k8s通过Service访问Pod
如何创建服务
1、创建Deployment
#启动三个pod，运行httpd镜像，label是run：mcw-httpd,Seveice将会根据这个label挑选Pod
apiVersion: apps/v1
[machangwei@mcwk8s-master ~]$ cat mcwHttpd.yml kind: Deployment metadata: name: mcw-httpd spec: replicas: 3 selector: matchLabels: run: mcw-httpd template: metadata: labels: run: mcw-httpd spec: containers: - name: mcw-httpd image: httpd ports: - containerPort: 80 [machangwei@mcwk8s-master ~]$ kubectl apply -f mcwHttpd.yml deployment.apps/mcw-httpd created [machangwei@mcwk8s-master ~]$ kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5rrkh 0/1 ContainerCreating 0 2m52s <none> mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-bqq58 0/1 ImagePullBackOff 0 2m52s 10.244.0.78 mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-j52ff 0/1 ImagePullBackOff 0 2m52s 10.244.0.70 mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~]$ kubectl get pod -o wide #过一会后，查看Pod分配了各自的IP，容器在创建的时候ip是none NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5rrkh 1/1 Running 0 6m28s 10.244.0.71 mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-bqq58 1/1 Running 0 6m28s 10.244.0.78 mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-j52ff 1/1 Running 0 6m28s 10.244.0.70 mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~]$ #这些ip只能被kubernates Cluster中的容器和节点访问
2、不通的情况，是不是就应该不通呢，答案是否。
[machangwei@mcwk8s-master ~]$ kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5rrkh 1/1 Running 0 16m 10.244.0.71 mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-bqq58 1/1 Running 0 16m 10.244.0.78 mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-j52ff 1/1 Running 0 16m 10.244.0.70 mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~]$ curl 10.244.0.78 #主节点访问节点2上的pod的ip，连接超时 curl: (7) Failed connect to 10.244.0.78:80; Connection timed out 节点2上访问节点2上的pod ip是能访问的 [root@mcwk8s-node2 ~]$ curl 10.244.0.78 <html><body><h1>It works!</h1></body></html> 怀疑是节点上flannel状态问题 [machangwei@mcwk8s-master ~]$ kubectl get pod --all-namespaces|grep flannel kube-system kube-flannel-ds-cn4m9 0/1 Error 233 (5m26s ago) 2d11h kube-system kube-flannel-ds-hpgkz 1/1 Running 0 6d23h kube-system kube-flannel-ds-nnjvj 0/1 CrashLoopBackOff 271 (15s ago) 6d23h 怀疑节点2上没添加-H的问题，然后添加重启docker daemon ，四个容器都重启了 [root@mcwk8s-node2 ~]$ vim /usr/lib/systemd/system/docker.service [root@mcwk8s-node2 ~]$ grep -i execstart /usr/lib/systemd/system/docker.service ExecStart=/usr/bin/dockerd -H fd:// -H tcp://0.0.0.0 --containerd=/run/containerd/containerd.sock [root@mcwk8s-node2 ~]$ systemctl daemon-reload [root@mcwk8s-node2 ~]$ systemctl restart docker [root@mcwk8s-node2 ~]$ docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES 88de5020b420 registry.aliyuncs.com/google_containers/pause:3.6 "/pause" 1 second ago Up 1 second k8s_POD_kube-flannel-ds-cn4m9_kube-system_ef070440-6778-430f-92b9-a1c48b755d2b_1 adf80a28c0be b46c42588d51 "/usr/local/bin/kube…" 2 seconds ago Up 1 second k8s_kube-proxy_kube-proxy-92g5c_kube-system_a69acf11-f51a-46d6-9472-d54b5383efef_1 46bceff879bd registry.aliyuncs.com/google_containers/pause:3.6 "/pause" 6 seconds ago Up 2 seconds k8s_POD_kube-proxy-92g5c_kube-system_a69acf11-f51a-46d6-9472-d54b5383efef_1 a5cdf7f6ef3b registry.aliyuncs.com/google_containers/pause:3.6 "/pause" 6 seconds ago Up 2 seconds k8s_POD_mcw-httpd-6fbf67d7d5-bqq58_default_4ceffe1e-df14-47dd-82f0-83cb68555de7_1 再次在主节点访问节点2上的pod ip ,还是无法访问，由于节点2上pod重启了，所以ip被重新分配了一个 [machangwei@mcwk8s-master ~]$ kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5rrkh 1/1 Running 0 30m 10.244.0.71 mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-bqq58 1/1 Running 1 (95s ago) 30m 10.244.0.79 mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-j52ff 1/1 Running 0 30m 10.244.0.70 mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~]$ [machangwei@mcwk8s-master ~]$ [machangwei@mcwk8s-master ~]$ curl 10.244.0.79 #虽然重新分配了ip，但是还是无法访问 curl: (7) Failed connect to 10.244.0.79:80; Connection timed out
3、创建Service
[machangwei@mcwk8s-master ~]$ cat mcwHttpdService.yml apiVersion: v1 kind: Service metadata: name: httpd-svc spec: selector: run: httpd ports: - protocol: TCP port: 8080 targetPort: 80 [machangwei@mcwk8s-master ~]$ kubectl apply -f mcwHttpdService.yml service/httpd-svc created [machangwei@mcwk8s-master ~]$ kubectl get service NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc ClusterIP 10.99.77.45 <none> 8080/TCP 14s kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 7d [machangwei@mcwk8s-master ~]$ curl 10.99.77.45:8080 curl: (7) Failed connect to 10.99.77.45:8080; Connection refused
dns访问Service
[machangwei@mcwk8s-master ~]$ kubectl get deployment --namespace=kube-system NAME READY UP-TO-DATE AVAILABLE AGE coredns 2/2 2 2 7d1h [machangwei@mcwk8s-master ~]$ kubectl get service -o wide NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE SELECTOR httpd-svc ClusterIP 10.99.77.45 <none> 8080/TCP 61m run=httpd kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 7d1h <none> [machangwei@mcwk8s-master ~]$ kubectl run mcwbusybox --rm -ti --image=busybox /bin/bash pod "mcwbusybox" deleted error: timed out waiting for the condition [machangwei@mcwk8s-master ~]$ kubectl get pod NAME READY STATUS RESTARTS AGE mcw-httpd-6fbf67d7d5-5rrkh 1/1 Running 0 124m mcw-httpd-6fbf67d7d5-bqq58 1/1 Running 1 (95m ago) 124m mcw-httpd-6fbf67d7d5-j52ff 1/1 Running 0 124m 只可以访问自己节点上的pod，不能跨主机同命名空间内的pod互相通信。不知道哪里的问题 [machangwei@mcwk8s-master ~]$ kubectl run mcwcentos3 -it --image=centos /bin/bash If you don't see a command prompt, try pressing enter. [root@mcwcentos3 /]# curl 10.244.0.79 curl: (7) Failed to connect to 10.244.0.79 port 80: Connection timed out [root@mcwcentos3 /]# curl 10.244.0.70 <html><body><h1>It works!</h1></body></html> [root@mcwcentos3 /]# curl 10.244.0.71 <html><body><h1>It works!</h1></body></html> route add -host 10.244.0.0 dev flannel.1 flannel.1 NMI watchdog: BUG: soft lockup - CPU#0 stuck for 22s 报错现象 系统或者网络占用过多CPU,造成内核软死锁(soft lockup)。Soft lockup名称解释:所谓,soft lockup就是说,这个bug没有让系统彻底死机,但是若干个进程(或...
