Table of Contents
The MySQL Performance Schema is a feature for monitoring MySQL Server execution at a low level. The Performance Schema is available as of MySQL 5.5.3 and has these characteristics:
The Performance Schema provides a way to inspect internal
execution of the server at runtime. It is implemented using the
PERFORMANCE_SCHEMA
storage engine
and the performance_schema
database. The
Performance Schema focuses primarily on performance data. This
differs from INFORMATION_SCHEMA
, which serves
for inspection of metadata.
The Performance Schema monitors server events. An “event” is anything the server does that takes time and has been instrumented so that timing information can be collected. In general, an event could be a function call, a wait for the operating system, a stage of an SQL statement execution such as parsing or sorting, or an entire statement or group of statements. Event collection provides access to information about synchronization calls (such as for mutexes) and file I/O calls for the server and for several storage engines.
Performance Schema events are distinct from events written to the server's binary log (which describe data modifications) and Event Scheduler events (which are a type of stored program).
Current events are available, as well as event histories and summaries. This enables you to determine how many times instrumented activities were performed and how much time they took. Event information is available to show the activities of specific threads, or activity associated with particular objects such as a mutex or file.
The PERFORMANCE_SCHEMA
storage
engine collects event data using “instrumentation
points” in server source code.
Collected events are stored in tables in the
performance_schema
database. These tables can
be queried using SELECT
statements like other tables.
Performance Schema configuration can be modified dynamically by
updating tables in the performance_schema
database through SQL statements. Configuration changes affect
data collection immediately.
Tables in the performance_schema
database are
views or temporary tables that use no persistent on-disk
storage.
Monitoring is available on all platforms supported by MySQL.
Some limitations might apply: The types of timers might vary per platform. Instruments that apply to storage engines might not be implemented for all storage engines. Instrumentation of each third-party engine is the responsibility of the engine maintainer. See also Section C.8, “Restrictions on Performance Schema”.
Data collection is implemented by modifying the server source code to add instrumentation. There are no separate threads associated with the Performance Schema, unlike other features such as replication or the Event Scheduler.
The Performance Schema is intended to provide access to useful information about server execution while having minimal impact on server performance. The implementation follows these design goals:
Activating the Performance Schema causes no changes in server
behavior. For example, it does not cause thread scheduling to
change, and it does not cause query execution plans (as shown by
EXPLAIN
) to change.
No memory allocation is done beyond that which occurs during server startup. By using early allocation of structures with a fixed size, it is never necessary to resize or reallocate them, which is critical for achieving good runtime performance.
Server monitoring occurs continuously and unobtrusively with very little overhead. Activating the Performance Schema does not make the server unusable.
The parser is unchanged. There are no new keywords or statements.
Execution of server code proceeds normally even if the Performance Schema fails internally.
When there is a choice between performing processing during event collection initially or during event retrieval later, priority is given to making collection faster. This is because collection is ongoing whereas retrieval is on demand and might never happen at all.
It is easy to add new instrumentation points.
Instrumentation is versioned. If the instrumentation implementation changes, previously instrumented code will continue to work. This benefits developers of third-party plugins because it is not necessary to upgrade each plugin to stay synchronized with the latest Performance Schema changes.
This section briefly introduces the Performance Schema with examples that show how to use it. For additional examples, see Section 22.12, “Using the Performance Schema to Diagnose Problems”.
For the Performance Schema to be available, support for it must
have been configured when MySQL was built. You can verify whether
this is the case by checking the server's help output. If the
Performance Schema is available, the output will mention several
variables with names that begin with
performance_schema
:
shell> mysqld --verbose --help
...
--performance_schema
Enable the performance schema.
--performance_schema_events_waits_history_long_size=#
Number of rows in events_waits_history_long.
...
If such variables do not appear in the output, your server has not been built to support the Performance Schema. In this case, see Section 22.2, “Performance Schema Configuration”.
Assuming that the Performance Schema is available, it is disabled
by default. To enable it, start the server with the
performance_schema
variable
enabled. For example, use these lines in your
my.cnf
file:
[mysqld] performance_schema
When the server starts, it sees
performance_schema
and attempts
to initialize the Performance Schema. To verify successful
initialization, use this statement:
mysql> SHOW VARIABLES LIKE 'performance_schema';
+--------------------+-------+
| Variable_name | Value |
+--------------------+-------+
| performance_schema | ON |
+--------------------+-------+
A value of ON
means that the Performance Schema
initialized successfully and is ready for use. A value of
OFF
means that some error occurred. Check the
server error log for information about what went wrong.
The Performance Schema is implemented as a storage engine. If this
engine is available (which you should already have checked
earlier), you should see it listed with a
SUPPORT
value of YES
in the
output from the
INFORMATION_SCHEMA.ENGINES
table or
the SHOW ENGINES
statement:
mysql>SELECT * FROM INFORMATION_SCHEMA.ENGINES
->WHERE ENGINE='PERFORMANCE_SCHEMA'\G
*************************** 1. row *************************** ENGINE: PERFORMANCE_SCHEMA SUPPORT: YES COMMENT: Performance Schema TRANSACTIONS: NO XA: NO SAVEPOINTS: NO mysql>SHOW ENGINES\G
... Engine: PERFORMANCE_SCHEMA Support: YES Comment: Performance Schema Transactions: NO XA: NO Savepoints: NO ...
The PERFORMANCE_SCHEMA
storage engine
operates on tables in the performance_schema
database. You can make performance_schema
the
default database so that references to its tables need not be
qualified with the database name:
mysql> USE performance_schema;
Many examples in this chapter assume
performance_schema
as the default database.
Performance Schema tables are stored in the
performance_schema
database. Information about
the structure of this database and its tables can be obtained, as
for any other database, by selecting from the
INFORMATION_SCHEMA
database or by using
SHOW
statements. For example, use
either of these statements to see what Performance Schema tables
exist:
mysql>SELECT TABLE_NAME FROM INFORMATION_SCHEMA.TABLES
->WHERE TABLE_SCHEMA = 'performance_schema';
+----------------------------------------------+ | TABLE_NAME | +----------------------------------------------+ | cond_instances | | events_waits_current | | events_waits_history | | events_waits_history_long | | events_waits_summary_by_instance | | events_waits_summary_by_thread_by_event_name | | events_waits_summary_global_by_event_name | | file_instances | | file_summary_by_event_name | | file_summary_by_instance | | mutex_instances | | performance_timers | | rwlock_instances | | setup_consumers | | setup_instruments | | setup_timers | | threads | +----------------------------------------------+ mysql>SHOW TABLES FROM performance_schema;
+----------------------------------------------+ | Tables_in_performance_schema | +----------------------------------------------+ | cond_instances | | events_waits_current | | events_waits_history | ...
The number of Performance Schema tables is expected to increase over time as implementation of additional instrumentation proceeds.
The name of the performance_schema
database is
lowercase, as are the names of tables within it. Queries should
specify the names in lowercase.
Before MySQL 5.5.8, the table names were uppercase, which caused
problems on some systems for certain values of the
lower_case_table_names
system
variable.
To see the structure of individual tables, use SHOW
CREATE TABLE
:
mysql> SHOW CREATE TABLE setup_timers\G
*************************** 1. row ***************************
Table: setup_timers
Create Table: CREATE TABLE `setup_timers` (
`NAME` varchar(64) NOT NULL,
`TIMER_NAME` enum('CYCLE','NANOSECOND','MICROSECOND','MILLISECOND','TICK')
NOT NULL
) ENGINE=PERFORMANCE_SCHEMA DEFAULT CHARSET=utf8
Table structure is also available by selecting from tables such as
INFORMATION_SCHEMA.COLUMNS
or by
using statements such as SHOW COLUMNS
.
Tables in the performance_schema
database can
be grouped according to the type of information in them: Current
events, event histories and summaries, object instances, and setup
(configuration) information. The following examples illustrate a
few uses for these tables. For detailed information about the
tables in each group, see
Section 22.7, “Performance Schema Table Descriptions”.
To see what the server is doing at the moment, examine the
events_waits_current
table. It
contains one row per thread showing each thread's most recent
monitored event:
mysql> SELECT * FROM events_waits_current\G
*************************** 1. row ***************************
THREAD_ID: 0
EVENT_ID: 5523
EVENT_NAME: wait/synch/mutex/mysys/THR_LOCK::mutex
SOURCE: thr_lock.c:525
TIMER_START: 201660494489586
TIMER_END: 201660494576112
TIMER_WAIT: 86526
SPINS: NULL
OBJECT_SCHEMA: NULL
OBJECT_NAME: NULL
OBJECT_TYPE: NULL
OBJECT_INSTANCE_BEGIN: 142270668
NESTING_EVENT_ID: NULL
OPERATION: lock
NUMBER_OF_BYTES: NULL
FLAGS: 0
...
This event indicates that thread 0 was waiting for 86,526
picoseconds to acquire a lock on
THR_LOCK::mutex
, a mutex in the
mysys
subsystem. The first few columns provide
the following information:
The ID columns indicate which thread the event comes from and the event number.
EVENT_NAME
indicates what was instrumented
and SOURCE
indicates which source file
contains the instrumented code.
The timer columns show when the event started and stopped and
how long it took. If an event is still in progress, the
TIMER_END
and TIMER_WAIT
values are NULL
. Timer values are
approximate and expressed in picoseconds. For information
about timers and event time collection, see
Section 22.2.3.1, “Performance Schema Event Timing”.
The history tables contain the same kind of rows as the
current-events table but have more rows and show what the server
has been doing “recently” rather than
“currently.” The
events_waits_history
and
events_waits_history_long
tables
contain the most recent 10 events per thread and most recent
10,000 events, respectively. For example, to see information for
recent events produced by thread 13, do this:
mysql>SELECT EVENT_ID, EVENT_NAME, TIMER_WAIT
->FROM events_waits_history WHERE THREAD_ID = 13
->ORDER BY EVENT_ID;
+----------+-----------------------------------------+------------+ | EVENT_ID | EVENT_NAME | TIMER_WAIT | +----------+-----------------------------------------+------------+ | 86 | wait/synch/mutex/mysys/THR_LOCK::mutex | 686322 | | 87 | wait/synch/mutex/mysys/THR_LOCK_malloc | 320535 | | 88 | wait/synch/mutex/mysys/THR_LOCK_malloc | 339390 | | 89 | wait/synch/mutex/mysys/THR_LOCK_malloc | 377100 | | 90 | wait/synch/mutex/sql/LOCK_plugin | 614673 | | 91 | wait/synch/mutex/sql/LOCK_open | 659925 | | 92 | wait/synch/mutex/sql/THD::LOCK_thd_data | 494001 | | 93 | wait/synch/mutex/mysys/THR_LOCK_malloc | 222489 | | 94 | wait/synch/mutex/mysys/THR_LOCK_malloc | 214947 | | 95 | wait/synch/mutex/mysys/LOCK_alarm | 312993 | +----------+-----------------------------------------+------------+
As new events are added to a history table, older events are discarded if the table is full.
