问题描述
使用CentOS7的 yum 包管理器安装了 Kubernetes 集群,使用 kubectl 创建服务成功后,执行 kubectl get pods
,发现AGE虽然在不断增加,但状态始终不变
本文内容
- 分析问题原因
- 给出直接解决此问题的方式 (不完美)
- 给出其他方案
且听我娓娓道来~
问题分析与解决
kubectl 提供了 describe 子命令来输出指定的一个/多个资源的详细信息。
执行 kubectl describe pod mytomcat-9lcq5
,查看问题 Pod 的状态信息,输出如下:
[root@kube-master app]# kubectl describe pod mytomcat-9lcq5
Name: mytomcat-9lcq5
Namespace: default
Node: kube-node-2/192.168.87.145
Start Time: Fri, 17 Apr 2020 15:53:50 +0800
Labels: app=mytomcat
Status: Pending
IP:
Controllers: ReplicationController/mytomcat
Containers:
mytomcat:
Container ID:
Image: tomcat:9-jre8-alpine
Image ID:
Port: 8080/TCP
State: Waiting
Reason: ContainerCreating
Ready: False
Restart Count: 0
Volume Mounts: <none>
Environment Variables: <none>
Conditions:
Type Status
Initialized True
Ready False
PodScheduled True
No volumes.
QoS Class: BestEffort
Tolerations: <none>
Events:
FirstSeen LastSeen Count From SubObjectPath Type Reason Message
--------- -------- ----- ---- ------------- -------- ------ -------
5m 5m 1 {default-scheduler } Normal Scheduled Successfully assigned mytomcat-9lcq5 to kube-node-2
4m 4m 1 {kubelet kube-node-2} Warning FailedSync Error syncing pod, skipping: failed to "StartContainer" for "POD" with ErrImagePull: "image pull failed for registry.access.redhat.com/rhel7/pod-infrastructure:latest, this may be because there are no credentials on this request. details: (Get https://registry.access.redhat.com/v1/_ping: net/http: TLS handshake timeout)"
3m 3m 1 {kubelet kube-node-2} Warning FailedSync Error syncing pod, skipping: failed to "StartContainer" for "POD" with ErrImagePull: "image pull failed for registry.access.redhat.com/rhel7/pod-infrastructure:latest, this may be because there are no credentials on this request. details: (Network timed out while trying to connect to https://registry.access.redhat.com/v1/repositories/rhel7/pod-infrastructure/images. You may want to check your internet connection or if you are behind a proxy.)"
2m 2m 1 {kubelet kube-node-2} Warning FailedSync Error syncing pod, skipping: failed to "StartContainer" for "POD" with ErrImagePull: "image pull failed for registry.access.redhat.com/rhel7/pod-infrastructure:latest, this may be because there are no credentials on this request. details: (Error: image rhel7/pod-infrastructure:latest not found)"
3m 1m 3 {kubelet kube-node-2} Warning FailedSync Error syncing pod, skipping: failed to "StartContainer" for "POD" with ImagePullBackOff: "Back-off pulling image "registry.access.redhat.com/rhel7/pod-infrastructure:latest""
通过查看最下方的输出信息,Successfully assigned mytomcat-9lcq5 to kube-node-2
说明这个 Pod 分配到 kube-node-2 这个主机上了,然后在这个主机上创建 Pod 失败,
原因是 image pull failed for registry.access.redhat.com/rhel7/pod-infrastructure:latest, this may be because there are no credentials on this request.
通过以上信息,我们了解到通过红帽自家的 docker 仓库 pull 镜像,需要使用 CA 证书进行认证,才能 pull 成功
docker的证书在 /etc/docker/certs.d
目录下,根据上边的错误提示域名是 registry.access.redhat.com
,证书在这个目录中
经过 ll
命令查看,发现 /etc/docker/certs.d/registry.access.redhat.com/redhat-ca.crt
是一个软链接(软链接是什么?),指向到 /etc/rhsm/ca/redhat-uep.pem
,
熟悉软连接的我们知道,处于红色闪烁状态的目标是不存在,需要生成 /etc/rhsm/ca/redhat-uep.pem
证书文件
生成证书:
# openssl s_client -showcerts -servername registry.access.redhat.com -connect registry.access.redhat.com:443 </dev/null 2>/dev/null | openssl x509 -text > /etc/rhsm/ca/redhat-uep.pem
生成证书命令执行有时会出现
unable to load certificate 139930742028176:error:0906D06C:PEM routines:PEM_read_bio:no start line:pem_lib.c:707:Expecting: TRUSTED CERTIFICATE
问题,重新执行就好
命令执行完毕后,查看软链接指向的证书文件:
[root@kube-node-2 registry.access.redhat.com]# ll /etc/rhsm/ca/redhat-uep.pem
-rw-r--r-- 1 root root 9233 Apr 17 16:55 /etc/rhsm/ca/redhat-uep.pem
证书文件已经存在,我们去 k8s 管理节点 kube-master 主机删除刚才的 Pods,等待 Pod 重新创建成功 (第二个节点因为网络问题没有拉成功镜像……)
至此完成 Pod 的创建
但是还有存在些问题的,当前国内网络环境访问外边的网络偶尔会有问题,导致创建 Pod 失败,通过 describe 描述还是同样的信息提示,但是查看证书文件却存在且有内容
原因分析与其他方案
k8s 管理节点分配创建 Pod 到执行节点,到达执行节点后,拉取红帽 docker 仓库的 Pod基础镜像 pod-infrastructure:latest
,由于其仓库使用 https 需要验证证书,证书不存在导致失败
另外就是因为拉取的镜像是红帽 docker 仓库中的,在国内网络环境下握手失败,无法下载镜像
所以问题就成了 如何解决 k8s pod-infrastructure 镜像拉取失败
,这里给出一个方案,步骤如下:
-
拉取 docker 官方仓库其他人上传的
pod-infrastructure
镜像,docker pull tianyebj/pod-infrastructure
-
添加tag标签,改为私有仓库地址,如:
docker tag tianyebj/pod-infrastructure 10.2.7.70:5000/dev/pod-infrastructure
-
push镜像到私有仓库,如:
docker push 10.2.7.70:5000/dev/pod-infrastructure
-
修改所有 worker 节点的
/etc/kubernetes/kubelet
,修改registry.access.redhat.com/rhel7/pod-infrastructure
为刚才设置的 tag 标签sed -i "s#registry.access.redhat.com/rhel7/pod-infrastructure#<私有仓库pod-infrastructure镜像tag>#" /etc/kubernetes/kubelet
-
重启所有 worker 节点的 kubelet,
systemctl restart kubelet
,即可
注意事项:
- 上传的镜像要设为公开镜像,否则 kubelet 自己没权限拉镜像的,另外也可以去 ssh 登录 worker 节点登录仓库,执行
docker pull <私有仓库pod-infrastructure镜像tag>
最后的效果:
参考
https://github.com/CentOS/sig-atomic-buildscripts/issues/329
https://cloud.tencent.com/developer/article/1156329
本文采用 CC BY 4.0 协议进行授权,转载请标注作者署名及来源。
https://www.cnblogs.com/hellxz/p/k8s-pod-always-container-creating-status-problem.html