fannel网络问题
网络是有问题的
两个节点上的fannel状态不对 [machangwei@mcwk8s-master ~]$ kubectl get pod --all-namespaces -o wide|grep flannel kube-system kube-flannel-ds-cn4m9 0/1 CrashLoopBackOff 286 (3m50s ago) 3d10h 10.0.0.6 mcwk8s-node2 <none> <none> kube-system kube-flannel-ds-hpgkz 1/1 Running 1 (22h ago) 7d22h 10.0.0.4 mcwk8s-master <none> <none> kube-system kube-flannel-ds-nnjvj 0/1 CrashLoopBackOff 325 (3m21s ago) 7d22h 10.0.0.5 mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~] 节点上查看对应容器的日志错误信，未成功注册cidr [root@mcwk8s-node2 ~]$ docker ps -a|grep flannel 2252229253e3 e6ea68648f0c "/opt/bin/flanneld -…" 10 seconds ago Exited (1) 2 seconds ago k8s_kube-flannel_kube-flannel-ds-cn4m9_kube-system_ef070440-6778-430f-92b9-a1c48b755d2b_284 [root@mcwk8s-node2 ~]$ docker logs 225 E0120 13:40:49.244765 1 main.go:325] Error registering network: failed to acquire lease: node "mcwk8s-node2" pod cidr not assigned W0120 13:40:49.245111 1 reflector.go:424] github.com/flannel-io/flannel/subnet/kube/kube.go:379: watch of *v1.Node ended with: an error on the server ("unable to decode an event from the watch stream: context canceled") has prevented the request from succeeding master初始化写进来的cidr [root@mcwk8s-master ~]$ grep cidr /etc/kubernetes/manifests/kube-controller-manager.yaml - --allocate-node-cidrs=true - --cluster-cidr=10.244.0.0/24 部署fannel用的网络 [machangwei@mcwk8s-master ~]$ grep -C 3 '"Network"' mm.yml } net-conf.json: | { "Network": "10.244.0.0/16", "Backend": { "Type": "vxlan" } 修改网络 [machangwei@mcwk8s-master ~]$ vim mm.yml [machangwei@mcwk8s-master ~]$ grep -C 3 '"Network"' mm.yml } net-conf.json: | { "Network": "10.244.0.0/24", "Backend": { "Type": "vxlan" } [machangwei@mcwk8s-master ~]
从头部署了k8s
修改网络后还是不行。直接从头部署k8s就好了，把flannel网络和初始化master都设置为10.244.0.0/16 网络 [machangwei@mcwk8s-master ~]$ kubectl get pod --all-namespaces NAMESPACE NAME READY STATUS RESTARTS AGE kube-system coredns-6d8c4cb4d-vctxx 1/1 Running 1 (5m57s ago) 33m kube-system coredns-6d8c4cb4d-xkv9h 1/1 Running 0 33m kube-system etcd-mcwk8s-master 1/1 Running 0 33m kube-system kube-apiserver-mcwk8s-master 1/1 Running 1 (4m42s ago) 33m kube-system kube-controller-manager-mcwk8s-master 1/1 Running 1 33m kube-system kube-flannel-ds-fvwgm 1/1 Running 0 22m kube-system kube-flannel-ds-l5fdg 1/1 Running 0 25m kube-system kube-flannel-ds-mzdcw 1/1 Running 0 21m kube-system kube-proxy-796l7 1/1 Running 0 21m kube-system kube-proxy-8wtxn 1/1 Running 0 22m kube-system kube-proxy-qr6b8 1/1 Running 0 33m kube-system kube-scheduler-mcwk8s-master 1/1 Running 1 33m [machangwei@mcwk8s-master ~]$ [machangwei@mcwk8s-master ~]$ kubectl get nodes NAME STATUS ROLES AGE VERSION mcwk8s-master Ready control-plane,master 35m v1.23.1 mcwk8s-node1 Ready <none> 23m v1.23.1 mcwk8s-node2 Ready <none> 22m v1.23.1
网络正常的三个节点，网卡和路由如下
跟之前相比，很明显的是多了路由了。之前应该是节点网络没好，所以路由也不全 主节点 [root@mcwk8s-master ~]$ ip a 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet6 ::1/128 scope host valid_lft forever preferred_lft forever 2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:a2:3a:b7 brd ff:ff:ff:ff:ff:ff inet 10.0.0.4/24 brd 10.0.0.255 scope global ens33 valid_lft forever preferred_lft forever inet6 fe80::6b7a:2214:bef3:5850/64 scope link valid_lft forever preferred_lft forever 3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000 link/ether 00:0c:29:a2:3a:c1 brd ff:ff:ff:ff:ff:ff 4: docker0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN link/ether 02:42:af:00:ac:08 brd ff:ff:ff:ff:ff:ff inet 172.17.0.1/16 brd 172.17.255.255 scope global docker0 valid_lft forever preferred_lft forever 5: flannel.1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN link/ether ca:9d:fb:2c:b7:22 brd ff:ff:ff:ff:ff:ff inet 10.244.0.0/32 brd 10.244.0.0 scope global flannel.1 valid_lft forever preferred_lft forever inet6 fe80::c89d:fbff:fe2c:b722/64 scope link valid_lft forever preferred_lft forever 6: cni0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UP qlen 1000 link/ether 3e:f7:13:d6:6b:5b brd ff:ff:ff:ff:ff:ff inet 10.244.0.1/24 brd 10.244.0.255 scope global cni0 valid_lft forever preferred_lft forever inet6 fe80::3cf7:13ff:fed6:6b5b/64 scope link valid_lft forever preferred_lft forever 7: vethf9bacd46@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue master cni0 state UP link/ether f6:9d:2c:f7:ff:4e brd ff:ff:ff:ff:ff:ff link-netnsid 0 inet6 fe80::f49d:2cff:fef7:ff4e/64 scope link valid_lft forever preferred_lft forever 8: veth59372586@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue master cni0 state UP link/ether 56:da:7b:cb:75:49 brd ff:ff:ff:ff:ff:ff link-netnsid 1 inet6 fe80::54da:7bff:fecb:7549/64 scope link valid_lft forever preferred_lft forever [root@mcwk8s-master ~]$ node1 [root@mcwk8s-node1 ~]$ route -n Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface 