kuberenetes 上使用helm部署rancher如何卸载干净

在已有的kuberenetes上用helm(部署文档)部署过rancher卸载之后在也别想再次部署rancher了。。。会出现无法登陆 资源正在删除中诸如此类问题,导致这个现象的原因是rancher创建的crd资源大部分都有finalizers字段,这会导致crd下的一系列资源一直在Terminating 下次安装rancher的时候不会初始化这部分资源,比如nodes.management.cattle.io这个crd是映射k8s的node节点的,下次安装rancher看到nodes.management.cattle.io正在Terminating 会“帮忙”把node节点从k8s集群中移除

如果还想在卸载过的环境再次部署rancher如需完整的清理一下rancher生成的所有资源

#!/bin/bash

# k8s=1.18 rchaner=2.5.5


api_server="https://192.168.10.229:6443"
token=`cat .kube/admin-token`

echo "处理删除中断"
curl -k $api_server/api/v1/namespaces/local/finalize 
 -H "Authorization: Bearer $token" 
 -H "Content-Type: application/json" 
 -XPUT 
 -d '{"apiVersion":"v1","kind":"Namespace","metadata":{"name":"local"},"spec":{"finalizers":[]}}'



kubectl patch --type='json' -p='[{"op": "remove", "path": "/metadata/finalizers"}]' ns $(kubectl get ns |egrep 'cattle|fleet|rancher|local|^p-|^user-' |awk '{print $1}')
project_ns=`kubectl get ns |egrep '^p-|local' |awk '{print $1}'`
project_crd=(clusterroletemplatebindings.management.cattle.io clusteralertgroups.management.cattle.io projects.management.cattle.io projectalertgroups.management.cattle.io projectalertrules.management.cattle.io projectroletemplatebindings.management.cattle.io nodes.management.cattle.io)


for ns in ${project_ns[@]}
do
  for crd in ${project_crd[@]}
  do
    kubectl patch --type='json' -p='[{"op": "remove", "path": "/metadata/finalizers"}]' $crd -n $ns $(kubectl get $crd -n $ns|grep -v NAME|awk '{print $1}')
  done
done


crd_list=`kubectl get crd |grep cattle.io |awk '{print $1}'`

for crd in $crd_list
do

kubectl patch --type='json' -p='[{"op": "remove", "path": "/metadata/finalizers"}]' $crd $(kubectl get $crd |grep -v NAME|awk '{print $1}')

done

echo "使用rancher清理工具 https://docs.rancher.cn/docs/rancher2/system-tools/_index/"
./system-tools remove -c .kube/config

kubectl get crd

echo "理论上所有rancher的crd都会被删除,如果还有说明有异常,继续会强制删除"
echo "按任意键继续"
read anykey


echo "helm卸载"
helm uninstall rancher -n cattle-system
helm uninstall rancher-operator -n rancher-operator-system
helm uninstall rancher-operator-crd -n rancher-operator-system
helm uninstall fleet -n fleet-system
helm uninstall fleet-crd -n fleet-system

echo "清理剩余资源"
sleep 5

for ns in $(kubectl get ns |grep -v NAME|awk '{print $1}')
do
  kubectl delete rolebinding $(kubectl get rolebinding -n $ns --selector='cattle.io/creator=norman' |grep -v NAME|awk '{print $1}') -n $ns
done


ns_list=`kubectl get ns |egrep 'cattle|fleet|rancher|local' |awk '{print $1}'`

kubectl delete ns $ns_list

kubectl patch --type='json' -p='[{"op": "remove", "path": "/metadata/finalizers"}]' crd $(kubectl get crd |grep 'cattle.io' |awk '{print $1}')
kubectl get crd |grep 'cattle.io' |awk '{print $1}' |xargs kubectl delete crd
kubectl get clusterrolebinding |egrep 'default-admin|fleet|pod-impersonation-helm-op|rancher' |awk '{print $1}' |xargs kubectl delete clusterrolebinding
kubectl get clusterrole |egrep 'fleet|pod-impersonation-helm-op' |awk '{print $1}' |xargs kubectl delete clusterrole

kubectl label ns $(kubectl get ns |egrep -v 'NAME' |awk '{print $1}') field.cattle.io/projectId-
原文地址:https://www.cnblogs.com/37yan/p/14275214.html