最近RHEL6.4系统老是自动关机，请各位大佬帮忙分析下，谢谢。

like310101

/var/log/messages:

Jan 16 04:00:06 mongoserver2 ntpdate[15874]: adjust time server 202.28.214.2 offset 0.109793 sec
Jan 16 05:00:06 mongoserver2 ntpdate[18801]: adjust time server 218.189.210.10 offset -0.140782 sec
Jan 16 06:00:05 mongoserver2 ntpdate[21837]: adjust time server 157.7.141.240 offset 0.072717 sec
Jan 16 07:00:05 mongoserver2 ntpdate[25310]: adjust time server 202.73.36.32 offset -0.027547 sec
Jan 16 08:00:05 mongoserver2 ntpdate[29454]: adjust time server 27.114.150.12 offset 0.027424 sec
Jan 16 09:00:05 mongoserver2 ntpdate[1316]: adjust time server 158.199.140.83 offset 0.002306 sec
Jan 16 09:18:30 mongoserver2 kernel: sky2 eth0: Link is down.
Jan 16 09:28:47 mongoserver2 init: tty (/dev/tty2) main process (229 killed by TERM signal
Jan 16 09:28:47 mongoserver2 init: tty (/dev/tty3) main process (2300) killed by TERM signal
Jan 16 09:28:47 mongoserver2 init: tty (/dev/tty4) main process (2302) killed by TERM signal
Jan 16 09:28:47 mongoserver2 init: tty (/dev/tty5) main process (2304) killed by TERM signal
Jan 16 09:28:47 mongoserver2 init: tty (/dev/tty6) main process (2306) killed by TERM signal
Jan 16 09:28:47 mongoserver2 init: tty (/dev/tty1) main process (2867) killed by TERM signal
Jan 16 09:28:49 mongoserver2 abrtd: Got signal 15, exiting
Jan 16 09:28:53 mongoserver2 acpid: exiting
Jan 16 09:28:53 mongoserver2 init: Disconnected from system bus
Jan 16 09:28:53 mongoserver2 rtkit-daemon[2507]: Demoting known real-time threads.
Jan 16 09:28:53 mongoserver2 rtkit-daemon[2507]: Demoted 0 threads.
Jan 16 09:28:53 mongoserver2 console-kit-daemon[2376]: WARNING: no sender#012
Jan 16 09:28:54 mongoserver2 rpcbind: rpcbind terminating on signal. Restart with "rpcbind -w"
Jan 16 09:28:54 mongoserver2 auditd[1517]: The audit daemon is exiting.
Jan 16 09:28:54 mongoserver2 kernel: type=1305 audit(1389835734.217:1155): audit_pid=0 old=1517 auid=4294967295 ses=4294967295 subj=system_u:system_r:auditd_t:s0 res=1
Jan 16 09:28:54 mongoserver2 kernel: type=1305 audit(1389835734.300:1156): audit_enabled=0 old=1 auid=4294967295 ses=4294967295 subj=system_u:system_r:auditctl_t:s0 res=1
Jan 16 09:28:54 mongoserver2 kernel: Kernel logging (proc) stopped.
Jan 16 09:28:54 mongoserver2 rsyslogd: [origin software="rsyslogd" swVersion="5.8.10" x-pid="1542" x-info="http://www.rsyslog.com"] exiting on signal 15.
Jan 17 08:37:18 mongoserver2 kernel: imklog 5.8.10, log source = /proc/kmsg started.

q1208c

18分 网卡掉线
28分 开始关机

log里的进程是收到 SINGLE TERM 才停掉的. 难道有人手工关机了? 或者按了电源 ?

liou146

目测楼主用的是双机环境，server2 被“Fencing”了，

nimysun

like310101 发表于 2014-01-17 08:55
/var/log/messages:

Jan 16 04:00:06 mongoserver2 ntpdate[15874]: adjust time server 202.28.214.2 o ...

我目测不出来，有可能是server硬件本身的问题，有可能是cluster把它给按死了。

netxjman_cu

网络问题引起的双机动作！节点被FENCE了！

humjb_1983

like310101 发表于 2014-01-17 08:55
/var/log/messages:

Jan 16 04:00:06 mongoserver2 ntpdate[15874]: adjust time server 202.28.214.2 o ...

可以肯定的是，这应该是软件发起的重启，从打印上看，由于没有shutdown和reboot的打印，应该是有业务主动调用了reboot的系统调用或是C库接口。

like310101

回复 2# q1208c


    可以确定的是没有手动关机。history 中没有相关记录。手动按电源是肯定没有的。现在还不知道原因。

like310101

回复 5# netxjman_cu


    这根网线比较差，我做根好点的换了再观察看看有没有解决。

q1208c

回复 7# like310101

关机不一定会在log里有记录的.

ssh host1 'shutdown -r now'

你要是有合适的机器, 可以试试.
   

like310101

已经确定是网线问题，导致有时网络不通，mongo程序执行了关机操作。

谢谢各位。

结贴。