k8s集群安装
环境说明：
k8s-Master-Centos8 ip：192.168.152.53
k8s-Node1-Centos7 ip：192.168.152.253
k8s-Node2-Centos8 ip：192.168.152.252
注意：
Master与Node节点操作步骤基本一致
Node节点只需配置到本文的第6步即可
1、前期准备
#关闭防火墙 [root@Centos8 ~]# systemctl stop firewalld.service #关闭SElinux，永久关闭可修改/etc/selinux/config文件 [root@Centos8 ~]# setenforce 0 #关闭Swap分区，防止将K8S安装至swap内存中 [root@Centos8 ~]# swapoff -a [root@Centos8 ~]# sed -i '/swap/ s/^/#/g' /etc/fstab #安装iptables并设置为空规则及开机自启 yum -y install iptables-services iptables systemctl enable --now iptables.service iptables -F #清空规则 service iptables save #保存 iptables -L #检查是否为空规则
2、升级内核，建议 >= 4.4
#—————————————— Centos8 ————————————————# #——————Centos8最好不要做升级，如果非要要建议也升级到4.4 ————————# #——————以下只是演示升级过程，真实不要升级到5.6，貌似本身不支持nf_conntrack_ipv4 ————————# #安装ELrepo仓库 rpm --import https://www.elrepo.org/RPM-GPG-KEY-elrepo.org yum install https:www.elrepo.org/elrepo-release-8.0-2.el8.elrepo.noarch.rpm #查看可用的系统内核包 yum --disablerepo="*" --enablerepo=elrepo-kernel" list available #安装内核 yum --enablerepo=elrepo-kernel install kernel-ml # 安装完毕后，查看现有所有内核 grubby --info=ALL #设置启动内核 grubby --set-default /boot/vmlinuz-5.6.2-1.el8.elrepo.x86_64 #———————————————— Centos7 ————————————————# #下载内核rpm包 rpm -Uvh http:www.elrepo.org/elrepo-release-7.0-3.el7.elrepo.noarch.rpm #指定包名安装新版kernel install -y kernel-lt #设置默认启动的内核 grub2-set-default CentOS Linux (4.4.218-1.el7.elrepo.x86_64) 7 (Core)' #———————————————— 以下相同 ————————————————# #重启生效 systemctl reboot #重新查看内核版本 [root@Centos8 ~]# uname -r 1.el8.elrepo.x86_64
3、设置kube-proxy开启ipvs的前置条件(centos7及以上一般默认都开启)
modprobe br_netfilter cat > /etc/sysconfig/modules/ipvs.modules <<EOF #!/bin/bash modprobe -- ip_vs ip_vs_rr ip_vs_wrr ip_vs_sh nf_conntrack_ipv4 EOF chmod 755 /etc/sysconfig/modules/ipvs.modules && bash /etc/sysconfig/modules/ipvs.modules
# 查看是否加载成功ipvs模块
lsmod | grep -e ip_vs -e nf_conntrack_ipv4
nf_conntrack_ipv4 20480 4
nf_defrag_ipv4 16384 1 nf_conntrack_ipv4
ip_vs_sh 16384 0
ip_vs_wrr 16384 0
ip_vs_rr 16384 0
ip_vs 147456 6 ip_vs_rr,ip_vs_sh,ip_vs_wrr
nf_conntrack 114688 9 ip_vs,nf_nat,nf_nat_ipv4,nf_nat_ipv6,xt_conntrack,nf_nat_masquerade_ipv4,nf_conntrack_netlink,nf_conntrack_ipv4,nf_conntrack_ipv6
libcrc32c 16384 2 xfs,ip_vs
4、安装Docker
#安装环境 install -y yum-utils device-mapper-persistent-data lvm2 #添加docker仓库 yum-config-manager --add-repo http:mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo #安装dpcker install docker-ce 报错： package docker-ce-3:19.03.8-3.el7.x86_64 requires containerd.io >= 1.2.3,but none of the providers can be installed 先安装containerd.io >= ： dnf download.docker.com/linux/centos/7/x86_64/stable/Packages/containerd.io-1.2.6-3.3.el7.x86_64.rpm 再次安装docker： ce #启动docker，并设置开机自启 systemctl enable --now docker.service #配置daemon.设置默认的cgroup组为systemd，并使docker的日志以json形式输出
#并将镜像仓库源更改为阿里云镜像源 cat > /etc/docker/daemon.json <<EOF { exec-opts": [native.cgroupdriver=systemd],log-driver": json-filelog-opts: { max-size100m },
