k8s在pod内无法ping通servicename和ClusterIP

需要使用 ipvs 替换iptables，操作是在所有节点上

1：开启内核支持

1 2 3 4 5 6 7

cat >> /etc/sysctl.conf << EOF net.ipv4.ip_forward = 1 net.bridge.bridge-nf-call-iptables = 1 net.bridge.bridge-nf-call-ip6tables = 1 EOF sysctl -p

　　

2：开启ipvs支持

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

yum -y install ipvsadm  ipset # 临时生效 modprobe -- ip_vs modprobe -- ip_vs_rr modprobe -- ip_vs_wrr modprobe -- ip_vs_sh modprobe -- nf_conntrack_ipv4 # 永久生效 cat > /etc/sysconfig/modules/ipvs.modules <<EOF modprobe -- ip_vs modprobe -- ip_vs_rr modprobe -- ip_vs_wrr modprobe -- ip_vs_sh modprobe -- nf_conntrack_ipv4 EOF

　　

3：配置kube-proxy，在master上操作，因使用kubeadmin安装，所以操作方式如下

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

[root@master] # kubectl edit cm kube-proxy -n kube-system configmap/kube-proxy edited #修改如下 kind: MasterConfiguration apiVersion: kubeadm.k8s.io/v1alpha1 ... ipvs:       excludeCIDRs: null       minSyncPeriod: 0s       scheduler: ""       syncPeriod: 30s     kind: KubeProxyConfiguration     metricsBindAddress: 127.0.0.1:10249     mode: "ipvs"                  #修改

　　

4：在master重启kube-proxy

1	kubectl  get pod -n kube-system | grep kube-proxy | awk '{print $1}' | xargs kubectl delete pod -n kube-system

　　

5：验证ipvs是否开启

1 2 3 4 5 6 7 8 9 10 11

[root@k8s-m mytest]# kubectl logs kube-proxy-cvzb4 -n kube-system I0409 03:37:29.194391       1 server_others.go:170] Using ipvs Proxier. W0409 03:37:29.194779       1 proxier.go:401] IPVS scheduler not specified, use rr by default I0409 03:37:29.194981       1 server.go:534] Version: v1.15.3 I0409 03:37:29.214255       1 conntrack.go:52] Setting nf_conntrack_max to 524288 I0409 03:37:29.216744       1 config.go:96] Starting endpoints config controller I0409 03:37:29.216812       1 controller_utils.go:1029] Waiting for caches to sync for endpoints config controller I0409 03:37:29.217445       1 config.go:187] Starting service config controller I0409 03:37:29.218320       1 controller_utils.go:1029] Waiting for caches to sync for service config controller I0409 03:37:29.318218       1 controller_utils.go:1036] Caches are synced for endpoints config controller I0409 03:37:29.318564       1 controller_utils.go:1036] Caches are synced for service config controller

　　

6：进入pod内，现在可以ping通servicename了，使用iptables时，发现ping的时候出现了如下错误，执行完上述操作，一切正常

1 2 3 4

root@xxxxxx-cb4c9cb8c-hpzdl:/opt# ping xxxxxx PING xxxxxx.xxxxx.svc.cluster.local (172.16.140.78) 56(84) bytes of data. From 172.16.8.1 (172.16.8.1) icmp_seq=1 Time to live exceeded From 172.16.8.1 (172.16.8.1) icmp_seq=2 Time to live exceeded