Issues resolving a requirement for Innodb (MySQL fails to restart, innodb installed)

stardotstar

Well-Known Member
Sep 14, 2009
68
0
56
Hi all,

I apparently have the innodb engine available on my whm server:

Code:
[email protected] [/etc/init.d]# mysqladmin variables | grep have_innodb
| have_innodb                             | YES
I have an application that is reporting no assignment of the innodb memory it wants (half the physical ram - I have 22G)

The ActiveCampaign email marketing solution has started complaining (I don't think it always did this!)

Your MySQL server is configured to use less than 64 megabytes of physical RAM, which may cause you to experience slow performance.

Please consider raising this amount in your server's my.cnf file, or ask your system administrator or ISP to do so for you. The variable which must be increased is innodb_buffer_pool_size, and ideally should be half of all the physical RAM your server has installed. (Example: innodb_buffer_pool_size=1G)
I find no skip_innodb in my my.cnf

Code:
[email protected] [/etc/init.d]# cat /etc/my.cnf
[mysqld]
local-infile=0
tmpdir = /dev/shm
#safe-show-database
max_connections = 650
max_user_connections=50
key_buffer_size = 512M
myisam_sort_buffer_size = 64M
join_buffer_size = 1M
read_buffer_size = 1M
sort_buffer_size = 2M
table_cache = 4000
thread_cache_size = 384
wait_timeout = 40
connect_timeout = 20
tmp_table_size = 64M
max_heap_table_size = 64M
max_allowed_packet = 64M
net_buffer_length = 16384
max_connect_errors = 10
thread_concurrency = 8
concurrent_insert = 2
#table_lock_wait_timeout only for mysql5
table_lock_wait_timeout = 40
read_rnd_buffer_size = 786432
bulk_insert_buffer_size = 8M
query_cache_limit = 7M
query_cache_size = 64M
query_cache_type = 1
query_prealloc_size = 262144
query_alloc_block_size = 65536
transaction_alloc_block_size = 8192
transaction_prealloc_size = 4096
max_write_lock_count = 16
skip-name-resolve
skip-locking

[mysqld_safe]
open_files_limit = 8192

[mysqldump]
quick
max_allowed_packet = 16M

[myisamchk]
key_buffer = 384M
sort_buffer = 384M
read_buffer = 256M
write_buffer = 256M
and therefore I tried what was suggested on the ActiveCampaign server optimisation pages after finding that my cpanel/whm server should support use of the innodb engine just fine...

but when I add:
Code:
innodb_flush_log_at_trx_commit=0
innodb_bugger_pool_size=8G
to the my.cnf MySQL won't restart complaining of a lock file:

Code:
[email protected] [/etc/init.d]# ./mysql status
MySQL running (2973)                                       [  OK  ]
[email protected] [/etc/init.d]# ./mysql
Usage: ./mysql  {start|stop|restart|reload|force-reload|status}  [ MySQL server options ]
[email protected] [/etc/init.d]# ./mysql reload
Reloading service MySQL                                    [  OK  ]
[email protected] [/etc/init.d]# ./mysql restart
Shutting down MySQL...                                     [  OK  ]
Starting MySQL.Manager of pid-file quit without updating fi[FAILED]
[email protected] [/etc/init.d]# ./mysql status
MySQL is not running, but lock exists                      [FAILED]
[email protected] [/etc/init.d]# ./mysql start

The error seems to be:
Code:
Starting MySQL.Manager of pid-file quit without updating fi
When I hash out the lines for innodb it starts fine...

Any ideas on this guys; I may have been trying an impossible thing but don't know it.

TIA
WIll
 

cPanelTristan

Quality Assurance Analyst
Staff member
Oct 2, 2010
7,607
40
248
somewhere over the rainbow
cPanel Access Level
Root Administrator
Please note that when you have an error on starting MySQL, the best place to find what is causing the failure would be the error log for MySQL at /var/lib/mysql/hostname.err location.

You can always do:

Code:
cd /var/lib/mysql
ls -lah *.err
This would output then the name of the error log for MySQL. At that point, you can tail it while trying to restart MySQL in another window to see the results for the error:

Code:
tail -f hostname.err
 

stardotstar

Well-Known Member
Sep 14, 2009
68
0
56
Thank you for that additional help. I find that setting the buffer size to 8G has allowed the MySQL Service to start normally but made Round Cube report Service Unavailable and triggered an update for Active Campaign that I was not expecting - needing the serialnumber for it took me to my webmail which was broken so I rolled back the config again.
Thank you for your help that is most appreciated!
WIll