"registry-mirrors": ["https://f1bhsuge.mirror.aliyuncs.com"] } EOF #创建配置文件目录 mkdir -p /etc/systemd/system/docker.service.d #重新加载daemon 及 重启docker systemctl daemon-reload && systemctl restart docker.service
5、安装Kubeadm（主从配置）
#配置yum仓库 cat <<EOF > /etc/yum.repos.d/kubernetes.repo [kubernetes] name=Kubernetes baseurl=http:mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64 enabled= gpgcheck= repo_gpgcheck= gpgkey=http:mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg EOF #安装kubeadm、kubectl、kubelet install kubeadm-1.15.1 kubectl-1 kubelet- #设置开机自启，暂时不开启服务 systemctl enable kubelet.service
6、拉取初始化镜像
（1）两种方法，第一种可以通过以下脚本完成导入
vim initimage.sh ... #!/usr/bin/env bash K8S_VERSION=v1.15.1 ETCD_VERSION=3.3.10 DASHBOARD_VERSION=v1.8.3 FLANNEL_VERSION=v0.10.0-amd64 DNS_VERSION=1.3. PAUSE_VERSION=3.1 # 基本组件 docker pull mirrorgooglecontainers/kube-apiserver-amd64:$K8S_VERSION docker pull mirrorgooglecontainers/kube-controller-manager-amd64:$K8S_VERSION docker pull mirrorgooglecontainers/kube-scheduler-amd64:$K8S_VERSION docker pull mirrorgooglecontainers/kube-proxy-amd64:$K8S_VERSION docker pull mirrorgooglecontainers/etcd-amd64:$ETCD_VERSION docker pull mirrorgooglecontainers/pause:$PAUSE_VERSION docker pull coredns/coredns:$DNS_VERSION # 修改tag docker tag mirrorgooglecontainers/kube-apiserver-amd64:$K8S_VERSION k8s.gcr.io/kube-apiserver:$K8S_VERSION docker tag mirrorgooglecontainers/kube-controller-manager-amd64:$K8S_VERSION k8s.gcr.io/kube-controller-manager:$K8S_VERSION docker tag mirrorgooglecontainers/kube-scheduler-amd64:$K8S_VERSION k8s.gcr.io/kube-scheduler:$K8S_VERSION docker tag mirrorgooglecontainers/kube-proxy-amd64:$K8S_VERSION k8s.gcr.io/kube-proxy:$K8S_VERSION docker tag mirrorgooglecontainers/etcd-amd64:$ETCD_VERSION k8s.gcr.io/etcd:$ETCD_VERSION docker tag mirrorgooglecontainers/pause:$PAUSE_VERSION k8s.gcr.io/pause:$PAUSE_VERSION docker tag coredns/coredns:$DNS_VERSION k8s.gcr.io/coredns:$DNS_VERSION #删除冗余的images docker rmi mirrorgooglecontainers/kube-apiserver-amd64:$K8S_VERSION docker rmi mirrorgooglecontainers/kube-controller-manager-amd64:$K8S_VERSION docker rmi mirrorgooglecontainers/kube-scheduler-amd64:$K8S_VERSION docker rmi mirrorgooglecontainers/kube-proxy-amd64:$K8S_VERSION docker rmi mirrorgooglecontainers/etcd-amd64:$ETCD_VERSION docker rmi mirrorgooglecontainers/pause:$PAUSE_VERSION docker rmi coredns/coredns:$DNS_VERSION ... chmod +x initimage. #执行此脚本，开始导入镜像... ./initimage. [root@Centos8 ]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE quay.io/coreos/flannel v0.12.0-amd64 4e9f801d2217 4 weeks ago 52.8MB k8s.gcr.io/kube-scheduler v1.15.1 b0b3c4c404da 8 months ago 81.1MB k8s.gcr.io/kube-controller-manager v1.15.1 d75082f1d121 8 months ago 159MB k8s.gcr.io/kube-proxy v1.15.1 89a062da739d 82.4MB k8s.gcr.io/kube-apiserver v1.15.1 68c3eb07bfc3 months ago 207MB k8s.gcr.io/coredns 1 eb516548c180 15 months ago 40.3MB k8s.gcr.io/etcd 10 2c4adeb21b4f 16 months ago 258MB k8s.gcr.io/pause 3.1 da86e6ba6ca1 2 years ago 742kB
