For a while, Drupal running on the server is normal, but then there is this error:

PDOException: SQLSTATE[HY000] [2002] Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2) in lock_may_be_available() (line 167 of /var/www/whydrupal.ru/public_html/includes/lock.inc).

earlier helped restart a droplet from the admin DO, but now this reboot does nothing.

Command in console service mysql restart answers:

sop: Unknown instance: 
start: Job failed to start

Logs:

/var/log/mysql/error.log:

140129  9:52:09 [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.
140129  9:52:09 [Note] Plugin 'FEDERATED' is disabled.
140129  9:52:09 InnoDB: The InnoDB memory heap is disabled
140129  9:52:09 InnoDB: Mutexes and rw_locks use GCC atomic builtins
140129  9:52:09 InnoDB: Compressed tables use zlib 1.2.3.4
140129  9:52:10 InnoDB: Initializing buffer pool, size = 64.0M
InnoDB: mmap(68681728 bytes) failed; errno 12
140129  9:52:10 InnoDB: Completed initialization of buffer pool
140129  9:52:10 InnoDB: Fatal error: cannot allocate memory for the buffer pool
140129  9:52:10 [ERROR] Plugin 'InnoDB' init function returned error.
140129  9:52:10 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
140129  9:52:10 [ERROR] Unknown/unsupported storage engine: InnoDB
140129  9:52:10 [ERROR] Aborting
140129  9:52:10 [Note] /usr/sbin/mysqld: Shutdown complete

I changed the parameters of many values ​​in /etc/mysql/my.cnf:

innodb_buffer_pool_size = 128М / 64М / 32M / 16М (does not help)
max_connections = 50    (500 default)
max_user_connections  = 10
innodb_buffer_pool_size = 32M     (128 default)
innodb_force_recovery = 4 (sometimes help but not permanently)

Full text my.cnf: http://pastebin.com/7bXWEtBu

tried to configure my-small.cnf (does not help)
rebooting server (does not help)
rebooting mysql (does not help)

more logs:

/var/log/syslog

Jan 30 07:15:36 localhost kernel: [165391.100711] select 27275 (apache2), adj 0, size 1306, to kill
Jan 30 07:15:36 localhost kernel: [165391.100713] select 21786 (apache2), adj 0, size 4084, to kill
Jan 30 07:15:36 localhost kernel: [165391.100714] select 21845 (apache2), adj 0, size 4454, to kill
Jan 30 07:15:36 localhost kernel: [165391.100715] select 21972 (apache2), adj 0, size 4586, to kill
Jan 30 07:15:36 localhost kernel: [165391.100729] select 1 (init), adj 0, size 365, to kill
Jan 30 07:15:36 localhost kernel: [165391.100732] select 817 (php5-fpm), adj 0, size 825, to kill

Jan 30 08:23:29 localhost kernel: [169464.172067] send sigkill to 24422 (apache2), adj 0, size 11829
Jan 30 08:23:29 localhost /etc/mysql/debian-start[24561]: Upgrading MySQL tables if necessary.
Jan 30 08:23:30 localhost /etc/mysql/debian-start[24564]: /usr/bin/mysql_upgrade: the '--basedir' option is always ignored
Jan 30 08:23:30 localhost /etc/mysql/debian-start[24564]: Looking for 'mysql' as: /usr/bin/mysql
Jan 30 08:23:30 localhost /etc/mysql/debian-start[24564]: Looking for 'mysqlcheck' as: /usr/bin/mysqlcheck
Jan 30 08:23:30 localhost /etc/mysql/debian-start[24564]: This installation of MySQL is already upgraded to 5.5.35, use --force if you still need to run mysql_upgrade
Jan 30 08:23:30 localhost /etc/mysql/debian-start[24575]: Checking for insecure root accounts.
Jan 30 08:23:30 localhost /etc/mysql/debian-start[24580]: Triggering myisam-recover for all MyISAM tables

in syslog incredible tons of errors that I find it hard to understand

/var/log/mysql/error.log:

140129 11:00:32 InnoDB: Initializing buffer pool, size = 16.0M
140129 11:00:32 InnoDB: Completed initialization of buffer pool
140129 11:00:32 InnoDB: highest supported file format is Barracuda.
InnoDB: Log scan progressed past the checkpoint lsn 71593559
140129 11:00:32  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...
InnoDB: Doing recovery: scanned up to log sequence number 71593569
140129 11:00:32  InnoDB: Waiting for the background threads to start
140129 11:00:33 InnoDB: 5.5.35 started; log sequence number 71593569

140215 13:10:48 [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.
140215 13:10:48 [Note] Plugin 'FEDERATED' is disabled.
140215 13:10:48 InnoDB: The InnoDB memory heap is disabled
140215 13:10:48 InnoDB: Mutexes and rw_locks use GCC atomic builtins
140215 13:10:48 InnoDB: Compressed tables use zlib 1.2.3.4
140215 13:10:48 InnoDB: Initializing buffer pool, size = 16.0M
140215 13:10:48 InnoDB: Completed initialization of buffer pool
InnoDB: Error: checksum mismatch in data file ./ibdata1
140215 13:10:48 InnoDB: Could not open or create data files.
140215 13:10:48 InnoDB: If you tried to add new data files, and it failed here,
140215 13:10:48 InnoDB: you should now edit innodb_data_file_path in my.cnf back
140215 13:10:48 InnoDB: to what it was, and remove the new ibdata files InnoDB created
140215 13:10:48 InnoDB: in this failed attempt. InnoDB only wrote those files full of
140215 13:10:48 InnoDB: zeros, but did not yet use them in any way. But be careful: do not
140215 13:10:48 InnoDB: remove old data files which contain your precious data!
140215 13:10:48 [ERROR] Plugin 'InnoDB' init function returned error.
140215 13:10:48 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
140215 13:10:48 [ERROR] Unknown/unsupported storage engine: InnoDB
140215 13:10:48 [ERROR] Aborting

140215 13:10:48 [Note] /usr/sbin/mysqld: Shutdown complete

Google does not help.
Similar unanswered question:
https://www.digitalocean.com/community/questions/mysql-keeping-crashings-on-my-drupal-site

Traffic in my all sites is very small: 200 users per day.
graph load of my server:

How to Raise MySQL?
How to make it work?