Optimization is a complex task because ultimately it requires understanding of the whole system. While it may be possible to perform some local optimizations with small knowledge of your system or application, the more optimal you want your system to become the more you will have to know about it.
This chapter tries to explain and give some examples of different ways to optimize MySQL. Remember, however, that there are always some (increasingly harder) additional ways to make the system even faster.
The most important factor in making a system fast is the basic design. You also need to know what kinds of things your system will be doing, and what your bottlenecks are.
The most common bottlenecks are:
When using the
MyISAM storage engine, MySQL uses extremely fast table
locking that allows multiple readers or a single writer. The biggest problem
with this storage engine occurs when you have a mix of a steady stream
of updates and slow selects on the same table. If this is a problem with
some tables, you can use another table type for these. See section 15 MySQL Storage Engines and Table Types.
MySQL can work with both transactional and non-transactional tables. To be able to work smoothly with non-transactional tables (which can't roll back if something goes wrong), MySQL has the following rules:
The implication of these rules is that you should not use MySQL to check column content. Instead, you should check values in the application before storing them in the database.
For more information about this, see section 1.8.6 How MySQL Deals with Constraints and section 14.1.4
Because all SQL servers implement different parts of SQL, it takes work to write portable SQL applications. It is very easy to achieve portability for very simple selects and inserts, but becomes more difficiult the more capabilities you require. If you want an application that is fast with many database systems, it becomes even harder!
To make a complex application portable, you need to choose a number of SQL servers that it should work with.
You can use the MySQL
to find functions,
types, and limits you can use with a selection of database
crash-me tests far from everything possible, but it
is still comprehensive with about 450 things tested.
An example of the type of information
crash-me can provide is that
you shouldn't have column names longer than 18 characters
if you want to be able to use Informix or DB2.
For information, visit http://www.mysql.com/information/crash-me.php.
Both the MySQL benchmarks and
crash-me programs are very database
independent. By taking a look at how we have written them, you can get
a feeling for what you have to do to make your own applications database
independent. The programs can be found in the `sql-bench' directory
in the MySQL source distribution. They are written in Perl and use the DBI
database interface. Use of DBI in itself solves part of the portability
problem because it provides database-independent access methods.
See http://www.mysql.com/information/benchmarks.html for the results from the benchmarks.
As you can see in the results, all database systems have some weak points. That is, they have different design compromises that lead to different behavior.
If you strive for database independence, you need to get a good feeling for each SQL server's bottlenecks. MySQL is very fast in retrieving and updating records, but will have a problem in mixing slow readers and writers on the same table. Oracle, on the other hand, has a big problem when you try to access rows that you have recently updated (until they are flushed to disk). Transactional databases in general are not very good at generating summary tables from log tables, as in this case row locking is almost useless.
To make your application really database-independent, you need to define an easily extendable interface through which you manipulate your data. As C++ is available on most systems, it makes sense to use a C++ class-based interface to the databases.
If you use some feature that is specific to a given database system (such
REPLACE statement, which is specific to MySQL), you should
implement the same feature for other SQL servers by coding an alternative
method. Although the alternative may be slower, it will allow the other
servers to perform the same tasks.
With MySQL, you can use the
/*! */ syntax to add MySQL-specific
keywords to a query. The code inside
/**/ will be treated as a
comment (and ignored) by most other SQL servers.
If high performance is more important than exactness, as in some Web applications, it is possibile to create an application layer that caches all results to give you even higher performance. By letting old results ``expire'' after a while, you can keep the cache reasonably fresh. This provides a method to handle high load spikes, in which case you can dynamically increase the cache and set the expiration timeout higher until things get back to normal.
In this case, the table creation information should contain information of the initial size of the cache and how often the table should normally be refreshed.
An alternative to implementing an application cache is to use the MySQL query cache. By enabling the query cache, the server handles the details of determining whether a query result can be reused. This simplifies your application. See section 5.10 The MySQL Query Cache.
During MySQL initial development, the features of MySQL were made to fit our largest customer. They handle data warehousing for a couple of the biggest retailers in Sweden.
From all stores, we get weekly summaries of all bonus card transactions, and we are expected to provide useful information for the store owners to help them find how their advertisement campaigns are affecting their own customers.
The volume of data is quite huge (about 7 million summary transactions per month), and we have data for 4-10 years that we need to present to the users. We got weekly requests from our customers, who want to get ``instant'' access to new reports from this data.
We solved this by storing all information per month in compressed 'transaction' tables. We have a set of simple macros (script) that generates summary tables grouped by different criteria (product group, customer id, store ...) from the transactional tables. The reports are Web pages that are dynamically generated by a small Perl script that parses a Web page, executes the SQL statements in it, and inserts the results. We would have used PHP or mod_perl instead, but they were not available at that time.
For graphical data, we wrote a simple tool in
C that can process SQL
query results and produce GIF images based on those results. This is also
dynamically executed from the Perl script that parses the Web pages.
In most cases, a new report can be done simply by copying an existing script and modifying the SQL query in it. In some cases, we will need to add more fields to an existing summary table or generate a new one, but this is also quite simple, as we keep all transactions tables on disk. (Currently we have at least 50GB of transactions tables and 200GB of other customer data.)
We also let our customers access the summary tables directly with ODBC so that the advanced users can experiment with the data themselves.
We haven't had any problems handling this with quite modest Sun Ultra SPARCstation hardware (2x200 Mhz). We recently upgraded one of our servers to a 2 CPU 400 Mhz UltraSPARC, and we are now planning to start handling transactions on the product level, which would mean a ten-fold increase of data. We think we can keep up with this by just adding more disk to our systems.
We are also experimenting with Intel-Linux to be able to get more CPU power cheaper. Now that we have the binary portable database format (new in Version 3.23), we will start to use this for some parts of the application.
Our initial feelings are that Linux will perform much better on low-to-medium load and Solaris will perform better when you start to get a high load because of extreme disk IO, but we don't yet have anything conclusive about this. After some discussion with a Linux kernel developer, this might be a side effect of Linux allocating so many resources to the batch job that the interactive performance gets very low. This makes the machine feel very slow and unresponsive while big batches are running. Hopefully this will be better handled in future Linux Kernels.
This section should contain a technical description of the MySQL
benchmark suite (and
crash-me), but that description is not
written yet. Currently, you can get a good idea of the benchmarks by
looking at the code and results in the `sql-bench' directory in any
MySQL source distribution.
This benchmark suite is meant to tell any user what operations a given SQL implementation performs well or poorly.
Note that this benchmark is single-threaded, so it measures the minimum time for the operations performed. We plan to add multi-threaded tests to the benchmark suite in the future.
The following tables show some comparative benchmark results for several database servers when accessed through ODBC on a Windows NT 4.0 machine.
|Reading 2000000 rows by index||Seconds||Seconds|
|Inserting 350768 rows||Seconds||Seconds|
For the preceding tests, MySQL was run with an index cache size of 8MB.
We have gathered some more benchmark results at http://www.mysql.com/information/benchmarks.html.
Note that Oracle is not included because they asked to be removed. All Oracle benchmarks have to be passed by Oracle! We believe that makes Oracle benchmarks very biased because the above benchmarks are supposed to show what a standard installation can do for a single client.
To use the benchmark suite, the following requirements must be satisified:
DBD::DB2modules installed. See section 20.6 MySQL Perl API.
The benchmark suite is located in the `sql-bench' directory of MySQL
To run the benchmark tests, change location into that directory and execute
shell> cd sql-bench shell> perl run-all-tests --server=server_name
server_name is one of supported servers. To get a list of
all options and supported servers, invoke
perl run-all-tests --help.
crash-me script also is located in the `sql-bench' directory.
crash-me tries to determine what features a database supports and
what its capabilities and limitations are by actually running
queries. For example, it determines:
VARCHARcolumn can be
We can find the results from
crash-me for many different database
servers at http://www.mysql.com/information/crash-me.php.
You should definitely benchmark your application and database to find out where the bottlenecks are. By fixing a bottleneck (or by replacing it with a ``dummy module'') you can then easily identify the next bottleneck. Even if the overall performance for your application currently is acceptable, you should at least make a plan for each bottleneck, and decide how to solve it if someday you really need the extra performance.
For an example of portable benchmark programs, look at the MySQL benchmark suite. See section 7.1.4 The MySQL Benchmark Suite. You can take any program from this suite and modify it for your needs. By doing this, you can try different solutions to your problem and test which is really fastest for you.
Another free benchmark suite is the
Open Source Database Benchmark,
available at http://osdb.sourceforge.net/.
It is very common for a problem to occur only when the system is very heavily loaded. We have had many customers who contact us when they have a (tested) system in production and have encountered load problems. In most cases, performance problems turn out to be due to issues of basic database design (for example, table scans are not good at high load) or problems with the operating system or libraries. Most of the time, these problems would be a lot easier to fix if the systems were not already in production.
To avoid problems like this, you should put some effort into benchmarking your whole application under the worst possible load! You can use Super Smack for this. It is available at http://jeremy.zawodny.com/mysql/super-smack/. As the name suggests, it can bring your system to its knees if you ask it, so make sure to use it only on your development systems.
SELECTStatements and Other Queries
First, one factor that affects all statements: The more complex your permission setup is, the more overhead you will have.
Using simpler permissions when you issue
GRANT statements enables
MySQL to reduce permission-checking overhead when clients execute
statements. For example, if you don't grant any table-level or column-level
privileges, the server need not ever check the contents of the
columns_priv tables. Similarly, if you place
no resource limits on any accounts, the server does not have to perform
resource counting. If you have a very high query volume, it may be worth
the time to use a simplified grant structure to reduce permission-checking
If your problem is with some specific MySQL expression or function, you can
BENCHMARK() function from the
mysql client program
to perform a timing test:
mysql> SELECT BENCHMARK(1000000,1+1); +------------------------+ | BENCHMARK(1000000,1+1) | +------------------------+ | 0 | +------------------------+ 1 row in set (0.32 sec)
This result was obtained on a Pentium II 400MHz system. It shows that MySQL can execute 1,000,000 simple addition expressions in 0.32 seconds on that system.
All MySQL functions should be very optimized, but there may be
BENCHMARK(loop_count,expression) is a
great tool to find out if this is a problem with your query.
EXPLAINSyntax (Get Information About a
EXPLAIN tbl_name or EXPLAIN SELECT select_options
EXPLAIN statement can be used either as a synonym for
DESCRIBE or as a way to obtain information about how MySQL will execute
EXPLAIN tbl_namesyntax is synonymous with
SHOW COLUMNS FROM tbl_name.
SELECTstatement with the keyword
EXPLAIN, MySQL explains how it would process the
SELECT, providing information about how tables are joined and in which order.
This section provides information about the second use of
With the help of
EXPLAIN, you can see when you must add indexes
to tables to get a faster
SELECT that uses indexes to find the
You should frequently run
ANALYZE TABLE to update table statistics
such as cardinality of keys which can affect the choices the optimizer
makes. See section 22.214.171.124
ANALYZE TABLE Syntax.