（2） 第二种可以在k8s配置文件初始化后，也就是下边的第七步之后，再执行导入镜像操作
7、初始化主节点
#打印k8s默认的初始化配置文件至kubeadm-conf.yaml中 kubeadm config print init-defaults > kubeadm-conf.yaml #修改kubeadm-conf.yaml vim kubeadm-conf.yaml ... apiVersion: kubeadm.k8s.io/v1beta2 bootstrapTokens: - groups: - system:bootstrappers:kubeadm:default-node-token token: abcdef.0123456789abcdef ttl: 24h0m0s usages: - signing - authentication kind: InitConfiguration localAPIEndpoint: advertiseAddress: 192.168.152.53 #此处需要修改为你的真实ip bindPort: 6443 nodeRegistration: criSocket: /var/run/dockershim.sock name: centos8 taints: - effect: NoSchedule key: node-role.kubernetes.io/master --- apiServer: timeoutForControlPlane: 4m0s apiVersion: kubeadm.k8s.io/v1beta2 certificatesDir: /etc/kubernetes/pki clusterName: kubernetes controllerManager: {} dns: type: CoreDNS etcd: local: dataDir: /var/lib/etcd imageRepository: registry.cn-hangzhou.aliyuncs.com/google_containers #镜像站点的更改，也可以不改 kind: ClusterConfiguration kubernetesVersion: v1. #版本信息修改一致 networking: dnsDomain: cluster.local podSubnet: 10.244.0.0/16 #添加此配置，用于指定flannel的默认PodNet网段 serviceSubnet: 10.96.0.0/12 scheduler: {} --- #以下为新增配置，将默认的调度方式改为ipvs apiVersion: kubeproxy.config.k8s.io/v1alpha1 kind: KubeProxyConfiguration featureGates: SupportIPVSProxyMode: true mode: ipvs ...
# 提前拉取配置文件种所需要的image，对应上边第6步的(2)
kubeadm config images pull --config kubeadm-conf.yaml
#开始初始化 kubeadm init --config=kubeadm-config.yaml --experimental-upload-certs | tee kubeadm-init.log 报错： [ERROR NumCPU]: the number of available CPUs 1 is less than the required 2 很明显，cpu数目小于所需的2，重新设置一个虚拟机CPU个数就ok 初始化成功回显： To start using your cluster,you need to run the following as a regular user: mkdir -p $HOME/.kube sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config chown $(id -u):$(id -g) $HOME/.kube/config kubeadm join 152.53:6443 --token abcdef.0123456789abcdef \ --discovery-token-ca-cert-hash sha256:50ca5375950abfa05cd4bd37dfb60e9ccd078083aeca49fa8bb6275c13d2a2cd #根据回显创建文件及目录 #目的为保存 kubectl 与 api server 交互时的缓存，交互过程为https协议 .kube config config #查看当前拥有节点，可见状态为NotReady，是因为没有添加网络 [root@Centos8 .kube]# kubectl get node NAME STATUS ROLES AGE VERSION centos8 NotReady master 4m1s v1.15.1
8、添加flannel网络
（1）第一种
mkdir -p install-k8s/plugin/flannel install-k8s/core cd core mv /etc/kubernetes/kubeadm-init.log /etc/kubernetes/kubeadm-config.yaml ./ cd ../plugin/flannel #下载flannel.yml wget https:raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml #执行yaml文件 kubectl create -f kube-flannel.yml