0.0.0.0 10.0.0.253 0.0.0.0 UG 100 0 0 ens33 10.0.0.0 0.0.0.0 255.255.255.0 U 100 0 0 ens33 10.244.0.0 10.244.0.0 255.255.255.0 UG 0 0 0 flannel.1 10.244.1.0 0.0.0.0 255.255.255.0 U 0 0 0 cni0 10.244.2.0 10.244.2.0 255.255.255.0 UG 0 0 0 flannel.1 172.17.0.0 0.0.0.0 255.255.0.0 U 0 0 0 docker0 [root@mcwk8s-node1 ~]$ ip a 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet6 ::1/128 scope host valid_lft forever preferred_lft forever 2: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000 link/ether 00:0c:29:0b:a3:15 brd ff:ff:ff:ff:ff:ff 3: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:0b:a3:0b brd ff:ff:ff:ff:ff:ff inet 10.0.0.5/24 brd 10.0.0.255 scope global ens33 valid_lft forever preferred_lft forever inet6 fe80::3516:c22b:d62:c43f/64 scope link valid_lft forever preferred_lft forever 4: docker0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN link/ether 02:42:5e:08:56:63 brd ff:ff:ff:ff:ff:ff inet 172.17.0.1/16 brd 172.17.255.255 scope global docker0 valid_lft forever preferred_lft forever 5: flannel.1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN link/ether 4e:92:a2:b0:6b:5a brd ff:ff:ff:ff:ff:ff inet 10.244.1.0/32 brd 10.244.1.0 scope global flannel.1 valid_lft forever preferred_lft forever inet6 fe80::4c92:a2ff:feb0:6b5a/64 scope link valid_lft forever preferred_lft forever 6: cni0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UP qlen 1000 link/ether e2:d3:b2:00:28:bf brd ff:ff:ff:ff:ff:ff inet 10.244.1.1/24 brd 10.244.1.255 scope global cni0 valid_lft forever preferred_lft forever inet6 fe80::e0d3:b2ff:fe00:28bf/64 scope link valid_lft forever preferred_lft forever 7: vethafc53bc5@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue master cni0 state UP link/ether 32:51:41:e9:e1:68 brd ff:ff:ff:ff:ff:ff link-netnsid 0 inet6 fe80::3051:41ff:fee9:e168/64 scope link valid_lft forever preferred_lft forever 8: veth8246ed1c@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue master cni0 state UP link/ether 12:67:28:44:fa:cd brd ff:ff:ff:ff:ff:ff link-netnsid 1 inet6 fe80::1067:28ff:fe44:facd/64 scope link valid_lft forever preferred_lft forever [root@mcwk8s-node1 ~]$ node2 [root@mcwk8s-node2 ~]$ route -n Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface 0.0.0.0 10.0.0.253 0.0.0.0 UG 100 0 0 ens33 10.0.0.0 0.0.0.0 255.255.255.0 U 100 0 0 ens33 10.244.0.0 10.244.0.0 255.255.255.0 UG 0 0 0 flannel.1 10.244.1.0 10.244.1.0 255.255.255.0 UG 0 0 0 flannel.1 10.244.2.0 0.0.0.0 255.255.255.0 U 0 0 0 cni0 172.17.0.0 0.0.0.0 255.255.0.0 U 0 0 0 docker0 You have new mail in /var/spool/mail/root [root@mcwk8s-node2 ~]$ ip a 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet6 ::1/128 scope host valid_lft forever preferred_lft forever 2: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000 link/ether 00:0c:29:eb:83:cd brd ff:ff:ff:ff:ff:ff 3: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:eb:83:c3 brd ff:ff:ff:ff:ff:ff inet 10.0.0.6/24 brd 10.0.0.255 scope global ens33 valid_lft forever preferred_lft forever inet6 fe80::fd02:359f:93a4:95af/64 scope link valid_lft forever preferred_lft forever 4: docker0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN link/ether 02:42:ff:3f:37:b0 brd ff:ff:ff:ff:ff:ff inet 172.17.0.1/16 brd 172.17.255.255 scope global docker0 valid_lft forever preferred_lft forever 5: flannel.1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN link/ether 56:16:44:85:ca:57 brd ff:ff:ff:ff:ff:ff inet 10.244.2.0/32 brd 10.244.2.0 scope global flannel.1 valid_lft forever preferred_lft forever inet6 fe80::5416:44ff:fe85:ca57/64 scope link valid_lft forever preferred_lft forever 6: cni0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UP qlen 1000 link/ether 06:1f:10:ee:32:6f brd ff:ff:ff:ff:ff:ff inet 10.244.2.1/24 brd 10.244.2.255 scope global cni0 valid_lft forever preferred_lft forever inet6 fe80::41f:10ff:feee:326f/64 scope link valid_lft forever preferred_lft forever 7: vethdf3032e1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue master cni0 state UP link/ether 1e:50:bf:95:46:4d brd ff:ff:ff:ff:ff:ff link-netnsid 0 inet6 fe80::1c50:bfff:fe95:464d/64 scope link valid_lft forever preferred_lft forever [root@mcwk8s-node2 ~]$
flannel正常之后，验证集群内pod跨主机访问
