和Windows相比，Linux提供了多种关机、重启的命令，包括reboot/halt/shutdown等，而如果在服务器上，BMC还会提供ipmitool chassis power reset/ipmitool chassis power cycle/ipmitool chassis power off等命令，那么面对这么多可行的命令，该如何选择呢？
由于Linux内核和OS版本众多，不同的OS和内核版本上的reboot/halt/shutdown等命令实现方式可能会有差异，这就需要基于具体的OS和内核进行分析。下面基于CentOS 7，分享一下认识reboot/halt/shutdown差异的过程。
1.确定是否都是链接到同一个命令
通过下面的命令可以看到，在CentOS 7上所有的reboot命令都链接到systemctl：
[root@localhost etc]# which shutdown
/usr/sbin/shutdown
[root@localhost etc]# which poweroff
/usr/sbin/poweroff
[root@localhost etc]# which reboot
/usr/sbin/reboot
[root@localhost etc]# cd /usr/sbin/
[root@localhost sbin]# ls -alrt shutdown
lrwxrwxrwx. 1 root root 16 Apr 5 13:07 shutdown -> ../bin/systemctl
[root@localhost sbin]# ls -alrt poweroff
lrwxrwxrwx. 1 root root 16 Apr 5 13:07 poweroff -> ../bin/systemctl
[root@localhost sbin]# ls -alrt reboot
lrwxrwxrwx. 1 root root 16 Apr 5 13:07 reboot -> ../bin/systemctl
[root@localhost sbin]# which systemctl
/usr/bin/systemctl
2.认识怎样关机保证不造成数据丢失
除了要求能按照期望实现开关机、重启之外，在服务器尤其是存储服务器的工作环境下，用户对关机和重启能否造成数据丢失也特别关心。为此，弄清楚哪种使用方式不会造成数据丢失很有意义。还是以CentOS为例，根据systemctl的说明，reboot、halt与poweroff一样，只要不带 –force --force后缀，就不会造成数据丢失。下面是关于这些命令的具体说明：
halt
Shut down and halt the system. This is mostly equivalent to start halt.target --irreversible, but also prints a wall message to all users. If combined with --force, shutdown of all running services is skipped, however all processes are killed and all file systems are unmounted or mounted read-only, immediately followed by the system halt. If --force is specified twice, the operation is immediately executed without terminating any processes or unmounting any file systems. This may result in data loss.
poweroff
Shut down and power-off the system. This is mostly equivalent to start poweroff.target --irreversible, but also prints a wall message to all users. If combined with --force, shutdown of all running services is skipped, however all processes are killed and all file systems are unmounted or mounted read-only, immediately followed by the powering off. If --force is specified twice, the operation is immediately executed without terminating any processes or unmounting any file systems. This may result in data loss.
reboot [arg]
Shut down and reboot the system. This is mostly equivalent to start reboot.target --irreversible, but also rints a wall message to all users. If combined with --force, shutdown of all running services is skipped, however all processes are killed and all file systems are unmounted or mounted read-only, immediately followed by the reboot. If --force is specified twice, the operation is immediately executed without terminating any processes or unmounting any file systems. This may result in data loss.
Shutdown
如果后面什么参数都不加，效果和shutdown -P +1相同，即先调用wall向所有登录的用户发出即将关机的信息，然后组织新的用户登录，接着等待60s，最后执行poweroff。因此，
区别于halt/poweroff,shutdown更安全，在不带任何参数的情况下，shutdown本质是delayed shutdown，当前我们的CentOS上默认是延迟60s后关机，具体请参考下面的命令和日志文件：
[root@localhost ~]# cat /var/log/messages | tail -n 2
Apr 28 22:50:30 localhost systemd: Started Session 195 of user root.
Apr 28 22:50:30 localhost systemd-logind: New session 195 of user root.
[root@localhost ~]# date
Thu Apr 28 22:51:25 CST 2016
[root@localhost ~]# shutdown -k
Shutdown scheduled for Thu 2016-04-28 22:52:30 CST, use 'shutdown -c' to cancel.
[root@localhost ~]#
Broadcast message from root@localhost.localdomain (Thu 2016-04-28 22:51:30 CST):
The system is going down for power-off at Thu 2016-04-28 22:52:30 CST!
[root@localhost ~]# cat /var/log/messages | tail -n 4
Apr 28 22:51:30 localhost systemd: Starting Delayed Shutdown Service...
Apr 28 22:51:30 localhost systemd: Started Delayed Shutdown Service.
Apr 28 22:51:30 localhost systemd-shutdownd: Shutting down at Thu 2016-04-28 22:52:30 CST (poweroff)...
Apr 28 22:51:30 localhost systemd-shutdownd: Creating /run/nologin, blocking further logins...
关于shutdown和poweroff/halt/reboot的关系可参考下面的说明：
SYNOPSIS
shutdown [OPTIONS...] [TIME] [WALL...]
DESCRIPTION
shutdown may be used to halt, power-off or reboot the machine.
The first argument may be a time string (which is usually "now"). Optionally, this may be followed by a wall
message to be sent to all logged-in users before going down.
The time string may either be in the format "hh:mm" for hour/minutes specifying the time to execute the
shutdown at, specified in 24h clock format. Alternatively it may be in the syntax "+m" referring to the
specified number of minutes m from now. "now" is an alias for "+0", i.e. for triggering an immediate
shutdown. If no time argument is specified, "+1" is implied.
Note that to specify a wall message you must specify a time argument, too.
If the time argument is used, 5 minutes before the system goes down the /run/nologin file is created to ensure
that further logins shall not be allowed.
OPTIONS
The following options are understood:
--help
Print a short help text and exit.
-H, --halt
Halt the machine.
-P, --poweroff
Power-off the machine (the default).
-r, --reboot
Reboot the machine.
-h
Equivalent to --poweroff, unless --halt is specified.
-k
Do not halt, power-off, reboot, just write wall message.
--no-wall
Do not send wall message before halt, power-off, reboot.
-c
Cancel a pending shutdown. This may be used cancel the effect of an invocation of shutdown with a time
argument that is not "+0" or "now".
3.利用IPMI命令开关机
如果机器支持BMC，可以通过BMC使用ipmitool去开关机。这种情况下,BMC直接关掉给host系统的DC电源，如果是chassis power reset/chassis power cycle还会间隔一会重新上电，因此在host系统上运行的应用程序和IO数据都来不及killed掉并同步到硬盘，这就会造成数据丢失。当是它的好处是：快捷，并且可以基于BMC Lan 接口去远程开关、重启机器。
这些命令包括：
ipmitool chassis power off ＃通过BMC关必DC
ipmitool chassis power on # 通过BMC给DC上电
ipmitool chassis power cycle # 先把DC掉店，间隔约1s后给DC上电
ipmitool chassis power reset # 先把DC掉店，然后给DC上电
本文转自存储之厨51CTO博客，原文链接：http://blog.51cto.com/xiamachao/1769160 ，如需转载请自行联系原作者