You can also see whether the optimizer joins the tables in an optimal
order. To force the optimizer to use a specific join order for a
SELECT statement, add a
For single-table joins,
EXPLAIN returns a row of information for each
table used in the
SELECT statement. The tables are listed in the output
in the order that MySQL would read them while processing the query.
MySQL resolves all joins using a single-sweep
multi-join method. This means that MySQL reads a row from the first
table, then finds a matching row in the second table, then in the third table
and so on. When all tables are processed, it outputs the selected columns and
backtracks through the table list until a table is found for which there are
more matching rows. The next row is read from this table and the process
continues with the next table.
In MySQL version 4.1, the
EXPLAIN output format was changed to work
better with constructs such as
UNION statements, subqueries, and
derived tables. Most notable is the addition of two new columns:
select_type. You will not see these columns when using servers
older than MySQL 4.1.
Each output row from
EXPLAIN provides information about one table, and
each row consists of the following columns:
SELECTidentifier, the sequential number of this
SELECTwithin the query.
SELECTclause, which can be any of the following:
SELECTstatements in a
SELECTstatements in a
UNION, dependent on outer subquery
SELECTin subquery, dependent on outer subquery
consttables are very fast as they are read only once!
constis used when you compare all parts of a
UNIQUEindex with constants:
SELECT * FROM const_table WHERE primary_key=1; SELECT * FROM const_table WHERE primary_key_part1=1 AND primary_key_part2=2;
consttypes, this is the best possible join type. It is used when all parts of an index are used by the join and the index is a
eq_refcan be used for indexed columns that are compared using the
=operator. The compared item may be a constant or an expression that uses columns from tables that are read before this table. In the following examples,
ref_tablewill be able to use
SELECT * FROM ref_table,other_table WHERE ref_table.key_column=other_table.column; SELECT * FROM ref_table,other_table WHERE ref_table.key_column_part1=other_table.column AND ref_table.key_column_part2=1;
refis used if the join uses only a leftmost prefix of the key, or if the key is not a
UNIQUEindex (in other words, if the join cannot select a single row based on the key value). If the key that is used matches only a few rows, this is a good join type.
refcan be used for indexed columns that are compared using the
=operator. In the following examples,
ref_tablewill be able to use
SELECT * FROM ref_table WHERE key_column=expr; SELECT * FROM ref_table,other_table WHERE ref_table.key_column=other_table.column; SELECT * FROM ref_table,other_table WHERE ref_table.key_column_part1=other_table.column AND ref_table.key_column_part2=1;
ref, but with the addition that MySQL will do an extra search for rows that contain
NULLvalues. In the following example,
ref_tablewill be able to use
SELECT * FROM ref_table WHERE key_column=expr OR key_column IS NULL;This join type optimization is new for MySQL 4.1.1 and is mostly used when resolving subqueries. See section 7.2.6 How MySQL Optimizes
Index Mergeoptimization is used. For more information, see section 7.2.5 How MySQL Optimizes
ORClauses. In this case, the
keycolumn contains a list of used indexes, and
key_lencontains a list of the longest key parts for the used indexes.
INsubqueries of the following form:
value IN (SELECT primary_key FROM single_table WHERE some_exp)
unique_subqueryis just an index lookup function that replaces the subquery completely for better efficiency.
unique_subquery. It replaces
INsubqueries, but it works for non-unique indexes in subqueries of the following form:
value IN (SELECT key_field FROM single_table WHERE some_exp)
keycolumn indicates which index is used. The
key_lencontains the longest key part that was used. The
refcolumn will be
NULLfor this type.
rangecan be used for when an key column is compared to a constant using any of the
SELECT * FROM range_table WHERE key_column = 10; SELECT * FROM range_table WHERE key_column BETWEEN 10 and 20; SELECT * FROM range_table WHERE key_column IN (10,20,30); SELECT * FROM range_table WHERE key_part1= 10 and key_part2 IN (10,20,30);
ALL, except that only the index tree is scanned. This is usually faster than
ALL, because the index file is usually smaller than the data file. MySQL can use this join type when the query uses only columns that are part of a single index.
const, and usually very bad in all other cases. Normally, you can avoid
ALLby adding indexes that allow row retrieval from the table based on constant values or column values from earlier tables.
possible_keyscolumn indicates which indexes MySQL could use to find the rows in this table. Note that this column is totally independent of the order of the tables as displayed in the output from
EXPLAIN. That means that some of the keys in
possible_keysmay not be usable in practice with the generated table order. If this column is
NULL, there are no relevant indexes. In this case, you may be able to improve the performance of your query by examining the
WHEREclause to see whether it refers to some column or columns that would be suitable for indexing. If so, create an appropriate index and check the query with
EXPLAINagain. See section 14.2.2
ALTER TABLESyntax. To see what indexes a table has, use
SHOW INDEX FROM tbl_name.
keycolumn indicates the key (index) that MySQL actually decided to use. The key is
NULLif no index was chosen. To force MySQL to use or ignore an index listed in the
USE INDEX, or
IGNORE INDEXin your query. See section 14.1.7
myisamchk --analyzeon the table will help the optimizer choose better indexes. See section 126.96.36.199
ANALYZE TABLESyntax and section 188.8.131.52
key_lencolumn indicates the length of the key that MySQL decided to use. The length is
NULL. Note that this tells us how many parts of a multi-part key MySQL will actually use.
refcolumn shows which columns or constants are used with the
keyto select rows from the table.
rowscolumn indicates the number of rows MySQL believes it must examine to execute the query.
LEFT JOINoptimization on the query and will not examine more rows in this table for the previous row combination after it finds one row that matches the
LEFT JOINcriteria. Here is an example of the type of query that can be optimized this way:
SELECT * FROM t1 LEFT JOIN t2 ON t1.id=t2.id WHERE t2.id IS NULL;Assume that
t2.idis defined with
NOT NULL. In this case, MySQL will scan
t1and look up the rows in
t2using the values of
t1.id. If MySQL finds a matching row in
t2, it knows that
t2.idcan never be
NULL, and will not scan through the rest of the rows in
t2that have the same
idvalue. In other words, for each row in
t1, MySQL only needs to do a single lookup in
t2, regardless of how many rows actually match in
range checked for each record (index map: #)
join typeand storing the sort key and pointer to the row for all rows that match the
WHEREclause. The keys then are sorted and the rows are retrieved in sorted order.
ORDER BYclauses that list columns differently.
WHEREclause will be used to restrict which rows will be matched against the next table or sent to the client. Unless you specifically intend to fetch or examine all rows from the table, you may have something wrong in your query if you don't have this information and the table join type is
You can get a good indication of how good a join is by multiplying all values
rows column of the
EXPLAIN output. This should tell you
roughly how many rows MySQL must examine to execute the query. This
number is also used when you restrict queries with the
See section 7.5.2 Tuning Server Parameters.
The following example shows how a
JOIN can be optimized progressively
using the information provided by
Suppose that you have the
SELECT statement shown here and you plan to
examine it using
EXPLAIN SELECT tt.TicketNumber, tt.TimeIn, tt.ProjectReference, tt.EstimatedShipDate, tt.ActualShipDate, tt.ClientID, tt.ServiceCodes, tt.RepetitiveID, tt.CurrentProcess, tt.CurrentDPPerson, tt.RecordVolume, tt.DPPrinted, et.COUNTRY, et_1.COUNTRY, do.CUSTNAME FROM tt, et, et AS et_1, do WHERE tt.SubmitTime IS NULL AND tt.ActualPC = et.EMPLOYID AND tt.AssignedPC = et_1.EMPLOYID AND tt.ClientID = do.CUSTNMBR;
For this example, make the following assumptions:
| || |
| || |
| || |
| || |
| || |
tt.ActualPCvalues aren't evenly distributed.
Initially, before any optimizations have been performed, the
statement produces the following information:
table type possible_keys key key_len ref rows Extra et ALL PRIMARY NULL NULL NULL 74 do ALL PRIMARY NULL NULL NULL 2135 et_1 ALL PRIMARY NULL NULL NULL 74 tt ALL AssignedPC, NULL NULL NULL 3872 ClientID, ActualPC range checked for each record (key map: 35)
ALL for each table, this output indicates
that MySQL is generating a Cartesian product of all the tables; that is,
every combination of rows. This will take quite a long time, because the
product of the number of rows in each table must be examined. For the case
at hand, this product is
74 * 2135 * 74 * 3872 = 45,268,558,720 rows.
If the tables were bigger, you can only imagine how long it would take.
One problem here is that MySQL can't (yet) use indexes on columns
efficiently if they are declared differently. In this context,
CHAR are the same unless they are declared as
different lengths. Because
tt.ActualPC is declared as
et.EMPLOYID is declared as
CHAR(15), there is a length
To fix this disparity between column lengths, use
ALTER TABLE to
ActualPC from 10 characters to 15 characters:
mysql> ALTER TABLE tt MODIFY ActualPC VARCHAR(15);
et.EMPLOYID are both
EXPLAIN statement again produces this result:
table type possible_keys key key_len ref rows Extra tt ALL AssignedPC, NULL NULL NULL 3872 Using ClientID, where ActualPC do ALL PRIMARY NULL NULL NULL 2135 range checked for each record (key map: 1) et_1 ALL PRIMARY NULL NULL NULL 74 range checked for each record (key map: 1) et eq_ref PRIMARY PRIMARY 15 tt.ActualPC 1
This is not perfect, but is much better: The product of the
values is now less by a factor of 74. This version is executed in a couple
A second alteration can be made to eliminate the column length mismatches
tt.AssignedPC = et_1.EMPLOYID and
mysql> ALTER TABLE tt MODIFY AssignedPC VARCHAR(15), -> MODIFY ClientID VARCHAR(15);
EXPLAIN produces the output shown here:
table type possible_keys key key_len ref rows Extra et ALL PRIMARY NULL NULL NULL 74 tt ref AssignedPC, ActualPC 15 et.EMPLOYID 52 Using ClientID, where ActualPC et_1 eq_ref PRIMARY PRIMARY 15 tt.AssignedPC 1 do eq_ref PRIMARY PRIMARY 15 tt.ClientID 1
This is almost as good as it can get.
The remaining problem is that, by default, MySQL assumes that values
tt.ActualPC column are evenly distributed, and that isn't the
case for the
tt table. Fortunately, it is easy to tell MySQL
mysql> ANALYZE TABLE tt;
Now the join is perfect, and
EXPLAIN produces this result:
table type possible_keys key key_len ref rows Extra tt ALL AssignedPC NULL NULL NULL 3872 Using ClientID, where ActualPC et eq_ref PRIMARY PRIMARY 15 tt.ActualPC 1 et_1 eq_ref PRIMARY PRIMARY 15 tt.AssignedPC 1 do eq_ref PRIMARY PRIMARY 15 tt.ClientID 1
Note that the
rows column in the output from
EXPLAIN is an
educated guess from the MySQL join optimizer. You should check whether the
numbers are even close to the truth. If not, you may get better performance
STRAIGHT_JOIN in your
SELECT statement and trying
to list the tables in a different order in the
In most cases, you can estimate the performance by counting disk seeks.