[machangwei@mcwk8s-master ~]$ kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5qfrl 0/1 ContainerCreating 0 5m19s <none> mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-98x8d 0/1 ContainerCreating 0 5m19s <none> mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-bpbq4 0/1 ContainerCreating 0 5m19s <none> mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~]$ kubectl describe pod mcw-httpd-6fbf67d7d5-5qfrl Events: #容器创建中的状态时间段，包括拉取镜像的过程 Type Reason Age From Message ---- ------ ---- ---- ------- Normal Scheduled 4m49s default-scheduler Successfully assigned default/mcw-httpd-6fbf67d7d5-5qfrl to mcwk8s-node1 Normal Pulling 4m35s kubelet Pulling image "httpd" mcwk8s-node1一直卡在拉取不下镜像，然后手动到节点1上docker pull,拉取完镜像之后，pod马上就运行状态了 [machangwei@mcwk8s-master ~]$ kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5qfrl 1/1 Running 0 21m 10.244.1.2 mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-98x8d 1/1 Running 0 21m 10.244.2.2 mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-bpbq4 1/1 Running 0 21m 10.244.1.3 mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~]$ 查看网络，现在分配的pod上分配的ip，可以正常的在集群其它节点上进行访问了。也就是Pod分配了各自的ip,这些ip只能被Kuernetes Cluster 中的容器和节点访问。前面那句话已经得到验证 [machangwei@mcwk8s-master ~]$ kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5qfrl 1/1 Running 0 21m 10.244.1.2 mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-98x8d 1/1 Running 0 21m 10.244.2.2 mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-bpbq4 1/1 Running 0 21m 10.244.1.3 mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~]$ curl 10.244.1.2 <html><body><h1>It works!</h1></body></html> [machangwei@mcwk8s-master ~]$ curl 10.244.1.3 <html><body><h1>It works!</h1></body></html> [machangwei@mcwk8s-master ~]$ hostname mcwk8s-master [machangwei@mcwk8s-master ~]$
部署service
部署service，但是没有后端服务
endpoint是空的，肯定不能通过服务ip加端口，去curl访问服务的响应数据 [machangwei@mcwk8s-master ~]$ cat mcwHttpdService.yml apiVersion: v1 kind: Service metadata: name: httpd-svc spec: selector: run: httpd ports: - protocol: TCP port: 8080 targetPort: 80 [machangwei@mcwk8s-master ~]$ kubectl apply -f mcwHttpdService.yml #创建Service service/httpd-svc created [machangwei@mcwk8s-master ~]$ kubectl get service #查看service。有服务名，集群ip,以及端口， NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc ClusterIP 10.102.232.38 <none> 8080/TCP 11s kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 140m [machangwei@mcwk8s-master ~]$ curl 10.102.232.38:8080 #通过curl集群ip冒号端口，无法访问，因为没有endpoint curl: (7) Failed connect to 10.102.232.38:8080; Connection refused [machangwei@mcwk8s-master ~]$ curl 10.244.1.2 #通过访问pod ip ，可以直接在节点上访问到集群中的pod的对应的80服务 <html><body><h1>It works!</h1></body></html> [machangwei@mcwk8s-master ~]$ [machangwei@mcwk8s-master ~]$ ping 10.244.1.2 #可以在节点上ping通集群中pod的ip。 PING 10.244.1.2 (10.244.1.2) 56(84) bytes of data. 64 bytes from 10.244.1.2: icmp_seq=1 ttl=63 time=1.04 ms [machangwei@mcwk8s-master ~]$ [machangwei@mcwk8s-master ~]$ kubectl get service NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc ClusterIP 10.102.232.38 <none> 8080/TCP 75s kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 141m [machangwei@mcwk8s-master ~]$ curl 10.102.232.38:8080 curl: (7) Failed connect to 10.102.232.38:8080; Connection refused [machangwei@mcwk8s-master ~]$ kubectl describe service httpd-svc #查看服务，可以看到endponts为none Name: httpd-svc Namespace: default Labels: <none> Annotations: <none> Selector: run=httpd Type: ClusterIP IP Family Policy: SingleStack IP Families: IPv4 IP: 10.102.232.38 IPs: 10.102.232.38 Port: <unset> 8080/TCP TargetPort: 80/TCP Endpoints: <none> Session Affinity: None Events: <none> [machangwei@mcwk8s-master ~]$
正常的有后端pod
[machangwei@mcwk8s-master ~]$ cat mcwHttpdService.yml #修改前的yml文件， apiVersion: v1 kind: Service metadata: name: httpd-svc spec: selector: run: httpd ports: - protocol: TCP port: 8080 targetPort: 80 [machangwei@mcwk8s-master ~]$ kubectl get pod --show-labels #查看服务对应的后端pod的标签，发现是run=mcw-httpd， NAME READY STATUS RESTARTS AGE LABELS mcw-httpd-6fbf67d7d5-5qfrl 1/1 Running 0 107m pod-template-hash=6fbf67d7d5,run=mcw-httpd mcw-httpd-6fbf67d7d5-98x8d 1/1 Running 0 107m pod-template-hash=6fbf67d7d5,run=mcw-httpd mcw-httpd-6fbf67d7d5-bpbq4 1/1 Running 0 107m pod-template-hash=6fbf67d7d5,run=mcw-httpd [machangwei@mcwk8s-master ~]$ kubectl delete -f mcwHttpdService.yml service "httpd-svc" deleted [machangwei@mcwk8s-master ~]$ vim mcwHttpdService.yml [machangwei@mcwk8s-master ~]$ cat mcwHttpdService.yml #修改选择器，纠正标签，指定service后端pod的标签是run: mcw-httpd apiVersion: v1 kind: Service metadata: name: httpd-svc spec: selector: run: mcw-httpd ports: - protocol: TCP port: 8080 targetPort: 80 [machangwei@mcwk8s-master ~]$ kubectl apply -f mcwHttpdService.yml #重新部署服务 service/httpd-svc created [machangwei@mcwk8s-master ~]$ kubectl get service #查看服务，集群已经换掉了 NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc ClusterIP 10.99.19.228 <none> 8080/TCP 7s kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 146m [machangwei@mcwk8s-master ~]$ curl 10.99.19.228:8080 #这下可以正常curl 集群IP:端口，来访问服务了。 <html><body><h1>It works!</h1></body></html> [machangwei@mcwk8s-master ~]$ kubectl describe service httpd-svc Name: httpd-svc Namespace: default Labels: <none> Annotations: <none> Selector: run=mcw-httpd Type: ClusterIP IP Family Policy: SingleStack IP Families: IPv4 IP: 10.99.19.228 IPs: 10.99.19.228 Port: <unset> 8080/TCP #查看到现在的服务，后endpoint的值 TargetPort: 80/TCP # endpoint的值，是代表有可以访问带有指定标签pod ip，指定pod目标端口的服务。 Endpoints: 10.244.1.2:80,10.244.1.3:80,10.244.2.2:80 Session Affinity: None #也就是访问集群ip:端口，应该是负载均衡路由到这带有服务选择对应标签的pod ip和目标端口的 Events: <none> #集群ip是创建服务时分配的，而选择带有什么样标签的pod,以及pod对应服务的端口,即目标端口，是在yml里面已经设置好了 [machangwei@mcwk8s-master ~]$
