Documentation Home
MySQL 8.0 Reference Manual
Related Documentation Download this Manual
PDF (US Ltr) - 33.6Mb
PDF (A4) - 33.6Mb
PDF (RPM) - 31.7Mb
HTML Download (TGZ) - 8.1Mb
HTML Download (Zip) - 8.1Mb
HTML Download (RPM) - 6.9Mb
Man Pages (TGZ) - 145.6Kb
Man Pages (Zip) - 206.4Kb
Info (Gzip) - 3.1Mb
Info (Zip) - 3.1Mb

Pre-General Availability Draft: 2017-10-20 How MySQL Uses Memory

MySQL allocates buffers and caches to improve performance of database operations. The default configuration is designed to permit a MySQL server to start on a virtual machine that has approximately 512MB of RAM. You can improve MySQL performance by increasing the values of certain cache and buffer-related system variables. You can also modify the default configuration to run MySQL on systems with limited memory.

The following list describes some of the ways that MySQL uses memory. Where applicable, relevant system variables are referenced. Some items are storage engine or feature specific.

  • The InnoDB buffer pool is a memory area that holds cached InnoDB data for tables, indexes, and other auxiliary buffers. For efficiency of high-volume read operations, the buffer pool is divided into pages that can potentially hold multiple rows. For efficiency of cache management, the buffer pool is implemented as a linked list of pages; data that is rarely used is aged out of the cache, using a variation of the LRU algorithm. For more information, see Section, “The InnoDB Buffer Pool”.

    The size of the buffer pool is important for system performance:

    • InnoDB allocates memory for the entire buffer pool at server startup, using malloc() operations. The innodb_buffer_pool_size system variable defines the buffer pool size. Typically, a recommended innodb_buffer_pool_size value is 50 to 75 percent of system memory. innodb_buffer_pool_size can be configured dynamically, while the server is running. For more information, see Section, “Configuring InnoDB Buffer Pool Size”.

    • On systems with a large amount of memory, you can improve concurrency by dividing the buffer pool into multiple buffer pool instances. The innodb_buffer_pool_instances system variable defines the number of buffer pool instances.

    • A buffer pool that is too small may cause excessive churning as pages are flushed from the buffer pool only to be required again a short time later.

    • A buffer pool that is too large may cause swapping due to competition for memory.

  • The storage engine interface enables the optimizer to provide information about the size of the record buffer to be used for scans that the optimizer estimates will read multiple rows. The buffer size can vary based on the size of the estimate. InnoDB uses this variable-size buffering capability to take advantage of row prefetching, and to reduce the overhead of latching and B-tree navigation.

  • All threads share the MyISAM key buffer. The key_buffer_size system variable determines its size.

    For each MyISAM table the server opens, the index file is opened once; the data file is opened once for each concurrently running thread that accesses the table. For each concurrent thread, a table structure, column structures for each column, and a buffer of size 3 * N are allocated (where N is the maximum row length, not counting BLOB columns). A BLOB column requires five to eight bytes plus the length of the BLOB data. The MyISAM storage engine maintains one extra row buffer for internal use.

  • The myisam_use_mmap system variable can be set to 1 to enable memory-mapping for all MyISAM tables.

  • If an internal in-memory temporary table becomes too large (as determined using the tmp_table_size and max_heap_table_size system variables), MySQL automatically converts the table from in-memory to on-disk format. On-disk temporary tables use the storage engine defined by the internal_tmp_disk_storage_engine system variable. You can increase the permissible temporary table size as described in Section 8.4.4, “Internal Temporary Table Use in MySQL”.

    For MEMORY tables explicitly created with CREATE TABLE, only the max_heap_table_size system variable determines how large the table is permitted to grow and there is no conversion to on-disk format.

  • The MySQL Performance Schema is a feature for monitoring MySQL server execution at a low level. The Performance Schema dynamically allocates memory incrementally, scaling its memory use to actual server load, instead of allocating required memory during server startup. Once memory is allocated, it is not freed until the server is restarted. For more information, see Section 26.16, “The Performance Schema Memory-Allocation Model”.

  • Each thread that the server uses to manage client connections requires some thread-specific space. The following list indicates these and which system variables control their size:

    The connection buffer and result buffer each begin with a size equal to net_buffer_length bytes, but are dynamically enlarged up to max_allowed_packet bytes as needed. The result buffer shrinks to net_buffer_length bytes after each SQL statement. While a statement is running, a copy of the current statement string is also allocated.