For small tables, you can usually find a row in one disk seek (as the
index is probably cached). For bigger tables, you can estimate that
(using B-tree indexes) you will need this many seeks to find a row:
log(index_block_length / 3 * 2 / (index_length + data_pointer_length)) +
In MySQL, an index block is usually 1024 bytes and the data
pointer is usually 4 bytes. A 500,000 row table with an
index length of 3 bytes (medium integer) gives you:
log(500,000)/log(1024/3*2/(3+4)) + 1 = 4 seeks.
This index would require storage of about 500,000 * 7 * 3/2 = 5.2M (assuming a typical index buffer fill ration of 2/3), so you will probably have much of the index in memory and you will probably need only one or two calls to read data to find the row.
For writes, however, you will need four seek requests (as above) to find where to place the new index and normally two seeks to update the index and write the row.
Note that the preceding discussion doesn't mean that your application will slowly degenerate by log N! As long as everything is cached by the OS or SQL server, things will go only marginally slower while the table gets bigger. After the data gets too big to be cached, things will start to go much slower until your applications is only bound by disk-seeks (which increase by log N). To avoid this, increase the index cache as the data grows. See section 7.5.2 Tuning Server Parameters.
In general, when you want to make a slow
SELECT ... WHERE query
faster, the first thing to check is whether you can add an index.
All references between different tables should usually be done with
indexes. You can use the
EXPLAIN statement to determine which
indexes are used for a
See section 7.4.5 How MySQL Uses Indexes and
EXPLAIN Syntax (Get Information About a
Some general tips for speeding up queries:
ANALYZE TABLEor run
myisamchk --analyzeon a table after it has been loaded with data. This updates a value for each index part that indicates the average number of rows that have the same value. (For unique indexes, this is always 1.) MySQL will use this to decide which index to choose when you join two tables based on a non-constant expression. You can check the result from the table analysis by using
SHOW INDEX FROM tbl_nameand examining the
myisamchk --description --verbosealso shows index distribution information.
myisamchk --sort-index --sort-records=1(if you want to sort on index 1). This is a good way to make queries faster if you have a unique index from which you want to read all records in order according to the index. Note that it may take a long time the first time you sort a large table this way.
WHERE optimizations are put in the
SELECT part here because
they are mostly used with
SELECT, but the same optimizations apply for
Note that this section is incomplete. MySQL does many optimizations, and we have not had time to document them all.
Some of the optimizations performed by MySQL are listed here:
((a AND b) AND c OR (((a AND b) AND (c AND d)))) -> (a AND b AND c) OR (a AND b AND c AND d)
(a<b AND b=c) AND a=5 -> b>5 AND b=c AND a=5
(B>=5 AND B=5) OR (B=6 AND 5=5) OR (B=7 AND 5=6) -> B=5 OR B=6
COUNT(*)on a single table without a
WHEREis retrieved directly from the table information for
HEAPtables. This is also done for any
NOT NULLexpression when used with only one table.
SELECTstatements are impossible and returns no rows.
HAVINGis merged with
WHEREif you don't use
GROUP BYor group functions (
WHEREis constructed to get a fast
WHEREevaluation for the table and also to skip records as soon as possible.
WHEREclause on a
PRIMARY KEYor a
UNIQUEindex, where all index parts are compared to constant expressions and are defined as
mysql> SELECT * FROM t WHERE primary_key=1; mysql> SELECT * FROM t1,t2 -> WHERE t1.primary_key=1 AND t2.primary_key=t1.id;
ORDER BYand in
GROUP BYcome from the same table, then this table is preferred first when joining.
ORDER BYclause and a different
GROUP BYclause, or if the
GROUP BYcontains columns from tables other than the first table in the join queue, a temporary table is created.
SQL_SMALL_RESULT, MySQL will use an in-memory temporary table.
HAVINGclause are skipped.
Some examples of queries that are very fast:
mysql> SELECT COUNT(*) FROM tbl_name; mysql> SELECT MIN(key_part1),MAX(key_part1) FROM tbl_name; mysql> SELECT MAX(key_part2) FROM tbl_name -> WHERE key_part_1=constant; mysql> SELECT ... FROM tbl_name -> ORDER BY key_part1,key_part2,... LIMIT 10; mysql> SELECT ... FROM tbl_name -> ORDER BY key_part1 DESC,key_part2 DESC,... LIMIT 10;
The following queries are resolved using only the index tree, assuming the indexed columns are numeric:
mysql> SELECT key_part1,key_part2 FROM tbl_name WHERE key_part1=val; mysql> SELECT COUNT(*) FROM tbl_name -> WHERE key_part1=val1 AND key_part2=val2; mysql> SELECT key_part2 FROM tbl_name GROUP BY key_part1;
The following queries use indexing to retrieve the rows in sorted order without a separate sorting pass:
mysql> SELECT ... FROM tbl_name -> ORDER BY key_part1,key_part2,... ; mysql> SELECT ... FROM tbl_name -> ORDER BY key_part1 DESC,key_part2 DESC,... ;
Index Merge method is used to retrieve rows with several
range scans and merge the
results into one.
This method is employed when the table condition is a disjunction of
conditions for which
could be used with different keys.
EXPLAIN output, this method appears as
index_merge in the
type column. In this case, the
key column contains a list of
used indexes, and
key_len contains a list of the longest key parts
for the used indexes.
SELECT * FROM table WHERE key_col1 = 10 OR key_col2 = 20; SELECT * FROM table WHERE (key_col1 = 10 OR key_col2 = 20) AND nonkey_col=30; SELECT * FROM t1,t2 WHERE (t1.key1 IN (1,2) OR t1.key2 LIKE 'value%') AND t2.key1=t1.some_col SELECT * FROM t1,t2 WHERE t1.key1=1 AND (t2.key1=t1.some_col OR t2.key2=t1.some_col2)
This ``join'' type optimization is new in MySQL 5.0.0, and represents a significant change in behavior with regard to indexes, because the old rule was that the server is only ever able to use at most one index for each referenced table.
MySQL can do the same optimization on
col_name IS NULL that it can do
col_name = constant_value. For example, MySQL can use
indexes and ranges to search for
SELECT * FROM tbl_name WHERE key_col IS NULL; SELECT * FROM tbl_name WHERE key_col <=> NULL; SELECT * FROM tbl_name WHERE key_col=const1 OR key_col=const2 OR key_col IS NULL;
WHERE clause includes a
col_name IS NULL condition for a
column that is declared as
NOT NULL, that expression will be
optimized away. This optimization does not occur in cases when the column
NULL anyway; for example, if it comes from a table on
the right side of a
MySQL 4.1.1 can additionally optimize the combination
expr AND col_name IS NULL, a form that is common in resolved subqueries.
EXPLAIN will show
ref_or_null when this
optimization is used.
This optimization can handle one
IS NULL for any key part.
Some examples of queries that are optimized, assuming that there is an index
SELECT * FROM t1 WHERE t1.a=expr OR t1.a IS NULL; SELECT * FROM t1,t2 WHERE t1.a=t2.a OR t2.a IS NULL; SELECT * FROM t1,t2 WHERE (t1.a=t2.a OR t2.a IS NULL) AND t2.b=t1.b; SELECT * FROM t1,t2 WHERE t1.a=t2.a AND (t2.b=t1.b OR t2.b IS NULL); SELECT * FROM t1,t2 WHERE (t1.a=t2.a AND t2.a IS NULL AND ...) OR (t1.a=t2.a AND t2.a IS NULL AND ...);
ref_or_null works by first doing a read on the reference key,
and after that a separate search for rows with a
NULL key value.
Note that the optimization can only handle one
IS NULL level.
SELECT * FROM t1,t2 WHERE (t1.a=t2.a AND t2.a IS NULL) OR (t1.b=t2.b AND t2.b IS NULL);
In the preceding case, MySQL will only use key lookups on the part
(t1.a=t2.a AND t2.a IS NULL) and not be able to use the key part on
DISTINCT combined with
ORDER BY will
need a temporary table in many cases.
Note that because
DISTINCT may use
GROUP BY, you should be
aware of how MySQL works with in fields in
ORDER BY or
that are not part of the selected fields. See section 13.9.3
GROUP BY with Hidden Fields.
LIMIT row_count with
DISTINCT, MySQL will stop
as soon as it finds
row_count unique rows.
If you don't use columns from all tables named in a query, MySQL will stop the scanning of the not used tables as soon as it has found the first match.
SELECT DISTINCT t1.a FROM t1,t2 where t1.a=t2.a;
In this case, assuming
t1 is used before
t2 (check with
EXPLAIN), then MySQL will stop reading from
t2 (for that
particular row in
t1) when the first row in
t2 is found.
A LEFT JOIN B join_condition in MySQL is implemented as follows:
Bis set to be dependent on table
Aand all tables on which
Ais set to be dependent on all tables (except
B) that are used in the
LEFT JOINcondition is used to decide how we should retrieve rows from table B. (In other words, any condition in the
WHEREclause is not used).
WHEREoptimizations are done.
Athat matches the
WHEREclause, but there is no row in
Bthat matches the
ONcondition, an extra
Brow is generated with all columns set to
LEFT JOINto find rows that don't exist in some table and you have the following test:
col_name IS NULLin the
col_nameis a column that is declared as
NOT NULL, then MySQL will stop searching for more rows (for a particular key combination) after it has found one row that matches the
RIGHT JOIN is implemented analogously to
The join optimizer calculates the order in which tables should be joined.
The table read order forced by
LEFT JOIN and
helps the join optimizer do its work much more quickly, because there are
fewer table permutations to check.
Note that this means that if you do a query of the following type,
MySQL will do a full scan on
b as the
LEFT JOIN will force
it to be read before
SELECT * FROM a,b LEFT JOIN c ON (c.key=a.key) LEFT JOIN d (d.key=a.key) WHERE b.key=d.key
The fix in this case is to change the query to:
SELECT * FROM b,a LEFT JOIN c ON (c.key=a.key) LEFT JOIN d (d.key=a.key) WHERE b.key=d.key
Starting from 4.0.14, MySQL does the following
LEFT JOIN optimization:
WHERE condition is always false for the generated
NULL row, the
LEFT JOIN is changed to a normal join.
For example, the
WHERE clause would be
false in the following query
if t2.column would be
SELECT * FROM t1 LEFT JOIN t2 ON (column) WHERE t2.column2=5;
Therefore, it's safe to convert the query to a normal join:
SELECT * FROM t1,t2 WHERE t2.column2=5 AND t1.column=t2.column;
This can be made faster as MySQL can now use table
t1 if this would result in a better query plan. To force a
specific table order, use
In some cases, MySQL can uses index to satisfy an
ORDER BY or
GROUP BY request without doing any extra sorting.