service文件介绍
[machangwei@mcwk8s-master ~]$ cat mcwHttpdService.yml apiVersion: v1 #Service的apiVersion kind: Service #资源类型 metadata: name: httpd-svc spec: selector: #指明挑选那些label为run: httpd的pod作为Service的后端 run: httpd ports: - protocol: TCP port: 8080 #将Servicede 8080端口映射到Pod的80端口，使用TCP协议 targetPort: 80
Cluster IP底层实现
跟service相关的防火墙规则，集群ip和pod ip相关 [machangwei@mcwk8s-master ~]$ kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5qfrl 1/1 Running 0 20h 10.244.1.2 mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-98x8d 1/1 Running 0 20h 10.244.2.2 mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-bpbq4 1/1 Running 0 20h 10.244.1.3 mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~]$ kubectl get service NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc ClusterIP 10.99.19.228 <none> 8080/TCP 18h kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 20h [machangwei@mcwk8s-master ~]$ 查看当前主机防火墙 [root@mcwk8s-master ~]$ iptables-save |grep httpd-svc -A KUBE-SEP-26GESA23ILIBJ6BG -s 10.244.1.2/32 -m comment --comment "default/httpd-svc" -j KUBE-MARK-MASQ -A KUBE-SEP-26GESA23ILIBJ6BG -p tcp -m comment --comment "default/httpd-svc" -m tcp -j DNAT --to-destination 10.244.1.2:80 -A KUBE-SEP-5MDWNIS6FGKOLKLF -s 10.244.1.3/32 -m comment --comment "default/httpd-svc" -j KUBE-MARK-MASQ -A KUBE-SEP-5MDWNIS6FGKOLKLF -p tcp -m comment --comment "default/httpd-svc" -m tcp -j DNAT --to-destination 10.244.1.3:80 -A KUBE-SEP-MZ7D7IEY543CBPN3 -s 10.244.2.2/32 -m comment --comment "default/httpd-svc" -j KUBE-MARK-MASQ -A KUBE-SEP-MZ7D7IEY543CBPN3 -p tcp -m comment --comment "default/httpd-svc" -m tcp -j DNAT --to-destination 10.244.2.2:80 -A KUBE-SERVICES -d 10.99.19.228/32 -p tcp -m comment --comment "default/httpd-svc cluster IP" -m tcp --dport 8080 -j KUBE-SVC-IYRDZZKXS5EOQ6Q6 -A KUBE-SVC-IYRDZZKXS5EOQ6Q6 ! -s 10.244.0.0/16 -d 10.99.19.228/32 -p tcp -m comment --comment "default/httpd-svc cluster IP" -m tcp --dport 8080 -j KUBE-MARK-MASQ -A KUBE-SVC-IYRDZZKXS5EOQ6Q6 -m comment --comment "default/httpd-svc" -m statistic --mode random --probability 0.33333333349 -j KUBE-SEP-26GESA23ILIBJ6BG -A KUBE-SVC-IYRDZZKXS5EOQ6Q6 -m comment --comment "default/httpd-svc" -m statistic --mode random --probability 0.50000000000 -j KUBE-SEP-5MDWNIS6FGKOLKLF -A KUBE-SVC-IYRDZZKXS5EOQ6Q6 -m comment --comment "default/httpd-svc" -j KUBE-SEP-MZ7D7IEY543CBPN3 [root@mcwk8s-master ~]$ 一： -A KUBE-SERVICES -d 10.99.19.228/32 -p tcp -m comment --comment "default/httpd-svc cluster IP" -m tcp --dport 8080 -j KUBE-SVC-IYRDZZKXS5EOQ6Q6 -A KUBE-SVC-IYRDZZKXS5EOQ6Q6 ! -s 10.244.0.0/16 -d 10.99.19.228/32 -p tcp -m comment --comment "default/httpd-svc cluster IP" -m tcp --dport 8080 -j KUBE-MARK-MASQ 1、其它源地址访问httpd-svc，则允许 2、如果Cluster内的pod(源地址来自10.244.0.0/16)要访问httpd-svc，则跳转到KUBE-SVC-IYRDZZKXS5EOQ6Q6 KUBE-SVC-IYRDZZKXS5EOQ6Q6规则之一如下： -A KUBE-SVC-IYRDZZKXS5EOQ6Q6 -m comment --comment "default/httpd-svc" -m statistic --mode random --probability 0.33333333349 -j KUBE-SEP-26GESA23ILIBJ6BG 二： -A KUBE-SVC-IYRDZZKXS5EOQ6Q6 -m comment --comment "default/httpd-svc" -m statistic --mode random --probability 0.33333333349 -j KUBE-SEP-26GESA23ILIBJ6BG -A KUBE-SVC-IYRDZZKXS5EOQ6Q6 -m comment --comment "default/httpd-svc" -m statistic --mode random --probability 0.50000000000 -j KUBE-SEP-5MDWNIS6FGKOLKLF -A KUBE-SVC-IYRDZZKXS5EOQ6Q6 -m comment --comment "default/httpd-svc" -j KUBE-SEP-MZ7D7IEY543CBPN3 1、1/3概率跳转到规则KUBE-SEP-26GESA23ILIBJ6BG 2、1/3概率（剩下2/3的一半）跳转到规则KUBE-SEP-5MDWNIS6FGKOLKLF 3、1/3概率跳转到规则KUBE-SEP-MZ7D7IEY543CBPN3 三： -A KUBE-SEP-26GESA23ILIBJ6BG -s 10.244.1.2/32 -m comment --comment "default/httpd-svc" -j KUBE-MARK-MASQ -A KUBE-SEP-26GESA23ILIBJ6BG -p tcp -m comment --comment "default/httpd-svc" -m tcp -j DNAT --to-destination 10.244.1.2:80 -A KUBE-SEP-5MDWNIS6FGKOLKLF -s 10.244.1.3/32 -m comment --comment "default/httpd-svc" -j KUBE-MARK-MASQ -A KUBE-SEP-5MDWNIS6FGKOLKLF -p tcp -m comment --comment "default/httpd-svc" -m tcp -j DNAT --to-destination 10.244.1.3:80 -A KUBE-SEP-MZ7D7IEY543CBPN3 -s 10.244.2.2/32 -m comment --comment "default/httpd-svc" -j KUBE-MARK-MASQ -A KUBE-SEP-MZ7D7IEY543CBPN3 -p tcp -m comment --comment "default/httpd-svc" -m tcp -j DNAT --to-destination 10.244.2.2:80 将请求分别转发到后端的三个pod.iptables将访问Service的流量转发到后端pod，而且使用类型轮询的负载均衡策略。集群的每一个结点都配置了相同的iptables规则，这样就确保了整个集群都能通过Service的集群ip访问服务。
DNS访问Service
访问本身namespace中的
