由于 Linux 不同发行版所采用的默认管理框架的不同，Docker 在不同环境下所使用的默认配置文件也有所不同。这给初学者带来了极大的困惑和不便。好消息是，Docker 官方也意识到了这个问题，所以从 V1.12 开始，引入了并行通用配置文件 /etc/docker/daemon.json 。从此以后，用户可以在所有平台统一通过调整该配置文件来调整 Docker Engine。本文对此进行简要说明。

配置文件

通用配置文件 /etc/docker/daemon.json

Docker Engine V1.12 之后版本，用户可以自行创建 daemon.json 文件对 Docker Engine 进行配置和调整。要点如下：

• 该文件作为 Docker Engine 的配置管理文件, 里面几乎涵盖了所有 docker 命令行启动可以配置的参数。
• 不管是在哪个平台以何种方式启动, Docker 默认都会来这里读取配置。使用户可以统一管理不同系统下的 docker daemon 配置。
• 相关参数的使用说明，可以参阅 man dockerd 帮助信息，或者参阅官方文档。
• 默认配置文件目录及支持的参数说明：
  • Linux： /etc/docker/daemon.json, 支持的完成参数配置示例如下（最新的支持参数列表可以参阅官方文档)：
    
{
"authorization-plugins": [],
"data-root": "",
"dns": [],
"dns-opts": [],
"dns-search": [],
"exec-opts": [],
"exec-root": "",
"experimental": false,
"storage-driver": "",
"storage-opts": [],
"labels": [],
"live-restore": true,
"log-driver": "",
"log-opts": {},
"mtu": 0,
"pidfile": "",
"cluster-store": "",
"cluster-store-opts": {},
"cluster-advertise": "",
"max-concurrent-downloads": 3,
"max-concurrent-uploads": 5,
"default-shm-size": "64M",
"shutdown-timeout": 15,
"debug": true,
"hosts": [],
"log-level": "",
"tls": true,
"tlsverify": true,
"tlscacert": "",
"tlscert": "",
"tlskey": "",
"swarm-default-advertise-addr": "",
"api-cors-header": "",
"selinux-enabled": false,
"userns-remap": "",
"group": "",
"cgroup-parent": "",
"default-ulimits": {},
"init": false,
"init-path": "/usr/libexec/docker-init",
"ipv6": false,
"iptables": false,
"ip-forward": false,
"ip-masq": false,
"userland-proxy": false,
"userland-proxy-path": "/usr/libexec/docker-proxy",
"ip": "0.0.0.0",
"bridge": "",
"bip": "",
"fixed-cidr": "",
"fixed-cidr-v6": "",
"default-gateway": "",
"default-gateway-v6": "",
"icc": false,
"raw-logs": false,
"allow-nondistributable-artifacts": [],
"registry-mirrors": [],
"seccomp-profile": "",
"insecure-registries": [],
"disable-legacy-registry": false,
"no-new-privileges": false,
"default-runtime": "runc",
"oom-score-adjust": -500,
"runtimes": {
"runc": {
"path": "runc"
},
"custom": {
"path": "/usr/local/bin/my-runc-replacement",
"runtimeArgs": [
"--debug"
]
}
}
}

  • Windows： %programdata%\docker\config\daemon.json，支持的完成参数配置示例如下（最新的支持参数列表可以参阅官方文档)：
    
{
"authorization-plugins": [],
"data-root": "",
"dns": [],
"dns-opts": [],
"dns-search": [],
"exec-opts": [],
"experimental": false,
"storage-driver": "",
"storage-opts": [],
"labels": [],
"log-driver": "",
"mtu": 0,
"pidfile": "",
"cluster-store": "",
"cluster-advertise": "",
"max-concurrent-downloads": 3,
"max-concurrent-uploads": 5,
"shutdown-timeout": 15,
"debug": true,
"hosts": [],
"log-level": "",
"tlsverify": true,
"tlscacert": "",
"tlscert": "",
"tlskey": "",
"swarm-default-advertise-addr": "",
"group": "",
"default-ulimits": {},
"bridge": "",
"fixed-cidr": "",
"raw-logs": false,
"allow-nondistributable-artifacts": [],
"registry-mirrors": [],
"insecure-registries": [],
"disable-legacy-registry": false
}