Summary tables provide aggregated information for all events over
time. The tables in this group summarize event data in different
ways. To see which instruments have been executed the most times
or have taken the most wait time, sort the
events_waits_summary_global_by_event_name
table on the COUNT_STAR
or
SUM_TIMER_WAIT
column, which correspond to a
COUNT(*)
or SUM(TIMER_WAIT)
value, respectively, calculated over all events:
mysql>SELECT EVENT_NAME, COUNT_STAR
->FROM events_waits_summary_global_by_event_name
->ORDER BY COUNT_STAR DESC LIMIT 10;
+---------------------------------------------------+------------+ | EVENT_NAME | COUNT_STAR | +---------------------------------------------------+------------+ | wait/synch/mutex/mysys/THR_LOCK_malloc | 6419 | | wait/io/file/sql/FRM | 452 | | wait/synch/mutex/sql/LOCK_plugin | 337 | | wait/synch/mutex/mysys/THR_LOCK_open | 187 | | wait/synch/mutex/mysys/LOCK_alarm | 147 | | wait/synch/mutex/sql/THD::LOCK_thd_data | 115 | | wait/io/file/myisam/kfile | 102 | | wait/synch/mutex/sql/LOCK_global_system_variables | 89 | | wait/synch/mutex/mysys/THR_LOCK::mutex | 89 | | wait/synch/mutex/sql/LOCK_open | 88 | +---------------------------------------------------+------------+ mysql>SELECT EVENT_NAME, SUM_TIMER_WAIT
->FROM events_waits_summary_global_by_event_name
->ORDER BY SUM_TIMER_WAIT DESC LIMIT 10;
+----------------------------------------+----------------+ | EVENT_NAME | SUM_TIMER_WAIT | +----------------------------------------+----------------+ | wait/io/file/sql/MYSQL_LOG | 1599816582 | | wait/synch/mutex/mysys/THR_LOCK_malloc | 1530083250 | | wait/io/file/sql/binlog_index | 1385291934 | | wait/io/file/sql/FRM | 1292823243 | | wait/io/file/myisam/kfile | 411193611 | | wait/io/file/myisam/dfile | 322401645 | | wait/synch/mutex/mysys/LOCK_alarm | 145126935 | | wait/io/file/sql/casetest | 104324715 | | wait/synch/mutex/sql/LOCK_plugin | 86027823 | | wait/io/file/sql/pid | 72591750 | +----------------------------------------+----------------+
These results show that the THR_LOCK_malloc
mutex is “hot,” both in terms of how often it is used
and amount of time that threads wait attempting to acquire it.
The THR_LOCK_malloc
mutex is used only in
debug builds. In production builds it is not hot because it is
nonexistent.
Instance tables document what types of objects are instrumented.
An instrumented object, when used by the server, produces an
event. These tables provide event names and explanatory notes or
status information. For example, the
file_instances
table lists instances
of instruments for file I/O operations and their associated files:
mysql> SELECT * FROM file_instances\G
*************************** 1. row ***************************
FILE_NAME: /opt/mysql-log/60500/binlog.000007
EVENT_NAME: wait/io/file/sql/binlog
OPEN_COUNT: 0
*************************** 2. row ***************************
FILE_NAME: /opt/mysql/60500/data/mysql/tables_priv.MYI
EVENT_NAME: wait/io/file/myisam/kfile
OPEN_COUNT: 1
*************************** 3. row ***************************
FILE_NAME: /opt/mysql/60500/data/mysql/columns_priv.MYI
EVENT_NAME: wait/io/file/myisam/kfile
OPEN_COUNT: 1
...
Setup tables are used to configure and display monitoring
characteristics. For example, to see which event timer is
selected, query the setup_timers
tables:
mysql> SELECT * FROM setup_timers;
+------+------------+
| NAME | TIMER_NAME |
+------+------------+
| wait | CYCLE |
+------+------------+
setup_instruments
lists the set of
instruments for which events can be collected and shows which of
them are enabled:
mysql> SELECT * FROM setup_instruments;
+------------------------------------------------------------+---------+-------+
| NAME | ENABLED | TIMED |
+------------------------------------------------------------+---------+-------+
...
| wait/synch/mutex/sql/LOCK_global_read_lock | YES | YES |
| wait/synch/mutex/sql/LOCK_global_system_variables | YES | YES |
| wait/synch/mutex/sql/LOCK_lock_db | YES | YES |
| wait/synch/mutex/sql/LOCK_manager | YES | YES |
...
| wait/synch/rwlock/sql/LOCK_grant | YES | YES |
| wait/synch/rwlock/sql/LOGGER::LOCK_logger | YES | YES |
| wait/synch/rwlock/sql/LOCK_sys_init_connect | YES | YES |
| wait/synch/rwlock/sql/LOCK_sys_init_slave | YES | YES |
...
| wait/io/file/sql/binlog | YES | YES |
| wait/io/file/sql/binlog_index | YES | YES |
| wait/io/file/sql/casetest | YES | YES |
| wait/io/file/sql/dbopt | YES | YES |
...
To understand how to interpret instrument names, see Section 22.4, “Performance Schema Instrument Naming Conventions”.
To control whether events are collected for an instrument, set its
ENABLED
value to YES
or
NO
. For example:
mysql>UPDATE setup_instruments SET ENABLED = 'NO'
->WHERE NAME = 'wait/synch/mutex/sql/LOCK_mysql_create_db';
The Performance Schema uses collected events to update tables in
the performance_schema
database, which act as
“consumers” of event information. The
setup_consumers
table lists the
available consumers and shows which of them are enabled:
mysql> SELECT * FROM setup_consumers;
+----------------------------------------------+---------+
| NAME | ENABLED |
+----------------------------------------------+---------+
| events_waits_current | YES |
| events_waits_history | YES |
| events_waits_history_long | YES |
| events_waits_summary_by_thread_by_event_name | YES |
| events_waits_summary_by_event_name | YES |
| events_waits_summary_by_instance | YES |
| file_summary_by_event_name | YES |
| file_summary_by_instance | YES |
+----------------------------------------------+---------+
To control whether the Performance Schema maintains a consumer as
a destination for event information, set its
ENABLED
value.
For more information about the setup tables and how to use them to control event collection, see Section 22.2.3.2, “Performance Schema Event Filtering”.
There are some miscellaneous tables that do not fall into any of
the previous groups. For example,
performance_timers
lists the
available event timers and their characteristics. For information
about timers, see Section 22.2.3.1, “Performance Schema Event Timing”.
To use the MySQL Performance Schema, these configuration considerations apply:
The Performance Schema must be configured into MySQL Server at build time to make it available. Performance Schema support is included in binary MySQL distributions. If you are building from source, you must ensure that it is configured into the build as described in Section 22.2.1, “Performance Schema Build Configuration”.
The Performance Schema must be enabled at server startup to enable event collection to occur. Specific Performance Schema features can be enabled at server startup or at runtime to control which types of event collection occur. See Section 22.2.2, “Performance Schema Startup Configuration”, Section 22.2.3, “Performance Schema Runtime Configuration”, and Section 22.2.3.2, “Performance Schema Event Filtering”.
For the Performance Schema to be available, it must be configured into the MySQL server at build time. Binary MySQL distributions provided by Oracle Corporation are configured to support the Performance Schema. If you use a binary MySQL distribution from another provider, check with the provider whether the distribution has been appropriately configured.
If you build MySQL from a source distribution, enable the
Performance Schema by running CMake with the
WITH_PERFSCHEMA_STORAGE_ENGINE
option enabled:
shell> cmake . -DWITH_PERFSCHEMA_STORAGE_ENGINE=1
Configuring MySQL with the
-DWITHOUT_PERFSCHEMA_STORAGE_ENGINE=1
option prevents inclusion of the Performance Schema, so if you
want it included, do not use this option. See
Section 2.9.4, “MySQL Source-Configuration Options”.
If you install MySQL over a previous installation that was
configured without the Performance Schema (or with an older
version of the Performance Schema that may not have all the
current tables), run mysql_upgrade after
starting the server to ensure that the
performance_schema
database exists with all
current tables. Then restart the server. One indication that you
need to do this is the presence of messages such as the
following in the error log:
[ERROR] Native table 'performance_schema'.'events_waits_history' has the wrong structure [ERROR] Native table 'performance_schema'.'events_waits_history_long' has the wrong structure ...
To verify whether a server was built with Performance Schema
support, check its help output. If the Performance Schema is
available, the output will mention several variables with names
that begin with performance_schema
:
shell> mysqld --verbose --help
...
--performance_schema
Enable the performance schema.
--performance_schema_events_waits_history_long_size=#
Number of rows in events_waits_history_long.
...
You can also connect to the server and look for a line that
names the PERFORMANCE_SCHEMA
storage engine in the output from SHOW
ENGINES
:
mysql> SHOW ENGINES\G
...
Engine: PERFORMANCE_SCHEMA
Support: YES
Comment: Performance Schema
Transactions: NO
XA: NO
Savepoints: NO
...
If the Performance Schema was not configured into the server at
build time, no row for
PERFORMANCE_SCHEMA
will appear in
the output from SHOW ENGINES
. You
might see performance_schema
listed in the
output from SHOW DATABASES
, but
it will have no tables and you will not be able to use it.
A line for PERFORMANCE_SCHEMA
in
the SHOW ENGINES
output means
that the Performance Schema is available, not that it is
enabled. To enable it, you must do so at server startup, as
described in the next section.
The Performance Schema is disabled by default. To enable it,
start the server with the
performance_schema
variable
enabled. For example, use these lines in your
my.cnf
file:
[mysqld] performance_schema
If the server is unable to allocate any internal buffer during
Performance Schema initialization, the Performance Schema
disables itself and sets
performance_schema
to
OFF
, and the server runs without
instrumentation.
