我有一个使用Ubuntu 14.04，PHP 5.5.9和mysql Ver 12.12 Distrib 5.5.40在VPS上运行的网站。
大约90％的时间，一切似乎都在使用MySQL服务正常运行。然而，偶尔MySQL服务似乎会崩溃，并且它不会自行恢复在线状态。
由于升级虚拟专用服务器实例以拥有更多RAM，因此此问题尚未再次出现。在一个盒子上运行Apache和MySQL服务器需要超过512MB的内存才能获得稳定的网站。
当我在Apache错误日志中看到以下错误时，我首先注意到了这个问题：
PHP Warning: mysqli::mysqli(): (HY000/2002): Connection refused in [...]
我检查了日志文件/var/log/mysql.err，var/log/mysql.log和/var/log/mysql/error.log，但它们都是空的。
我检查了日志文件/var/log/mysql/error.log.1.gz并找到了：
141218 14:30:20 [Warning] Using unique option prefix myisam-recover instead of myisam-recover-options is deprecated and will be removed in a future release. Please use the full name instead.
141218 14:30:20 [Note] Plugin 'FEDERATED' is disabled.
141218 14:30:20 InnoDB: The InnoDB memory heap is disabled
141218 14:30:20 InnoDB: Mutexes and rw_locks use GCC atomic builtins
141218 14:30:20 InnoDB: Compressed tables use zlib 1.2.8
141218 14:30:20 InnoDB: Using Linux native AIO
141218 14:30:20 InnoDB: Initializing buffer pool, size = 128.0M
InnoDB: mmap(137363456 bytes) failed; errno 12
141218 14:30:20 InnoDB: Completed initialization of buffer pool
141218 14:30:20 InnoDB: Fatal error: cannot allocate memory for the buffer pool
141218 14:30:20 [ERROR] Plugin 'InnoDB' init function returned error.
141218 14:30:20 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
141218 14:30:20 [ERROR] Unknown/unsupported storage engine: InnoDB
141218 14:30:20 [ERROR] Aborting
我现在从命令行检查与mysql的连接：
mysql -u root -p并返回ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (111)
无赖。我尝试telnet 127.0.0.1 3306并获得telnet: Unable to connect to remote host: Connection refused
我还查看了netstat -a | grep mysql并且没有输出
运行sudo service mysql start后，一切恢复正常。 netstat显示正在运行的mysql服务，telnet连接，mysql命令行连接工作，并且不再有PHP警告。网站恢复正常功能。
我查看/var/log/mysql/error.log并看到以下内容：
141219 12:26:31 [Warning] Using unique option prefix myisam-recover instead of myisam-recover-options is deprecated and will be removed in a future release. Please use the full name instead.
141219 12:26:31 [Note] Plugin 'FEDERATED' is disabled.
141219 12:26:31 InnoDB: The InnoDB memory heap is disabled
141219 12:26:31 InnoDB: Mutexes and rw_locks use GCC atomic builtins
141219 12:26:31 InnoDB: Compressed tables use zlib 1.2.8
141219 12:26:31 InnoDB: Using Linux native AIO
141219 12:26:31 InnoDB: Initializing buffer pool, size = 128.0M
141219 12:26:31 InnoDB: Completed initialization of buffer pool
141219 12:26:31 InnoDB: highest supported file format is Barracuda.
InnoDB: The log sequence number in ibdata files does not match
InnoDB: the log sequence number in the ib_logfiles!
141219 12:26:31 InnoDB: Database was not shut down normally!
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files...
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer...
141219 12:26:31 InnoDB: Waiting for the background threads to start
141219 12:26:32 InnoDB: 5.5.40 started; log sequence number 19000727
141219 12:26:32 [Note] Server hostname (bind-address): '127.0.0.1'; port: 3306
141219 12:26:32 [Note] - '127.0.0.1' resolves to '127.0.0.1';
141219 12:26:32 [Note] Server socket created on IP: '127.0.0.1'.
141219 12:26:32 [Note] Event Scheduler: Loaded 0 events
141219 12:26:32 [Note] /usr/sbin/mysqld: ready for connections.
Version: '5.5.40-0ubuntu0.14.04.1' socket: '/var/run/mysqld/mysqld.sock' port: 3306 (Ubuntu)
没有让MySQL服务失效显然是理想的，但至少我想要一些方法来检测它是否发生故障，然后在发生这种情况时自动重启它。我考虑过设置一个cron作业来运行一个尝试连接到本地mysql服务的脚本，然后在连接失败时重新启动服务。但是，如果我能够解决根本原因而不是坚持使用创可贴，那将会好得多。
现在我发现的唯一线索是我的调试步骤中步骤3中日志文件/var/log/mysql/error.log.1.gz的最后7行。我的理论是我的VPS有时超载导致MySQL服务崩溃和烧毁，但我真的不想在不知道可以解决问题的情况下增加我的VPS大小。我希望有人遇到类似的问题，并且可以说明服务是否因为内存不足而导致服务中断，以及内存错误是否是一个不同的问题。
其他StackOverflow问题
我还查看了PHP MYSQL could not connect (HY000/2002)和mysqli::mysqli(): (HY000/2002): Can't connect to local MySQL server through socket 'MySQL' (2)但这两个问题似乎只是解决了从PHP连接到MySQL的问题，而不是调试MySQL服务本身的问题。