search

cloudnativer / kube-install

一键安装k8s(kubernetes)：二進位離線方式部署，支持定時安裝、添加與銷毀node、銷毀與修復master、一鍵卸載集羣等。Install k8s (kubernetes): one click offline installation of highly available multiple kubernetes cluster, supports schedule installation, addition of nodes, rebuild of kubernetes master, and uninstallation of clusters.
Apache License 2.0
422 stars 118 forks source link

不知道哪里有问题,安装出现5个ignore问题,service启动不了 #28

Closed yingpwu closed 2 years ago

yingpwu commented 2 years ago

TASK [/root/kube-install/data/output/cluster/sys/0x0000000000base/runtime : 1.Cleaning up garbage files left in history.] *** fatal: [10.10.10.52]: FAILED! => {"changed": false, "msg": "Could not find the requested service kubelet: host"} ...ignoring fatal: [10.10.10.51]: FAILED! => {"changed": false, "msg": "Could not find the requested service kubelet: host"} ...ignoring

TASK [/root/kube-install/data/output/cluster/sys/0x0000000000base/runtime : service] *** fatal: [10.10.10.52]: FAILED! => {"changed": false, "msg": "Could not find the requested service kube-proxy: host"} ...ignoring fatal: [10.10.10.51]: FAILED! => {"changed": false, "msg": "Could not find the requested service kube-proxy: host"} ...ignoring

TASK [/root/kube-install/data/output/cluster/sys/0x0000000000base/runtime : service] * fatal: [10.10.10.52]: FAILED! => {"changed": false, "msg": "Could not find the requested service containerd: host"} ...ignoring fatal: [10.10.10.51]: FAILED! => {"changed": false, "msg": "Could not find the requested service containerd: host"} ...ignoring ........................ PLAY RECAP *** 10.10.10.50 : ok=128 changed=115 unreachable=0 failed=0 skipped=0 rescued=0 ignored=5
10.10.10.51 : ok=85 changed=68 unreachable=0 failed=0 skipped=0 rescued=0 ignored=5
10.10.10.52 : ok=85 changed=68 unreachable=0 failed=0 skipped=0 rescued=0 ignored=5

[Info] 2022-01-09 08:41:26.594549917 -0500 EST m=+1013.176189743 Kubernetes cluster install completed!

系统上centos7 node上docker等service都启动不了 pod无法部署

yingpwu commented 2 years ago

试了3遍都这样.....

houseonline commented 2 years ago

你可以試試選擇k8s v1. 23或以下版本的k8s進行部署。

20220110125114

如果你安裝的是k8s 1.24版本，也許會存在一些未知問題。 因為cncf官方還未發佈正式版本的k8s v1. 24，所以k8s v1. 24現時還處於測試階段，難免會存在不穩定。

yingpwu commented 2 years ago

非常感谢你的回复,测试过1.23和1.24,都是一样的这个错误,每个机器上都有5个ignore,基本上都是关于service无法注册

houseonline commented 2 years ago

凡是標記為ignoring的都可以忽略。 （1）你看看日誌裡面是否還有其他沒有ignoring的錯誤？ （2）你安裝的是什麼版本的k8s？ cnicni挿件安裝的是什麼？ （3）你的機器是虛擬機器還是硬體服務器？ 你的機器上是否以前就已經安裝過k8s或docker？ （4）是master和node都啟動不了？ 還是只有node啟動不了，但是master正常。

kubectl get cs #查看master狀態
kubectl get node #查看node狀態

（5）如果docker或kubelet啟動不了，可以提供一下docker或kubelet的錯誤日誌資訊。

yingpwu commented 2 years ago

TASK [/root/kube-install/data/output/cluster/sys/0x0000000network/flannel : 1.Create cniplugin directory] *** changed: [10.10.10.50] => (item=flannel) changed: [10.10.10.50] => (item=calico) changed: [10.10.10.50] => (item=kuberouter) changed: [10.10.10.50] => (item=weave) changed: [10.10.10.50] => (item=cilium)

TASK [/root/kube-install/data/output/cluster/sys/0x0000000network/flannel : 2.Make docker options] *** fatal: [10.10.10.50]: FAILED! => {"changed": false, "checksum": "155274ebca961f16fb2cf9d87a3e66a540ec6c03", "msg": "Destination directory /opt/cni/bin does not exist"}

PLAY RECAP ***** 10.10.10.50 : ok=91 changed=81 unreachable=0 failed=1 skipped=0 rescued=0 ignored=3 10.10.10.51 : ok=73 changed=51 unreachable=0 failed=0 skipped=0 rescued=0 ignored=10 10.10.10.52 : ok=73 changed=51 unreachable=0 failed=0 skipped=0 rescued=0 ignored=10

yingpwu commented 2 years ago

0.7.3版本测试日志

yingpwu commented 2 years ago

机器是虚拟机,全新centos7系统 测试状态

[root@master4k8s ~]# kubectl get cs Warning: v1 ComponentStatus is deprecated in v1.19+ NAME STATUS MESSAGE ERROR scheduler Healthy ok etcd-0 Healthy {"health":"true"} controller-manager Healthy ok [root@master4k8s ~]# kubectl get nod error: the server doesn't have a resource type "nod" [root@master4k8s ~]# kubectl get node No resources found [root@master4k8s ~]# kubectl get pod No resources found in default namespace. [root@master4k8s ~]# kubectl run nginx --image=nginx pod/nginx created [root@master4k8s ~]# kubectl get pods NAME READY STATUS RESTARTS AGE nginx 0/1 Pending 0 15s [root@master4k8s ~]# ^C [root@master4k8s ~]# kubectl get pods NAME READY STATUS RESTARTS AGE nginx 0/1 Pending 0 46s [root@master4k8s ~]#

cloudnativer commented 2 years ago

**TASK [/root/kube-install/data/output/cluster/sys/0x0000000network/flannel : 1.Create cniplugin directory] *** changed: [10.10.10.50] => (item=flannel) changed: [10.10.10.50] => (item=calico) changed: [10.10.10.50] => (item=kuberouter) changed: [10.10.10.50] => (item=weave) changed: [10.10.10.50] => (item=cilium)

TASK [/root/kube-install/data/output/cluster/sys/0x0000000network/flannel : 2.Make docker options] *** fatal: [10.10.10.50]: FAILED! => {"changed": false, "checksum": "155274ebca961f16fb2cf9d87a3e66a540ec6c03", "msg": "Destination directory /opt/cni/bin does not exist"}

Prompt that a directory ( /opt/cni/bin ) does not exist, so the installation is not successful.

This bug is already Fixed in v0.7.4.

0.7.3版本测试日志

You can try v0.7.4: https://github.com/cloudnativer/kube-install/releases/tag/v0.7.4 Installation documentation: https://github.com/cloudnativer/kube-install/blob/master/README0.7.md

houseonline commented 2 years ago

如果是中國大陸的用戶，你還可以加入微信技術交流群，進一步溝通技術問題。

20220110134655

houseonline commented 2 years ago

[root@master4k8s ~]# kubectl get node No resources found

這說明你的node不正常，所以導致nginx是Pending的。 如果docker或kubelet啟動不了，可以提供一下docker或kubelet的錯誤日誌資訊。

systemctl status docker -l
systemctl status kubelet -l
cloudnativer commented 2 years ago

0.7.3版本测试日志

Both kube-install v0.7.4 and kube-install v0.8.* versions are installed without this problem.

yingpwu commented 2 years ago

问题已解决,0.8.0beta安装1.24版本,1.24版本不支持docker. 如果依然想使用docker作为k8s的默认容器运行时，可以安装k8s 1.23或以下版本