The index can also be used even if the
ORDER BY doesn't match the
index exactly, as long as all the unused index parts and all the extra
ORDER BY columns are constants in the
clause. The following queries will use the index to resolve the
ORDER BY /
GROUP BY part:
SELECT * FROM t1 ORDER BY key_part1,key_part2,... SELECT * FROM t1 WHERE key_part1=constant ORDER BY key_part2 SELECT * FROM t1 WHERE key_part1=constant GROUP BY key_part2 SELECT * FROM t1 ORDER BY key_part1 DESC,key_part2 DESC SELECT * FROM t1 WHERE key_part1=1 ORDER BY key_part1 DESC,key_part2 DESC
In some cases, MySQL can not use indexes to resolve the
BY (note that MySQL will still use indexes to find the rows that
ORDER BYon different keys:
SELECT * FROM t1 ORDER BY key1,key2;
ORDER BYon non-consecutive key parts:
SELECT * FROM t1 WHERE key2=constant ORDER BY key_part2;
SELECT * FROM t1 ORDER BY key_part1 DESC,key_part2 ASC;
SELECT * FROM t1 WHERE key2=constant ORDER BY key1
ORDER BYare not all from the first not-constant table that is used to retrieve rows. (This is the first table in the
EXPLAINoutput that doesn't have a
HASHindex in a
In those cases where MySQL must sort the result, it uses the following algorithm:
WHEREclause are skipped.
MERGEBUFF(7) regions to one block in another temporary file. Repeat until all blocks from the first file are in the second file.
MERGEBUFF2(15) blocks left.
read_rnd_buffer_size. The code for this step is in the `sql/records.cc' source file.
EXPLAIN SELECT ... ORDER BY, you can check whether MySQL can use
indexes to resolve the query. It cannot if you see
Using filesort in
See section 7.2.1
EXPLAIN Syntax (Get Information About a
If you want to increase
ORDER BY speed, first see whether you can get
MySQL to use indexes instead of using an extra sorting phase. If this is not
possible, you can try the following strategies:
tmpdirto point to a dedicated disk with lots of empty space. If you use MySQL 4.1 or later, this option accepts several paths that are used in round-robin fashion. Paths should be separated by colon characters (`:') on Unix and semicolon characters (`;') on Windows, NetWare, and OS/2. You can use this feature to spread load across several directories. Note: The paths should be for directories in filesystems that are on different physical disks, not different partitions of the same disk.
By default, MySQL sorts all
GROUP BY x,y,... queries as if you
ORDER BY x,y,... in the query as well. If you include the
ORDER BY clause explicitly that contains the same column list, MySQL
optimizes it away without any speed penalty, though the sorting still
occurs. If a query includes
GROUP BY but you want to avoid the
overhead of sorting the result, you can supress sorting by specifying
ORDER BY NULL:
INSERT INTO foo SELECT a,COUNT(*) FROM bar GROUP BY a ORDER BY NULL;
In some cases, MySQL will handle the query differently when you are
LIMIT row_count and not using
LIMIT, MySQL uses indexes in some cases when it normally would prefer to do a full table scan.
ORDER BY, MySQL ends the sorting as soon as it has found the first
row_countlines, rather than sorting the whole table.
DISTINCT, MySQL stops as soon as it finds
GROUP BYcan be resolved by reading the key in order (or doing a sort on the key) and then calculating summaries until the key value changes. In this case,
LIMIT row_countwill not calculate any unnecessary
LIMIT 0always quickly returns an empty set. This is useful to check the query or to get the column types of the result columns.
LIMIT row_countis used to calculate how much space is required.
The time to insert a record is determined by the following factors, where the numbers indicate approximate proportions:
This does not take into consideration the initial overhead to open tables, which is done once for each concurrently running query.
The size of the table slows down the insertion of indexes by log N (assuming B-tree indexes).
You can use the following methods to speed up inserts:
INSERTstatements with multiple
VALUESlists to insert several rows at a time. This is much faster (many times faster in some cases) than using separate single-row
INSERTstatements. If you are adding data to non-empty table, you may tune up the
bulk_insert_buffer_sizevariable to make it even faster. See section 5.2.3 Server System Variables.
INSERT DELAYEDstatement. See section 14.1.4
MyISAMtables you can insert rows at the same time
SELECTstatements are running if there are no deleted rows in the tables.
LOAD DATA INFILE. This is usually 20 times faster than using a lot of
INSERTstatements. See section 14.1.5
LOAD DATA INFILESyntax.
LOAD DATA INFILErun even faster when the table has many indexes. Use the following procedure:
FLUSH TABLESstatement or a
myisamchk --keys-used=0 -rq /path/to/db/tbl_name. This will remove all use of all indexes for the table.
LOAD DATA INFILE. This will not update any indexes and will therefore be very fast.
myisampackto make it smaller. See section 184.108.40.206 Compressed Table Characteristics.
myisamchk -r -q /path/to/db/tbl_name. This will create the index tree in memory before writing it to disk, which is much faster because it avoids lots of disk seeks. The resulting index tree is also perfectly balanced.
FLUSH TABLESstatement or a
LOAD DATA INFILEalso performs the preceding optimization if you insert into an empty table; the main difference is that you can let
myisamchkallocate much more temporary memory for the index creation that you might want the server to allocate for index re-creation when it executes the
LOAD DATA INFILEstatement. As of MySQL 4.0, you can also use
ALTER TABLE tbl_name DISABLE KEYSinstead of
myisamchk --keys-used=0 -rq /path/to/db/tbl_nameand
ALTER TABLE tbl_name ENABLE KEYSinstead of
myisamchk -r -q /path/to/db/tbl_name. This way you can also skip
INSERToperations that are done with multiple statements by locking your tables:
mysql> LOCK TABLES a WRITE; mysql> INSERT INTO a VALUES (1,23),(2,34),(4,33); mysql> INSERT INTO a VALUES (8,26),(6,29); mysql> UNLOCK TABLES;A performance benefit occurs because the index buffer is flushed to disk only once, after all
INSERTstatements have completed. Normally there would be as many index buffer flushes as there are different
INSERTstatements. Locking is not needed if you can insert all rows with a single statement. For transactional tables, you should use
LOCK TABLESto get a speedup. Locking also lowers the total time of multiple-connection tests, but the maximum wait time for individual threads might go up because they wait for locks. For example:
Thread 1 does 1000 inserts Threads 2, 3, and 4 do 1 insert Thread 5 does 1000 insertsIf you don't use locking, 2, 3, and 4 will finish before 1 and 5. If you use locking, 2, 3, and 4 probably will not finish before 1 or 5, but the total time should be about 40% faster.
DELETEoperations are very fast in MySQL, but you will obtain better overall performance by adding locks around everything that does more than about 5 inserts or updates in a row. If you do very many inserts in a row, you could do a
LOCK TABLESfollowed by an
UNLOCK TABLESonce in a while (about each 1000 rows) to allow other threads access to the table. This would still result in a nice performance gain.
INSERTis still much slower for loading data than
LOAD DATA INFILE, even when using the strategies just outlined.
LOAD DATA INFILEand
INSERT, enlarge the key buffer. See section 7.5.2 Tuning Server Parameters.
Update queries are optimized as a
SELECT query with the additional
overhead of a write. The speed of the write is dependent on the size of
the data that is being updated and the number of indexes that are
updated. Indexes that are not changed will not be updated.
Also, another way to get fast updates is to delay updates and then do many updates in a row later. Doing many updates in a row is much quicker than doing one at a time if you lock the table.
Note that for a
MyISAM table that uses dynamic record format,
updating a record to a longer total length may split the record. If you do
this often, it is very important to use
OPTIMIZE TABLE occasionally.
See section 220.127.116.11
OPTIMIZE TABLE Syntax.
The time to delete individual records is exactly proportional to the number of indexes. To delete records more quickly, you can increase the size of the key cache. See section 7.5.2 Tuning Server Parameters.
If you want to delete all rows in the table, use
tbl_name rather than
DELETE FROM tbl_name.
See section 14.1.9
This section lists a number of miscellaneous tips for improving query processing speeed:
thread_cache_sizevariable. See section 7.5.2 Tuning Server Parameters.
EXPLAINstatement. See section 7.2.1
EXPLAINSyntax (Get Information About a
MyISAMtables that are updated frequently. This is to avoid problems with table locking that occur due to contention between readers and writers.
MyISAMtables that have no deleted rows, you can insert rows at the end at the same time another query is reading from the table. If this is important for you, you should consider using the table in ways that avoid deleting rows. Another possibility is to run
OPTIMIZE TABLEafter you have deleted a lot of rows.
ALTER TABLE ... ORDER BY expr1,expr2...if you mostly retrieve rows in
expr1,expr2...order. By using this option after extensive changes to the table, you may be able to get higher performance.
SELECT * FROM tbl_name WHERE hash=MD5(CONCAT(col1,col2)) AND col_1='constant' AND col_2='constant'
MyISAMtables that change a lot, you should try to avoid all variable-length columns (
TEXT). The table will use dynamic record format if it includes a single variable-length column. See section 15 MySQL Storage Engines and Table Types.
MyISAMtable with dynamic record format (see above) that you can change to a fixed record size, or if you very often need to scan the table and don't need most of the columns. See section 15 MySQL Storage Engines and Table Types.
UPDATE table SET count=count+1 WHERE index_column=constant;This is really important when you use MySQL storage engines like
ISAMthat only have table-level locking (multiple readers / single writers). This will also give better performance with most databases, as the row locking manager in this case will have less to do.
BLOBcolumn. In this case, you must add some extra code in your application to pack and unpack information in the
BLOBvalues, but this may save a lot of accesses at some stage. This is practical when you have data that doesn't conform to a rows-and-columns table structure.
INSERT /*! DELAYED */when you do not need to know when your data is written. This speeds things up because many records can be written with a single disk write.
INSERT /*! LOW_PRIORITY */when you want to give
SELECTstatements higher priority than your inserts.
SELECT /*! HIGH_PRIORITY */to get retrievals that jump the queue. That is, the
SELECTis done even if there is another client waiting to do a write.
INSERTstatements to store many rows with one SQL statement (many SQL servers supports this).
LOAD DATA INFILEto load bigger amounts of data. This is faster than using
AUTO_INCREMENTcolumns to generate unique values.
OPTIMIZE TABLEonce in a while to avoid fragmentation when using a dynamic table format with
MyISAMtables. See section 18.104.22.168
HEAPtables when possible to get more speed. See section 15 MySQL Storage Engines and Table Types.
customer, use a column name of
customer_name. To make your names portable to other SQL servers, you should keep them shorter than 18 characters.
MyISAMstorage engine directly, you could get a speed increase of 2-5 times compared to using the SQL interface. To be able to do this the data must be on the same server as the application, and usually it should only be accessed by one process (because external file locking is really slow). One could eliminate the above problems by introducing low-level
MyISAMcommands in the MySQL server (this could be one easy way to get more performance if needed). By carefully designing the database interface, it should be quite easy to support this types of optimization.
MyISAMtable with the
DELAY_KEY_WRITE=1table option makes index updates faster because they are not flushed to disk until the table is closed. The downside is that if something kills the server while such tables are open, you should ensure that they are okay by running
myisamchkbefore restarting the server. (However, even in this case, you should not lose anything by using
DELAY_KEY_WRITE, because the key information can always be generated from the data rows.)
You can find a discussion about different locking methods in the appendix. See section D.4 Locking methods.
