centos - Can't connect to local MySQL server, Lost connection, Server has gone away -

i have strange problem server.

first, server crashed. use cpanel , whm. , not able access used ssh restart server. , worked, normal.

but after began new problem: "can not connect local mysql server through socket '/var/lib/mysql/mysql.sock' (2)" "mysql server has gone away" "lost connection mysql server @ 'reading initial communication packet', system error: 104" "lost connection mysql server @ 'reading authorization packet', system error: 104"

but these errors not always, site online , working, having problem @ 1 time or another, or in cpanel , mysqltuner error shown.

the first thing did: search.

i changed my.cnf (added socks), not resolved. rebooted server, not solved. came old my.cnf (which had backup), not resolved. tried changing 'localhost' '127.0.0.1' created new problem "can not connect mysql server on '127.0.0.1'."

i tried find. nothing solved.

furthermore, process of "mysqld" restarting! before, process spent hours / days / weeks without restarting. @ "top" command shown mysqld on 50 hours, 100 hours in "time +". now, no more 10 minutes. in other words, process restarting!

in "-m free", have server 1303 free, , "top" using processor 4.5% "cpu (s)".

and mysterious thing, "cron jobs" not working. have cron job truncate , insert data. truncate works, insert / select not.

need in order fix problem. '(

i have server, centos system.

edit:

140910 12:32:01  innodb: error: space object of table 'w_w/srands', innodb: space id 274796 did not exist in memory. retrying open. innodb: error: tried read 16384 bytes @ offset 0 0. innodb: able read 0. 140910 12:32:01  innodb: operating system error number 0 in file operation. innodb: error number 0 means 'success'. innodb: operating system error numbers described @ innodb: http://dev.mysql.com/doc/refman/5.5/en/operating-system-error-codes.html innodb: file operation call: 'read'. innodb: cannot continue operation. 140910 12:32:01 mysqld_safe number of processes running now: 0 140910 12:32:01 mysqld_safe mysqld restarted 140910 12:32:02 [warning] using unique option prefix join_buffer instead of join_buffer_size deprecated , removed in future release. please use full name instead. 140910 12:32:02 [note] plugin 'federated' disabled. 140910 12:32:02 innodb: innodb memory heap disabled 140910 12:32:02 innodb: mutexes , rw_locks use gcc atomic builtins 140910 12:32:02 innodb: compressed tables use zlib 1.2.3 140910 12:32:02 innodb: using linux native aio 140910 12:32:02 innodb: initializing buffer pool, size = 119.0m 140910 12:32:02 innodb: completed initialization of buffer pool 140910 12:32:02 innodb: highest supported file format barracuda. innodb: log scan progressed past checkpoint lsn 395690234452 140910 12:32:02  innodb: database not shut down normally! innodb: starting crash recovery. innodb: reading tablespace information .ibd files... innodb: error: size of single-table tablespace file ./w_w/srands.ibd innodb: 0 0, should @ least 65536!innodb: restoring possible half-written data pages doublewrite innodb: buffer... innodb: doing recovery: scanned log sequence number 395690846835 140910 12:32:02  innodb: starting apply batch of log records database... innodb: progress in percents: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 5$ innodb: apply batch completed 140910 12:32:03  innodb: error: table 'w_w/srands' innodb: in innodb data dictionary has tablespace id 274796, innodb: tablespace id or name not exist. have innodb: deleted or moved .ibd files? innodb: may table created create temporary table innodb: .ibd , .frm files mysql automatically removed, innodb: table still exists in innodb internal data dictionary. innodb: please refer innodb: http://dev.mysql.com/doc/refman/5.5/en/innodb-troubleshooting-datadict.html innodb: how resolve issue. 140910 12:32:03  innodb: waiting background threads start 140910 12:32:04 innodb: 5.5.37 started; log sequence number 395690846835 140910 12:32:04 [note] server hostname (bind-address): '0.0.0.0'; port: 3306 140910 12:32:04 [note]   - '0.0.0.0' resolves '0.0.0.0'; 140910 12:32:04 [note] server socket created on ip: '0.0.0.0'. 140910 12:32:04 [note] event scheduler: loaded 1 event 140910 12:32:04 [note] /usr/sbin/mysqld: ready connections. version: '5.5.37-cll'  socket: '/var/lib/mysql/mysql.sock'  port: 3306  mysql community server (gpl)