.3.101:6443/apis/traefik.containo.us/v1alpha1/middlewares?limit=500&resourceVersion=0: traefik_1 | E0421 12:30:12.625492 1 reflector.go:125] pkg/mod/k8s.io/client-go@v0.0.0-20190718183610-8e956561bbf5/tools/cache/reflector.go:98: Failed to list *v1alpha1.IngressRoute: Get https://192.168.3.101:6443/apis/traefik.containo.us/v1alpha1/ingressroutes?limit=500&resourceVersion=0: traefik_1 | E0421 12:30:12.625531 1 reflector.go:125] pkg/mod/k8s.io/client-go@v0.0.0-20190718183610-8e956561bbf5/tools/cache/reflector.go:98: Failed to list *v1alpha1.TraefikService: Get https://192.168.3.101:6443/apis/traefik.containo.us/v1alpha1/traefikservices?limit=500&resourceVersion=0: traefik_1 | E0421 12:30:12.625572 1 reflector.go:125] pkg/mod/k8s.io/client-go@v0.0.0-20190718183610-8e956561bbf5/tools/cache/reflector.go:98: Failed to list *v1alpha1.TLSOption: Get https://192.168.3.101:6443/apis/traefik.containo.us/v1alpha1/tlsoptions?limit=500&resourceVersion=0: traefik_1 | E0421 12:30:12.625610 1 reflector.go:125] pkg/mod/k8s.io/client-go@v0.0.0-20190718183610-8e956561bbf5/tools/cache/reflector.go:98: Failed to list *v1alpha1.IngressRouteTCP: Get https://192.168.3.101:6443/apis/traefik.containo.us/v1alpha1/ingressroutetcps?limit=500&resourceVersion=0:
f) 记录K3s本身
安装脚本将自动检测你的操作系统是使用systemd还是openrc并启动服务。使用openrc运行时,将在/var/log/k3s.log中创建日志。使用systemd运行时,将在/var/log/syslog中创建日志,并使用journalctl -u k3s查看。
在那里,你可能会得到一些提示,例如:
кві 21 15:42:44 u18d k3s[612]: E0421 15:42:44.936960 612 authentication.go:104] Unable to authenticate the request due to an error: invalid bearer token
这将为你提供有关K8s Traefik起初使用时出现问题的线索,Enjoy your journey!
相关代码你可以在以下链接中找到:
以上所述就是小编给大家介绍的《使用Kubernetes、K3s和Traefik2进行本地开发》,希望对大家有所帮助,如果大家有任何疑问请给我留言,小编会及时回复大家的。在此也非常感谢大家对 码农网 的支持!
猜你喜欢:- 流程篇 | 避免 evalString 进行回调,使用 JSB 进行手动绑定
- 使用python进行短信轰炸
- 使用 Docker 进行交叉编译
- 使用 Gomock 进行单元测试
- 使用容器Docker进行开发
- 使用DeepLab进行语义分割
本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们。
TensorFlow实战
黄文坚、唐源 / 电子工业出版社 / 2017-2-1 / 79
Google近日发布了TensorFlow 1.0候选版,这个稳定版将是深度学习框架发展中的里程碑的一步。自TensorFlow于2015年底正式开源,距今已有一年多,这期间TensorFlow不断给人以惊喜,推出了分布式版本,服务框架TensorFlow Serving,可视化工具TensorFlow,上层封装TF.Learn,其他语言(Go、Java、Rust、Haskell)的绑定、Wind......一起来看看 《TensorFlow实战》 这本书的介绍吧!