BDB storage engines, all locking in
MySQL is deadlock-free for storage engines that use table-level locking.
This include the
engines. Deadlock avoidance is managed by always requesting all needed locks
at once at the beginning of a query and always locking the tables in the
InnoDB uses row locks and
BDB uses page locks. For the
BDB storage engines, deadlock is possible. This is
InnoDB automatically acquires row locks and
acquires page locks during the processing of SQL statements, not at the
start of the transaction.
The locking method MySQL uses for
WRITE locks works as follows:
The locking method MySQL uses for
READ locks works as follows:
When a lock is released, the lock is made available to the threads in the write lock queue, then to the threads in the read lock queue.
This means that if you have many updates for a table,
statements will wait until there are no more updates.
To work around this for the case where you want to do many
SELECT operations on a table, you can insert rows in a temporary
table and update the real table with the records from the temporary table
once in a while.
This can be done with the following code:
mysql> LOCK TABLES real_table WRITE, insert_table WRITE; mysql> INSERT INTO real_table SELECT * FROM insert_table; mysql> TRUNCATE TABLE insert_table; mysql> UNLOCK TABLES;
You can use the
LOW_PRIORITY options with
SELECT if you want to prioritize retrieval in some specific
cases. You can also start
to get the same behavior.
SQL_BUFFER_RESULT can also help to make the duration of table
See section 14.1.7
You could also change the locking code in `mysys/thr_lock.c' to use a single queue. In this case, write locks and read locks would have the same priority, which might help some applications.
The table locking code in MySQL is deadlock free.
To achieve a very high lock speed, MySQL uses table locking (instead of
page, row, or column locking) for all storage engines except
BDB. For large tables, table locking is much better than row
locking for most applications, but there are some pitfalls.
BDB tables, MySQL only uses table
locking if you explicitly lock the table with
For these table types we recommend you to not use
LOCK TABLES at all, because
InnoDB uses automatic
row level locking and
BDB uses page level locking to
ensure transaction isolation.
As of MySQL Version 3.23.7 (3.23.25 for Windows), you can insert rows into a
MyISAM table at the same time other threads are reading from it.
Note that currently this works at the time the insert is made only if there
are no holes resulting from rows having been deleted from the middle of the
table. When all holes have been filled with new data, concurrent inserts
are re-enabled automatically.
Table locking enables many threads to read from a table at the same time, but if a thread wants to write to a table, it must first get exclusive access. During the update, all other threads that want to access this particular table must wait until the update is done.
Table updates normally are considered to be more important than table
retrievals, so they are given higher priority. This should ensure that
updates to a table are not 'starved' even if there is heavy
activity for the table. You can change this behavior by using
LOW_PRIORITY with update statements or
Starting from MySQL Version 3.23.7, you can use the
max_write_lock_count variable to force MySQL to temporarily elevate
the priority of all
SELECT statements that are waiting for a table,
after a specific number of inserts to the table occur.
Table locking causes problems in cases such as when a thread is waiting because the disk is full and free space needs to become available before the thread can proceed. In this case, all threads that want to access the problem table will also be put in a waiting state until more disk space is made available.
Table locking is also disadvantageous under the following scenario:
SELECTthat takes a long time to run.
UPDATEon a used table. This client will wait until the
SELECTstatement on the same table. As
UPDATEhas higher priority than
SELECTwill wait for the
UPDATEto finish. It will also wait for the first
The following list describes some ways to avoid or reduce contention caused by table locking:
SELECTstatements to run faster. You might have to create some summary tables to do this.
--low-priority-updates. This will give all statements that update (modify) a table lower priority than
SELECTstatements. In this case, the second
SELECTstatement in the previous scenario would execute before the
INSERTstatement, and would not need to wait for the first
DELETEstatement lower priority with the
mysqldwith a low value for
READlocks after a certain number of
SET LOW_PRIORITY_UPDATES=1. See section 22.214.171.124
SELECTis very important with the
HIGH_PRIORITYattribute. See section 14.1.7
SELECT, switch to using
MyISAMtables, which support concurrent
INSERTwill probably solve your problems. See section 14.1.4
DELETEmay help. See section 14.1.1
MySQL keeps row data and index data in separate files. Many (almost all) other databases mix row and index data in the same file. We believe that the MySQL choice is better for a very wide range of modern systems.
Another way to store the row data is to keep the information for each column in a separate area (examples are SDBM and Focus). This will cause a performance hit for every query that accesses more than one column. Because this degenerates so quickly when more than one column is accessed, we believe that this model is not good for general purpose databases.
The more common case is that the index and data are stored together (as in Oracle/Sybase et al). In this case, you will find the row information at the leaf page of the index. The good thing with this layout is that it, in many cases, depending on how well the index is cached, saves a disk read. The bad things with this layout are:
One of the most basic optimizations is to design your tables to take as little space on the disk as possible. This can give huge improvements because disk reads are faster, and smaller tables normally require less main memory while their contents are being actively processed during query execution. Indexing also is a smaller resource burden if done on smaller columns.
MySQL supports a lot of different table types and row formats. For each table, you can decide which storage/index method to use. Choosing the right table format for your application may give you a big performance gain. See section 15 MySQL Storage Engines and Table Types.
You can get better performance on a table and minimize storage space using the techniques listed here:
MEDIUMINTis often better than
NOT NULLif possible. It makes everything faster and you save one bit per column. If you really need
NULLin your application, you should definitely use it. Just avoid having it on all columns by default.
MyISAMtables, if you don't have any variable-length columns (
BLOBcolumns), a fixed-size record format is used. This is faster but unfortunately may waste some space. See section 15.1.3
MyISAMTable Storage Formats.
All MySQL column types can be indexed. Use of indexes on the
relevant columns is the best way to improve the performance of
The maximum number of indexes per table and the maximum index length is defined per storage engine. See section 15 MySQL Storage Engines and Table Types. All storage engines support at least 16 indexes per table and a total index length of at least 256 bytes. Most storage engines have higher limits.
VARCHAR columns, you can index a prefix of a
column. This is much faster and requires less disk space than indexing the
whole column. The syntax to use in the
CREATE TABLE statement to
index a column prefix looks like this:
INDEX index_name (col_name(length))
The example here creates an index for the first 10 bytes of the
mysql> CREATE TABLE test ( -> name CHAR(200) NOT NULL, -> INDEX index_name (name(10)));
TEXT columns can be indexed as well, but for these
types an index prefix is mandatory, not optional. The prefix may be up to
255 bytes long.
As of MySQL Version 3.23.23, you can also create
They are used for full-text search. Only the
MyISAM table type
FULLTEXT indexes and only for
TEXT columns. Indexing always happens over the entire column and
partial (prefix) indexing is not supported. See section 13.6 Full-Text Search Functions for
As of MySQL Version 4.1.0, you can create indexes on spatial column types.
Currently, spatial types are supported only by the
engine. Spatial indexes use R-trees.
HEAP) storage engine supports hash indexes.
As of MySQL Version 4.1.0, the engine also supports B-tree indexes.
MySQL can create indexes on multiple columns. An index may consist of up to 15 columns. For certain column types, you can index a prefix of the column (see section 7.4.3 Column Indexes).
A multiple-column index can be considered a sorted array containing values that are created by concatenating the values of the indexed columns.
MySQL uses multiple-column indexes in such a way that queries are
fast when you specify a known quantity for the first column of the index in a
WHERE clause, even if you don't specify values for the other columns.
Suppose that a table has the following specification:
mysql> CREATE TABLE test ( -> id INT NOT NULL, -> last_name CHAR(30) NOT NULL, -> first_name CHAR(30) NOT NULL, -> PRIMARY KEY (id), -> INDEX name (last_name,first_name));
name index is an index over
first_name. The index can used for queries that specify
values in a known range for
last_name, or for both
name index will be used in the following queries:
mysql> SELECT * FROM test WHERE last_name='Widenius'; mysql> SELECT * FROM test -> WHERE last_name='Widenius' AND first_name='Michael'; mysql> SELECT * FROM test -> WHERE last_name='Widenius' -> AND (first_name='Michael' OR first_name='Monty'); mysql> SELECT * FROM test -> WHERE last_name='Widenius' -> AND first_name >='M' AND first_name < 'N';
name index will not be used in the following
mysql> SELECT * FROM test WHERE first_name='Michael'; mysql> SELECT * FROM test -> WHERE last_name='Widenius' OR first_name='Michael';
The manner in which MySQL uses indexes to improve query performance is discussed further in the next section.
Indexes are used to find rows with specific column values fast. Without an index, MySQL has to start with the first record and then read through the whole table to find the relevant rows. The bigger the table, the more this costs. If the table has an index for the columns in question, MySQL can quickly determine the position to seek to in the middle of the data file without having to look at all the data. If a table has 1000 rows, this is at least 100 times faster than reading sequentially. Note that if you need to access almost all 1000 rows, it is faster to read sequentially, because that minimizes disk seeks.
Most MySQL indexes (
FULLTEXT) are stored in B-trees. Exceptions are that indexes on
spatial column types use R-trees, and
support hash indexes.
Strings are automatically prefix-
and end-space compressed. See section 14.2.4
CREATE INDEX Syntax.
In general, indexes are used as described in the following discussion.
Characteristics specific to hash indexes (as used in
are described at the end of this section.
MAX()value for a specific indexed column
key_col. This is optimized by a preprocessor that checks if you are using
WHERE key_part_# = constanton all key parts that occur before
key_colin the index. In this case, MySQL will do a single key lookup for each
MAX()expression and replace it with a constant. If all expressions are replaced with constants, the query will return at once:
SELECT MIN(key_part2),MAX(key_part2) FROM tbl_name where key_part1=10;
ORDER BY key_part_1,key_part_2). If all key parts are followed by
DESC, the key is read in reverse order. See section 7.2.9 How MySQL Optimizes
SELECT key_part3 FROM tbl_name WHERE key_part1=1
Suppose that you issue the following
mysql> SELECT * FROM tbl_name WHERE col1=val1 AND col2=val2;
If a multiple-column index exists on
appropriate rows can be fetched directly. If separate single-column
indexes exist on
col2, the optimizer tries to
find the most restrictive index by deciding which index will find fewer
rows and using that index to fetch the rows.
If the table has a multiple-column index, any leftmost prefix of the
index can be used by the optimizer to find rows. For example, if you
have a three-column index on
(col1, col2, col3), you have indexed
search capabilities on
(col1, col2), and
(col1, col2, col3).
MySQL can't use a partial index if the columns don't form a
leftmost prefix of the index. Suppose that you have the
statements shown here:
mysql> SELECT * FROM tbl_name WHERE col1=val1; mysql> SELECT * FROM tbl_name WHERE col2=val2; mysql> SELECT * FROM tbl_name WHERE col2=val2 AND col3=val3;
If an index exists on
(col1, col2, col3), only the first of the preceding
queries uses the index. The second and third queries do involve
indexed columns, but
(col2, col3) are not
leftmost prefixes of
(col1, col2, col3).
