Overlay Network：覆盖网络，在基础网络上叠加的一种虚拟网络技术模式，该网络中的主机通过虚拟链路连接起来。
VXLAN：将源数据包封装到UDP中，并使用基础网络的IP/MAC作为外层报文头进行封装，然后在以太网上传输，到达目的地后由隧道端点解封并将数据发送给目的地址。
Fannel：Overlay网络的一种，也是将源数据包封装在另一种网络包里面进行路由转发和通信，目前已经支持UDP、VXLAN、AWS VPC和GCE路由等数据转发方式。
多主机容器网络通信其他主流方案：隧道方案（Weave、OpenvSwitch）、路由方案（Calico）等。
部署Flannel网络
1、写入分配的子网段到etcd，供flanneld使用
[root@master ~]# /opt/kubernetes/bin/etcdctl --ca-file=/opt/kubernetes/ssl/ca.pem --cert-file=/opt/kubernetes/ssl/server.pem --key-file=/opt/kubernetes/ssl/server-key.pem --endpoints="https://192.168.238.130:2379,https://192.168.238.129:2379,https://192.168.238.128:2379" set /coreos.com/network/config '{"Network":"172.17.0.0/16","Backend":{"Type":"vxlan"}}'
2、下载二进制包
[root@master ~]# wget https://github.com/coreos/flannel/releases/download/v0.11.0/flannel-v0.11.0-linux-amd64.tar.gz
[root@master ~]# ls
flannel-v0.11.0-linux-amd64.tar.gz
[root@master ~]# tar -zxf flannel-v0.11.0-linux-amd64.tar.gz
[root@master ~]# ls
mk-docker-opts.sh flanneld
[root@master ~]# mv flanneld mk-docker-opts.sh /opt/kubernetes/bin
[root@master ~]# ls /opt/kubernetes/bin/
etcd etcdctl flanneld mk-docker-opts.sh
在node01和node02重复上述操作。
3、配置flannel
[root@node01 ~]# cat /opt/kubernetes/cfg/flanneld
FIANNEL_OPTIONS="--etcd-endpoints=https://192.168.238.129:2380,https://192.168.238.128:2380,https://192.168.238.130:2380 -etcd-cafile=/opt/kubernetes/ssl/ca.pem -etcd-certfile=/opt/kubernetes/ssl/server.pem --etcd-keyfile=/opt/kubernetes/ssl/server-key.pem"
4、systemd管理flannel
[root@node01 ~]# cat /usr/lib/systemd/system/flanneld.service
[Unit]
Description=Flanneld overlay address etcd agent
After=network-online.target network.target
Before=docker.service
[Service]
Type=notify
EnvironmentFile=/opt/kubernetes/cfg/flanneld
ExecStart=/opt/kubernetes/bin/flanneld --ip-masq $FLANNEL_OPTIOS
ExecStartPost=/opt/kubernetes/bin/mk-docker-opts.sh -k DOCKER_NETWORK_OPTIONS -d /run/flannel/subnet.env
Restart=on-failure
[Install]
WantedBy=multi-user.target
5、配置docker启动指定子网段
6、启动
加载配置
[root@node01 ~]# systemctl daemon-reload
[root@node01 ~]# systemctl start flanneld
Job for flanneld.service failed because the control process exited with error code. See "systemctl status flanneld.service" and "journalctl -xe" for details.