The Performance Schema includes several system variables that provide configuration information:
mysql> SHOW VARIABLES LIKE 'perf%';
+---------------------------------------------------+---------+
| Variable_name | Value |
+---------------------------------------------------+---------+
| performance_schema | ON |
| performance_schema_events_waits_history_long_size | 10000 |
| performance_schema_events_waits_history_size | 10 |
| performance_schema_max_cond_classes | 80 |
| performance_schema_max_cond_instances | 1000 |
| performance_schema_max_file_classes | 50 |
| performance_schema_max_file_handles | 32768 |
| performance_schema_max_file_instances | 10000 |
| performance_schema_max_mutex_classes | 200 |
| performance_schema_max_mutex_instances | 1000000 |
| performance_schema_max_rwlock_classes | 30 |
| performance_schema_max_rwlock_instances | 1000000 |
| performance_schema_max_table_handles | 100000 |
| performance_schema_max_table_instances | 50000 |
| performance_schema_max_thread_classes | 50 |
| performance_schema_max_thread_instances | 1000 |
+---------------------------------------------------+---------+
The performance_schema
variable
is ON
or OFF
to indicate
whether the Performance Schema is enabled or disabled. The other
variables indicate table sizes (number of rows) or memory
allocation values.
With the Performance Schema enabled, the number of Performance Schema instances affects the server memory footprint, perhaps to a large extent. It may be necessary to tune the values of Performance Schema system variables to find the number of instances that balances insufficient instrumentation against excessive memory consumption.
To change the value of Performance Schema system variables, set
them at server startup. For example, put the following lines in
a my.cnf
file to change the sizes of the
history tables for wait events:
[mysqld] performance_schema performance_schema_events_waits_history_size=20 performance_schema_events_waits_history_long_size=15000
Performance Schema setup tables contain information about monitoring configuration:
mysql>SELECT TABLE_NAME FROM INFORMATION_SCHEMA.TABLES
->WHERE TABLE_SCHEMA = 'performance_schema'
->AND TABLE_NAME LIKE 'setup%';
+-------------------+ | TABLE_NAME | +-------------------+ | setup_consumers | | setup_instruments | | setup_timers | +-------------------+
You can examine the contents of these tables to obtain
information about Performance Schema monitoring characteristics.
If you have the UPDATE
privilege,
you can change Performance Schema operation by modifying setup
tables to affect how monitoring occurs. For additional details
about these tables, see
Section 22.7.2, “Performance Schema Setup Tables”.
To see which event timer is selected, query the
setup_timers
tables:
mysql> SELECT * FROM setup_timers;
+------+------------+
| NAME | TIMER_NAME |
+------+------------+
| wait | CYCLE |
+------+------------+
The NAME
value indicates the type of
instrument to which the timer applies, and
TIMER_NAME
indicates which timer applies to
those instruments. The timer applies to instruments where their
name begins with a component matching the
NAME
value. There are only
“wait” instruments, so this table has only one row
and the timer applies to all instruments.
To change the timer, update the NAME
value.
For example, to use the NANOSECOND
timer:
mysql>UPDATE setup_timers SET TIMER_NAME = 'NANOSECOND'
->WHERE NAME = 'wait';
mysql>SELECT * FROM setup_timers;
+------+------------+ | NAME | TIMER_NAME | +------+------------+ | wait | NANOSECOND | +------+------------+
For discussion of timers, see Section 22.2.3.1, “Performance Schema Event Timing”.
The setup_instruments
and
setup_consumers
tables list the
instruments for which events can be collected and the types of
consumers for which event information actually is collected,
respectively. Section 22.2.3.2, “Performance Schema Event Filtering”,
discusses how you can modify these tables to affect event
collection.
If there are Performance Schema configuration changes that must
be made at runtime using SQL statements and you would like these
changes to take effect each time the server starts, put the
statements in a file and start the server with the
--init-file=
option. This strategy can also be useful if you have multiple
monitoring configurations, each tailored to produce a different
kind of monitoring, such as casual server health monitoring,
incident investigation, application behavior troubleshooting,
and so forth. Put the statements for each monitoring
configuration into their own file and specify the appropriate
file as the file_name
--init-file
argument
when you start the server.
Events are collected by means of instrumentation added to the server source code. Instruments time events, which is how the Performance Schema provides an idea of how long events take. It is also possible to configure instruments not to collect timing information. This section discusses the available timers and their characteristics, and how timing values are represented in events.
Two Performance Schema tables provide timer information:
performance_timers
lists
the available timers and their characteristics.
setup_timers
indicates
which timers are used for which instruments.
Each timer row in setup_timers
must refer to one of the timers listed in
performance_timers
.
Timers vary in precision and amount of overhead. To see what
timers are available and their characteristics, check the
performance_timers
table:
mysql> SELECT * FROM performance_timers;
+-------------+-----------------+------------------+----------------+
| TIMER_NAME | TIMER_FREQUENCY | TIMER_RESOLUTION | TIMER_OVERHEAD |
+-------------+-----------------+------------------+----------------+
| CYCLE | 2389029850 | 1 | 72 |
| NANOSECOND | 1000000000 | 1 | 112 |
| MICROSECOND | 1000000 | 1 | 136 |
| MILLISECOND | 1036 | 1 | 168 |
| TICK | 105 | 1 | 2416 |
+-------------+-----------------+------------------+----------------+
The columns have these meanings:
The TIMER_NAME
column shows the names
of the available timers. CYCLE
refers
to the timer that is based on the CPU (processor) cycle
counter. The timers in
setup_timers
that you can
use are those that do not have NULL
in the other columns. If the values associated with a
given timer name are NULL
, that timer
is not supported on your platform.
TIMER_FREQUENCY
indicates the number
of timer units per second. For a cycle timer, the
frequency is generally related to the CPU speed. The
value shown was obtained on a system with a 2.4GHz
processor. The other timers are based on fixed fractions
of seconds. For TICK
, the frequency
may vary by platform (for example, some use 100
ticks/second, others 1000 ticks/second).
TIMER_RESOLUTION
indicates the number
of timer units by which timer values increase at a time.
If a timer has a resolution of 10, its value increases
by 10 each time.
TIMER_OVERHEAD
is the minimal number
of cycles of overhead to obtain one timing with the
given timer. The overhead per event is twice the value
displayed because the timer is invoked at the beginning
and end of the event.
To see which timer is in effect or to change the timer,
access the setup_timers
table:
mysql>SELECT * FROM setup_timers;
+------+------------+ | NAME | TIMER_NAME | +------+------------+ | wait | CYCLE | +------+------------+ mysql>UPDATE setup_timers SET TIMER_NAME = 'MICROSECOND'
->WHERE NAME = 'wait';
mysql>SELECT * FROM setup_timers;
+------+-------------+ | NAME | TIMER_NAME | +------+-------------+ | wait | MICROSECOND | +------+-------------+
By default, the Performance Schema uses the best timer
available for each instrument type, but you can select a
different one. Generally the best timer is
CYCLE
, which uses the CPU cycle counter
whenever possible to provide high precision and low
overhead.
The precision offered by the cycle counter depends on
processor speed. If the processor runs at 1 GHz (one billion
cycles/second) or higher, the cycle counter delivers
sub-nanosecond precision. Using the cycle counter is much
cheaper than getting the actual time of day. For example,
the standard gettimeofday()
function can
take hundreds of cycles, which is an unacceptable overhead
for data gathering that may occur thousands or millions of
times per second.
Cycle counters also have disadvantages:
End users expect to see timings in wall-clock units, such as fractions of a second. Converting from cycles to fractions of seconds can be expensive. For this reason, the conversion is a quick and fairly rough multiplication operation.
Processor cycle rate might change, such as when a laptop goes into power-saving mode or when a CPU slows down to reduce heat generation. If a processor's cycle rate fluctuates, conversion from cycles to real-time units is subject to error.
Cycle counters might be unreliable or unavailable
depending on the processor or the operating system. For
example, on Pentiums, the instruction is
RDTSC
(an assembly-language rather
than a C instruction) and it is theoretically possible
for the operating system to prevent user-mode programs
from using it.
Some processor details related to out-of-order execution or multiprocessor synchronization might cause the counter to seem fast or slow by up to 1000 cycles.
MySQL works with cycle counters on x386 (Windows, OS X, Linux, Solaris, and other Unix flavors), PowerPC, and IA-64.
Rows in Performance Schema tables that store current events
and historical events have three columns to represent timing
information: TIMER_START
and
TIMER_END
indicate when the event started
and finished, and TIMER_WAIT
indicates
the event duration.
The setup_instruments
table has
an ENABLED
column to indicate the
instruments for which to collect events. The table also has
a TIMED
column to indicate which
instruments are timed. If an instrument is not enabled, it
produces no events. If an enabled instrument is not timed,
events produced by the instrument have
NULL
for the
TIMER_START
,
TIMER_END
, and
TIMER_WAIT
timer values. This in turn
causes those values to be ignored when calculating the sum,
minimum, maximum, and average time values in summary tables.
Within events, times are stored in picoseconds (trillionths of a second) to normalize them to a standard unit, regardless of which timer is selected. The timer used for an event is the one in effect when event timing begins. This timer is used to convert start and end values to picoseconds for storage in the event.
Modifications to the
setup_timers
table affect
monitoring immediately. Events already measured are stored
using the original timer unit, and events in progress may
use the original timer for the begin time and the new timer
for the end time. To avoid unpredictable results if you make
timer changes, use TRUNCATE
TABLE
to reset Performance Schema statistics.
The timer baseline (“time zero”) occurs at
Performance Schema initialization during server startup.
TIMER_START
and
TIMER_END
values in events represent
picoseconds since the baseline.
TIMER_WAIT
values are durations in
picoseconds.
Picosecond values in events are approximate. Their accuracy
is subject to the usual forms of error associated with
conversion from one unit to another. If the
CYCLE
timer is used and the processor
rate varies, there might be drift. For these reasons, it is
not reasonable to look at the TIMER_START
value for an event as an accurate measure of time elapsed
since server startup. On the other hand, it is reasonable to
use TIMER_START
or
TIMER_WAIT
values in ORDER
BY
clauses to order events by start time or
duration.
