mysql启动失败又一例

搭的wordpress报错：

后台用的mysql，之前也崩过，原因是虚拟内存耗尽，通过增加swap空间最终让数据重新启动。

但仅过一晚上，数据库再次崩溃。看来要查一查是什么程序耗尽资源。

执行top，切到VIRT列，以虚拟内存使用量进行排序。

root@ns-xxzx-svr:/var/log/mysql# top
top - 10:51:24 up 81 days, 19:52,  2 users,  load average: 1.47, 1.14, 0.89
Tasks: 774 total,   3 running, 770 sleeping,   0 stopped,   1 zombie
%Cpu(s):  1.8 us, 32.4 sy,  0.0 ni, 36.4 id, 28.0 wa,  0.0 hi,  1.4 si,  0.0 st
KiB Mem:   1026284 total,   941336 used,    84948 free,      188 buffers
KiB Swap:  3952632 total,  3952632 used,        0 free,     4152 cached
  PID USER      PR  NI  VIRT  RES  SHR S  %CPU %MEM    TIME+  COMMAND
 9065 root      20   0  328m  50m    0 S   2.0  5.0  28:07.37 java
26155 hyang0    20   0  170m  24m  660 S   0.0  2.4   0:03.38 check-new-relea
28176 hyang0    20   0  170m  24m  624 S   0.0  2.4   0:04.07 check-new-relea
26782 hyang0    20   0  170m  23m  516 S   0.0  2.3   0:03.35 check-new-relea
25522 hyang0    20   0  169m  23m  652 S   0.0  2.3   0:04.92 check-new-relea
27531 hyang0    20   0  170m  23m  704 S   0.0  2.3   0:05.11 check-new-relea
28832 hyang0    20   0  169m  23m  516 S   0.0  2.3   0:07.84 check-new-relea
19835 hyang0    20   0  170m  22m  604 S   0.0  2.3   0:03.34 check-new-relea
20530 hyang0    20   0  170m  22m  704 S   0.0  2.2   0:03.81 check-new-relea
24742 hyang0    20   0  168m  21m  848 S   0.0  2.2   0:04.99 check-new-relea
11952 hyang0    20   0  170m  21m  856 S   0.0  2.1   0:03.28 check-new-relea
23943 hyang0    20   0  170m  21m  748 S   0.0  2.1   0:04.97 check-new-relea
 6850 hyang0    20   0  168m  20m  876 S   0.0  2.0   0:03.00 check-new-relea
17144 hyang0    20   0  170m  19m  648 S   0.0  1.9   0:03.24 check-new-relea
18423 hyang0    20   0  170m  19m  640 S   0.0  1.9   0:03.23 check-new-relea
 3105 hyang0    20   0  106m  18m 1796 S  41.5  1.9   1:40.46 check-new-relea
 2387 hyang0    20   0  170m  18m  624 S   0.0  1.9   0:03.18 check-new-relea

发现有几十个check-new-relea的进程，每个占用170多兆虚拟内存。再查看swap使用量：

root@ns-xxzx-svr:/var/log/mysql# sudo swapon  -s
Filename                                Type            Size    Used    Priority
/dev/sda5                               partition         -
/dev/sdd1                               partition         -

像上次一样，已经耗尽。在网上也找到类拟的问题：

下面就是处理这些进程了。用pstree, pgrep查找这些进程，确认不会误杀，pkill干掉这些进程。

root@ns-xxzx-svr:~# pgrep check-new- -l
 check-new-relea
 check-new-relea
 check-new-relea
 check-new-relea
 check-new-relea
 check-new-relea
 check-new-relea
 check-new-relea
 check-new-relea
 check-new-relea
 check-new-relea
 check-new-relea

root@ns-xxzx-svr:~# pkill check-new-
root@ns-xxzx-svr:~# pgrep check-new- -l
root@ns-xxzx-svr:~#

再次通过top 和 swapon工具查看资源情况，已经都释放了。mysql 再次复活。

root@ns-xxzx-svr:/var/log/mysql# sudo swapon  -s
Filename                                Type            Size    Used    Priority
/dev/sda5                               partition          -
/dev/sdd1                               partition          -
root@ns-xxzx-svr:/var/log/mysql# /etc/init.d/mysql start
Rather than invoking init scripts through /etc/init.d, use the service()
utility, e.g. service mysql start

Since the script you are attempting to invoke has been converted to an
Upstart job, you may also use the start() utility, e.g. start mysql
mysql start/running, process
root@ns-xxzx-svr:/var/log/mysql#