[machangwei@mcwk8s-master ~]$ kubectl get deployment --namespace=kube-system #查看 dns组件 NAME READY UP-TO-DATE AVAILABLE AGE coredns 2/2 2 2 21h [machangwei@mcwk8s-master ~]$ [machangwei@mcwk8s-master ~]$ kubectl run mcwcentos --rm -ti --image=centos /bin/bash #创建pod并进入 If you don't see a command prompt, try pressing enter. [root@mcwcentos /]# wget httpd-svc.default:8080 #在pod中通过<SERVICE_NAME>.<NAMESPACE_NAME>来访问Service bash: wget: command not found [root@mcwcentos /]# curl httpd-svc.default:8080 <html><body><h1>It works!</h1></body></html> [root@mcwcentos /]# [root@mcwcentos /]# curl httpd-svc:8080 #因为pod和httpd-svc同属于default namespace,因此可以省略default命名空间，直接访问服务 <html><body><h1>It works!</h1></body></html> [root@mcwcentos /]# yum -y install bind-utils #安装工具，以方便使用nslookup命令 Failed to set locale, defaulting to C.UTF-8 Last metadata expiration check: 0:02:36 ago on Fri Jan 21 12:45:19 2022. Dependencies resolved. [root@mcwcentos /]# ls /etc/yum.repos.d/ 查看cenos镜像中的repo文件 CentOS-Linux-AppStream.repo CentOS-Linux-ContinuousRelease.repo CentOS-Linux-Devel.repo CentOS-Linux-FastTrack.repo CentOS-Linux-Media.repo CentOS-Linux-PowerTools.repo CentOS-Linux-BaseOS.repo CentOS-Linux-Debuginfo.repo CentOS-Linux-Extras.repo CentOS-Linux-HighAvailability.repo CentOS-Linux-Plus.repo CentOS-Linux-Sources.repo [root@mcwcentos /]# nslookup httpd-svc #用命令查看httpd-svc的DNS信息 Server: 10.96.0.10 Address: 10.96.0.10#53 Name: httpd-svc.default.svc.cluster.local #这个是DNS服务器，dns组件，这个是httpd的完整域名 Address: 10.99.19.228 #可以看到这个是service的cluster ip [root@mcwcentos /]# [machangwei@mcwk8s-master ~]$ kubectl get service 集群ip NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc ClusterIP 10.99.19.228 <none> 8080/TCP 19h kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 21h [machangwei@mcwk8s-master ~]$ 查看这个新创建的pod [machangwei@mcwk8s-master ~]$ kubectl get pod -o wide|grep mcwcentos mcwcentos 1/1 Running 0 8m8s 10.244.2.3 mcwk8s-node2 <none> <none> 到pod对应节点上找到这个容器 [root@mcwk8s-node2 ~]$ docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES 0b11a9d9ac35 centos "/bin/bash" 8 minutes ago Up 8 minutes k8s_mcwcentos_mcwcentos_default_05aaed6d-57b0-4f6b-807d-1cd0e9c39ad9_0 然后查看容器的日志，可以看到，记录了我进入容器后做的所有操作包括yum安装的过程 [root@mcwk8s-node2 ~]$ docker logs 0b1|tail CentOS-Linux-AppStream.repo CentOS-Linux-ContinuousRelease.repo CentOS-Linux-Devel.repo CentOS-Linux-FastTrack.repo CentOS-Linux-Media.repo CentOS-Linux-PowerTools.repo CentOS-Linux-BaseOS.repo CentOS-Linux-Debuginfo.repo CentOS-Linux-Extras.repo CentOS-Linux-HighAvailability.repo CentOS-Linux-Plus.repo CentOS-Linux-Sources.repo [root@mcwcentos /]# [root@mcwcentos /]# nslookup httpd-svc Server: 10.96.0.10 Address: 10.96.0.10#53 Name: httpd-svc.default.svc.cluster.local Address: 10.99.19.228 [root@mcwk8s-node2 ~]$ ‘ 主节点进入的pod中： [root@mcwcentos /]# ping httpd-svc PING httpd-svc.default.svc.cluster.local (10.99.19.228) 56(84) bytes of data. ^C --- httpd-svc.default.svc.cluster.local ping statistics --- 2 packets transmitted, 0 received, 100% packet loss, time 1001ms [root@mcwcentos /]# curl httpd-svc curl: (7) Failed to connect to httpd-svc port 80: Connection timed out [root@mcwcentos /]# curl httpd-svc:8080 <html><body><h1>It works!</h1></body></html> [root@mcwcentos /]# curl httpd-svc.default.svc.cluster.local:8080 #也可以curl完整域名，如果不加端口，不行，因为服务里显示用到这个端口了 <html><body><h1>It works!</h1></body></html> [root@mcwcentos /]#
访问其它namespace中的service
查看已存在的namespcace [machangwei@mcwk8s-master ~]$ kubectl get namespace #这几个都是部署好集群时就已经创建好的namespace NAME STATUS AGE default Active 21h kube-node-lease Active 21h kube-public Active 21h kube-system Active 21h [machangwei@mcwk8s-master ~]$ 查看之前部署的文件 [machangwei@mcwk8s-master ~]$ cat mcwHttpd.yml apiVersion: apps/v1 kind: Deployment metadata: name: mcw-httpd spec: replicas: 3 selector: matchLabels: run: mcw-httpd template: metadata: labels: run: mcw-httpd spec: containers: - name: mcw-httpd image: httpd ports: - containerPort: 80 [machangwei@mcwk8s-master ~]$ cat mcwHttpdService.yml apiVersion: v1 kind: Service metadata: name: httpd-svc spec: selector: run: mcw-httpd ports: - protocol: TCP port: 8080 targetPort: 80 [machangwei@mcwk8s-master ~]$ cat mcwhttpd2quanyml #查看上两个文件合并修改名称，标签等信息，添加指定的命名空间。