    Each connection thread uses memory for computing statement digests. The server allocates max_digest_length bytes per session. See Section 26.9, “Performance Schema Statement Digests and Sampling”.

  • All threads share the same base memory.

  • When a thread is no longer needed, the memory allocated to it is released and returned to the system unless the thread goes back into the thread cache. In that case, the memory remains allocated.

  • Each request that performs a sequential scan of a table allocates a read buffer. The read_buffer_size system variable determines the buffer size.

  • When reading rows in an arbitrary sequence (for example, following a sort), a random-read buffer may be allocated to avoid disk seeks. The read_rnd_buffer_size system variable determines the buffer size.

  • All joins are executed in a single pass, and most joins can be done without even using a temporary table. Most temporary tables are memory-based hash tables. Temporary tables with a large row length (calculated as the sum of all column lengths) or that contain BLOB columns are stored on disk.

  • Most requests that perform a sort allocate a sort buffer and zero to two temporary files depending on the result set size. See Section B.5.3.5, “Where MySQL Stores Temporary Files”.

  • Almost all parsing and calculating is done in thread-local and reusable memory pools. No memory overhead is needed for small items, thus avoiding the normal slow memory allocation and freeing. Memory is allocated only for unexpectedly large strings.

  • For each table having BLOB columns, a buffer is enlarged dynamically to read in larger BLOB values. If you scan a table, the buffer grows as large as the largest BLOB value.

  • MySQL requires memory and descriptors for the table cache. Handler structures for all in-use tables are saved in the table cache and managed as First In, First Out (FIFO). The table_open_cache system variable defines the initial table cache size; see Section, “How MySQL Opens and Closes Tables”.

    MySQL also requires memory for the table definition cache. The table_definition_cache system variable defines the number of table definitions that can be stored in the table definition cache. If you use a large number of tables, you can create a large table definition cache to speed up the opening of tables. The table definition cache takes less space and does not use file descriptors, unlike the table cache.

  • A FLUSH TABLES statement or mysqladmin flush-tables command closes all tables that are not in use at once and marks all in-use tables to be closed when the currently executing thread finishes. This effectively frees most in-use memory. FLUSH TABLES does not return until all tables have been closed.

  • The server caches information in memory as a result of GRANT, CREATE USER, CREATE SERVER, and INSTALL PLUGIN statements. This memory is not released by the corresponding REVOKE, DROP USER, DROP SERVER, and UNINSTALL PLUGIN statements, so for a server that executes many instances of the statements that cause caching, there will be an increase in memory use. This cached memory can be freed with FLUSH PRIVILEGES.

ps and other system status programs may report that mysqld uses a lot of memory. This may be caused by thread stacks on different memory addresses. For example, the Solaris version of ps counts the unused memory between stacks as used memory. To verify this, check available swap with swap -s. We test mysqld with several memory-leakage detectors (both commercial and Open Source), so there should be no memory leaks.

Monitoring MySQL Memory Usage

The following example demonstrates how to use Performance Schema and sys schema to monitor MySQL memory usage.

Most Performance Schema memory instrumentation is disabled by default. Instruments can be enabled by updating the ENABLED column of the Performance Schema setup_instruments table. Memory instruments have names in the form of memory/code_area/instrument_name, where code_area is a value such as sql or innodb, and instrument_name is the instrument detail.

  1. To view available MySQL memory instruments, query the Performance Schema setup_instruments table. The following query returns hundreds of memory instruments for all code areas.

    mysql> SELECT * FROM performance_schema.setup_instruments
           WHERE NAME LIKE '%memory%';

    You can narrow results by specifying a code area. For example, you can limit results to InnoDB memory instruments by specifying innodb as the code area.

    mysql> SELECT * FROM performance_schema.setup_instruments
           WHERE NAME LIKE '%memory/innodb%';
    | NAME                                      | ENABLED | TIMED |
    | memory/innodb/adaptive hash index         | NO      | NO    |
    | memory/innodb/buf_buf_pool                | NO      | NO    |
    | memory/innodb/dict_stats_bg_recalc_pool_t | NO      | NO    |
    | memory/innodb/dict_stats_index_map_t      | NO      | NO    |
    | memory/innodb/dict_stats_n_diff_on_level  | NO      | NO    |
    | memory/innodb/other                       | NO      | NO    |
    | memory/innodb/row_log_buf                 | NO      | NO    |
    | memory/innodb/row_merge_sort              | NO      | NO    |
    | memory/innodb/std                         | NO      | NO    |
    | memory/innodb/trx_sys_t::rw_trx_ids       | NO      | NO    |

    Depending on your MySQL installation, code areas may include performance_schema, sql, client, innodb, myisam, csv, memory, blackhole, archive, partition, and others.

