使用kubernetes1.20.*部署踩坑记

[forgottener]

照着非Quick Start指南一步步操作，nfs都是没问题的，翻来覆去就是不成功，nacos的pods报错kubectl logs -f nacos-0 0/4 nodes are available: 4 pod has unbound immediate PersistentVolumeClaims.
这一行错误全网都没有解决方案，kubectl get pvc 3个都是pending,照这条路怎么也查不到，灵机一动想到看看nfs-client-provisioner日志吧，嘿好家伙错误日志如下：

I1231 09:05:32.221800       1 controller.go:987] provision "default/logdir-nacos-0" class "managed-nfs-storage": started
I1231 09:05:32.221840       1 controller.go:987] provision "default/plugindir-nacos-0" class "managed-nfs-storage": started
I1231 09:05:32.221965       1 controller.go:987] provision "default/datadir-nacos-0" class "managed-nfs-storage": started
E1231 09:05:32.235211       1 controller.go:1004] provision "default/plugindir-nacos-0" class "managed-nfs-storage": unexpected error getting claim reference: selfLink was empty, can't make reference
E1231 09:05:32.235253       1 controller.go:1004] provision "default/logdir-nacos-0" class "managed-nfs-storage": unexpected error getting claim reference: selfLink was empty, can't make reference
E1231 09:05:32.235464       1 controller.go:1004] provision "default/datadir-nacos-0" class "managed-nfs-storage": unexpected error getting claim reference: selfLink was empty, can't make reference

这一顿操作就明了了click here，原来是kubernetes1.20.0废弃了selfLink解决办法是重新加回来

vi /etc/kubernetes/manifests/kube-apiserver.yaml

## 增加一行
- --feature-gates=RemoveSelfLink=false
## 使之生效
kubectl apply -f /etc/kubernetes/manifests/kube-apiserver.yaml

[mingcheng]

感谢，同样碰到一模一样的问题，根据这个 Issues 解决了。

补充下，如果是控制面集群只需要修改一台然后 apply 就行（简单粗暴的直接 delete，然后会自动重建）。

[only21tender]

我的kubernets是1.20.2，创建snapshot的时候提示：
kubectl describe volumesnapshot rbd-pvc-snapshot |grep Message
Message: Failed to create snapshot: selfLink was empty, can't make reference
参考咱们这里的方法修改了kube-apiserver.yaml，后面k8s集群直接崩溃无法使用了

[forgottener]

我的kubernets是1.20.2，创建snapshot的时候提示：
kubectl describe volumesnapshot rbd-pvc-snapshot |grep Message
Message: Failed to create snapshot: selfLink was empty, can't make reference
参考咱们这里的方法修改了kube-apiserver.yaml，后面k8s集群直接崩溃无法使用了

只修改主节点的kube-apiserver.yaml，然后执行kubectl apply -f /etc/kubernetes/manifests/kube-apiserver.yaml就可以了，这个加回SelfLink应该不会导致集群崩溃的

[DapengJiao]

直接修改完静态Pod的manifest文件后，K8S应该就会自动更新这些静态Pod的。
应该不需要在额外运行kubectl apply命令。

[oneinstack]

nfs deployment.yaml image change to : gcr.io/k8s-staging-sig-storage/nfs-subdir-external-provisioner:v4.0.2

[paderlol]

感谢一路踩坑并记录，我把这个issue置顶了，方便其他人查看。

[duiniwukenaihe]

我想知道 k8s 部署nacos为什么日志会那么大....有没有变量或者参数修改一下？打印的raft 下面log太多了

[helloxk617]

1.24.1的版本，这个方法似乎已经不适用了，logs报Error: invalid argument "RemoveSelfLink=false" for "--feature-gates" flag: cannot set feature gate RemoveSelfLink to false, feature is locked to true“

[Hypo5279]

Kubernetes v1.24.9，deploy/nfs/deployment.yaml image change to ：registry.cn-shenzhen.aliyuncs.com/shuhui/nfs-subdir-external-provisioner:v4.0.2.