The choice of picoseconds in events rather than a value such
as microseconds has a performance basis. One implementation
goal was to show results in a uniform time unit, regardless
of the timer. In an ideal world this time unit would look
like a wall-clock unit and be reasonably precise; in other
words, microseconds. But to convert cycles or nanoseconds to
microseconds, it would be necessary to perform a division
for every instrumentation. Division is expensive on many
platforms. Multiplication is not expensive, so that is what
is used. Therefore, the time unit is an integer multiple of
the highest possible TIMER_FREQUENCY
value, using a multiplier large enough to ensure that there
is no major precision loss. The result is that the time unit
is “picoseconds.” This precision is spurious,
but the decision enables overhead to be minimized.
Events are processed in a producer/consumer fashion:
Instrumented code is the source for events and produces
events to be collected. The
setup_instruments
table lists
the instruments for which events can be collected, whether
they are enabled, and (for enabled instruments) whether to
collect timing information:
mysql> SELECT * FROM setup_instruments;
+------------------------------------------------------------+---------+-------+
| NAME | ENABLED | TIMED |
+------------------------------------------------------------+---------+-------+
...
| wait/synch/mutex/sql/LOCK_global_read_lock | YES | YES |
| wait/synch/mutex/sql/LOCK_global_system_variables | YES | YES |
| wait/synch/mutex/sql/LOCK_lock_db | YES | YES |
| wait/synch/mutex/sql/LOCK_manager | YES | YES |
...
| wait/synch/rwlock/sql/LOCK_grant | YES | YES |
| wait/synch/rwlock/sql/LOGGER::LOCK_logger | YES | YES |
| wait/synch/rwlock/sql/LOCK_sys_init_connect | YES | YES |
| wait/synch/rwlock/sql/LOCK_sys_init_slave | YES | YES |
...
| wait/io/file/sql/binlog | YES | YES |
| wait/io/file/sql/binlog_index | YES | YES |
| wait/io/file/sql/casetest | YES | YES |
| wait/io/file/sql/dbopt | YES | YES |
...
Performance Schema tables are the destinations for events
and consume events. The
setup_consumers
table lists
the types of consumers to which event information can be
sent:
mysql> SELECT * FROM setup_consumers;
+----------------------------------------------+---------+
| NAME | ENABLED |
+----------------------------------------------+---------+
| events_waits_current | YES |
| events_waits_history | YES |
| events_waits_history_long | YES |
| events_waits_summary_by_thread_by_event_name | YES |
| events_waits_summary_by_event_name | YES |
| events_waits_summary_by_instance | YES |
| file_summary_by_event_name | YES |
| file_summary_by_instance | YES |
+----------------------------------------------+---------+
Filtering can be done at different stages of performance monitoring:
Pre-filtering. This is done by modifying Performance Schema configuration so that only certain types of events are collected from producers, and collected events update only certain consumers. To do this, enable or disable instruments or consumers. Pre-filtering is done by the Performance Schema and has a global effect that applies to all users.
Reasons to use pre-filtering:
To reduce overhead. Performance Schema overhead should be minimal even with all instruments enabled, but perhaps you want to reduce it further. Or you do not care about timing events and want to disable the timing code to eliminate timing overhead.
To avoid filling the current-events or history tables with events in which you have no interest. Pre-filtering leaves more “room” in these tables for instances of rows for enabled instrument types. If you enable only file instruments with pre-filtering, no rows are collected for nonfile instruments. With post-filtering, nonfile events are collected, leaving fewer rows for file events.
To avoid maintaining some kinds of event tables. If you disable a consumer, the server does not spend time maintaining destinations for that consumer. For example, if you do not care about event histories, you can disable the history table consumers to improve performance.
Post-filtering. This
involves the use of WHERE
clauses in
queries that select information from Performance Schema
tables, to specify which of the available events you want
to see. Post-filtering is performed on a per-user basis
because individual users select which of the available
events are of interest.
Reasons to use post-filtering:
To avoid making decisions for individual users about which event information is of interest.
To use the Performance Schema to investigate a performance issue when the restrictions to impose using pre-filtering are not known in advance.
The following sections provide more detail about pre-filtering and provide guidelines for naming instruments or consumers in filtering operations. For information about writing queries to retrieve information (post-filtering), see Section 22.3, “Performance Schema Queries”.
Pre-filtering is done by modifying Performance Schema configuration so that only certain types of events are collected from producers, and collected events update only certain consumers. This type of filtering is done by the Performance Schema and has a global effect that applies to all users.
Pre-filtering can be applied to either the producer or consumer stage of event processing:
To affect pre-filtering at the producer stage, modify the
setup_instruments
table. An
instrument can be enabled or disabled by setting its
ENABLED
value to YES
or NO
. An instrument can be configured
whether to collect timing information by setting its
TIMED
value to YES
or NO
.
To affect pre-filtering at the consumer stage, modify the
setup_consumers
table. A
consumer can be enabled or disabled by setting its
ENABLED
value to YES
or NO
.
Here are some examples that show the types of pre-filtering operations available:
Disable all instruments:
mysql> UPDATE setup_instruments SET ENABLED = 'NO';
Now no events will be collected. This change, like other pre-filtering operations, affects other users as well, even if they want to see event information.
Disable all file instruments, adding them to the current set of disabled instruments:
mysql>UPDATE setup_instruments SET ENABLED = 'NO'
->WHERE NAME LIKE 'wait/io/file/%';
Disable only file instruments, enable all other instruments:
mysql>UPDATE setup_instruments
->SET ENABLED = IF(NAME LIKE 'wait/io/file/%', 'NO', 'YES');
The preceding queries use the
LIKE
operator and the pattern
'wait/io/file/%'
to match all
instrument names that begin with
'wait/io/file/
. For additional
information about specifying patterns to select
instruments, see
Section 22.2.3.4, “Naming Instruments or Consumers for Filtering Operations”.
Enable all but those instruments in the
mysys
library:
mysql>UPDATE setup_instruments
->SET ENABLED = CASE WHEN NAME LIKE '%/mysys/%' THEN 'YES' ELSE 'NO' END;
Disable a specific instrument:
mysql>UPDATE setup_instruments SET ENABLED = 'NO'
->WHERE NAME = 'wait/synch/mutex/mysys/TMPDIR_mutex';
To toggle the state of an instrument, “flip”
its ENABLED
value:
mysql>UPDATE setup_instruments
->SET ENABLED = IF(ENABLED = 'YES', 'NO', 'YES')
->WHERE NAME = 'wait/synch/mutex/mysys/TMPDIR_mutex';
Disable timing for all events:
mysql> UPDATE setup_instruments SET TIMED = 'NO';
Setting the TIMED
column for instruments
affects Performance Schema table contents as described in
Section 22.2.3.1, “Performance Schema Event Timing”.
When you change the monitoring configuration, the Performance
Schema does not flush the history tables. Events already
collected remain in the current-events and history tables
until displaced by newer events. If you disable instruments,
you might need to wait a while before events for them are
displaced by newer events of interest. Alternatively, use
TRUNCATE TABLE
to empty the
history tables.
After making instrumentation changes, you might want to
truncate the summary tables to clear aggregate information for
previously collected events. The effect of
TRUNCATE TABLE
for summary
tables is to reset the summary columns to 0 or
NULL
, not to remove rows.
If you disable a consumer, the server does not spend time maintaining destinations for that consumer. For example, if you do not care about historical event information, disable the history consumers:
mysql>UPDATE setup_consumers
->SET ENABLED = 'NO' WHERE NAME LIKE '%history%';
Names given for filtering operations can be as specific or general as required. To indicate a single instrument or consumer, specify its name in full:
mysql>UPDATE setup_instruments
->SET ENABLED = 'NO'
->WHERE NAME = 'wait/synch/mutex/myisammrg/MYRG_INFO::mutex';
mysql>UPDATE setup_consumers
->SET ENABLED = 'NO' WHERE NAME = 'file_summary_by_instance';
To specify a group of instruments or consumers, use a pattern that matches the group members:
mysql>UPDATE setup_instruments
->SET ENABLED = 'NO'
->WHERE NAME LIKE 'wait/synch/mutex/%';
mysql>UPDATE setup_consumers
->SET ENABLED = 'NO' WHERE NAME LIKE '%history%';
If you use a pattern, it should be chosen so that it matches all the items of interest and no others. For example, to select all file I/O instruments, it is better to use a pattern that includes the entire instrument name prefix:
... WHERE NAME LIKE 'wait/io/file/%';
A pattern of '%/file/%'
will match other
instruments that have a component of
'/file/'
anywhere in the name. Even less
suitable is the pattern '%file%'
because it
will match instruments with 'file'
anywhere
in the name, such as
wait/synch/mutex/sql/LOCK_des_key_file
.
To check which instrument or consumer names a pattern matches, perform a simple test:
mysql>SELECT NAME FROM setup_instruments WHERE NAME LIKE '
mysql>pattern
';SELECT NAME FROM setup_consumers WHERE NAME LIKE '
pattern
';
For information about the types of names that are supported, see Section 22.4, “Performance Schema Instrument Naming Conventions”.
It is always possible to determine what instruments the
Performance Schema includes by checking the
setup_instruments
table. For
example, to see what file-related events are instrumented for
the InnoDB
storage engine, use this query:
mysql> SELECT * FROM setup_instruments WHERE NAME LIKE 'wait/io/file/innodb/%';
+--------------------------------------+---------+-------+
| NAME | ENABLED | TIMED |
+--------------------------------------+---------+-------+
| wait/io/file/innodb/innodb_data_file | YES | YES |
| wait/io/file/innodb/innodb_log_file | YES | YES |
| wait/io/file/innodb/innodb_temp_file | YES | YES |
+--------------------------------------+---------+-------+
An exhaustive description of precisely what is instrumented is not given in this documentation, for several reasons:
What is instrumented is the server code. Changes to this code occur often, which also affects the set of instruments.
It is not practical to list all the instruments because there are hundreds of them.
As described earlier, it is possible to find out by
querying the
setup_instruments
table. This
information is always up to date for your version of
MySQL, also includes instrumentation for instrumented
plugins you might have installed that are not part of the
core server, and can be used by automated tools.
Pre-filtering limits which event information is collected and is
independent of any particular user. By contrast, post-filtering is
performed by individual users through the use of queries with
appropriate WHERE
clauses that restrict what
event information to select from the events available after
pre-filtering has been applied.
