Optimize MySQL table_open_cache

·
optimize mysql table_open_cache

Optimize MySQL table_open_cache

The number of open tables for all threads. Increasing this value increases the number of file descriptors that mysqld requires. You can check whether you need to increase the table cache by checking the Opened_tables status variable.  If the value of Opened_tables is large and you do not use FLUSH TABLES often (which just forces all tables to be closed and reopened), then you should increase the value of the table_open_cache variable.  To optimize MySQL table_open_cache you need to consider lot of factors. However, these are the general procedure to decide whether this variable should be tuned.

1. Find current value of open_tables and opened_tables

mysql> show global status  like 'open%';
+--------------------------+---------+
| Variable_name            | Value   |
+--------------------------+---------+
| Open_files               | 1583    |
| Open_streams             | 0       |
| Open_table_definitions   | 1400    |
| Open_tables              | 2000    |
| Opened_files             | 2619222 |
| Opened_table_definitions | 110583  |
| Opened_tables            | 482099  |
+--------------------------+---------+
7 rows in set (0.00 sec)

2. Find out Table cache hit rate

Table cache hit rate = table_open_cache*100/Opened_tables.  
                 = 2000*100/482099
                 = 0.41%

In general it should be more than 50%. So you need to increase value of table_open_cache, though there are lots of reasons to have a high value of Opened_tables. Like FLUSH TABLES will close all open tables and reopen it which significantly increases Opened_tables value.

At this stage you are almost sure table_open_cache system variable is not tuned properly. Now you have to optimize MySQL table_open_cache and  find out perfect value for this.  To find tuned value of  table_open_cache value follow the steps:

1. Find out total tables of your database

mysql> SELECT COUNT(*) FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_TYPE='BASE TABLE';
+----------+
| COUNT(*) |
+----------+
|     2020 |
+----------+
1 row in set (0.05 sec)
  1. Find threads currently connected to your database.
mysql> show global status like '%Threads_connected%';
+-------------------+-------+
| Variable_name | Value |
+-------------------+-------+
| Threads_connected | 39 |
+-------------------+-------+
1 row in set (0.00 sec)

It would be best if you take threads connected at busiest time of your database or take several times at different time and make an average.

3. Calculate  the tune value of table_open_cache and set it

Table_open_cache = total_tables*Threads_connected
         = 2020*39
         = 78780

As all the threads (user) are not generally access all tables. I think you should set 50% of the value calculated.  Because too big value of this variable has some other side effects.  So the formula becomes

Table_open_cache = total_tables*Threads_connected*.50

4. Along with table_open_cache you should also tune open_files_limit system variable.

In general it is 2x of  table_open_cache.

open_files_limit= Table_open_cache*2

open_files_limit is not a dynamic variable. So you should set it in my.cnf file and restart MySQL.

*Make sure that your operating system can cope with the number of open file descriptors required by the table_open_cache setting.

  1.  Go to your Mysql configuration file (in linux it is /etc/my.cnf) and set the table_open_cache and open_files_limit
vi /etc/my.cnf
table_open_cache=39390
open_files_limit=78780

6. Restart the MySQL ( In Linux it is like)

/etc/init.d/mysqld restart
or
/etc/init.d/mysql restart

The database I have taken has 2020 tables so the value of table_open_cache is little big. For your case it may be significantly small.



Leave a Reply