Upgrading MySQL from MySQL 5.0.95 to 5.5

[Shawnosaurus]

I just upgraded Plesk from 12.0.18 to 12.5. After which, I tried from the control panel to upgrade MySQL from 5.0 to 5.5. Now I can't login to Plesk control panel and am getting the following error:

ERROR: PleskDBException: Unable to connect to database: mysql_connect(): No such file or directory /var/lib/mysql/mysql.sock (Error code: 2002). Please check that database server is started and accessible. (Abstract.php:69)

Any ideas on how to get this fixed? Any help would be appreciated. I'm far from a Plesk / Linux expert.

 

[IgorG]

Check these articles

https://kb.odin.com/en/121057
https://kb.odin.com/en/124151

 

[Shawnosaurus]

Thanks but these aren't the exact errors. Logging in to Power Panel works and says I have a lot of space left. And I'm not sure how to tell if there's a hung process. I checked my psa_service.log and it shows:

"Starting mysqld: [FAILED]"

 

[IgorG]

"Starting mysqld: [FAILED]"

So, your mysql server is not started. You need to check mysql log, try to start it manually, check available free disk space with df -h command, etc.
I'd suggest you hire experienced Linux administrator or contact Plesk Support Team if you are not familiar in Linux command line interface.

 

[Shawnosaurus]

Yes, I know the server hasn't started. I've tried to manually start it via SSH without success.

Disk space shows the following:

Filesystem Size Used Avail Use% Mounted on
/dev/vzfs 29G 8.1G 21G 29% /
/dev/simfs 29G 8.1G 21G 29% /tmp
/dev/simfs 29G 8.1G 21G 29% /var/tmp
none 973M 16K 973M 1% /dev
/dev/vzfs 29G 8.1G 21G 29% /var/named/run-root/var/run/dbus

And end of mysqld.log is as follows:

160128 01:20:58 mysqld_safe mysqld from pid file /var/run/mysqld/mysqld.pid ended
160128 01:54:53 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
160128 1:54:53 [Note] /usr/libexec/mysqld (mysqld 5.5.46) starting as process 29469 ...
160128 1:54:53 [Note] Plugin 'FEDERATED' is disabled.
160128 1:54:53 InnoDB: The InnoDB memory heap is disabled
160128 1:54:53 InnoDB: Mutexes and rw_locks use GCC atomic builtins
160128 1:54:53 InnoDB: Compressed tables use zlib 1.2.3
160128 1:54:53 InnoDB: Using Linux native AIO
160128 1:54:53 InnoDB: Initializing buffer pool, size = 128.0M
160128 1:54:53 InnoDB: Completed initialization of buffer pool
160128 1:54:53 InnoDB: highest supported file format is Barracuda.
160128 1:54:53 InnoDB: Waiting for the background threads to start
160128 1:54:54 InnoDB: 5.5.46 started; log sequence number 612003931
160128 1:54:54 [ERROR] /usr/libexec/mysqld: unknown variable 'set-variable=local-infile=0'
160128 1:54:54 [ERROR] Aborting

160128 1:54:54 InnoDB: Starting shutdown...
160128 1:54:55 InnoDB: Shutdown completed; log sequence number 612003931
160128 1:54:55 [Note] /usr/libexec/mysqld: Shutdown complete

 

[IgorG]

What about

# /etc/init.d/mysqld start
# /etc/init.d/mysqld status

?

 

[Shawnosaurus]

I fixed it.

The errors were caused by the variables in MySQL config file located at /etc/my.cnf.

This is what I ended up commenting out:

• # variable 'set-variable=innodb_buffer_pool_size=2M'
• # variable 'set-variable=innodb_thread_concurrency=2'
• # variable 'set-variable=innodb_additional_mem_pool_size=500K'

After that, I was able to start the database server. That being said, I have no idea what those variables do or mean. Anyone know if I need them, or should replace with something else?

 

[gennolo]

I had the same problem upgrading time ago.
Those variables were parts of some mysql "tuning" made on the server on the old 5.0 version.

In mysql 5.5 version there is no need to specify the "variable set-variable=" header anymore.
So if you need you can paste again those line/instructions with "name=value" syntax :

• innodb_buffer_pool_size=2M
• innodb_thread_concurrency=2
• innodb_additional_mem_pool_size=500K

and then the server should start properly with the specified custom settings.

 

[Shawnosaurus]

Sweet, thanks!