查看系统日志
[root@node01 ~]# tail -n 20 /var/log/messages
Jul 4 20:15:24 localhost etcd: c858c42725f38881 received MsgVoteResp from c858c42725f38881 at term 16130
Jul 4 20:15:24 localhost etcd: c858c42725f38881 [logterm: 7765, index: 18] sent MsgVote request to a7e9807772a004c5 at term 16130
Jul 4 20:15:24 localhost etcd: c858c42725f38881 [logterm: 7765, index: 18] sent MsgVote request to 203750a5948d27da at term 16130
Jul 4 20:15:25 localhost etcd: c858c42725f38881 is starting a new election at term 16130
Jul 4 20:15:25 localhost etcd: c858c42725f38881 became candidate at term 16131
Jul 4 20:15:25 localhost etcd: c858c42725f38881 received MsgVoteResp from c858c42725f38881 at term 16131
Jul 4 20:15:25 localhost etcd: c858c42725f38881 [logterm: 7765, index: 18] sent MsgVote request to 203750a5948d27da at term 16131
Jul 4 20:15:25 localhost etcd: c858c42725f38881 [logterm: 7765, index: 18] sent MsgVote request to a7e9807772a004c5 at term 16131
Jul 4 20:15:27 localhost etcd: c858c42725f38881 is starting a new election at term 16131
Jul 4 20:15:27 localhost etcd: c858c42725f38881 became candidate at term 16132
Jul 4 20:15:27 localhost etcd: c858c42725f38881 received MsgVoteResp from c858c42725f38881 at term 16132
Jul 4 20:15:27 localhost etcd: c858c42725f38881 [logterm: 7765, index: 18] sent MsgVote request to 203750a5948d27da at term 16132
Jul 4 20:15:27 localhost etcd: c858c42725f38881 [logterm: 7765, index: 18] sent MsgVote request to a7e9807772a004c5 at term 16132
Jul 4 20:15:28 localhost etcd: c858c42725f38881 is starting a new election at term 16132
Jul 4 20:15:28 localhost etcd: c858c42725f38881 became candidate at term 16133
Jul 4 20:15:28 localhost etcd: c858c42725f38881 received MsgVoteResp from c858c42725f38881 at term 16133
Jul 4 20:15:28 localhost etcd: c858c42725f38881 [logterm: 7765, index: 18] sent MsgVote request to 203750a5948d27da at term 16133
Jul 4 20:15:28 localhost etcd: c858c42725f38881 [logterm: 7765, index: 18] sent MsgVote request to a7e9807772a004c 5 at term 16133
Jul 4 20:15:28 localhost etcd: health check for peer 203750a5948d27da could not connect: dial tcp 192.168.238.128:2380: getsockopt: no route to host
Jul 4 20:15:28 localhost etcd: health check for peer a7e9807772a004c5 could not connect: dial tcp 192.168.238.130:2380: i/o timeout
初步判定防火墙导致，关闭防火墙
[root@node01 ~]# systemctl stop firewalld.service
[root@node01 ~]# systemctl start flanneld
Job for flanneld.service failed because a timeout was exceeded. See "systemctl status flanneld.service" and "journalctl -xe" for details.
网络故障原因
[root@node01 ~]# tail -n 20 /var/log/messages
Jul 6 08:49:15 localhost systemd: flanneld.service failed.
Jul 6 08:49:15 localhost systemd: flanneld.service holdoff time over, scheduling restart.
Jul 6 08:49:15 localhost systemd: Stopped Flanneld overlay address etcd agent.
Jul 6 08:49:15 localhost systemd: Starting Flanneld overlay address etcd agent...
Jul 6 08:49:15 localhost flanneld: I0706 08:49:15.831267 8741 main.go:514] Determining IP address of default interface
Jul 6 08:49:15 localhost flanneld: I0706 08:49:15.831870 8741 main.go:527] Using interface with name eno16777736 and address 192.168.238.129
Jul 6 08:49:15 localhost flanneld: I0706 08:49:15.831905 8741 main.go:544] Defaulting external address to interface address (192.168.238.129)
Jul 6 08:49:15 localhost flanneld: I0706 08:49:15.831987 8741 main.go:244] Created subnet manager: Etcd Local Manager with Previous Subnet: None
Jul 6 08:49:15 localhost flanneld: I0706 08:49:15.831992 8741 main.go:247] Installing signal handlers
Jul 6 08:49:15 localhost flanneld: E0706 08:49:15.834924 8741 main.go:382] Couldn't fetch network config: 100: Key not found (/coreos.com) [16]
Jul 6 08:49:16 localhost flanneld: timed out
Jul 6 08:49:16 localhost flanneld: E0706 08:49:16.837394 8741 main.go:382] Couldn't fetch network config: 100: Key not found (/coreos.com) [16]
Jul 6 08:49:17 localhost flanneld: timed out
Jul 6 08:49:17 localhost flanneld: E0706 08:49:17.840183 8741 main.go:382] Couldn't fetch network config: 100: Key not found (/coreos.com) [16]
Jul 6 08:49:18 localhost flanneld: timed out
Jul 6 08:49:18 localhost flanneld: E0706 08:49:18.842579 8741 main.go:382] Couldn't fetch network config: 100: Key not found (/coreos.com) [16]
Jul 6 08:49:19 localhost flanneld: timed out
Jul 6 08:49:19 localhost flanneld: E0706 08:49:19.845302 8741 main.go:382] Couldn't fetch network config: 100: Key not found (/coreos.com) [16]
Jul 6 08:49:20 localhost flanneld: timed out
Jul 6 08:49:20 localhost flanneld: E0706 08:49:20.848554 8741 main.go:382] Couldn't fetch network config: 100: Key not found (/coreos.com) [16]
测试网络是否正常
[root@node01 ~]# telnet 192.168.238.130 2379
Trying 192.168.238.130...
