docker 网络源码探究

希望能帮助到对源码感兴趣的同学

架构

图片来自 docker 官网文档

源码

dockerd https://github.com/moby/moby

docker https://github.com/docker/cli

相关目录

api

cmd

daemon

libnetwork

调用关系

源码版本 master commit 6834304,时间 2023年1月7日。

cmd/dockerd/docker.go main
cmd/dockerd/daemon.go start
重要的
daemon.NewDaemon 后面有讲,先略过
initRouter
主要关心 container 相关接口
container.NewRouter

api/server/router/container/container.go NewRouter 
注意参数 backend 即 daemon 指针
重要函数 initRoutes

api/server/router/container/backend.go Backend interface

感兴趣的 docker start 调用
daemon/start.go ContainerStart
重要的
daemon.conditionalMountOnStart
daemon.initializeNetworking
libcontainerd.ReplaceContainer

网络
daemon/container_operations.go initializeNetworking
daemon/container_operations.go allocateNetwork
重要的
daemon.netController.NewSandbox
daemon.connectToNetwork

daemon/container_operations.go connectToNetwork
ep, err := n.CreateEndpoint(endpointName, createOptions...)
err := ep.Join(sb, joinOptions...)
err = d.ProgramExternalConnectivity(n.ID(), ep.ID(), sb.Labels())

netController 是什么 回到前面的 daemon.NewDaemon
daemon/daemon.go NewDaemon
daemon/daemon.go restore
关注的
daemon.initNetworkController

daemon\daemon_unix.go initNetworkController
daemon.netController, err = libnetwork.New(netOptions...)
err := configureNetworking(daemon.netController, daemon.configStore)

initBridgeDriver
Initialize default network on "bridge" with the same name
controller.NewNetwork

网络 CNM

Design

Sandbox

容器网络栈包含路由表、dns、loopback 等。

Endpoint

实现 sandbox 与 network 的连接,如 veth。

Network

类似交换机实现多个 endpoint 互相通信,如 bridge docker0。

源码
bridge
libnetwork\controller.go New
drvRegistry.AddDriver(i.ntype, i.fn, dcfg)
libnetwork\drvregistry\drvregistry.go AddDriver
fn(r, config) 即 bridge.Init
libnetwork\drivers\bridge\bridge.go Init

initIPAMDrivers 备注 ipam 指 IP Address Management

libnetwork\controller.go NewNetwork
err = c.addNetwork(nw)
err := d.CreateNetwork(n.id, n.generic, n, n.getIPData(4), n.getIPData(6))

libnetwork\drivers\bridge\bridge.go CreateNetwork
bridgeIface, err := newInterface(d.nlh, config)
创建 bridge
libnetwork\drivers\bridge\setup_ipv4.go setupBridgeIPv4
配置 IP 等
libnetwork\drivers\bridge\setup_ip_tables.go setupIP4Tables
配置 iptables 
loopback
libnetwork\controller.go NewSandbox
libnetwork\osl\namespace_linux.go NewSandbox
命令行调 netns-create 创建 network namespace

github.com/vishvananda/netns
sboxNs, err := netns.GetFromPath(n.path) 获取 network namespace

github.com/vishvananda/netlink
n.nlHandle, err = netlink.NewHandleAt(sboxNs, syscall.NETLINK_ROUTE)
创建 NETLINK_ROUTE socket 操作 network device

libnetwork\osl\namespace_linux.go loopbackUp
iface, err := n.nlHandle.LinkByName("lo") 创建 loopback
dns
libnetwork\network.go CreateEndpoint
err = n.addEndpoint(ep)
调用相关 driver 的 CreateEndpoint
err = d.CreateEndpoint(n.id, ep.id, ep.Interface(), ep.generic)

libnetwork\endpoint.go Join
libnetwork\sandbox.go populateNetworkResources

libnetwork\sandbox_dns_unix.go startResolver
重要
resolverIPSandbox = "127.0.0.11"
err = sb.osSbox.InvokeFunc(sb.resolver.SetupFunc(0))
err = sb.resolver.Start()