In Section 22.2.3.3, “Event Pre-Filtering”, an example
showed how to pre-filter for file instruments. If the event tables
contain both file and nonfile information, post-filtering is
another way to see information only for file events. Add a
WHERE
clause to queries to restrict event
selection appropriately:
mysql>SELECT THREAD_ID, NUMBER_OF_BYTES
->FROM events_waits_history
->WHERE EVENT_NAME LIKE 'wait/io/file/%'
->AND NUMBER_OF_BYTES IS NOT NULL;
+-----------+-----------------+ | THREAD_ID | NUMBER_OF_BYTES | +-----------+-----------------+ | 11 | 66 | | 11 | 47 | | 11 | 139 | | 5 | 24 | | 5 | 834 | +-----------+-----------------+
An instrument name consists of a sequence of components separated
by '/'
characters. Example names:
wait/io/file/myisam/log wait/io/file/mysys/charset wait/synch/cond/mysys/COND_alarm wait/synch/cond/sql/BINLOG::update_cond wait/synch/mutex/mysys/BITMAP_mutex wait/synch/mutex/sql/LOCK_delete wait/synch/rwlock/innodb/trx_sys_lock wait/synch/rwlock/sql/Query_cache_query::lock
The instrument name space has a tree-like structure. The components of an instrument name from left to right provide a progression from more general to more specific. The number of components a name has depends on the type of instrument.
The interpretation of a given component in a name depends on the
components to the left of it. For example,
myisam
appears in both of the following names,
but myisam
in the first name is related to file
I/O, whereas in the second it is related to a synchronization
instrument:
wait/io/file/myisam/log wait/synch/cond/myisam/MI_SORT_INFO::cond
Instrument names consist of a prefix with a structure defined by
the Performance Schema implementation and a suffix defined by the
developer implementing the instrument code. The top-level
component of an instrument prefix indicates the type of
instrument. This component also determines which event timer in
the setup_timers
table applies to the
instrument. For the prefix part of instrument names, the top level
indicates the type of instrument.
The suffix part of instrument names comes from the code for the instruments themselves. Suffixes may include levels such as these:
A name for the major component (a server module such as
myisam
, innodb
,
mysys
, or sql
) or a
plugin name.
The name of a variable in the code, in the form
XXX
(a global variable) or
(a member CCC
::MMM
MMM
in class
CCC
). Examples:
COND_thread_cache
,
THR_LOCK_myisam
,
BINLOG::LOCK_index
.
In MySQL 5.5, there is a single top-level component,
wait
, indicating a wait instrument. The naming
tree for wait instruments has this structure:
wait/io
An instrumented I/O operation.
wait/io/file
An instrumented file I/O operation. For files, the wait is
the time waiting for the file operation to complete (for
example, a call to fwrite()
). Due to
caching, the physical file I/O on the disk might not
happen within this call.
wait/synch
An instrumented synchronization object. For synchronization
objects, the TIMER_WAIT
time includes the
amount of time blocked while attempting to acquire a lock on
the object, if any.
wait/synch/cond
A condition is used by one thread to signal to other threads that something they were waiting for has happened. If a single thread was waiting for a condition, it can wake up and proceed with its execution. If several threads were waiting, they can all wake up and compete for the resource for which they were waiting.
wait/synch/mutex
A mutual exclusion object used to permit access to a resource (such as a section of executable code) while preventing other threads from accessing the resource.
wait/synch/rwlock
A read/write lock object used to lock a specific variable for access while preventing its use by other threads. A shared read lock can be acquired simultaneously by multiple threads. An exclusive write lock can be acquired by only one thread at a time.
There are several status variables associated with the Performance Schema:
mysql> SHOW STATUS LIKE 'perf%';
+------------------------------------------+-------+
| Variable_name | Value |
+------------------------------------------+-------+
| Performance_schema_cond_classes_lost | 0 |
| Performance_schema_cond_instances_lost | 0 |
| Performance_schema_file_classes_lost | 0 |
| Performance_schema_file_handles_lost | 0 |
| Performance_schema_file_instances_lost | 0 |
| Performance_schema_locker_lost | 0 |
| Performance_schema_mutex_classes_lost | 0 |
| Performance_schema_mutex_instances_lost | 0 |
| Performance_schema_rwlock_classes_lost | 0 |
| Performance_schema_rwlock_instances_lost | 0 |
| Performance_schema_table_handles_lost | 0 |
| Performance_schema_table_instances_lost | 0 |
| Performance_schema_thread_classes_lost | 0 |
| Performance_schema_thread_instances_lost | 0 |
+------------------------------------------+-------+
The Performance Schema status variables provide information about instrumentation that could not be loaded or created due to memory constraints. Names for these variables have several forms:
Performance_schema_
indicates how many instruments of type
xxx
_classes_lostxxx
could not be loaded.
Performance_schema_
indicates how many instances of object type
xxx
_instances_lostxxx
could not be created.
Performance_schema_
indicates how many instances of object type
xxx
_handles_lostxxx
could not be opened.
Performance_schema_locker_lost
indicates
how many events are “lost” or not recorded.
For example, if a mutex is instrumented in the server source but
the server cannot allocate memory for the instrumentation at
runtime, it increments
Performance_schema_mutex_classes_lost
.
The mutex still functions as a synchronization object (that is,
the server continues to function normally), but performance data
for it will not be collected. If the instrument can be allocated,
it can be used for initializing instrumented mutex instances. For
a singleton mutex such as a global mutex, there will be only one
instance. Other mutexes have an instance per connection, or per
page in various caches and data buffers, so the number of
instances varies over time. Increasing the maximum number of
connections or the maximum size of some buffers will increase the
maximum number of instances that might be allocated at once. If
the server cannot create a given instrumented mutex instance, it
increments
Performance_schema_mutex_instances_lost
.
Suppose that the following conditions hold:
The server was started with the
--performance_schema_max_mutex_classes=200
option and thus has room for 200 mutex instruments.
150 mutex instruments have been loaded already.
The plugin named plugin_a
contains 40 mutex
instruments.
The plugin named plugin_b
contains 20 mutex
instruments.
The server allocates mutex instruments for the plugins depending on how many they need and how many are available, as illustrated by the following sequence of statements:
INSTALL PLUGIN plugin_a
The server now has 150+40 = 190 mutex instruments.
UNINSTALL PLUGIN plugin_a;
The server still has 190 instruments. All the historical data generated by the plugin code is still available, but new events for the instruments are not collected.
INSTALL PLUGIN plugin_a;
The server detects that the 40 instruments are already defined, so no new instruments are created, and previously assigned internal memory buffers are reused. The server still has 190 instruments.
INSTALL PLUGIN plugin_b;
The server has room for 200-190 = 10 instruments (in this case,
mutex classes), and sees that the plugin contains 20 new
instruments. 10 instruments are loaded, and 10 are discarded or
“lost.” The
Performance_schema_mutex_classes_lost
indicates the number of instruments (mutex classes) lost:
mysql> SHOW STATUS LIKE "perf%mutex_classes_lost";
+---------------------------------------+-------+
| Variable_name | Value |
+---------------------------------------+-------+
| Performance_schema_mutex_classes_lost | 10 |
+---------------------------------------+-------+
1 row in set (0.10 sec)
The instrumentation still works and collects (partial) data for
plugin_b
.
When the server cannot create a mutex instrument, these results occur:
No row for the instrument is inserted into the
setup_instruments
table.
Performance_schema_mutex_classes_lost
increases by 1.
Performance_schema_mutex_instances_lost
does not change. (When the mutex instrument is not created, it
cannot be used to create instrumented mutex instances later.)
The pattern just described applies to all types of instruments, not just mutexes.
A value of
Performance_schema_mutex_classes_lost
greater than 0 can happen in two cases:
To save a few bytes of memory, you start the server with
--performance_schema_max_mutex_classes=
,
where N
N
is less than the default
value. The default value is chosen to be sufficient to load
all the plugins provided in the MySQL distribution, but this
can be reduced if some plugins are never loaded. For example,
you might choose not to load some of the storage engines in
the distribution.
You load a third-party plugin that is instrumented for the
Performance Schema but do not allow for the plugin's
instrumentation memory requirements when you start the server.
Because it comes from a third party, the instrument memory
consumption of this engine is not accounted for in the default
value chosen for
performance_schema_max_mutex_classes
.
If the server has insufficient resources for the plugin's
instruments and you do not explicitly allocate more using
--performance_schema_max_mutex_classes=
,
loading the plugin leads to starvation of instruments.
N
If the value chosen for
performance_schema_max_mutex_classes
is too small, no error is reported in the error log and there is
no failure at runtime. However, the content of the tables in the
performance_schema
database will miss events.
The
Performance_schema_mutex_classes_lost
status variable is the only visible sign to indicate that some
events were dropped internally due to failure to create
instruments.
If an instrument is not lost, it is known to the Performance
Schema, and is used when instrumenting instances. For example,
wait/synch/mutex/sql/LOCK_delete
is the name of
a mutex instrument in the
setup_instruments
table. This single
instrument is used when creating a mutex in the code (in
THD::LOCK_delete
) however many instances of the
mutex are needed as the server runs. In this case,
LOCK_delete
is a mutex that is per connection
(THD
), so if a server has 1000 connections,
there are 1000 threads, and 1000 instrumented
LOCK_delete
mutex instances
(THD::LOCK_delete
).
If the server does not have room for all these 1000 instrumented
mutexes (instances), some mutexes are created with
instrumentation, and some are created without instrumentation. If
the server can create only 800 instances, 200 instances are lost.
The server continues to run, but increments
Performance_schema_mutex_instances_lost
by 200 to indicate that instances could not be created.
A value of
Performance_schema_mutex_instances_lost
greater than 0 can happen when the code initializes more mutexes
at runtime than were allocated for
--performance_schema_max_mutex_instances=
.
N
The bottom line is that if
SHOW STATUS LIKE
'perf%'
says that nothing was lost (all values are
zero), the Performance Schema data is accurate and can be relied
upon. If something was lost, the data is incomplete, and the
Performance Schema could not record everything given the
insufficient amount of memory it was given to use. In this case,
the specific
Performance_schema_
variable indicates the problem area.
xxx
_lost
It might be appropriate in some cases to cause deliberate instrument starvation. For example, if you do not care about performance data for file I/O, you can start the server with all Performance Schema parameters related to file I/O set to 0. No memory will be allocated for file-related classes, instances, or handles, and all file events will be lost.
Use SHOW ENGINE
PERFORMANCE_SCHEMA STATUS
to inspect the internal
operation of the Performance Schema code:
mysql> SHOW ENGINE PERFORMANCE_SCHEMA STATUS\G
...