Connected to 192.168.238.130.
Escape character is '^]'.
quit
Connection closed by foreign host.
检查key是否存在
[root@master ~]# /opt/kubernetes/bin/etcdctl --ca-file=/opt/kubernetes/ssl/ca.pem --cert-file=/opt/kubernetes/ssl/server.pem --key-file=/opt/kubernetes/ssl/server-key.pem --endpoints="https://192.168.238.130:2379,https://192.168.238.129:2379,https://192.168.238.128:2379" get /coreos.com/network/config
Error: 100: Key not found (/coreos.com) [16]
主节点重新添加网络步骤一
[root@master ~]# /opt/kubernetes/bin/etcdctl --ca-file=/opt/kubernetes/ssl/ca.pem --cert-file=/opt/kubernetes/ssl/server.pem --key-file=/opt/kubernetes/ssl/server-key.pem --endpoints="https://192.168.238.130:2379,https://192.168.238.129:2379,https://192.168.238.128:2379" set /coreos.com/network/config '{"Network":"172.17.0.0/16","Backend":{"Type":"vxlan"}}'
{"Network":"172.17.0.0/16","Backend":{"Type":"vxlan"}}
再次启动
[root@node01 ~]# systemctl start flanneld
[root@node01 ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
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: eno16777736: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:29:11:0e brd ff:ff:ff:ff:ff:ff
inet 192.168.238.129/24 brd 192.168.238.255 scope global dynamic eno16777736
valid_lft 1633sec preferred_lft 1633sec
inet6 fe80::20c:29ff:fe29:110e/64 scope link
valid_lft forever preferred_lft forever
3: docker0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN
link/ether 02:42:aa:0a:b1:a5 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
4: flannel.1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN
link/ether 16:22:a1:7a:3a:99 brd ff:ff:ff:ff:ff:ff
inet 172.17.64.0/32 scope global flannel.1
valid_lft forever preferred_lft forever
inet6 fe80::1422:a1ff:fe7a:3a99/64 scope link
valid_lft forever preferred_lft forever
查看flannel分配的子网信息
[root@node01 ~]# cat /run/flannel/subnet.env
DOCKER_OPT_BIP="--bip=172.17.64.1/24"
DOCKER_OPT_IPMASQ="--ip-masq=false"
DOCKER_OPT_MTU="--mtu=1450"
DOCKER_NETWORK_OPTIONS=" --bip=172.17.64.1/24 --ip-masq=false --mtu=1450"
配置docker，注释相同选项，新增如下内容
[root@node01 ~]# vi /usr/lib/systemd/system/docker.service
EnvironmentFile=/run/flannel/subnet.env
ExecStart=/usr/bin/dockerd $DOCKER_NETWORK_OPTIONS
重启docker
[root@node01 ~]# systemctl daemon-reload
[root@node01 ~]# systemctl restart docker
[root@node01 ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
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: eno16777736: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:29:11:0e brd ff:ff:ff:ff:ff:ff
inet 192.168.238.129/24 brd 192.168.238.255 scope global dynamic eno16777736
valid_lft 1400sec preferred_lft 1400sec
inet6 fe80::20c:29ff:fe29:110e/64 scope link
valid_lft forever preferred_lft forever
3: docker0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN
link/ether 02:42:aa:0a:b1:a5 brd ff:ff:ff:ff:ff:ff
inet 172.17.64.1/24 brd 172.17.64.255 scope global docker0
valid_lft forever preferred_lft forever
4: flannel.1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN
link/ether 16:22:a1:7a:3a:99 brd ff:ff:ff:ff:ff:ff
inet 172.17.64.0/32 scope global flannel.1
valid_lft forever preferred_lft forever
inet6 fe80::1422:a1ff:fe7a:3a99/64 scope link
valid_lft forever preferred_lft forever
此时docker0与flannel.1在同一网段内
节点2重复上述操作进行配置
[root@node02 ~]# systemctl start flanneld
[root@node02 ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
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: eno16777736: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:5a:c2:eb brd ff:ff:ff:ff:ff:ff
inet 192.168.238.128/24 brd 192.168.238.255 scope global dynamic eno16777736
valid_lft 1496sec preferred_lft 1496sec
inet6 fe80::20c:29ff:fe5a:c2eb/64 scope link
