1、安装etcd
yum install etcd
systemctl stop etcd
systemctl start etcd
systemctl status etcd
systemctl enable etcd ##开机自启动
查看etcd是否启动成功,etcd默认端口为2379
$ lsof -i:2379
测试etcd可用性
$ etcd --version
$ etcdctl mkdir /test
$ etcdctl ls /
2、安装flannel
$ wget https://github.com/coreos/flannel/releases/download/v0.5.5/flannel-0.5.5-linux-amd64.tar.gz
$ tar xvf flannel-0.5.5-linux-amd64.tar.gz
$ cd flannel-0.5.5
$ mv flanneld /usr/bin
$ mv mk-docker-opts.sh /usr/bin
启动flannel方法一:
$ flanneld &
查看flannel是否启动
$ ps -ef|grep flannel
启动flannel方法二(推荐):添加到systemd服务中
a、创建配置/etc/flannel/flanneld.conf
$ mkdir -p /etc/flannel/
$ vim /etc/flannel/flanneld.conf
# Flanneld configuration options
# etcd url location. Point this to the server where etcd runs
FLANNEL_ETCD_ENDPOINTS="-etcd-endpoints=http://localhost:2379"
# etcd集群配置
FLANNEL_ETCD_ENDPOINTS="-etcd-endpoints=http://172.16.0.6:2379,http://172.16.0.7:2379"
# etcd config key. This is the configuration key that flannel queries
# For address range assignment
FLANNEL_ETCD_PREFIX="-etcd-prefix=/coreos.com/network"
# Any additional options that you want to pass
#FLANNEL_OPTIONS=""
创建flanneld服务文件/usr/lib/systemd/system/flanneld.service
$ vim /usr/lib/systemd/system/flanneld.service
[Unit]
Description=Flanneld overlay address etcd agent
After=network.target
After=network-online.target
Wants=network-online.target
After=etcd.service
Before=docker.service
[Service]
Type=notify
EnvironmentFile=-/etc/flannel/flanneld.conf
ExecStart=/usr/bin/flanneld $FLANNEL_ETCD_ENDPOINTS $FLANNEL_ETCD_PREFIX $FLANNEL_OPTIONS
Restart=on-failure
[Install]
WantedBy=multi-user.target
RequiredBy=docker.service
启动flanneld
$ systemctl daemon-reload
$ systemctl start flanneld
$ systemctl restart flanneld
$ systemctl enable flanneld ##开机自启动
查看启动日志:
$ systemctl status flanneld -l
发现报错了:Couldn't fetch network config: 100: Key not found
想起来etcd里还没写入网络配置信息呢;
flanneld网络配置
docker默认用172.17.0.0/16这个网段,我想换个网段,以明确知道是在使用flanneld提供的网络,于是就用了172.10.0.0/16。
$ etcdctl mk /coreos.com/network/config '{"Network":"172.10.0.0/16", "SubnetLen": 24, "SubnetMin": "172.10.1.0", "SubnetMax": "172.10.254.0", "Backend": {"Type": "vxlan"}}' $ etcdctl get /coreos.com/network/config
{"Network":"172.10.0.0/16", "SubnetLen": 24, "SubnetMin": "172.10.1.0", "SubnetMax": "172.10.254.0", "Backend": {"Type": "vxlan"}}
因为flannel一直在尝试从etcd读配置信息,所以不需要重启flanneld,再看启动日志,flanneld已经获取到想要的子网地址了;
此时,在/run/flannel/subnet.env位置会生成这个host所分配到的子网信息:
FLANNEL_NETWORK=172.10.0.0/16
FLANNEL_SUBNET=172.10.60.1/24
FLANNEL_MTU=1450
FLANNEL_IPMASQ=false
正常情况下,执行ifconfig或ip a可以看到flannel0桥接网卡了,etcd上执行etcdctl ls /coreos.com/network/subnets也能够看到flannel申请的网段。
而这个文件正是mk-docker-opts.sh脚本的输入文件:
$ /usr/bin/mk-docker-opts.sh -c
$ cat /run/docker_opts.env
DOCKER_OPTS=" --bip=172.10.60.1/24 --ip-masq=true --mtu=1450 "
这时看一下etcd,多了一条信息:
$ etcdctl get /coreos.com/network/subnets/172.10.60.0-24
{"PublicIP":"172.16.92.255","BackendType":"vxlan","BackendData":{"VtepMAC":"1e:36:0a:72:6b:c7"}}
看一下主机网卡
$ ip a |grep flannel
2036: flannel.1:
inet 172.10.60.0/16 scope global flannel.1
看一下网卡详情
$ ip -d link show dev flannel.1
2036: flannel.1:
link/ether 1e:36:0a:72:6b:c7 brd ff:ff:ff:ff:ff:ff promiscuity 0
vxlan id 1 local 172.16.92.255 dev eth0 srcport 0 0 dstport 8472 nolearning ageing 300 noudpcsum noudp6zerocsumtx noudp6zerocsumrx addrgenmode eui64 numtxqueues 1 numrxqueues 1 gso_max_size 65536 gso_max_segs 65535
3、Docker使用flannel网络
再重新生成一次docker_opts.env:
# /usr/bin/mk-docker-opts.sh -c
# cat /run/docker_opts.env
DOCKER_OPTS=" --bip=172.10.60.1/24 --ip-masq=true --mtu=1450 "
修改vim /lib/systemd/system/docker.service:
EnvironmentFile=/run/docker_opts.env # 在ExecStart参数之前新增这句话
ExecStart=/usr/bin/dockerd $DOCKER_OPTS #在ExecStart末尾添加$DOCKER_OPTS,$DOCKER_OPTS添加网桥的参数
重启docker服务:
# systemctl daemon-reload
# systemctl restart docker
查看bridge网络的信息:
# docker network ls |grep bridge
8d4e33066eb0 bridge bridge local
db61b72ccaa7 metabase_default bridge local
# docker network inspect 8d4e33066eb0 | grep Subnet
"Subnet": "172.18.0.0/16"
这就已经是在flannel提供的网段内了;
测试
然后在另一台主机上也照做一遍,分别启动一个container,互ping一下,测试通过。
docker run -itd --name mycentos docker.io/centos /bin/bash
4、flannel生成的vxlan设备的删除方法
停止flanneld服务并确认
# systemctl stop flanneld
# status flanneld
使用ifconfig将设备停止
# ip addr s flannel.1
# ifconfig flannel.1 down
使用ip link del删除vxlan设备
# ip link del flannel.1
# ip addr s flannel.1
==========================================================================
温馨提示:
如上面操作后,发现各容器内分配的ip之间相互ping不通,基本就是由于防火墙问题引起的!
可是明明已经在前面部署的时候,通过"systemctl stop firewalld.service"关闭了防火墙,为什么还有防火墙问题??
这是因为linux还有底层的iptables,所以解决办法是在各节点上执行下面操作:
$ systemctl stop firewalld.service ##只关闭一次防火墙
$ systemctl disable firewalld.service ##永久关闭一次防火墙
[root@node-1 ~]# iptables -P INPUT ACCEPT
[root@node-1 ~]# iptables -P FORWARD ACCEPT
[root@node-1 ~]# iptables -F
执行上面操作后,基本各容器间就能相互ping通了。
docker通过Flannel可以实现各容器间的相互通信,即宿主机和容器,容器和容器之间都能相互通信。
如果服务器重启,这时docker是基于flannel网络的,docker自启动会失败,此时我们可以把etcd、flannel设置自启动,然后执行如下命令进行docker重启:
/usr/bin/mk-docker-opts.sh -c
systemctl daemon-reload
systemctl restart docker