160126
-
memory hangs, poor performance, unruly processes
- restarted server after +240 days of uptime
-
nginx interfered preventing apache from starting
- brought all sites down
-
manually stopped nginx service / started apache
- sites back up
-
attempted database backup
-
mysql user ‘jerm’ lost ability to
mysqldump -
attempted password change through ‘phpmyadmin’ (PMA from here out)
- found user ‘admin@localhost’ lost privileges in PMA
- PMA throwing error of unknown column ‘plugin’
- thinking it was an accidental add in previous db repair, dropped plugin table / flaw
-
considering mysql_upgrade but concerned problems may escalate
-
user ‘harrison’ password issue
- attempted manual entry via PMA
- where ‘admin’ had lost rights in PMA became apparent
- attempted manual entry via PMA
-
added root account no pass sketchy
-
could not get site to come up without running with skip-grant-tables in
/etc/my.cnf
need to
- backup server (complete)
- backup database (complete)
- verify size / accuracy
- repair errors in PMA
- remove nginx
- confirm safe before removal
160127
- attempted dump of all databases in
~/dump - resulted in ~51mb file, not convinced that’s all of them, doesnt seem large enough
- attempt PMA login with different user
- looking for previous backup of sql to extract ‘user’ table
- doesnt exist. fuck.
/var/log/mysqld.log
- Version: ‘5.5.43’ socket: ‘/var/lib/mysql/mysql.sock’ port: 3306 MySQL Community Server (GPL) by Remi
- mysql: 5.5.43-1.el5.remi (plesk components list)
- many errors reporting: ‘performance_schema’ has wrong structure im hoping this is the reason there are communication errors between EE and SQL server
attempting mysql_upgrade
- made complete backup of server ~1.3gb / executed through ‘plesk/backup manager’
- saved to: /var/lib/psa/dumps/domains/ocmdhotels.com
- make complete backup of sql
- confirm backup (50 mb correct) appears so, difficult to verify
- ran
$ tail -f /var/log/mysqldto monitor issues while running the upgrade - upgrade ran fine but wouldnt print log file
- restarted mysql but log file was generating errors about user ‘harrison’ having incorrect password
- retrieved password from
~/httpdocs/baltave/expressionengine/config/database.php - updated pw manually through
mysql>running asroot- mysql -uroot;
- use mysql;
- mysql> UPDATE user SET Password=password(‘xyz’) WHERE User=‘harrison’;
- mysql> FLUSH PRIVILEGES;
- mysql> exit;
- restart again
- sudo
/etc/init.d/mysqld restart
- sudo
- tailed /var/log/mysqld.log / appears all is well slight DNS issue
need to
- restore
rootpassword or remove user…- restored password Q-$
- remove nginx
- yum remove nginx / successful
- create “clean” backup
- all but petty ‘events’ table (mysql bug/ of no consequence)