*************************** 3. row ***************************
Type: performance_schema
Name: events_waits_history.row_size
Status: 76
*************************** 4. row ***************************
Type: performance_schema
Name: events_waits_history.row_count
Status: 10000
*************************** 5. row ***************************
Type: performance_schema
Name: events_waits_history.memory
Status: 760000
...
*************************** 57. row ***************************
Type: performance_schema
Name: performance_schema.memory
Status: 26459600
...
This statement is intended to help the DBA understand the effects that different Performance Schema options have on memory requirements. For a description of the field meanings, see Section 13.7.5.16, “SHOW ENGINE Syntax”.
The name of the performance_schema
database is
lowercase, as are the names of tables within it. Queries should
specify the names in lowercase.
Most tables in the performance_schema
database
are read only and cannot be modified. Some of the setup tables
have columns that can be modified to affect Performance Schema
operation. Truncation is permitted to clear collected events, so
TRUNCATE TABLE
can be used on
tables containing those kinds of information, such as tables named
with a prefix of events_waits_
.
TRUNCATE TABLE
can also be used
with summary tables, but the effect is to reset the summary
columns to 0 or NULL
, not to remove rows.
Privileges are as for other databases and tables:
Tables in the performance_schema
database can
be grouped as follows:
Setup tables. These tables are used to configure and display monitoring characteristics.
Current events table. The
events_waits_current
table
contains the most recent event for each thread.
History tables. These tables have the same structure as
events_waits_current
but contain
more rows. The
events_waits_history
table
contains the most recent 10 events per thread.
events_waits_history_long
contains the most recent 10,000 events.
To change the sizes of these tables, set the
performance_schema_events_waits_history_size
and
performance_schema_events_waits_history_long_size
system variables at server startup.
Summary tables. These tables contain information aggregated over groups of events, including those that have been discarded from the history tables.
Instance tables. These tables document what types of objects are instrumented. An instrumented object, when used by the server, produces an event. These tables provide event names and explanatory notes or status information.
Miscellaneous tables. These do not fall into any of the other table groups.
The following table lists each Performance Schema table and provides a short description of each one.
Table 22.1 Performance Schema Tables
Table Name | Description |
---|---|
cond_instances | synchronization object instances |
events_waits_current | Current wait events |
events_waits_history | Most recent wait events for each thread |
events_waits_history_long | Most recent wait events overall |
events_waits_summary_by_instance | Wait events per instance |
events_waits_summary_by_thread_by_event_name | Wait events per thread and event name |
events_waits_summary_global_by_event_name | Wait events per event name |
file_instances | File instances |
file_summary_by_event_name | File events per event name |
file_summary_by_instance | File events per file instance |
mutex_instances | Mutex synchronization object instances |
performance_timers | Which event timers are available |
rwlock_instances | Lock synchronization object instances |
setup_consumers | Consumers for which event information can be stored |
setup_instruments | Classes of instrumented objects for which events can be collected |
setup_timers | Current event timer |
threads | Information about server threads |
The setup tables provide information about the current
instrumentation and enable the monitoring configuration to be
changed. For this reason, some columns in these tables can be
changed if you have the UPDATE
privilege.
The use of tables rather than individual variables for setup information provides a high degree of flexibility in modifying Performance Schema configuration. For example, you can use a single statement with standard SQL syntax to make multiple simultaneous configuration changes.
These setup tables are available:
setup_consumers
: The types of
consumers for which event information can be stored
setup_instruments
: The classes
of instrumented objects for which events can be collected
setup_timers
: The current event
timer
The setup_consumers
table lists
the types of consumers for which event information can be
stored:
mysql> SELECT * FROM setup_consumers;
+----------------------------------------------+---------+
| NAME | ENABLED |
+----------------------------------------------+---------+
| events_waits_current | YES |
| events_waits_history | YES |
| events_waits_history_long | YES |
| events_waits_summary_by_thread_by_event_name | YES |
| events_waits_summary_by_event_name | YES |
| events_waits_summary_by_instance | YES |
| file_summary_by_event_name | YES |
| file_summary_by_instance | YES |
+----------------------------------------------+---------+
The setup_consumers
table has
these columns:
NAME
The consumer name.
ENABLED
Whether the consumer is enabled. The value is
YES
or NO
. This
column can be modified. If you disable a consumer, the
server does not spend time adding event information to it.
Disabling the events_waits_current
consumer
disables everything else that depends on waits, such as the
events_waits_history
and
events_waits_history_long
tables,
and all summary tables.
The setup_instruments
table lists
classes of instrumented objects for which events can be
collected:
mysql> SELECT * FROM setup_instruments;
+------------------------------------------------------------+---------+-------+
| NAME | ENABLED | TIMED |
+------------------------------------------------------------+---------+-------+
...
| wait/synch/mutex/sql/LOCK_global_read_lock | YES | YES |
| wait/synch/mutex/sql/LOCK_global_system_variables | YES | YES |
| wait/synch/mutex/sql/LOCK_lock_db | YES | YES |
| wait/synch/mutex/sql/LOCK_manager | YES | YES |
...
| wait/synch/rwlock/sql/LOCK_grant | YES | YES |
| wait/synch/rwlock/sql/LOGGER::LOCK_logger | YES | YES |
| wait/synch/rwlock/sql/LOCK_sys_init_connect | YES | YES |
| wait/synch/rwlock/sql/LOCK_sys_init_slave | YES | YES |
...
| wait/io/file/sql/binlog | YES | YES |
| wait/io/file/sql/binlog_index | YES | YES |
| wait/io/file/sql/casetest | YES | YES |
| wait/io/file/sql/dbopt | YES | YES |
...
Each instrument added to the source code provides a row for
this table, even when the instrumented code is not executed.
When an instrument is enabled and executed, instrumented
instances are created, which are visible in the
*_instances
tables.
The setup_instruments
table has
these columns:
NAME
The instrument name. Instrument names may have multiple
parts and form a hierarchy, as discussed in
Section 22.4, “Performance Schema Instrument Naming Conventions”.
Events produced from execution of an instrument have an
EVENT_NAME
value that is taken from the
instrument NAME
value. (Events do not
really have a “name,” but this provides a way
to associate events with instruments.)
ENABLED
Whether the instrument is enabled. The value is
YES
or NO
. This
column can be modified. A disabled instrument produces no
events.
TIMED
Whether the instrument is timed. This column can be modified.
If an enabled instrument is not timed, the instrument code
is enabled, but the timer is not. Events produced by the
instrument have NULL
for the
TIMER_START
,
TIMER_END
, and
TIMER_WAIT
timer values. This in turn
causes those values to be ignored when calculating the
sum, minimum, maximum, and average time values in summary
tables.
The setup_timers
table shows the
currently selected event timer:
mysql> SELECT * FROM setup_timers;
+------+------------+
| NAME | TIMER_NAME |
+------+------------+
| wait | CYCLE |
+------+------------+
The setup_timers.TIMER_NAME
value can be
changed to select a different timer. The value can be any of
the values in the
performance_timers.TIMER_NAME
column. For
an explanation of how event timing occurs, see
Section 22.2.3.1, “Performance Schema Event Timing”.
Modifications to the setup_timers
table affect monitoring immediately. Events already in
progress use the original timer for the begin time and the new
timer for the end time, which leads to unpredictable results.
If you make timer changes, you may want to use
TRUNCATE TABLE
to reset
Performance Schema statistics.
The setup_timers
table has these
columns:
NAME
The type of instrument the timer is used for.
TIMER_NAME
The timer that applies to the instrument type. This column can be modified.
Instance tables document what types of objects are instrumented. They provide event names and explanatory notes or status information:
cond_instances
: Condition
synchronization object instances
file_instances
: File instances
mutex_instances
: Mutex
synchronization object instances
rwlock_instances
: Lock
synchronization object instances
These tables list instrumented synchronization objects and
files. There are three types of synchronization objects:
cond
, mutex
, and
rwlock
. Each instance table has an
EVENT_NAME
or NAME
column
to indicate the instrument associated with each row. Instrument
names may have multiple parts and form a hierarchy, as discussed
in Section 22.4, “Performance Schema Instrument Naming Conventions”.
The mutex_instances.LOCKED_BY_THREAD_ID
and
rwlock_instances.WRITE_LOCKED_BY_THREAD_ID
columns are extremely important for investigating performance
bottlenecks or deadlocks. For examples of how to use them for
this purpose, see Section 22.12, “Using the Performance Schema to Diagnose Problems”
The cond_instances
table lists
all the conditions seen by the Performance Schema while the
server executes. A condition is a synchronization mechanism
used in the code to signal that a specific event has happened,
so that a thread waiting for this condition can resume work.
When a thread is waiting for something to happen, the condition name is an indication of what the thread is waiting for, but there is no immediate way to tell which other thread, or threads, will cause the condition to happen.
The cond_instances
table has
these columns:
NAME
The instrument name associated with the condition.
OBJECT_INSTANCE_BEGIN
The address in memory of the instrumented condition.
The file_instances
table lists
all the files seen by the Performance Schema when executing
file I/O instrumentation. If a file on disk has never been
opened, it will not be in
file_instances
. When a file is
deleted from the disk, it is also removed from the
file_instances
table.
The file_instances
table has
these columns:
FILE_NAME
The file name.
EVENT_NAME
The instrument name associated with the file.
OPEN_COUNT
The count of open handles on the file. If a file was
opened and then closed, it was opened 1 time, but
OPEN_COUNT
will be 0. To list all the
files currently opened by the server, use WHERE
OPEN_COUNT > 0
.
The mutex_instances
table lists
all the mutexes seen by the Performance Schema while the
server executes. A mutex is a synchronization mechanism used
in the code to enforce that only one thread at a given time
can have access to some common resource. The resource is said
to be “protected” by the mutex.
When two threads executing in the server (for example, two user sessions executing a query simultaneously) do need to access the same resource (a file, a buffer, or some piece of data), these two threads will compete against each other, so that the first query to obtain a lock on the mutex will cause the other query to wait until the first is done and unlocks the mutex.
The work performed while holding a mutex is said to be in a “critical section,” and multiple queries do execute this critical section in a serialized way (one at a time), which is a potential bottleneck.
The mutex_instances
table has
these columns:
NAME
The instrument name associated with the mutex.
OBJECT_INSTANCE_BEGIN
The address in memory of the instrumented mutex.