  2. To enable memory instruments, add a performance-schema-instrument rule to your MySQL configuration file. For example, to enable all memory instruments, add this rule to your configuration file and restart the server:


    Enabling memory instruments at startup ensures that memory allocations that occur at startup are counted.

    After restarting the server, the ENABLED column of the Performance Schema setup_instruments table should report YES for memory instruments that you enabled. The TIMED column in the setup_instruments table is ignored for memory instruments because memory operations are not timed.

    mysql> SELECT * FROM performance_schema.setup_instruments
           WHERE NAME LIKE '%memory/innodb%';
    | NAME                                      | ENABLED | TIMED |
    | memory/innodb/adaptive hash index         | NO      | NO    |
    | memory/innodb/buf_buf_pool                | NO      | NO    |
    | memory/innodb/dict_stats_bg_recalc_pool_t | NO      | NO    |
    | memory/innodb/dict_stats_index_map_t      | NO      | NO    |
    | memory/innodb/dict_stats_n_diff_on_level  | NO      | NO    |
    | memory/innodb/other                       | NO      | NO    |
    | memory/innodb/row_log_buf                 | NO      | NO    |
    | memory/innodb/row_merge_sort              | NO      | NO    |
    | memory/innodb/std                         | NO      | NO    |
    | memory/innodb/trx_sys_t::rw_trx_ids       | NO      | NO    |
  3. Query memory instrument data. In this example, memory instrument data is queried in the Performance Schema memory_summary_global_by_event_name table, which summarizes data by EVENT_NAME. The EVENT_NAME is the name of the instrument.

    The following query returns memory data for the InnoDB buffer pool. For column descriptions, see Section, “Memory Summary Tables”.

    mysql> SELECT * FROM performance_schema.memory_summary_global_by_event_name
           WHERE EVENT_NAME LIKE 'memory/innodb/buf_buf_pool'\G
                      EVENT_NAME: memory/innodb/buf_buf_pool
                     COUNT_ALLOC: 1
                      COUNT_FREE: 0
       SUM_NUMBER_OF_BYTES_ALLOC: 137428992
                  LOW_COUNT_USED: 0
              CURRENT_COUNT_USED: 1
                 HIGH_COUNT_USED: 1
       HIGH_NUMBER_OF_BYTES_USED: 137428992

    The same underlying data can be queried using the sys schema memory_global_by_current_bytes table, which shows current memory usage within the server globally, broken down by allocation type.

    mysql> SELECT * FROM sys.memory_global_by_current_bytes
           WHERE event_name LIKE 'memory/innodb/buf_buf_pool'\G
    *************************** 1. row ***************************
           event_name: memory/innodb/buf_buf_pool
        current_count: 1
        current_alloc: 131.06 MiB
    current_avg_alloc: 131.06 MiB
           high_count: 1
           high_alloc: 131.06 MiB
       high_avg_alloc: 131.06 MiB

    This sys schema query aggregates currently allocated memory (current_alloc) by code area:

    mysql> SELECT SUBSTRING_INDEX(event_name,'/',2) AS
           code_area, sys.format_bytes(SUM(current_alloc))
           AS current_alloc
           FROM sys.x$memory_global_by_current_bytes
           GROUP BY SUBSTRING_INDEX(event_name,'/',2)
           ORDER BY SUM(current_alloc) DESC;
    | code_area                 | current_alloc |
    | memory/innodb             | 843.24 MiB    |
    | memory/performance_schema | 81.29 MiB     |
    | memory/mysys              | 8.20 MiB      |
    | memory/sql                | 2.47 MiB      |
    | memory/memory             | 174.01 KiB    |
    | memory/myisam             | 46.53 KiB     |
    | memory/blackhole          | 512 bytes     |
    | memory/federated          | 512 bytes     |
    | memory/csv                | 512 bytes     |
    | memory/vio                | 496 bytes     |

    For more information about sys schema, see Chapter 27, MySQL sys Schema.