#构建完成，查看当前名称空间为kube-system的pod状态，-n 指定名称空间 [root@Centos8 core]# kubectl get pod -n kube-system NAME READY STATUS RESTARTS AGE coredns-5c98db65d4-5gwmj 0/1 CrashLoopBackOff 22 91m coredns-5c98db65d4-c277w 91m etcd-centos8 1/1 Running 90m kube-apiserver-centos8 90m kube-controller-manager-centos8 90m kube-flannel-ds-amd64-ggghn 8m45s kube-proxy-gslw2 91m kube-scheduler-centos8 90m #构建完成后，网卡界面会显示flannel信息 [root@Centos8 core]# ifconfig flannel.1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1450 inet 10.244.0.0 netmask 255.255.255.255 broadcast 0.0.0.0 inet6 fe80::4019:beff:fe7c:5582 prefixlen 64 scopeid 0x20<link> ether 42:19:be:7c:55:82 txqueuelen (Ethernet) RX packets 0 bytes 0 ( B) RX errors 0 dropped 0 overruns 0 frame TX packets B) TX errors 13 overruns 0 carrier 0 collisions #网络加载成功，状态变为Ready [root@Centos8 ~]# kubectl get node NAME STATUS ROLES AGE VERSION centos8 Ready master 53d v1.15.1
（2）第二种，当 kubectl create -f kube-flannel.yml 时，镜像导入不成功，或者因为网络问题导致失败时
# 修改主机hosts文件 echo 199.232.28.133 raw.githubusercontent.com" >> /etc/hosts # 然后下载flannel文件 curl -o kube-flannel.yml https: # 编辑镜像源，默认的镜像地址我们修改一下。把yaml文件中所有的quay.io 修改为quay-mirror.qiniu.com s/quay.io/quay-mirror.qiniu.com/g' kube-flannel.yml # 最后再执行创建命令就可以啦 kubectl apply -f kube-flannel.yml
9、配置从节点与主节点关联
# 直接将kubeadm-init.log的最后一句回显在node节点运行即可 [root@kube-node2 ~]# kubeadm token abcdef.0123456789abcdef \ > --discovery-token-ca-cert-hash sha256:c291b4fc646b5925299f8cdf7fafe33ad9c0505a1609041d8c8214d104eb08da [root@kube-node2 ~]# kubeadm token abcdef.0123456789abcdef \ > --discovery-token-ca-cert-hash sha256:c291b4fc646b5925299f8cdf7fafe33ad9c0505a1609041d8c8214d104eb08da
遇到的问题及启动服务报错：
1、W0412 04:57:19.803140 846 watcher.go:87] Error while processing event ("/sys/fs/cgroup/devices/libcontainer_30695_systemd_test_default.slice": 0x40000100 == IN_CREATE|IN_ISDIR): inotify_add_watch /sys/fs/cgroup/devices/libcontainer_30695_systemd_test_default.slice: no such file or directory
## 问题的原因是docker与kubelet的cgroup组不一致
## 查看kubeadm的配置文件位置
[root@Centos8 ~]# rpm -ql kubeadm
/usr/bin/kubeadm
/usr/lib/systemd/system/kubelet.service.d/10-kubeadm.conf
## 修改配置文件,将kubelet的cgroup改为systemd
vim /usr/lib/systemd/system/kubelet.service.d/10-kubeadm.conf
...
Environment="KUBELET_CONFIG_ARGS=--config=/var/lib/kubelet/config.yaml --cgroup-driver=systemd"
...
## 重新加载配置文件及重启服务
[root@Centos8 ~]# systemctl daemon-reload
[root@Centos8 ~]# systemctl restart kubelet.service
2、node2节点的flannel一直报错
## 还未找到根本原因，正在排错，如有大佬路过，希望指点
kube-flannel-ds-amd64-b47l9 0/1 Init:ErrImagePull 0 3m50s
3、如果需要重新安装master节点，执行：
# kubeadm reset
然后再进行重新安装
4、如果需要重新安装node节点，执行：
# kubectl drain <node name> --delete-local-data --force --ignore-daemonsets
# kubectl delete node <node name>
或
kubeadm reset #直接重新join
5、node节点加入master时报错：error execution phase preflight: couldn't validate the identity of the API Server: abort connecting to API servers after timeout of 5m0s
原因：master的token过期了，需要重新创建
解决：
Master：
[root@Centos8 ~]# kubeadm token create
blopur.fn8gtr06gsjlq7yi
Node：
kubeadm join 192.168.152.53:6443 --token blopur.fn8gtr06gsjlq7yi --discovery-token-ca-cert-hash sha256:c291b4fc646b5925299f8cdf7fafe33ad9c0505a1609041d8c8214d104eb08da
6、node2节点加入master后，systemctl status kubelet.service报错：Unable to update cni config: No networks found in /etc/cni/net.d
vim /usr/lib/systemd/system/kubelet.service.d/10-kubeadm.conf
...
