k8s集群安装过程中的相关问题和解决

前言

断点续传模式~

记录

我用的是ubuntu16.04，首先要做的是配置apt源，这里推荐阿里云的源地址 https://mirrors.aliyun.com/kubernetes/apt/doc/apt-key.gpg，centos的在这儿https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64，用梯子的或者海外的用谷歌源是最方便的，后面可以省很多事。

安装docker和kubeadm、kubectl、kubelet基本没啥问题，照着来就行，安装完工具就需要下载镜像了。

由于镜像在谷歌仓库（k8s.grc.io），config image pull会报以下错误：

[preflight/images] You can also perform this action in beforehand using 'kubeadm config images pull'
[preflight] Some fatal errors occurred:
    [ERROR ImagePull]: failed to pull image [k8s.gcr.io/kube-apiserver-amd64:v1.11.3]: exit status

这个很明显是墙的缘故，手动从docker的国内托管站点下载：

docker pull mirrorgooglecontainers/kube-apiserver-amd64:v1.11.3
docker pull mirrorgooglecontainers/kube-controller-manager-amd64:v1.11.3
docker pull mirrorgooglecontainers/kube-scheduler-amd64:v1.11.3
docker pull mirrorgooglecontainers/kube-proxy-amd64:v1.11.3
docker pull mirrorgooglecontainers/pause:3.1
docker pull mirrorgooglecontainers/etcd-amd64:3.2.18

相关的镜像就上面几个，但是下下来之后仍然无法kubeadm init，依然会去谷歌仓库拉镜像，卡了很久才发现从docker仓库拉下来的镜像和谷歌仓库中的不一样，因此在kubeadm init时会认为这些镜像不存在，接着去谷歌仓库拉，docker images可以查看这些镜像现在的名字，所以需要执行如下命令：

docker tag docker.io/mirrorgooglecontainers/kube-proxy-amd64:v1.11.3 k8s.gcr.io/kube-proxy-amd64:v1.11.3
docker tag docker.io/mirrorgooglecontainers/kube-scheduler-amd64:v1.11.3 k8s.gcr.io/kube-scheduler-amd64:v1.11.3
docker tag docker.io/mirrorgooglecontainers/kube-apiserver-amd64:v1.11.3 k8s.gcr.io/kube-apiserver-amd64:v1.11.3
docker tag docker.io/mirrorgooglecontainers/kube-controller-manager-amd64:v1.11.3 k8s.gcr.io/kube-controller-manager-amd64:v1.11.3
docker tag docker.io/mirrorgooglecontainers/etcd-amd64:3.2.18 k8s.gcr.io/etcd-amd64:3.2.18
docker tag docker.io/mirrorgooglecontainers/pause:3.1 k8s.gcr.io/pause:3.1
docker tag docker.io/coredns/coredns:1.1.3 k8s.gcr.io/coredns:1.1.3

接着出现了CPU数量少于2（VM）和swap分区的问题，调整了CPU的数量，并禁用了swap分区，这两个问题解决，搜了一下为啥要禁止swap分区，大概就是不使用虚拟内存以提高性能，将实例紧密包装到尽可能接近百分之百的意思吧，接着继续执行kubeadm init，报错如下：

[wait-control-plane] Waiting for the kubelet to boot up the control plane as static Pods from directory "/etc/kubernetes/manifests". This can take up to 4m0s
[kubelet-check] Initial timeout of 40s passed.

Unfortunately, an error has occurred:
	timed out waiting for the condition

This error is likely caused by:
	- The kubelet is not running
	- The kubelet is unhealthy due to a misconfiguration of the node in some way (required cgroups disabled)

If you are on a systemd-powered system, you can try to troubleshoot the error with the following commands:
	- 'systemctl status kubelet'
	- 'journalctl -xeu kubelet'

Additionally, a control plane component may have crashed or exited when started by the container runtime.
To troubleshoot, list all containers using your preferred container runtimes CLI, e.g. docker.

开始以为是kubelet的版本和apiserver版本不兼容，于是又重新装了一遍工具，发现端口被占用，这是因为之前master启动时占了6443端口，kubeadm reset后解决，但是中间又很二的执行了一个改ip的操作-_-!，依然报上面的错误，在日志里看到重复报以下错误：

1006 02:44:41.050125   19805 reflector.go:123] k8s.io/client-go/informers/factor
1006 02:44:41.129531   19805 kubelet.go:2267] node "ubuntu" not found
1006 02:44:41.230347   19805 kubelet.go:2267] node "ubuntu" not found
1006 02:44:41.331174   19805 kubelet.go:2267] node "ubuntu" not found
1006 02:44:41.431984   19805 kubelet.go:2267] node "ubuntu" not found
1006 02:44:41.532748   19805 kubelet.go:2267] node "ubuntu" not found
1006 02:44:41.596687   19805 controller.go:135] failed to ensure node lease exist
1006 02:44:41.633573   19805 kubelet.go:2267] node "ubuntu" not found
1006 02:44:41.734381   19805 kubelet.go:2267] node "ubuntu" not found
1006 02:44:41.745881   19805 kubelet_node_status.go:94] Unable to register node with apiserver ~~
1006 02:44:41.835220   19805 kubelet.go:2267] node "ubuntu" not found
1006 02:44:41.936016   19805 kubelet.go:2267] node "ubuntu" not found

因为改地址导致旧的ip无法匹配，最简单的解决方案依然是kubeadm reset，推荐修改conf参数的方式来解决，将conf文件中的旧地址修改为新地址，重启kubelet服务，问题解决。

因为整个安装操作都是在root下进行，所以还要复制下配置文件到home目录，然后改个权限，做完这些master节点安装kubernetes就成功了，这里要把安装成功后的token记下来，用于之后node的添加：

已达总线带宽上限，后面的下次续传。