libnetwork\resolver.go SetupFunc
参数 0 随机端口
r.conn, err = net.ListenUDP("udp", addr)
r.tcpListen, err = net.ListenTCP("tcp", tcpaddr)

libnetwork\resolver.go Start
err := r.setupIPTable()

libnetwork\resolver_unix.go setupIPTable
命令行调用参数 setup-resolver
libnetwork\resolver_unix.go reexecSetupResolver

github.com/miekg/dns
docker-proxy
libnetwork\drivers\bridge\bridge.go ProgramExternalConnectivity
network.allocatePorts(endpoint, network.config.DefaultBindingIP, d.config.EnableUserlandProxy)

libnetwork\portmapper\mapper.go MapRange
m.userlandProxy, err = newProxy(proto, hostIP, allocatedHostPort, t.IP, t.Port, pm.proxyPath)

libnetwork\portmapper\proxy_linux.go newProxyCommand
命令行启动 docker-proxy

实际环境观测
loopback
[root@computer02 /home/user_00]# docker ps | grep nginx
164b577c53df   nginx-test:a35b0f87    "/docker-entrypoint.…"   2 days ago     Up 2 days  0.0.0.0:80->80/tcp, 0.0.0.0:443->443/tcp nginx-test

[root@computer02 /home/user_00]# docker inspect --format '{{ .State.Pid }}'  164b577c53df
11297

[root@computer02 /home/user_00]# nsenter -t 11297  -n ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    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
279249: eth0@if279250: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default 
    link/ether 02:42:c0:a8:60:2c brd ff:ff:ff:ff:ff:ff link-netnsid 0
    inet 192.168.96.44/20 brd 192.168.111.255 scope global eth0
       valid_lft forever preferred_lft forever

dns
[root@computer02 /home/user_00]# nsenter -t 11297  -n netstat -lnp
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address           Foreign Address         State       PID/Program name    
tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN      11297/nginx: master 
tcp        0      0 127.0.0.11:34738        0.0.0.0:*               LISTEN      1693/dockerd        
udp        0      0 127.0.0.11:33184        0.0.0.0:*                           1693/dockerd

[root@computer02 /home/user_00]# nsenter -t 11297  -n iptables -t nat -L
Chain PREROUTING (policy ACCEPT)
target     prot opt source               destination         

Chain INPUT (policy ACCEPT)
target     prot opt source               destination         

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination         
DOCKER_OUTPUT  all  --  anywhere             127.0.0.11          

Chain POSTROUTING (policy ACCEPT)
target     prot opt source               destination         
DOCKER_POSTROUTING  all  --  anywhere             127.0.0.11          

Chain DOCKER_OUTPUT (1 references)
target     prot opt source               destination         
DNAT       tcp  --  anywhere             127.0.0.11           tcp dpt:domain to:127.0.0.11:34738
DNAT       udp  --  anywhere             127.0.0.11           udp dpt:domain to:127.0.0.11:33184

Chain DOCKER_POSTROUTING (1 references)
target     prot opt source               destination         
SNAT       tcp  --  127.0.0.11           anywhere             tcp spt:34738 to::53
SNAT       udp  --  127.0.0.11           anywhere             udp spt:33184 to::53
docker-proxy

相关 iptables 就不列了

[root@computer02 /home/user_00]# ps -aux | grep 192.168.96.44
root     10749  0.0  0.0 217044  3236 ?        Sl   Jan05   0:00 /usr/bin/docker-proxy -proto tcp -host-ip 0.0.0.0 -host-port 443 -container-ip 192.168.96.44 -container-port 443
root     10966  0.0  0.0 364508  3236 ?        Sl   Jan05   0:00 /usr/bin/docker-proxy -proto tcp -host-ip 0.0.0.0 -host-port 80 -container-ip 192.168.96.44 -container-port 80

网络架构

Linux 组件

bridge
veth

  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值