多个资源用---来分割 apiVersion: apps/v1 kind: Deployment metadata: name: mcw-httpd2 namespace: kube-public spec: replicas: 3 selector: matchLabels: run: mcw-httpd2 template: metadata: labels: run: mcw-httpd2 spec: containers: - name: mcw-httpd2 image: httpd ports: - containerPort: 80 --- apiVersion: v1 kind: Service metadata: name: httpd2-svc namespace: kube-public spec: selector: run: mcw-httpd2 ports: - protocol: TCP port: 8080 targetPort: 80 [machangwei@mcwk8s-master ~]$ 查看mcwcentos的ip [root@mcwcentos /]# ip a 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever 3: eth0@if8: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UP group default link/ether c6:c1:ce:94:49:24 brd ff:ff:ff:ff:ff:ff link-netnsid 0 inet 10.244.2.3/24 brd 10.244.2.255 scope global eth0 valid_lft forever preferred_lft forever [root@mcwcentos /]# hostname -i 10.244.2.3 [root@mcwcentos /]# [machangwei@mcwk8s-master ~]$ kubectl apply -f mcwhttpd2quanyml #部署httpd2服务 deployment.apps/mcw-httpd2 created service/httpd2-svc created [machangwei@mcwk8s-master ~]$ kubectl get service --namespace=kube-public #查看服务2，需要指定命名空间 NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd2-svc ClusterIP 10.101.134.243 <none> 8080/TCP 39s [machangwei@mcwk8s-master ~]$ kubectl get service #不指定命名空间，无法看到kube-public中的服务 NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc ClusterIP 10.99.19.228 <none> 8080/TCP 19h kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 21h [machangwei@mcwk8s-master ~]$ [machangwei@mcwk8s-master ~]$ kubectl run mcwcentos2 --rm -ti --image=centos /bin/bash #运行第二个，mcwcentos2 If you don't see a command prompt, try pressing enter. [root@mcwcentos2 /]# curl httpd2-svc:8080 #pod属于default命名空间，访问其它命名空间的服务，不指定命名空间，直接访问服务名加端口，是无法访问到 curl: (6) Could not resolve host: httpd2-svc [root@mcwcentos2 /]# curl httpd2-svc.kube-public:8080 #加上服务名称.命名空间:服务端口，就可以实现在pod中跨命名空间访问服务 <html><body><h1>It works!</h1></body></html> [root@mcwcentos2 /]# ping -c 2 10.244.2.3 #pod中能直接ping同一命名空间内的pod PING 10.244.2.3 (10.244.2.3) 56(84) bytes of data. 64 bytes from 10.244.2.3: icmp_seq=1 ttl=64 time=0.252 ms 64 bytes from 10.244.2.3: icmp_seq=2 ttl=64 time=0.072 ms --- 10.244.2.3 ping statistics --- 2 packets transmitted, 2 received, 0% packet loss, time 1001ms rtt min/avg/max/mdev = 0.072/0.162/0.252/0.090 ms [root@mcwcentos2 /]# [root@mcwcentos2 /]# exit #当ctrl+d退出时，这个命令创建的pod就被删除掉了，get pod也看不见这个容器 Session ended, resume using 'kubectl attach mcwcentos2 -c mcwcentos2 -i -t' command when the pod is running pod "mcwcentos2" deleted [machangwei@mcwk8s-master ~]$ 如下，默认只能看到default下的pod [machangwei@mcwk8s-master ~]$ kubectl get pod NAME READY STATUS RESTARTS AGE mcw-httpd-6fbf67d7d5-5qfrl 1/1 Running 0 21h mcw-httpd-6fbf67d7d5-98x8d 1/1 Running 0 21h mcw-httpd-6fbf67d7d5-bpbq4 1/1 Running 0 21h [machangwei@mcwk8s-master ~]$ [machangwei@mcwk8s-master ~]$ [machangwei@mcwk8s-master ~]$ kubectl get pod --all-namespaces NAMESPACE NAME READY STATUS RESTARTS AGE default mcw-httpd-6fbf67d7d5-5qfrl 1/1 Running 0 21h default mcw-httpd-6fbf67d7d5-98x8d 1/1 Running 0 21h default mcw-httpd-6fbf67d7d5-bpbq4 1/1 Running 0 21h kube-public mcw-httpd2-6b98bfbbbf-69jb5 1/1 Running 0 21m kube-public mcw-httpd2-6b98bfbbbf-qv7g9 1/1 Running 0 21m kube-public mcw-httpd2-6b98bfbbbf-ztddf 1/1 Running 0 21m kube-system coredns-6d8c4cb4d-vctxx 1/1 Running 1 (21h ago) 22h kube-system coredns-6d8c4cb4d-xkv9h 1/1 Running 0 22h
外网如何访问Service
1、集群ip加端口，可以提供集群内部访问服务。
这里是在三个节点上都通过curl集群ip：端口，正常访问。如果是pod中，不同命名空间的pod中访问，结果如何，以后验证，应该也是没有问题的。 [machangwei@mcwk8s-master ~]$ kubectl get service NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc NodePort 10.107.208.46 <none> 8080:30450/TCP 14m kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 22h [machangwei@mcwk8s-master ~]$ curl 10.107.208.46:8080 <html><body><h1>It works!</h1></body></html> [root@mcwk8s-node1 ~]$ curl 10.107.208.46:8080 <html><body><h1>It works!</h1></body></html> [root@mcwk8s-node2 ~]$ curl 10.107.208.46:8080 <html><body><h1>It works!</h1></body></html> [root@mcwk8s-node2 ~]$
2、这里演示node port 方式将应用的service暴露给cluster外部。
[machangwei@mcwk8s-master ~]$ ls #查看当前有的yml文件 mcwhttpd2quanyml mcwHttpdService.yml mcwHttpd.yml mm.yml [machangwei@mcwk8s-master ~]$ kubectl get service #查看当前服务 NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc ClusterIP 10.99.19.228 <none> 8080/TCP 19h kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 22h [machangwei@mcwk8s-master ~]$ kubectl get pod #查看当前的pod NAME READY STATUS RESTARTS AGE mcw-httpd-6fbf67d7d5-5qfrl 1/1 Running 0 21h mcw-httpd-6fbf67d7d5-98x8d 1/1 Running 0 21h mcw-httpd-6fbf67d7d5-bpbq4 1/1 Running 0 21h [machangwei@mcwk8s-master ~]$ kubectl delete -f mcwHttpdService.yml #把已有的服务删除掉，一会创建新的服务，还用以前的pod service "httpd-svc" deleted [machangwei@mcwk8s-master ~]$ kubectl get service #查看当前有的服务，之前的一个服务已经成功删除 NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 22h [machangwei@mcwk8s-master ~]$ vim mcwHttpdService.yml #编辑服务yml [machangwei@mcwk8s-master ~]$ cat mcwHttpdService.yml #将服务yml增添键值对 type: NodePort。 