上周六凌晨协助一客户进行服务器数据库升级。
数据库升级完成后，为了检验服务器重启后数据库能否正常启动并主从同步，故对系统进行了reboot操作。
没想到reboot后，服务器一直都能PING通，但SSH却登录不过去，且各类服务端口也无法连接。
焦急中等待了十多分钟，服务器正常启动了，业务恢复正常。

事后查阅服务器日志（/var/log/messages）发现关机时出现异常慢：
Aug 4 02:14:19 localhost init: tty (/dev/tty1) main process (2380) killed by TERM signal
Aug 4 02:14:19 localhost init: tty (/dev/tty2) main process (2383) killed by TERM signal
Aug 4 02:14:19 localhost init: tty (/dev/tty3) main process (2385) killed by TERM signal
Aug 4 02:14:19 localhost init: tty (/dev/tty4) main process (2387) killed by TERM signal
Aug 4 02:14:19 localhost init: tty (/dev/tty5) main process (2389) killed by TERM signal
Aug 4 02:14:19 localhost init: tty (/dev/tty6) main process (2391) killed by TERM signal
Aug 4 02:14:19 localhost avahi-daemon[1363]: Got SIGTERM, quitting.
Aug 4 02:14:19 localhost avahi-daemon[1363]: Leaving mDNS multicast group on interface eth0.IPv4 with address 172.31.251.67.
Aug 4 02:14:20 localhost abrtd: Got signal 15, exiting
Aug 4 02:27:00 localhost acpid: exiting
Aug 4 02:27:01 localhost init: Disconnected from system bus
Aug 4 02:27:01 localhost rpcbind: rpcbind terminating on signal. Restart with "rpcbind -w"
Aug 4 02:27:01 localhost auditd[1210]: The audit daemon is exiting.
Aug 4 02:27:01 localhost kernel: type=1305 audit(1375554421.205:1706): audit_pid=0 old=1210 auid=4294967295 ses=4294967295 res=1
Aug 4 02:27:01 localhost kernel: type=1305 audit(1375554421.309:1707): audit_enabled=0 old=1 auid=4294967295 ses=4294967295 res=1
Aug 4 02:27:01 localhost kernel: Kernel logging (proc) stopped.
Aug 4 02:27:01 localhost rsyslogd: [origin software="rsyslogd" swVersion="5.8.10" x-pid="1235" x-info="http://www.rsyslog.com"] exiting on signal 15.

解决方案（还有待时间验证）：
尝试关闭acpid服务：

# chkconfig acpid off

引用百度百科中acpid词条：
Advanced Configuration and PowerInterface，为替代传统的APM电源管理标准而推出的新型电源管理标准。建议所有的笔记本用户开启它。一些服务器可能不需要acpi。支持的通用操作有：“电源开关“，”电池监视“，”笔记本 Lid 开关“，“笔记本显示屏亮度“，“休眠”， “挂机”，等等。