添加：
Environment="KUBELET_NETWORK_ARGS=--network-plugin=cni --cni-conf-dir=/etc/cni/ --cni-bin-dir=/opt/cni/bin"
...
systemctl daemon-reload && systemctl restart kubelet.service
7、node2节点加入master后，systemctl status kubelet.service报错：open /run/flannel/subnet.env: no such file or directory
## 将master节点的/run/flannel/subnet.env文件拷贝到此处即可
[root@Centos8 flannel]# scp subnet.env kubenode2:/run/flannel/
## 重启
systemctl restart kubelet.service
补充：
node节点为了方便以后的部署和扩展，我采用Ansible自动化安装docker及k8s。
ansible服务端为k8s的MASTER端：192.168.152.53
客户端根据node节点需求来控制
#### 注意：所有的操作及文件都在192.168.152.53中 #### [root@Centos8 ansible]# pwd /root/ansible *** 先编辑好导入导出images的脚本 *** [root@Centos8 ansible]# cat saveImages. #导出镜像脚本 #!/usr/bin/ bash IMAGESNAME=(`docker images | awk /ago/{print $1}`) IMAGESTAG=(`docker images | /ago/{print $2}`) IPADDR1=192.168.152.253 IPADDR2=192.168.152.252' if [[ -d /root/images ]];then for i in `seq 0 6`;do docker save > /root/images/${i}.tar.gz ${IMAGESNAME[$i]}:${IMAGESTAG[$i]} scp /root/images/${i}.tar.gz ${IPADDR1}:/root/ tar.gz ${IPADDR2}:/root/ done else mkdir -p /root/images docker save > /root/images/${IMAGESNAME[$i]}.fi [root@Centos8 ansible]# cat loadImages. #导入镜像脚本 #!/usr/bin/ bash docker load < /root/${i}.tar.gz && rm -f /root/${i}..gz done *** 在本地执行saveImages.sh *** ./saveImages.sh *** 创建PlayBook *** [root@Centos8 ansible]# vim kuber.yaml ... --- - name: Install docker and k8s hosts: all tasks: - block: - name: Add repository shell: mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo - name: ass yum: name: - yum-utils - device-mapper-persistent-data - lvm2 - https:download.docker.com/linux/centos/7/x86_64/stable/Packages/containerd.io-1.2.6-3.3.el7.x86_64.rpm - docker-ce state: latest - name: mkdir docker.service.d file: path: {{ item }} state: directory loop: - /etc/docker - /etc/systemd/system/docker.service.d - name: Copy daemon.json copy: src: /etc/docker/daemon.json dest: /etc/docker/daemon.json - name: daemon-reload shell: systemctl daemon-reload - name: Start docker service: name: docker state: restarted enabled: yes rescue: - debug: msg: docker Installation failed!' - name: Copy using inline content copy: content: [kubernetes] name=Kubernetes baseurl=http:mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64 enabled= gpgcheck= repo_gpgcheck= gpgkey=http:mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg' dest: /etc/kubernetes.repo - name: Install k8s : name: - kubeadm-1 - kubectl-1 - kubelet- state: present - name: Start kubelet service: name: kubelet state: started enabled: yes - name: Run script load Images script: /root/ansible/loadImages. ... ansible-playbook kuber.yaml #执行，完成，最后在node节点再执行以上第8步即可
文章浏览阅读3.8k次。上篇文章详细介绍了弹性云混部的落地历...
文章浏览阅读897次。对于cpu来说，这种分配方式并不会有太大...
文章浏览阅读763次。但是此时如果配置成 NONE, 租户创建成功...