An index is used for columns that you compare with the
MySQL also uses indexes for
LIKE comparisons if the argument
LIKE is a constant string that doesn't start with a wildcard
character. For example, the following
SELECT statements use indexes:
mysql> SELECT * FROM tbl_name WHERE key_col LIKE 'Patrick%'; mysql> SELECT * FROM tbl_name WHERE key_col LIKE 'Pat%_ck%';
In the first statement, only rows with
'Patrick' <= key_col <
'Patricl' are considered. In the second statement, only rows with
'Pat' <= key_col < 'Pau' are considered.
SELECT statements will not use indexes:
mysql> SELECT * FROM tbl_name WHERE key_col LIKE '%Patrick%'; mysql> SELECT * FROM tbl_name WHERE key_col LIKE other_col;
In the first statement, the
LIKE value begins with a wildcard
character. In the second statement, the
LIKE value is not a
MySQL 4.0 and up performs an additional
LIKE optimization. If you use
... LIKE '%string%' and
string is longer than 3 characters,
MySQL will use the
Turbo Boyer-Moore algorithm to initialize the
pattern for the string and then use this pattern to perform the search
col_name IS NULL will use indexes if
Any index that doesn't span all
AND levels in the
is not used to optimize the query. In other words, to be able to use an
index, a prefix of the index must be used in every
WHERE clauses use indexes:
... WHERE index_part1=1 AND index_part2=2 AND other_column=3 /* index = 1 OR index = 2 */ ... WHERE index=1 OR A=10 AND index=2 /* optimized like "index_part1='hello'" */ ... WHERE index_part1='hello' AND index_part_3=5 /* Can use index on index1 but not on index2 or index 3 */ ... WHERE index1=1 AND index2=2 OR index1=3 AND index3=3;
WHERE clauses do not use indexes:
/* index_part_1 is not used */ ... WHERE index_part2=1 AND index_part3=2 /* Index is not used in both AND parts */ ... WHERE index=1 OR A=10 /* No index spans all rows */ ... WHERE index_part1=1 OR index_part2=10
Sometimes MySQL will not use an index, even if one is available. One way
this occurs is when the optimizer estimates that using the index would
require MySQL to access more than 30% of the rows in the table. (In this
case, a table scan is probably much faster, because it will require many fewer
seeks.) However, if such a query uses
LIMIT to only retrieve part of
the rows, MySQL will use an index anyway, because it can much more quickly
find the few rows to return in the result.
Hash indexes have somewhat different characteristics than those just discussed:
<=>comparisons (but are VERY fast).
ORDER BYoperations. (The index cannot be used to search for the next entry in order.)
MyISAMtable to a
To minimize disk I/O, the
MyISAM storage engine employs a strategy that
is used by many database management systems. It exploits a cache
mechanism to keep the most frequently accessed table blocks in memory:
This section first describes the basic operation of the
cache. Then it discusses changes made in MySQL 4.1 that improve key cache
performance and that enable you to better control cache operation:
The key cache mechanism also is used for
ISAM tables. However, the
significance of this fact is on the wane.
ISAM table use has been
decreasing since MySQL 3.23 when
MyISAM was introduced. MySQL 4.1
carries this trend further; the
ISAM storage engine is disabled by
You can control the size of the key cache by means of the
key_buffer_size system variable. If this variable is set equal
to zero, no key cache is used. The key cache also is not used if the
key_buffer_size value is too small to allocate the minimal number
of block buffers (8).
When the key cache is not operational, index files are accessed using only the native filesystem buffering provided by the operating system. (In other words, table index blocks are accessed using the same strategy as that employed for table data blocks.)
An index block is a contiguous unit of access to the
MyISAM index files.
Usually the size of an index block is equal to the size of nodes of the
index B-tree. (Indexes are represented on disk using a B-tree data structure.
Nodes at the bottom of the tree are leaf nodes. Nodes above the leaf nodes
are non-leaf nodes.)
All block buffers in a key cache structure are the same size. This size can be equal to, greater than, or less than the size of a table index block. Usually one these two values is a multiple of the other.
When data from any table index block must be accessed, the server first checks whether it is available in some block buffer of the key cache. If it is, the server accesses data in the key cache rather than on disk. That is, it reads from the cache or writes into it rather than reading from or writing to disk. Otherwise, the server chooses a cache block buffer containing a different table index block (or blocks) and replaces the data there by a copy of required table index block. As soon as the new index block is in the cache, the index data can be accessed.
If it happens that a block selected for replacement has been modified, the block is considered ``dirty.'' In this case, before being replaced, its contents are flushed to the table index from which it came.
Usually the server follows a LRU (Least Recently Used) strategy: When choosing a block for replacement, it selects the least recently used index block. To be able to make such a choice easy, the key cache module maintains a special queue (LRU chain) of all used blocks. When a block is accessed, it is placed at the end of the queue. When blocks need to be replaced, blocks at the beginning of the queue are the least recently used and become the first candidates for eviction.
Prior to MySQL 4.1, access to the key cache is serialized: No two threads can access key cache buffers simultaneously. The server processes a request for an index block only after it has finished processing the previous request. As a result, a request for an index block not present in any key cache buffer blocks access by other threads while a buffer is being updated to contain the requested index block.
Starting from version 4.1.0, the server supports shared access to the key cache:
Shared access to the key cache allows the server to improve throughput significantly.
Shared access to the key cache improves performance but does not eliminate contention among threads entirely. They still compete for control structures that manage access to the key cache buffers. To reduce key cache access contention further, MySQL 4.1.1 offers the feature of multiple key caches. This allows you to assign different table indexes to different key caches.
When there may be multiple key caches, the server must know which cache to
use when processing queries for a given
MyISAM table. By default, all
MyISAM table indexes are cached in the default key cache. To assign
table indexes to a specific key cache, use the
CACHE INDEX statement.
For example, the following statement assigns indexes from the tables
t3 to the key cache named
mysql> CACHE INDEX t1, t2, t3 IN hot_cache; +---------+--------------------+----------+----------+ | Table | Op | Msg_type | Msg_text | +---------+--------------------+----------+----------+ | test.t1 | assign_to_keycache | status | OK | | test.t2 | assign_to_keycache | status | OK | | test.t3 | assign_to_keycache | status | OK | +---------+--------------------+----------+----------+
Note: If the server has been built with the
ISAM tables use the key cache mechanism. However,
ISAM indexes use only the default key cache and cannot be reassigned to
a different cache.
The key cache referred to in a
CACHE INDEX statement can be created
by setting its size with a
SET GLOBAL parameter setting statement or by
using server startup options. For example:
mysql> SET GLOBAL keycache1.key_buffer_size=128*1024;
To destroy a key cache, set its size to zero:
mysql> SET GLOBAL keycache1.key_buffer_size=0;
Key cache variables are structured system variables that have a name and
keycache1 is the
cache variable name and
key_buffer_size is the cache component.
See section 10.4.1 Structured System Variables for a description of the syntax used
for referring to structured key cache system variables.
By default, table indexes are assigned to the main (default) key cache created at the server startup. When a key cache is destroyed, all indexes assigned to it are reassigned to the default key cache.
For a busy server, we recommend a strategy that uses three key caches:
One reason the use of three key caches be beneficial is that access to one key cache structure does not block access to the others. Queries that access tables assigned to one cache do not compete with queries that access tables assigned to another cache. Performance gains occur for other reasons as well:
By default, the key cache management system of MySQL 4.1 uses the LRU strategy for choosing key cache blocks to be evicted, but it also supports a more sophisticated method called the midpoint insertion strategy.
When using the midpoint insertion strategy, the LRU chain is divided into
two parts: A hot sub-chain and a warm sub-chain. The division point between
two parts is not fixed, but the key cache management system takes care that
the warm part is not ``too short,'' always containing at least
key_cache_division_limit percent of the key cache blocks.
key_cache_division_limit is a component of structured key cache
variables, so its value is a parameter that can be set per cache.
When an index block is read from a table into the key cache, it is placed at the end of the warm sub-chain. After a certain number of hits (accesses of the block) it is promoted to the hot sub-chain. At present, the number of hits required to promote a block (3) is the same for all index blocks. In the future, we will allow the hit count to depend on the B-tree level of the node corresponding to an index block: Fewer hits will be required for promotion of an index block if it contains a non-leaf node from the upper levels of the index B-tree than if it contains a leaf node.
A block promoted into the hot sub-chain is placed at the end of the chain.
The block then circulates within this sub-chain. If the block stays at the
beginning of the sub-chain for a long enough time, it is demoted to the warm
chain. This time is determined by the value of the
key_cache_age_threshold component of the key cache.
The threshold value prescribes that, for a key cache containing
blocks, the block at the beginning of the hot sub-chain not accessed within
N*key_cache_age_threshold/100 hits is to be moved to the
beginning of the warm sub-chain. It then becomes the first candidate for
eviction, because blocks for replacement always are taken from the beginning
of the warm sub-chain.
The midpoint insertion strategy allows you to keep more valued blocks
always in the cache. If you prefer to use the plain LRU strategy, leave the
key_cache_division_limit value set to its default of 100.
The midpoint insertion strategy helps to improve performance when execution
of a query that requires an index scan effectively pushes out of the cache
all the index blocks corresponding to valuable high level B-tree nodes. To
avoid this, you must use a midpoint insertion strategy with the
key_cache_division_limit set to much less than 100. Then valuable
frequently hit nodes will be preserved in the hot sub-chain during an index
scan operation as well.
If there are enough blocks in a key cache to hold blocks of an entire index, or at least the blocks corresponding to its non-leaf nodes, then it makes sense to preload the key cache with index blocks before starting to use it. Preloading allows you to put the table index blocks into a key cache buffer in the most efficient way: By reading the index blocks from disk sequentially.
Without preloading, the blocks still will be placed into the key cache as needed by queries. Although the blocks will stay in the cache as there are enough buffers for all of them, they will be fetched from disk in a random order, not sequentially.
To preload an index into a cache, use the
LOAD INDEX INTO CACHE
statement. For example,
the following statement preloads nodes (index blocks) of indexes of the
mysql> LOAD INDEX INTO CACHE t1, t2 IGNORE LEAVES; +---------+--------------+----------+----------+ | Table | Op | Msg_type | Msg_text | +---------+--------------+----------+----------+ | test.t1 | preload_keys | status | OK | | test.t2 | preload_keys | status | OK | +---------+--------------+----------+----------+
IGNORE LEAVES modifier causes only blocks for the non-leaf
nodes of the index to be preloaded. Thus,
the statement shown preloads all index blocks from
t1, but only
blocks for the non-leaf nodes from
If an index has been assigned to a key cache using a
CACHE INDEX statement, preloading places index blocks into that cache.
Otherwise, the index is loaded into the default key cache.
MySQL 4.1 introduces a new
key_cache_block_size variable on
a per-key cache basis. This variable specifies the size of the block
buffers for a key cache. It is intended to allow tuning of the performance
of I/O operations for index files.
The best performance for I/O operations is achieved when the size of read buffers is equal to the size of the native operating system I/O buffers. But setting the size of key nodes equal to the size of I/O buffer does not always ensure the best overall performance. When reading the big leaf nodes the server pulls in a lot of unnecessary data, effectively preventing reading other leaf nodes.