不同框架的配置文件

不同 Linux 发行版本默认使用的管理框架可能存在不同。而不同管理框架下配置文件的架构差异，最终导致了不同环境下默认 Docker 配置文件目录和文件名的不同。Linux 常见管理框架下的默认配置文件说明如下：

| 管理框架类型 | 默认使用该框架的操作系统 | Docker 默认配置文件 | 相关操作指令 |
| - | - | - | - |
| Sysvinit | CentOS 6.x, Ubuntu 12.04 等| /etc/default/docker | CentOS: service docker restart
Ubuntu: /etc/init.d/docker restart |
| Upstart | Ubuntu 14.x,15.x 等| /etc/default/docker| restart docker |
| Systemd | CentOS 7+，Ubunt 16+，Debian，RHEl 7, Fedora, Archlinux 等| Ubuntu/CentOS/: /lib/systemd/system/docker.service
RHEL/CentOS: /usr/lib/systemd/system/docker.service | systemctl restart docker.service |

配置文件修改说明

参阅前述说明，对配置文件调整后，注意如下要点：

• 修改配置后，需要重启 docker 服务生效。当前主流的 Sytemd 框架相关平台（CentOS 7+，Ubunt 16+ 等）的重启指令如下：
  
systemctl daemon-reolad
systemctl restart docker.service

• 配置冲突
  如果通用配置文件 /etc/docker/daemon.json 和上述默认配置文件中有配置冲突（相同配置项分别配置了不同值），那么 Docker daemon 启动时会报错。比如，如果在 Ubuntu 16.04 系统中，同时在 /etc/docker/daemon.json 和 /lib/systemd/system/docker.service 中指定了 --dns= 属性，那么 Docker daemon 会启动失败，相关错误信息示例如下：

  [root@node3 ~]# systemctl start docker.service
  Job for docker.service failed because the control process exited with error code. See "systemctl status docker.service" and "journalctl -xe" for details.
  
  [root@node3 ~]# systemctl status docker.service
  ● docker.service - Docker Application Container Engine
     Loaded: loaded (/lib/systemd/system/docker.service; enabled; vendor preset: enabled)
     Active: failed (Result: exit-code) since Wed 2017-11-15 15:57:16 CST; 2min 29s ago
       Docs: https://docs.docker.com
    Process: 30610 ExecStart=/usr/bin/dockerd -H tcp://0.0.0.0 -H fd:// --dns=223.5.5.5 --dns=223.6.6.6 (code=exited,
   Main PID: 30610 (code=exited, status=1/FAILURE)
  
  Nov 15 15:57:16 c400e32b9e7cd46aea0d484ad260f4e4a-node3 systemd[1]: Starting Docker Application Container Engine...
  Nov 15 15:57:16 c400e32b9e7cd46aea0d484ad260f4e4a-node3 dockerd[30610]: unable to configure the Docker daemon with file /etc/docker/daemon.json: the following directives are specified both as a flag and in the configuration file: dns: (from flag: [223.5.5.5  223.6.6.6], from file: [114.114.114.114 8.8.8.8])
  Nov 15 15:57:16 c400e32b9e7cd46aea0d484ad260f4e4a-node3 systemd[1]: docker.service: Main process exited, code=exite
  Nov 15 15:57:16 c400e32b9e7cd46aea0d484ad260f4e4a-node3 systemd[1]: Failed to start Docker Application Container En
  Nov 15 15:57:16 c400e32b9e7cd46aea0d484ad260f4e4a-node3 systemd[1]: docker.service: Unit entered failed state.
  Nov 15 15:57:16 c400e32b9e7cd46aea0d484ad260f4e4a-node3 systemd[1]: docker.service: Failed with result 'exit-code'.
  lines 1-13/13 (END)