Ljohn

I found that there is this phenomenon, that is, this test sometimes takes effect and sometimes fails. I guess there is something wrong with the alpine system, but I have...

I added debugging and rebuilt an image. Two servers executed the same command and returned different results. Here is the execution result. ## dockerfile and kube-dump ``` (base) ➜ 1.1.1...

> I added debugging and rebuilt an image. Two servers executed the same command and returned different results. Here is the execution result. > > ## dockerfile and kube-dump >...

我也遇到了这个问题，kubeasz 3.1.1，集群版本：v1.22.2，系统版本centos7.9，内核5.4.207；这个集群其他ubuntu18.04 没问题。

我发现Kops社区的有人提了个[Issue](https://github.com/kubernetes/kops/issues/11607)，目前为已解决状态，[Fix PR ](https://github.com/kubernetes/kops/pull/12917)。处理方案：1、要么使用Ubuntu20.04 以上的操作系统。2、使用Centos7.9 k8s要升级到1.23.X。3、规避方案，enforceNodeAllocatable字段删除kube-reserved： `enforceNodeAllocatable: "pods,kube-reserved" enforceNodeAllocatable: "pods"`；或者不设置Qos