Out of memory: Kill process 2075 (mysqld) score 100 or sacrifice child

[nilldot]

Environment: Zabbix 2.0 Appliance.

Problem: After 2 days of running I got a message: can't connect to Mysql server () and etc.

While troubleshooting i found this (Please find more from log file attached) :

Jun 7 18:02:12 Perseus kernel: [106462.583486] zabbix_server invoked oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0
Jun 7 18:02:12 Perseus kernel: [106462.583493] zabbix_server cpuset=/ mems_allowed=0

Jun 7 18:02:12 Perseus kernel: [106462.584937] Out of memory: Kill process 2075 (mysqld) score 100 or sacrifice child
Jun 7 18:02:12 Perseus kernel: [106462.584942] Killed process 2075 (mysqld) total-vm:320780kB, anon-rss:50996kB, file-rss:32kB
Jun 7 18:02:14 Perseus mysql[14527]: Shutting down service MySQL ..done

Only several days of running on fairly loaded system ( 2 hosts with 10 items!).

[nilldot]

After a week of running, got the same problem >> Out of memory: Kill process mysqld

The bad thing is that now I'm now able to start the mysqld daemon resulting the following errors in my mysqld.log file:

120618 10:49:51 mysqld_safe mysqld from pid file /var/run/mysql/mysqld.pid ended
120618 10:51:07 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
120618 10:51:07 InnoDB: The InnoDB memory heap is disabled
120618 10:51:07 InnoDB: Mutexes and rw_locks use GCC atomic builtins
120618 10:51:07 InnoDB: Compressed tables use zlib 1.2.5
120618 10:51:07 InnoDB: Initializing buffer pool, size = 128.0M
InnoDB: mmap(135987200 bytes) failed; errno 12
120618 10:51:07 InnoDB: Completed initialization of buffer pool
120618 10:51:07 InnoDB: Fatal error: cannot allocate memory for the buffer pool
120618 10:51:07 [ERROR] Plugin 'InnoDB' init function returned error.
120618 10:51:07 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
120618 10:51:07 [ERROR] Unknown/unsupported storage engine: InnoDB
120618 10:51:07 [ERROR] Aborting

120618 10:51:07 [Note] /usr/sbin/mysqld: Shutdown complete

P.S.
A really sick thing to happen on Friday evening


Some update:
I have uncommented the innodb_buffer_pool_size = 16M line in my /etc/my.cnf thus I was able to start the service, let's see how it goes.