User Comments
  Posted by sheila yao on October 2, 2007
I got this formula from mysql error log complaining it doesn't have enough memory to start mysqld:
key_buffer_size + (read_buffer_size + sort_buffer_size) * max_connections = K bytes of memory

I hope this document could be straight forward by providing a formula to calculate the memory usage for mysqld.


  Posted by Guy Baconniere on May 6, 2009
I use the following SQL query to guess MySQL memory usage
of MySQL unfortunately innodb_* and thread_stack are not
part of MySQL system variables so you need to fill them

Best Regards,
Guy Baconniere


SHOW VARIABLES LIKE 'innodb_buffer_pool_size';
SHOW VARIABLES LIKE 'innodb_additional_mem_pool_size';
SHOW VARIABLES LIKE 'innodb_log_buffer_size';
SHOW VARIABLES LIKE 'thread_stack';
SET @kilo_bytes = 1024;
SET @mega_bytes = @kilo_bytes * 1024;
SET @giga_bytes = @mega_bytes * 1024;
SET @innodb_buffer_pool_size = 2 * @giga_bytes;
SET @innodb_additional_mem_pool_size = 16 * @mega_bytes;
SET @innodb_log_buffer_size = 8 * @mega_bytes;
SET @thread_stack = 192 * @kilo_bytes;
( @@key_buffer_size + @@query_cache_size + @@tmp_table_size
+ @innodb_buffer_pool_size + @innodb_additional_mem_pool_size
+ @innodb_log_buffer_size
+ @@max_connections * (
@@read_buffer_size + @@read_rnd_buffer_size + @@sort_buffer_size
+ @@join_buffer_size + @@binlog_cache_size + @thread_stack
) ) / @giga_bytes AS MAX_MEMORY_GB;
| 3.7002 |
1 row in set (0.00 sec)
  Posted by Lig Isler-Turmelle on March 19, 2010
I disagree with how the previous comment handles the tmp_table_size value. They treat it as a single allocation on the global scope when for memory consumption purposes it is more in line with a per thread buffer.

A single connection/query can use a single or multiple temporary tables in the duration of its processing. The connections do not use a single temporary table "area" reserved just for that purpose.

If you are going to use a formula for memory consumption, the tmp_table-size should be located with the other per thread buffers - not in the single allocation listing.
  Posted by Christopher Schultz on January 30, 2013
Based upon the previous two comments, I re-worked the max-mem-usage query and made it work (there weren't enough @'s on some variables and @giga_bytes isn't defined). This returns a non-null value on MySQL 5.5.29:

SELECT ( @@key_buffer_size
+ @@query_cache_size
+ @@innodb_buffer_pool_size
+ @@innodb_additional_mem_pool_size
+ @@innodb_log_buffer_size
+ @@max_connections * ( @@read_buffer_size
+ @@read_rnd_buffer_size
+ @@sort_buffer_size
+ @@join_buffer_size
+ @@binlog_cache_size
+ @@thread_stack
+ @@tmp_table_size )
) / (1024 * 1024 * 1024) AS MAX_MEMORY_GB;
  Posted by Eduardo Franceschi on December 26, 2013
I wrote a shell script based on above examples. I've added a min/max memory suggestions also. The minimum memory is estimated using the Max_used_connections variable from SHOW STATUS.


mysql -e "show variables; show status" | awk '
MAX_CONN = VAR["max_connections"]
MAX_USED_CONN = VAR["Max_used_connections"]
BASE_MEM=VAR["key_buffer_size"] + VAR["query_cache_size"] + VAR["innodb_buffer_pool_size"] + VAR["innodb_additional_mem_pool_size"] + VAR["innodb_log_buffer_size"]
MEM_PER_CONN=VAR["read_buffer_size"] + VAR["read_rnd_buffer_size"] + VAR["sort_buffer_size"] + VAR["join_buffer_size"] + VAR["binlog_cache_size"] + VAR["thread_stack"] + VAR["tmp_table_size"]