Currently, you cannot control the size of the index blocks in a table. This
size is set by the server when the `.MYI' index file is created,
depending on the size of the keys in the indexes present in the table
definition. In most cases, it is set equal to the I/O buffer size. In the
future this will be changed and then
will be fully employed.
A key cache can be restructured at any time by updating its parameter values. For example:
mysql> SET GLOBAL cold_cache.key_buffer_size=4*1024*1024;
If you assign to either the
key_cache_block_size key cache component a value that differs from
the component's current value, the server destroys the cache's old
structure and creates a new one based on the new values. If the cache
contains any dirty blocks, the server saves them to disk before destroying
and re-creating the cache. Restructuring does not occur if you set other key
When restructuring a key cache, the server first flushes the contents of any dirty buffers to disk. After that, the cache contents become unavailable. However, restructuring does not block queries that need to use indexes assigned to the cache. Instead, the server directly accesses the table indexes using native filesystem caching. Filesystem caching is not as efficient as using a key cache, so although queries will execute, a slowdown can be anticipated. Once the cache has been restructured, it becomes available again for caching indexes assigned to it, and the use of filesystem caching for the indexes ceases.
When you execute a
mysqladmin status command, you'll see something
Uptime: 426 Running threads: 1 Questions: 11082 Reloads: 1 Open tables: 12
Open tables value of 12 can be somewhat puzzling if you have
only 6 tables.
MySQL is multi-threaded, so there may be many clients issuing queries for a
given simultaneously. To minimize the problem with multiple client threads
having different states on the same file, the table is opened independently
by each concurrent thread. This takes some memory but normally increases
MyISAM tables, one extra file descriptor is
required for the data file for each client that has the table open. (By
contrast, the index file descriptor is shared between all threads.) The
ISAM storage engine shares this behavior.
You can read more about this topic in the next section. See section 7.4.8 How MySQL Opens and Closes Tables.
system variables affect the maximum number of files the server keeps open.
If you increase one or more of these values, you may run up against a limit
imposed by your operating system on the per-process number of open file
descriptors. Many operating systems allow you to increase the open-files
limit, though the method varies widely from system to system. Consult your
operating system documentation to determine whether it is possible to
increase the limit and how to do so.
table_cache is related to
max_connections. For example,
for 200 concurrent running connections, you should have a table cache size of
200 * n, where
n is the maximum number of tables
in a join. You also need to reserve some extra file descriptors for
temporary tables and files.
Make sure that your operating system can handle the number of open file
descriptors implied by the
table_cache setting. If
table_cache is set too high, MySQL may run out of file
descriptors and refuse connections, fail to perform queries, and be very
unreliable. You also have to take into account that the
engine needs two file descriptors for each unique open table. You can
increase the number of file descriptors available for MySQL with
--open-files-limit startup option to
See section A.2.18 File Not Found.
The cache of open tables will be kept at a level of
entries. The default value is 64; this can be changed with the
--table_cache option to
mysqld). Note that MySQL may
temporarily open even more tables to be able to execute queries.
An unused table is closed and removed from the table cache under the following circumstances:
table_cacheentries and a thread is no longer using a table.
FLUSH TABLESstatement or executes a
When the table cache fills up, the server uses the following procedure to locate a cache entry to use:
When the cache is in a temporarily extended state and a table goes from a used to unused state, the table is closed and released from the cache.
A table is opened for each concurrent access. This means the table needs to be opened twice if two threads access the same table or if a thread accesses the table twice in the same query (for example, by joining the table to itself). The first open of any table takes two file descriptors, one for the data file and one for the index file. Each additional use of the table takes only one file descriptor, for the data file. The index file descriptor is shared among all threads.
If you are opening a table with the
HANDLER tbl_name OPEN
statement, a dedicated table object is allocated for the thread.
This table object is not shared by other threads and is not closed
until the thread calls
HANDLER tbl_name CLOSE or the thread
terminates. When this happens, the table is put back in the table cache
(if the cache isn't full).
See section 14.1.3
You can determine whether your table cache is too small by checking the
mysqld status variable
mysql> SHOW STATUS LIKE 'Opened_tables'; +---------------+-------+ | Variable_name | Value | +---------------+-------+ | Opened_tables | 2741 | +---------------+-------+
If the value is quite big, even when you haven't done a lot of
TABLES, you should increase your table cache size.
See section 126.96.36.199
SHOW STATUS Syntax and
section 5.2.4 Server Status Variables.
If you have many
ISAM tables in a database
directory, open, close, and create operations will be slow. If you execute
SELECT statements on many different tables, there will be a little
overhead when the table cache is full, because for every table that has to
be opened, another must be closed. You can reduce this overhead by making
the table cache larger.
We start with system level factors, because some of these decisions must be made very early to achieve large performance gains. In other cases, a quick look at this section may suffice. However, it is always nice to have a sense for how much can be gained by changing things at this level.
The default operating system to use is really important! To get the best use of multiple-CPU machines, you should use Solaris (because its threads implementation works really well) or Linux (because the 2.2 kernel has really good SMP support). Also, older Linux kernels have a 2GB file-size limit by default. If you have such a kernel and a desperate need for files larger than 2GB, you should get the LFS (large filesystem) patch for the ext2 filesystem. Other filesystems such as ReiserFS and XFS do not have this 2GB limitation.
Because we have not run MySQL in production on that many platforms, we advise you to test your intended platform before choosing it, if possible.
--skip-external-lockingMySQL option to avoid external locking. This option is on by default as of MySQL 4.0. Before that, it is on by default when compiling with MIT-pthreads, because
flock()isn't fully supported by MIT-pthreads on all platforms. It's also on default for Linux as Linux file locking is not yet safe. Note that the
--skip-external-lockingoption will not impact MySQL's functionality as long as you only run one server. Just remember to take down the server (or lock and flush the relevant tables) before you run
myisamchk. On some systems this option is mandatory, because the external locking does not work in any case. The only case when you can't use
--skip-external-lockingis if you run multiple MySQL servers (not clients) on the same data, or if you run
myisamchkto check (not repair) a table without telling the server to flush and lock the tables first. You can still use
UNLOCK TABLESeven if you are using
You can determine the default buffer sizes used by the
with this command (prior to MySQL 4.1, omit
shell> mysqld --verbose --help
This command produces a list of all
mysqld options and configurable
system variables. The output includes the default variable values and
looks something like this:
back_log current value: 5 bdb_cache_size current value: 1048540 binlog_cache_size current value: 32768 connect_timeout current value: 5 delayed_insert_limit current value: 100 delayed_insert_timeout current value: 300 delayed_queue_size current value: 1000 flush_time current value: 0 interactive_timeout current value: 28800 join_buffer_size current value: 131072 key_buffer_size current value: 1048540 long_query_time current value: 10 lower_case_table_names current value: 0 max_allowed_packet current value: 1048576 max_binlog_cache_size current value: 4294967295 max_connect_errors current value: 10 max_connections current value: 100 max_delayed_threads current value: 20 max_heap_table_size current value: 16777216 max_join_size current value: 4294967295 max_sort_length current value: 1024 max_tmp_tables current value: 32 max_write_lock_count current value: 4294967295 myisam_sort_buffer_size current value: 8388608 net_buffer_length current value: 16384 net_read_timeout current value: 30 net_retry_count current value: 10 net_write_timeout current value: 60 read_buffer_size current value: 131072 read_rnd_buffer_size current value: 262144 slow_launch_time current value: 2 sort_buffer current value: 2097116 table_cache current value: 64 thread_concurrency current value: 10 thread_stack current value: 131072 tmp_table_size current value: 1048576 wait_timeout current value: 28800
If there is a
mysqld server currently running, you can see what
values it actually is using for the system variables by connecting to it and
issuing this statement:
mysql> SHOW VARIABLES;
You can also see some statistical and status indicators for a running server by issuing this statement:
mysql> SHOW STATUS;
System variable and status information also can be obtained using
shell> mysqladmin variables shell> mysqladmin extended-status
You can find a full description for all system and status variables in section 5.2.3 Server System Variables and section 5.2.4 Server Status Variables.
MySQL uses algorithms that are very scalable, so you can usually run with very little memory. However, normally you will get better performance by giving MySQL more memory.
When tuning a MySQL server, the two most important variables to configure
table_cache. You should first feel
confident that you have these set appropriately before trying to change
any other variables.
The following examples indicate some typical variable values for different
runtime configurations. The examples use the
mysqld_safe script and
--var_name=value syntax to set the variable
var_name to the value
value. This syntax is available as of MySQL 4.0. For older versions
of MySQL, take the following differences into account:
_size, you may need to specify them without
_size. For example, the old name for
sort_buffer. The old name for
record_buffer. To see which variables your version of the server recognizes, use
If you have at least 256MB of memory and many tables and want maximum performance with a moderate number of clients, you should use something like this:
shell> mysqld_safe --key_buffer_size=64M --table_cache=256 \ --sort_buffer_size=4M --read_buffer_size=1M &
If you have only 128MB of memory and only a few tables, but you still do a lot of sorting, you can use something like this:
shell> mysqld_safe --key_buffer_size=16M --sort_buffer_size=1M
If there are very many simultaneous connections, swapping problems
may occur unless
mysqld has been configured to use very
little memory for each connection.
mysqld performs better
if you have enough memory for all connections.
With little memory and lots of connections, use something like this:
shell> mysqld_safe --key_buffer_size=512K --sort_buffer_size=100K \ --read_buffer_size=100K &
Or even this:
shell> mysqld_safe --key_buffer_size=512K --sort_buffer_size=16K \ --table_cache=32 --read_buffer_size=8K \ -O net_buffer_length=1K &
If you are doing
GROUP BY or
ORDER BY operations on tables
that are much larger than your available memory, you should increase the
read_rnd_buffer_size to speed up the reading of rows after
When you have installed MySQL, the `support-files' directory will contain some different `my.cnf' example files, `my-huge.cnf', `my-large.cnf', `my-medium.cnf', and `my-small.cnf'. You can use these as a basis for optimizing your system.
Note that if you specify an option on the command line for
mysqld_safe, it remains in effect only for that invocation of the server.
To use the option every time the server runs, put it in an option file.
To see the effects of a parameter change, do something like this (prior to
MySQL 4.1, omit
shell> mysqld --key_buffer_size=32m --verbose --help
The variable values are listed near the end of the output.
Make sure that the
--help options are
last. Otherwise, the effect of any options listed after them on the command
line will not be reflected in the output.
Most of the following tests were performed on Linux with the MySQL benchmarks, but they should give some indication for other operating systems and workloads.
You get the fastest executables when you link with
On Linux, you will get the fastest code when compiling with
-O3. You need about 200MB memory to compile `sql_yacc.cc'
with these options, because
gcc/pgcc needs a lot of memory to
make all functions inline. You should also set
configuring MySQL to avoid inclusion of the
library (it is not needed). Note that with some versions of
the resulting code will run only on true Pentium processors, even if you
use the compiler option indicating that you want the resulting code to work on
all x586-type processors (such as AMD).