LOCKED_BY_THREAD_ID
When a thread currently has a mutex locked,
LOCKED_BY_THREAD_ID
is the
THREAD_ID
of the locking thread,
otherwise it is NULL
.
For every mutex instrumented in the code, the Performance Schema provides the following information.
The setup_instruments
table
lists the name of the instrumentation point, with the
prefix wait/synch/mutex/
.
When some code creates a mutex, a row is added to the
mutex_instances
table. The
OBJECT_INSTANCE_BEGIN
column is a
property that uniquely identifies the mutex.
When a thread attempts to lock a mutex, the
events_waits_current
table
shows a row for that thread, indicating that it is waiting
on a mutex (in the EVENT_NAME
column),
and indicating which mutex is waited on (in the
OBJECT_INSTANCE_BEGIN
column).
When a thread succeeds in locking a mutex:
events_waits_current
shows that the wait on the mutex is completed (in the
TIMER_END
and
TIMER_WAIT
columns)
The completed wait event is added to the
events_waits_history
and
events_waits_history_long
tables
mutex_instances
shows
that the mutex is now owned by the thread (in the
THREAD_ID
column).
When a thread unlocks a mutex,
mutex_instances
shows that
the mutex now has no owner (the
THREAD_ID
column is
NULL
).
When a mutex object is destroyed, the corresponding row is
removed from mutex_instances
.
By performing queries on both of the following tables, a monitoring application or a DBA can detect bottlenecks or deadlocks between threads that involve mutexes:
events_waits_current
, to see
what mutex a thread is waiting for
mutex_instances
, to see which
other thread currently owns a mutex
The rwlock_instances
table lists
all the rwlock
instances (read write locks)
seen by the Performance Schema while the server executes. An
rwlock
is a synchronization mechanism used
in the code to enforce that threads at a given time can have
access to some common resource following certain rules. The
resource is said to be “protected” by the
rwlock
. The access is either shared (many
threads can have a read lock at the same time) or exclusive
(only one thread can have a write lock at a given time).
Depending on how many threads are requesting a lock, and the nature of the locks requested, access can be either granted in shared mode, granted in exclusive mode, or not granted at all, waiting for other threads to finish first.
The rwlock_instances
table has
these columns:
NAME
The instrument name associated with the lock.
OBJECT_INSTANCE_BEGIN
The address in memory of the instrumented lock.
WRITE_LOCKED_BY_THREAD_ID
When a thread currently has an rwlock
locked in exclusive (write) mode,
WRITE_LOCKED_BY_THREAD_ID
is the
THREAD_ID
of the locking thread,
otherwise it is NULL
.
READ_LOCKED_BY_COUNT
When a thread currently has an rwlock
locked in shared (read) mode,
READ_LOCKED_BY_COUNT
is incremented by
1. This is a counter only, so it cannot be used directly
to find which thread holds a read lock, but it can be used
to see whether there is a read contention on an
rwlock
, and see how many readers are
currently active.
By performing queries on both of the following tables, a monitoring application or a DBA may detect some bottlenecks or deadlocks between threads that involve locks:
events_waits_current
, to see
what rwlock
a thread is waiting for
rwlock_instances
, to see
which other thread currently owns an
rwlock
There is a limitation: The
rwlock_instances
can be used only
to identify the thread holding a write lock, but not the
threads holding a read lock.
These tables store wait events:
events_waits_current
: Current
wait events
events_waits_history
: The most
recent wait events for each thread
events_waits_history_long
: The
most recent wait events overall
The following sections describe those tables. There are also summary tables that aggregate information about wait events; see Section 22.7.5.1, “Event Wait Summary Tables”.
To enable collection of wait events, enable the relevant instruments and consumers.
The setup_instruments
table
contains instruments with names that begin with
wait
. For example:
mysql>SELECT * FROM setup_instruments
->WHERE NAME LIKE 'wait/io/file/innodb%';
+--------------------------------------+---------+-------+ | NAME | ENABLED | TIMED | +--------------------------------------+---------+-------+ | wait/io/file/innodb/innodb_data_file | YES | YES | | wait/io/file/innodb/innodb_log_file | YES | YES | | wait/io/file/innodb/innodb_temp_file | YES | YES | +--------------------------------------+---------+-------+ mysql>SELECT * FROM setup_instruments WHERE
->NAME LIKE 'wait/io/socket/%';
+----------------------------------------+---------+-------+ | NAME | ENABLED | TIMED | +----------------------------------------+---------+-------+ | wait/io/socket/sql/server_tcpip_socket | NO | NO | | wait/io/socket/sql/server_unix_socket | NO | NO | | wait/io/socket/sql/client_connection | NO | NO | +----------------------------------------+---------+-------+
To modify collection of wait events, change the
ENABLED
and TIMING
columns
of the relevant instruments. For example:
mysql>UPDATE setup_instruments SET ENABLED = 'YES', TIMED = 'YES'
->WHERE NAME LIKE 'wait/io/socket/sql/%';
The setup_consumers
table contains
consumer values with names corresponding to the current and
recent wait event table names. These consumers may be used to
filter collection of wait events. The wait consumers are
disabled by default:
mysql> SELECT * FROM setup_consumers WHERE NAME LIKE '%waits%';
+---------------------------+---------+
| NAME | ENABLED |
+---------------------------+---------+
| events_waits_current | NO |
| events_waits_history | NO |
| events_waits_history_long | NO |
+---------------------------+---------+
To enable all wait consumers, do this:
mysql>UPDATE setup_consumers SET ENABLED = 'YES'
->WHERE NAME LIKE '%waits%';
The setup_timers
table contains a
row with a NAME
value of
wait
that indicates the unit for wait event
timing. The default unit is CYCLE
.
mysql> SELECT * FROM setup_timers WHERE NAME = 'wait';
+------+------------+
| NAME | TIMER_NAME |
+------+------------+
| wait | CYCLE |
+------+------------+
To change the timing unit, modify the
TIMER_NAME
value:
mysql>UPDATE setup_timers SET TIMER_NAME = 'NANOSECOND'
->WHERE NAME = 'wait';
For additional information about configuring event collection, see Section 22.2, “Performance Schema Configuration”.
The events_waits_current
table
contains current wait events, one row per thread showing the
current status of the thread's most recent monitored wait
event.
The events_waits_current
table
can be truncated with TRUNCATE
TABLE
.
Of the tables that contain wait event rows,
events_waits_current
is the most
fundamental. Other tables that contain wait event rows are
logically derived from the current events. For example, the
events_waits_history
and
events_waits_history_long
tables
are collections of the most recent wait events, up to a fixed
number of rows.
For information about configuration of wait event collection, see Section 22.7.4, “Performance Schema Wait Event Tables”.
The events_waits_current
table
has these columns:
THREAD_ID
, EVENT_ID
The thread associated with the event and the thread
current event number when the event starts. The
THREAD_ID
and
EVENT_ID
values taken together form a
primary key that uniquely identifies the row. No two rows
will have the same pair of values.
EVENT_NAME
The name of the instrument that produced the event. This
is a NAME
value from the
setup_instruments
table.
Instrument names may have multiple parts and form a
hierarchy, as discussed in
Section 22.4, “Performance Schema Instrument Naming Conventions”.
SOURCE
The name of the source file containing the instrumented code that produced the event and the line number in the file at which the instrumentation occurs. This enables you to check the source to determine exactly what code is involved. For example, if a mutex or lock is being blocked, you can check the context in which this occurs.
TIMER_START
,
TIMER_END
,
TIMER_WAIT
Timing information for the event. The unit for these
values is picoseconds (trillionths of a second). The
TIMER_START
and
TIMER_END
values indicate when event
timing started and ended. TIMER_WAIT
is
the event elapsed time (duration).
If an event has not finished, TIMER_END
and TIMER_WAIT
are
NULL
.
If an event is produced from an instrument that has
TIMED = NO
, timing information is not
collected, and TIMER_START
,
TIMER_END
, and
TIMER_WAIT
are all
NULL
.
For discussion of picoseconds as the unit for event times and factors that affect time values, see Section 22.2.3.1, “Performance Schema Event Timing”.
SPINS
For a mutex, the number of spin rounds. If the value is
NULL
, the code does not use spin rounds
or spinning is not instrumented.
OBJECT_SCHEMA
,
OBJECT_NAME
,
OBJECT_TYPE
,
OBJECT_INSTANCE_BEGIN
These columns identify the object “being acted on.” What that means depends on the object type.
For a synchronization object (cond
,
mutex
, rwlock
):
OBJECT_SCHEMA
,
OBJECT_NAME
, and
OBJECT_TYPE
are
NULL
.
OBJECT_INSTANCE_BEGIN
is the
address of the synchronization object in memory.
For a file I/O object:
OBJECT_SCHEMA
is
NULL
.
OBJECT_NAME
is the file name.
OBJECT_TYPE
is
FILE
.
OBJECT_INSTANCE_BEGIN
is an address
in memory.
An OBJECT_INSTANCE_BEGIN
value itself
has no meaning, except that different values indicate
different objects.
OBJECT_INSTANCE_BEGIN
can be used for
debugging. For example, it can be used with GROUP
BY OBJECT_INSTANCE_BEGIN
to see whether the load
on 1,000 mutexes (that protect, say, 1,000 pages or blocks
of data) is spread evenly or just hitting a few
bottlenecks. This can help you correlate with other
sources of information if you see the same object address
in a log file or another debugging or performance tool.
NESTING_EVENT_ID
Always NULL
.
OPERATION
The type of operation performed, such as
lock
, read
, or
write
.
NUMBER_OF_BYTES
The number of bytes read or written by the operation.
FLAGS
Reserved for future use.
The events_waits_history
table
contains the most recent 10 wait events per thread. To change
the table size, modify the
performance_schema_events_waits_history_size
system variable at server startup. Wait events are not added
to the table until they have ended. As new events are added,
older events are discarded if the table is full.
The events_waits_history
table
has the same structure as
events_waits_current
. See
Section 22.7.4.1, “The events_waits_current Table”.
The events_waits_history
table
can be truncated with TRUNCATE
TABLE
.
For information about configuration of wait event collection, see Section 22.7.4, “Performance Schema Wait Event Tables”.
The events_waits_history_long
table contains the most recent 10,000 wait events. To change
the table size, modify the
performance_schema_events_waits_history_long_size
system variable at server startup. Wait events are not added
to the table until they have ended. As new events are added,
older events are discarded if the table is full. When a thread
ends, its rows are removed from the table.