valid_lft forever preferred_lft forever
3: docker0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN
link/ether 02:42:63:4f:0b:45 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
4: flannel.1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN
link/ether ea:b7:55:da:3b:a7 brd ff:ff:ff:ff:ff:ff
inet 172.17.89.0/32 scope global flannel.1
valid_lft forever preferred_lft forever
inet6 fe80::e8b7:55ff:feda:3ba7/64 scope link
valid_lft forever preferred_lft forever
设置docker
[root@node02 ~]# vim /usr/lib/systemd/system/docker.service
EnvironmentFile=/run/flannel/subnet.env
ExecStart=/usr/bin/dockerd $DOCKER_NETWORK_OPTIONS
[root@node02 ~]# systemctl daemon-reload
[root@node02 ~]# systemctl restart docker
[root@node02 ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
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: eno16777736: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:5a:c2:eb brd ff:ff:ff:ff:ff:ff
inet 192.168.238.128/24 brd 192.168.238.255 scope global dynamic eno16777736
valid_lft 1191sec preferred_lft 1191sec
inet6 fe80::20c:29ff:fe5a:c2eb/64 scope link
valid_lft forever preferred_lft forever
3: docker0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN
link/ether 02:42:63:4f:0b:45 brd ff:ff:ff:ff:ff:ff
inet 172.17.89.1/24 brd 172.17.89.255 scope global docker0
valid_lft forever preferred_lft forever
4: flannel.1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN
link/ether ea:b7:55:da:3b:a7 brd ff:ff:ff:ff:ff:ff
inet 172.17.89.0/32 scope global flannel.1
valid_lft forever preferred_lft forever
inet6 fe80::e8b7:55ff:feda:3ba7/64 scope link
valid_lft forever preferred_lft forever
测试网络是否正常
[root@node02 ~]# ping 172.17.64.1
PING 172.17.64.1 (172.17.64.1) 56(84) bytes of data.
64 bytes from 172.17.64.1: icmp_seq=1 ttl=64 time=0.508 ms
64 bytes from 172.17.64.1: icmp_seq=2 ttl=64 time=0.336 ms
[root@node01 ~]# ping 172.17.64.1
PING 172.17.64.1 (172.17.64.1) 56(84) bytes of data.
64 bytes from 172.17.64.1: icmp_seq=1 ttl=64 time=0.032 ms
64 bytes from 172.17.64.1: icmp_seq=2 ttl=64 time=0.030 ms
启用防火墙的情况下需要配置防火墙策略
[root@master ~]# iptables -I INPUT -s 192.168.0.0/24 -j ACCEPT
列出存储的信息
[root@master ~]# /opt/kubernetes/bin/etcdctl --ca-file=/opt/kubernetes/ssl/ca.pem --cert-file=/opt/kubernetes/ssl/server.pem --key-file=/opt/kubernetes/ssl/server-key.pem --endpoints="https://192.168.238.130:2379,https://192.168.238.129:2379,https://192.168.238.128:2379" ls /coreos.com/network/
/coreos.com/network/subnets
/coreos.com/network/config
列出配置的网络
[root@master ~]# /opt/kubernetes/bin/etcdctl --ca-file=/opt/kubernetes/ssl/ca.pem --cert-file=/opt/kubernetes/ssl/server.pem --key-file=/opt/kubernetes/ssl/server-key.pem --endpoints="https://192.168.238.130:2379,https://192.168.238.129:2379,https://192.168.238.128:2379" ls /coreos.com/network/subnets
/coreos.com/network/subnets/172.17.64.0-24
/coreos.com/network/subnets/172.17.89.0-24
获取key
[root@master ~]# /opt/kubernetes/bin/etcdctl --ca-file=/opt/kubernetes/ssl/ca.pem --cert-file=/opt/kubernetes/ssl/server.pem --key-file=/opt/kubernetes/ssl/server-key.pem --endpoints="https://192.168.238.130:2379,https://192.168.238.129:2379,https://192.168.238.128:2379" get /coreos.com/network/subnets/172.17.64.0-24
{"PublicIP":"192.168.238.129","BackendType":"vxlan","BackendData":{"VtepMAC":"16:22:a1:7a:3a:99"}}
查看路由表信息
[root@node01 ~]# ip route show
default via 192.168.238.2 dev eno16777736 proto static metric 100
172.17.64.0/24 dev docker0 proto kernel scope link src 172.17.64.1
172.17.89.0/24 via 172.17.89.0 dev flannel.1 onlink
192.168.238.0/24 dev eno16777736 proto kernel scope link src 192.168.238.129 metric 100