apiVersion: v1 kind: Service metadata: name: httpd-svc spec: type: NodePort selector: run: mcw-httpd ports: - protocol: TCP port: 8080 targetPort: 80 [machangwei@mcwk8s-master ~]$ kubectl apply -f mcwHttpdService.yml #创建服务，好像不用删除，直接重新执行，会重新创建服务，有时间验证 service/httpd-svc created [machangwei@mcwk8s-master ~]$ kubectl get service #查看新创建的服务，可以看到端口部分，多了一个端口，30450， NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc NodePort 10.107.208.46 <none> 8080:30450/TCP 21s kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 22h [machangwei@mcwk8s-master ~]$ netstat -an|grep 30450 #查看要暴露的端口，这个端口在所有节点（node）上都有，节点ip:这个端口，都能供外网访问，供集群外的访问 tcp 0 0 0.0.0.0:30450 0.0.0.0:* LISTEN [machangwei@mcwk8s-master ~]$ curl 10.107.208.46:8080 #集群ip 端口访问服务 <html><body><h1>It works!</h1></body></html> [machangwei@mcwk8s-master ~]$ curl 10.107.208.46:30450 #集群ip访问这个端口是不行的 curl: (7) Failed connect to 10.107.208.46:30450; Connection timed out [machangwei@mcwk8s-master ~]$ kubectl get pod -o wide #查看pod的ip NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5qfrl 1/1 Running 0 21h 10.244.1.2 mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-98x8d 1/1 Running 0 21h 10.244.2.2 mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-bpbq4 1/1 Running 0 21h 10.244.1.3 mcwk8s-node1 <none> <none> [machangwei@mcwk8s-master ~]$ curl 10.244.1.2 #直接在节点上curl pod的ip，就能访问到服务 <html><body><h1>It works!</h1></body></html> [machangwei@mcwk8s-master ~]$ curl 10.244.1.2:30450 #直接在节点上curl ip:要暴露的端口，是不行的 curl: (7) Failed connect to 10.244.1.2:30450; Connection refused [machangwei@mcwk8s-master ~]$ hostname -i 10.0.0.4 [machangwei@mcwk8s-master ~]$ curl 10.0.0.4:30450 #应该用curl 节点ip:要暴露的端口 <html><body><h1>It works!</h1></body></html> [machangwei@mcwk8s-master ~]$ curl 10.0.0.5:30450 #节点ip:这个端口，还可以在外面浏览器上访问。 <html><body><h1>It works!</h1></body></html> #只不过三个 节点ip加这个端口，外网访问这个服务的时候， [machangwei@mcwk8s-master ~]$ curl 10.0.0.6:30450 #是否后端实现了负载均衡，调度到三个后端pod上呢 <html><body><h1>It works!</h1></body></html> [machangwei@mcwk8s-master ~]$ 主节点上找到服务的三个pod [machangwei@mcwk8s-master ~]$ kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES mcw-httpd-6fbf67d7d5-5qfrl 1/1 Running 0 21h 10.244.1.2 mcwk8s-node1 <none> <none> mcw-httpd-6fbf67d7d5-98x8d 1/1 Running 0 21h 10.244.2.2 mcwk8s-node2 <none> <none> mcw-httpd-6fbf67d7d5-bpbq4 1/1 Running 0 21h 10.244.1.3 mcwk8s-node1 <none> <none> 根据三个pod信息，去节点上找到这三个容器，将容器内容修改，添加上自己的pod ip [root@mcwk8s-node1 ~]$ netstat -an|grep 30450 tcp 0 0 0.0.0.0:30450 0.0.0.0:* LISTEN [root@mcwk8s-node1 ~]$ docker ps |grep mcw-httpd-6fbf67d7d5-bpbq4 c978d2770826 httpd "httpd-foreground" 22 hours ago Up 22 hours k8s_mcw-httpd_mcw-httpd-6fbf67d7d5-bpbq4_default_1380d70d-e2b1-4276-b80f-813bcd3bae10_0 eddd4b542888 registry.aliyuncs.com/google_containers/pause:3.6 "/pause" 22 hours ago Up 22 hours k8s_POD_mcw-httpd-6fbf67d7d5-bpbq4_default_1380d70d-e2b1-4276-b80f-813bcd3bae10_0 [root@mcwk8s-node1 ~]$ docker exec -it c978 bash root@mcw-httpd-6fbf67d7d5-bpbq4:/usr/local/apache2# ls bin build cgi-bin conf error htdocs icons include logs modules root@mcw-httpd-6fbf67d7d5-bpbq4:/usr/local/apache2# cd htdocs/ root@mcw-httpd-6fbf67d7d5-bpbq4:/usr/local/apache2/htdocs# cat index.html <html><body><h1>It works!</h1></body></html> root@mcw-httpd-6fbf67d7d5-bpbq4:/usr/local/apache2/htdocs# echo '<html><body><h1>It works!</h1></body></html> 10.244.1.3'>index.html root@mcw-httpd-6fbf67d7d5-bpbq4:/usr/local/apache2/htdocs# 在主节点上直接访问各个pod ip，查看都修改成功 [machangwei@mcwk8s-master ~]$ curl 10.244.1.3 <html><body><h1>It works!</h1></body></html> 10.244.1.3 [machangwei@mcwk8s-master ~]$ 然后根据节点ip加要暴露出去的端口，进行访问。可以发现，在节点上无论访问哪个节点ip:端口，都能实现后端三个pod的负载均衡。也就是说实现了负载均衡的。访问任意一个结点ip:加要暴露的端口，都等于访问这个服务，而这个服务后端三个pod是实现负载均衡的。在浏览器上访问，也是能看出了的, [machangwei@mcwk8s-master ~]$ kubectl get service NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE httpd-svc NodePort 10.107.208.46 <none> 8080:30450/TCP 46m kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 23h [machangwei@mcwk8s-master ~]$ hostname -i 10.0.0.4 [machangwei@mcwk8s-master ~]$ curl 10.0.0.4:30450 <html><body><h1>It works!</h1></body></html> 10.244.2.2 [machangwei@mcwk8s-master ~]$ curl 10.0.0.4:30450 <html><body><h1>It works!</h1></body></html> 10.244.1.2 [machangwei@mcwk8s-master ~]$ curl 10.0.0.4:30450 <html><body><h1>It works!</h1></body></html> 10.244.2.2 [machangwei@mcwk8s-master ~]$ curl 10.0.0.4:30450 <html><body><h1>It works!</h1></body></html> 10.244.2.2 [machangwei@mcwk8s-master ~]$ curl 10.0.0.4:30450 <html><body><h1>It works!</h1></body></html> 10.244.2.2 [machangwei@mcwk8s-master ~]$ curl 10.0.0.4:30450 <html><body><h1>It works!</h1></body></html> 10.244.1.2 [machangwei@mcwk8s-master ~]$ curl 10.0.0.4:30450 <html><body><h1>It works!</h1></body></html> 10.244.1.3 [machangwei@mcwk8s-master ~]$ curl 10.0.0.4:30450 <html><body><h1>It works!</h1></body></html> 10.244.2.2 [machangwei@mcwk8s-master ~]$ curl 10.0.0.5:30450 <html><body><h1>It works!</h1></body></html> 10.244.1.2 [machangwei@mcwk8s-master ~]$ curl 10.0.0.5:30450 <html><body><h1>It works!</h1></body></html> 10.244.1.3
3、loadbalancer
以后添加