printf "+------------------------------------------+--------------------+\n"
printf "| %40s | %15.3f MB |\n", "key_buffer_size", VAR["key_buffer_size"]/1048576
printf "| %40s | %15.3f MB |\n", "query_cache_size", VAR["query_cache_size"]/1048576
printf "| %40s | %15.3f MB |\n", "innodb_buffer_pool_size", VAR["innodb_buffer_pool_size"]/1048576
printf "| %40s | %15.3f MB |\n", "innodb_additional_mem_pool_size", VAR["innodb_additional_mem_pool_size"]/1048576
printf "| %40s | %15.3f MB |\n", "innodb_log_buffer_size", VAR["innodb_log_buffer_size"]/1048576
printf "+------------------------------------------+--------------------+\n"
printf "| %40s | %15.3f MB |\n", "BASE MEMORY", BASE_MEM/1048576
printf "+------------------------------------------+--------------------+\n"
printf "| %40s | %15.3f MB |\n", "sort_buffer_size", VAR["sort_buffer_size"]/1048576
printf "| %40s | %15.3f MB |\n", "read_buffer_size", VAR["read_buffer_size"]/1048576
printf "| %40s | %15.3f MB |\n", "read_rnd_buffer_size", VAR["read_rnd_buffer_size"]/1048576
printf "| %40s | %15.3f MB |\n", "join_buffer_size", VAR["join_buffer_size"]/1048576
printf "| %40s | %15.3f MB |\n", "thread_stack", VAR["thread_stack"]/1048576
printf "| %40s | %15.3f MB |\n", "binlog_cache_size", VAR["binlog_cache_size"]/1048576
printf "| %40s | %15.3f MB |\n", "tmp_table_size", VAR["tmp_table_size"]/1048576
printf "+------------------------------------------+--------------------+\n"
printf "| %40s | %15.3f MB |\n", "MEMORY PER CONNECTION", MEM_PER_CONN/1048576
printf "+------------------------------------------+--------------------+\n"
printf "| %40s | %18d |\n", "Max_used_connections", MAX_USED_CONN
printf "| %40s | %18d |\n", "max_connections", MAX_CONN
printf "+------------------------------------------+--------------------+\n"
printf "| %40s | %15.3f MB |\n", "TOTAL (MIN)", MEM_TOTAL_MIN/1048576
printf "| %40s | %15.3f MB |\n", "TOTAL (MAX)", MEM_TOTAL_MAX/1048576
printf "+------------------------------------------+--------------------+\n"

  Posted by Shane Bester on September 2, 2014
folks above using their own formulas, don't forget about memory used by performance schema (SHOW ENGINE PERFORMANCE_SCHEMA STATUS) and global variable innodb_ft_total_cache_size.
  Posted by kedar vaijanapurkar on February 27, 2015
Long back I had created a stored procedure which resides in mysql and should be as easy as giving a call to the procedure to estimate the memory usage based on Global and Per Thread Variables.

mysql> call my_memory();
| Parameter | Value (M) |
| Global Buffers | 531 M |
| Per Thread | 1.890625 M |
| Maximum Connections | 160 |
| Total Memory Usage | 833.5 M |
| + Per Heap Table | 16 M |
| + Per Temp Table | 26 M |


(Do read the cursor declaration note in the post which may cause an error due to bug)


CREATE PROCEDURE `my_memory` ()


#Variables for storing calculations

#Cursor for Global Variables

#### For < MySQL 5.1

#### For MySQL 5.1+
#### Ref:




IF done=1 THEN
LEAVE mylp;
IF var in ('key_buffer_size','innodb_buffer_pool_size','innodb_additional_mem_pool_size','innodb_log_buffer_size','query_cache_size') THEN
#Summing Up Global Memory Usage
ELSEIF var in ('read_buffer_size','read_rnd_buffer_size','sort_buffer_size','join_buffer_size','thread_stack','max_allowed_packet','net_buffer_length') THEN
#Summing Up Per Thread Memory Variables
ELSEIF var in ('max_connections') THEN
#Maximum allowed connections
ELSEIF var in ('max_heap_table_size') THEN
#Size of Max Heap tables created
#Size of possible Temporary Table = Maximum of tmp_table_size / max_heap_table_size.
ELSEIF var in ('tmp_table_size','max_heap_table_size') THEN

select "Global Buffers" as "Parameter",CONCAT(GLOBAL_SUM/(1024*1024),' M') as "Value" union
select "Per Thread",CONCAT(PER_THREAD_SUM/(1024*1024),' M') union
select "Maximum Connections",MAX_CONN union
select "Total Memory Usage",CONCAT((GLOBAL_SUM + (MAX_CONN * PER_THREAD_SUM))/(1024*1024),' M') union
select "+ Per Heap Table",CONCAT(HEAP_TABLE / (1024*1024),' M') union
select "+ Per Temp Table",CONCAT(TEMP_TABLE / (1024*1024),' M') ;

END $$
Sign Up Login You must be logged in to post a comment.