By just using a better compiler and better compiler options, you can get a 10-30% speed increase in your application. This is particularly important if you compile the SQL server yourself.
We have tested both the Cygnus CodeFusion and Fujitsu compilers, but when we tested them, neither was sufficiently bug free to allow MySQL to be compiled with optimizations on.
The standard MySQL binary distributions are compiled with support for all
character sets. When you compile MySQL yourself, you should include support
only for the character sets that you are going to use. This is controlled by
Here is a list of some measurements that we have made:
pgccand compile everything with
mysqldserver is 1% faster than with
-static), the result is 13% slower on Linux. Note that you still can use a dynamically linked MySQL library for your client applications. It is the server that is most critical for performance.
strip mysqld, the resulting binary can be up to 4% faster.
localhost, MySQL uses a socket file by default.)
--with-debug=full, most queries will be 20% slower. Some queries may take substantially longer; for example, the MySQL benchmarks ran 35% slower. If you use
=full), the slowdown will be only 15%. For a version of
mysqldthat has been compiled with
--with-debug=full, you can disable memory checking at runtime by starting it with the
--skip-safemallocoption. The execution speed should then be close to when configuring with
gcc2.95.2 for UltraSPARC with the option
-mcpu=v8 -Wa,-xarch=v8plusagives 4% more performance.
gccwithout frame pointers
Binary MySQL distributions for Linux that are provided by MySQL AB used
to be compiled with
pgcc. We had to go back to regular
due to a bug in
pgcc that would generate code that does
not run on AMD. We will continue using
gcc until that bug is resolved.
In the meantime, if you have a non-AMD machine, you can get a faster
binary by compiling with
pgcc. The standard MySQL
Linux binary is linked statically to make it faster and more portable.
The following list indicates some of the ways that the
uses memory. Where applicable, the name of the system variable relevant
to the memory use is given:
key_buffer_size) is shared by all threads; other buffers used by the server are allocated as needed. See section 7.5.2 Tuning Server Parameters.
thread_stack), a connection buffer (variable
net_buffer_length), and a result buffer (variable
net_buffer_length). The connection buffer and result buffer are dynamically enlarged up to
max_allowed_packetwhen needed. While a query is running, a copy of the current query string is also allocated.
MyISAMtables are memory mapped. This is because the 32-bit memory space of 4GB is not large enough for most big tables. When systems with a 64-bit address space become more common, we may add general support for memory mapping.
HEAP) tables. Temporary tables with a large record length (calculated as the sum of all column lengths) or that contain
BLOBcolumns are stored on disk. One problem in MySQL before Version 3.23.2 is that if an internal in-memory heap table exceeds the size of
tmp_table_size, the error
The table tbl_name is fulloccurs. From 3.23.2 on, this is handled automatically by changing the in-memory heap table to a disk-based
MyISAMtable as necessary. To work around this problem for older servers, you can increase the temporary table size by setting the
mysqld, or by setting the SQL option
SQL_BIG_TABLESin the client program. See section 188.8.131.52
SETSyntax. In MySQL Version 3.20, the maximum size of the temporary table is
record_buffer*16; if you are using this version, you have to increase the value of
record_buffer. You can also start
--big-tablesoption to always store temporary tables on disk. However, this will affect the speed of many complicated queries.
ISAMtable that is opened, the index file is opened once and the data file is opened once for each concurrently running thread. For each concurrent thread, a table structure, column structures for each column, and a buffer of size
3 * nis allocated (where
nis the maximum row length, not counting
BLOBcolumn requires five to eight bytes plus the length of the
ISAMstorage engines maintain one extra row buffer for internal use.
BLOBcolumns, a buffer is enlarged dynamically to read in larger
BLOBvalues. If you scan a table, a buffer as large as the largest
BLOBvalue is allocated.
FLUSH TABLESstatement or
mysqladmin flush-tablescommand closes all tables that are not in use and marks all in-use tables to be closed when the currently executing thread finishes. This effectively frees most in-use memory.
ps and other system status programs may report that
uses a lot of memory. This may be caused by thread-stacks on different
memory addresses. For example, the Solaris version of
the unused memory between stacks as used memory. You can verify this by
checking available swap with
swap -s. We have tested
mysqld with commercial memory-leakage detectors, so there should
be no memory leaks.
When a new client connects to
mysqld spawns a
new thread to handle the request. This thread first checks if the
hostname is in the hostname cache. If not, the thread attempts to resolve the
gethostbyname_r()calls, the thread uses them to perform hostname resolution.
gethostbyname()instead. In this case, no other thread can resolve hostnames that are not in the hostname cache until the first thread unlocks the mutex.
You can disable DNS hostname lookups by starting
mysqld with the
--skip-name-resolve option. However, in this case, you can use only IP
numbers in the MySQL grant tables.
If you have a very slow DNS and many hosts, you can get more performance by
either disabling DNS lookups with
--skip-name-resolve or by
HOST_CACHE_SIZE define (default value: 128) and
You can disable the hostname cache by starting the server with the
--skip-host-cache option. To clear the hostname cache, issue a
FLUSH HOSTS statement or execute the
If you want to disallow TCP/IP connections entirely, start
MyISAMtables, you symlink the index file and/or data file from their usual location in the data directory to another disk (that may also be striped). This makes both the seek and read times better, assuming that the disk is not used for other purposes as well. See section 7.6.1 Using Symbolic Links.
hdparmto configure your disk's interface. (Up to 100% under load is not uncommon.) The following
hdparmoptions should be quite good for MySQL, and probably for many other applications:
hdparm -m 16 -d 1Note that performance and reliability when using this command depends on your hardware, so we strongly suggest that you test your system thoroughly after using
hdparm. Please consult the
hdparmmanual page for more information. If
hdparmis not used wisely, filesystem corruption may result, so back up everything before experimenting!
-o noatimeoption. That skips updates to the last access time in inodes on the filesystem, which avoids some disk seeks. On many operating systems, you can set a filesystem to be updated asynchronously by mounting it with the
-o asyncoption. If your computer is reasonably stable, this should give you more performance without sacrificing too much reliability. (This flag is on by default on Linux.)
You can move tables and databases from the database directory to other locations and replace them with symbolic links to the new locations. You might want to do this, for example, to move a database to a file system with more free space or increase the speed of your system by spreading your tables to different disk.
The recommended way to do this is to just symlink databases to a different disk. Symlink tables only as a last resort.
On Unix, the way to symlink a database is to first create a directory on some disk where you have free space and then create a symlink to it from the MySQL data directory.
shell> mkdir /dr1/databases/test shell> ln -s /dr1/databases/test /path/to/datadir
MySQL doesn't support linking one directory to multiple
databases. Replacing a database directory with a symbolic link will
work fine as long as you don't make a symbolic link between databases.
Suppose that you have a database
db1 under the MySQL data
directory, and then make a symlink
db2 that points to
shell> cd /path/to/datadir shell> ln -s db1 db2
Now, for any table
db1, there also appears to be
db2. If one thread updates
and another thread updates
db2.tbl_a, there will be problems.
If you really need to do this, you must change the following code in `mysys/mf_format.c':
if (flag & 32 || (!lstat(to,&stat_buff) && S_ISLNK(stat_buff.st_mode)))
On Windows, you can use internal symbolic links to directories by compiling
-DUSE_SYMDIR. This allows you to put different
databases on different disks. See section 184.108.40.206 Using Symbolic Links for Databases on Windows.
Before MySQL 4.0 you should not symlink tables unless you are
very careful with them. The problem is that if you run
REPAIR TABLE, or
OPTIMIZE TABLE on a symlinked
table, the symlinks will be removed and replaced by the original
files. This happens because these statements work by creating a
temporary file in the database directory and replacing the original
file with the temporary file when the statement operation is complete.
You should not symlink tables on systems that don't have a fully
realpath() call. (At least Linux and Solaris support
realpath()). You can check whether your system supports symbolic links
SHOW VARIABLES LIKE 'have_symlink'.
In MySQL 4.0, symlinks are fully supported only for
tables. For other table types, you will probably get strange problems
if you try to use symbolic links on files in the operating system with
any of the preceding statements.
The handling of symbolic links for
MyISAM tables in MySQL 4.0 works
the following way:
mysqldis not running. With SQL, you can instruct the server to perform the symlinking by using the
INDEX DIRECTORYoptions to
CREATE TABLE. See section 14.2.5
myisamchkwill not replace a symlink with the data file or index file. It works directly on the file a symlink points to. Any temporary files are created in the directory where the data file or index file is located.
rootor allow users to have write access to the MySQL database directories.
ALTER TABLE RENAMEand you don't move the table to another database, the symlinks in the database directory are renamed to the new names and the data file and index file are renamed accordingly.
ALTER TABLE RENAMEto move a table to another database, the table is moved to the other database directory and the old symlinks and the files to which they pointed are deleted. (In other words, the new table will not be symlinked.)
mysqldto ensure that no one can use
mysqldto drop or rename a file outside of the data directory.
SHOW CREATE TABLE doesn't report if the table has symbolic links
prior to MySQL 4.0.15. This is also true for
mysqldump, which uses
SHOW CREATE TABLE to generate
CREATE TABLE statements.
Table symlink operations that are not yet supported:
ALTER TABLEignores the
INDEX DIRECTORYtable options.
RESTORE TABLEdon't respect symbolic links.
db1under the MySQL data directory, a table
tbl1in this database, and in the
db1directory you make a symlink
tbl2that points to
shell> cd /path/to/datadir/db1 shell> ln -s tbl1.frm tbl2.frm shell> ln -s tbl1.MYD tbl2.MYD shell> ln -s tbl1.MYI tbl2.MYINow there will be problems if one thread reads
db1.tbl1and another thread updates
tbl1has not been updated so will return out-of-date results).
tbl2will also fail.
Beginning with MySQL Version 3.23.16, the
mysql-max-nt servers in the MySQL distribution are
compiled with the
-DUSE_SYMDIR option. This allows you to put a
database directory on a different disk by setting up a symbolic link to it.
This is similar to the way that symbolic links work on Unix, though the
procedure for setting up the link is different.
On Windows, you make a symbolic link to a MySQL database by creating a file
in the data directory that contains the path to the destination directory.
The file should be named `db_name.sym', where
db_name is the
For example, if the MySQL data directory is `C:\mysql\data'
and you want to have database
foo located at `D:\data\foo',
you should create the file `C:\mysql\data\foo.sym' that contains
D:\data\foo\. After that, all tables created in
foo will be created in `D:\data\foo'.
(The `D:\data\foo' directory must already exist for this to work.) Also,
note that the symbolic link will not be used if a directory with
the database name exists in the MySQL data directory. This means
that if you already have a database directory named `foo' in
the data directory, you must move it to `D:\data' before the
symbolic link will be effective. (To avoid problems, the server
should not be running when you move the database directory.)
In MySQL 4.0, symbolic links are enabled by default. If you don't need them,
you can disable them with the
Before MySQL 4.0, symbolic links are disabled by default. To enable them, you should put the following entry in your `my.cnf' or `my.ini' file:
Go to the first, previous, next, last section, table of contents.