The events_waits_history_long
table has the same structure as
events_waits_current
. See
Section 22.7.4.1, “The events_waits_current Table”.
The events_waits_history_long
table can be truncated with TRUNCATE
TABLE
.
For information about configuration of wait event collection, see Section 22.7.4, “Performance Schema Wait Event Tables”.
Summary tables provide aggregated information for terminated events over time. The tables in this group summarize event data in different ways.
Event Wait Summaries:
events_waits_summary_global_by_event_name
:
Wait events summarized per event name
events_waits_summary_by_instance
:
Wait events summarized per instance
events_waits_summary_by_thread_by_event_name
:
Wait events summarized per thread and event name
File I/O Summaries:
file_summary_by_event_name
:
File events summarized per event name
file_summary_by_instance
: File
events summarized per file instance
The events_waits_summary_global_by_event_name
table was named
EVENTS_WAITS_SUMMARY_BY_EVENT_NAME
before
MySQL 5.5.7.
Each summary table has grouping columns that determine how to group the data to be aggregated, and summary columns that contain the aggregated values. Tables that summarize events in similar ways often have similar sets of summary columns and differ only in the grouping columns used to determine how events are aggregated.
Summary tables can be truncated with
TRUNCATE TABLE
. The effect is to
reset the summary columns to 0 or NULL
, not
to remove rows. This enables you to clear collected values and
restart aggregation. That might be useful, for example, after
you have made a runtime configuration change.
The Performance Schema maintains tables for collecting current and recent wait events, and aggregates that information in summary tables. Section 22.7.4, “Performance Schema Wait Event Tables” describes the events on which wait summaries are based. See that discussion for information about the content of wait events, the current and recent wait event tables, and how to control wait event collection.
Each event waits summary table has one or more grouping
columns to indicate how the table aggregates events. Event
names refer to names of event instruments in the
setup_instruments
table.
events_waits_summary_global_by_event_name
has an EVENT_NAME
column. Each row
summarizes events for a given event name. An instrument
might be used to create multiple instances of the
instrumented object. For example, if there is an
instrument for a mutex that is created for each
connection, there are as many instances as there are
connections. The summary row for the instrument summarizes
over all these instances.
events_waits_summary_by_instance
has EVENT_NAME
and
OBJECT_INSTANCE_BEGIN
columns. Each row
summarizes events for a given event name and object. If an
instrument is used to create multiple instances, each
instance has a unique
OBJECT_INSTANCE_BEGIN
value, so these
instances are summarized separately in this table.
events_waits_summary_by_thread_by_event_name
has THREAD_ID
and
EVENT_NAME
columns. Each row summarizes
events for a given thread and event name.
Each event waits summary table has these summary columns containing aggregated values:
COUNT_STAR
The number of summarized events. This value includes all events, whether timed or nontimed.
SUM_TIMER_WAIT
The total wait time of the summarized timed events. This
value is calculated only for timed events because nontimed
events have a wait time of NULL
. The
same is true for the other
values.
xxx
_TIMER_WAIT
MIN_TIMER_WAIT
The minimum wait time of the summarized timed events.
AVG_TIMER_WAIT
The average wait time of the summarized timed events.
MAX_TIMER_WAIT
The maximum wait time of the summarized timed events.
Example wait event summary information:
mysql> SELECT * FROM events_waits_summary_global_by_event_name\G
...
*************************** 6. row ***************************
EVENT_NAME: wait/synch/mutex/sql/BINARY_LOG::LOCK_index
COUNT_STAR: 8
SUM_TIMER_WAIT: 2119302
MIN_TIMER_WAIT: 196092
AVG_TIMER_WAIT: 264912
MAX_TIMER_WAIT: 569421
...
*************************** 9. row ***************************
EVENT_NAME: wait/synch/mutex/sql/hash_filo::lock
COUNT_STAR: 69
SUM_TIMER_WAIT: 16848828
MIN_TIMER_WAIT: 0
AVG_TIMER_WAIT: 244185
MAX_TIMER_WAIT: 735345
...
TRUNCATE TABLE
is permitted for
wait summary tables. It resets the summary columns to zero
rather than removing rows.
The file I/O summary tables aggregate information about I/O operations.
Each file I/O summary table has one or more grouping columns
to indicate how the table aggregates events. Event names refer
to names of event instruments in the
setup_instruments
table.
file_summary_by_event_name
has an EVENT_NAME
column. Each row
summarizes events for a given event name.
file_summary_by_instance
has
FILE_NAME
and
EVENT_NAME
columns. Each row summarizes
events for a given file instrument instance.
Each file I/O summary table has these summary columns containing aggregated values:
COUNT_READ
The number of read operations in the summarized events.
COUNT_WRITE
The number of write operations in the summarized events.
SUM_NUMBER_OF_BYTES_READ
The number of bytes read in the summarized events.
SUM_NUMBER_OF_BYTES_WRITE
The number of bytes written in the summarized events.
Example file I/O event summary information:
mysql> SELECT * FROM file_summary_by_instance\G
...
*************************** 2. row ***************************
FILE_NAME: /var/mysql/share/english/errmsg.sys
EVENT_NAME: wait/io/file/sql/ERRMSG
COUNT_READ: 3
COUNT_WRITE: 0
SUM_NUMBER_OF_BYTES_READ: 42211
SUM_NUMBER_OF_BYTES_WRITE: 0
...
*************************** 6. row ***************************
FILE_NAME: /var/mysql/data/binlog.000001
EVENT_NAME: wait/io/file/sql/binlog
COUNT_READ: 0
COUNT_WRITE: 0
SUM_NUMBER_OF_BYTES_READ: 0
SUM_NUMBER_OF_BYTES_WRITE: 0
...
TRUNCATE TABLE
is permitted for
file I/O summary tables. It resets the summary columns to zero
rather than removing rows.
The MySQL server uses several techniques to avoid I/O operations by caching information read from files, so it is possible that statements you might expect to result in I/O events will not. You may be able to ensure that I/O does occur by flushing caches or restarting the server to reset its state.
The following sections describe tables that do not fall into the table categories discussed in the preceding sections:
performance_timers
: Which event
timers are available
threads
: Information about
server threads
The performance_timers
table
shows which event timers are available:
mysql> SELECT * FROM performance_timers;
+-------------+-----------------+------------------+----------------+
| TIMER_NAME | TIMER_FREQUENCY | TIMER_RESOLUTION | TIMER_OVERHEAD |
+-------------+-----------------+------------------+----------------+
| CYCLE | 2389029850 | 1 | 72 |
| NANOSECOND | NULL | NULL | NULL |
| MICROSECOND | 1000000 | 1 | 585 |
| MILLISECOND | 1035 | 1 | 738 |
| TICK | 101 | 1 | 630 |
+-------------+-----------------+------------------+----------------+
If the values associated with a given timer name are
NULL
, that timer is not supported on your
platform. The rows that do not contain NULL
indicate which timers you can use in
setup_timers
.
The performance_timers
table has
these columns:
TIMER_NAME
The name by which to refer to the timer when configuring
the setup_timers
table.
TIMER_FREQUENCY
The number of timer units per second. For a cycle timer,
the frequency is generally related to the CPU speed. For
example, on a system with a 2.4GHz processor, the
CYCLE
may be close to 2400000000.
TIMER_RESOLUTION
Indicates the number of timer units by which timer values increase. If a timer has a resolution of 10, its value increases by 10 each time.
TIMER_OVERHEAD
The minimal number of cycles of overhead to obtain one timing with the given timer. The Performance Schema determines this value by invoking the timer 20 times during initialization and picking the smallest value. The total overhead really is twice this amount because the instrumentation invokes the timer at the start and end of each event. The timer code is called only for timed events, so this overhead does not apply for nontimed events.
The threads
table contains a row
for each server thread:
mysql> SELECT * FROM threads;
+-----------+----------------+----------------------------------------+
| THREAD_ID | PROCESSLIST_ID | NAME |
+-----------+----------------+----------------------------------------+
| 0 | 0 | thread/sql/main |
| 1 | 0 | thread/innodb/io_handler_thread |
| 16 | 0 | thread/sql/signal_handler |
| 23 | 7 | thread/sql/one_connection |
| 5 | 0 | thread/innodb/io_handler_thread |
| 12 | 0 | thread/innodb/srv_lock_timeout_thread |
| 22 | 6 | thread/sql/one_connection |
...
For
INFORMATION_SCHEMA.PROCESSLIST
and SHOW PROCESSLIST
,
information about threads for other users is shown only if
the current user has the
PROCESS
privilege. That is
not true of the threads
table;
all rows are shown to any user who has the
SELECT
privilege for the table. Users who
should not be able to see threads for other users should not
be given that privilege.
The threads
table has these
columns:
THREAD_ID
This is the unique identifier of an instrumented thread.
PROCESSLIST_ID
For threads that are displayed in the
INFORMATION_SCHEMA.PROCESSLIST
table, this is the same value displayed in the
ID
column of that table. It is also the
value displayed in the Id
column of
SHOW PROCESSLIST
output,
and the value that
CONNECTION_ID()
would
return within that thread.
For background threads (threads not associated with a user
connection), PROCESSLIST_ID
is 0, so
the values are not unique.
This column was named ID
before MySQL
5.5.8.
NAME
NAME
is the name associated with the
instrumentation of the code in the server. For example,
thread/sql/one_connection
corresponds
to the thread function in the code responsible for
handling a user connection, and
thread/sql/main
stands for the
main()
function of the server.
The threads
table was named
PROCESSLIST
before MySQL 5.5.6.
Table 22.2 Performance Schema Variable Reference
The Performance Schema implements several system variables that provide configuration information:
mysql> SHOW VARIABLES LIKE 'perf%';
+---------------------------------------------------+---------+
| Variable_name | Value |
+---------------------------------------------------+---------+
| performance_schema | ON |
| performance_schema_events_waits_history_long_size | 10000 |
| performance_schema_events_waits_history_size | 10 |
| performance_schema_max_cond_classes | 80 |
| performance_schema_max_cond_instances | 1000 |
| performance_schema_max_file_classes | 50 |
| performance_schema_max_file_handles | 32768 |
| performance_schema_max_file_instances | 10000 |
| performance_schema_max_mutex_classes | 200 |
| performance_schema_max_mutex_instances | 1000000 |
| performance_schema_max_rwlock_classes | 30 |
| performance_schema_max_rwlock_instances | 1000000 |
| performance_schema_max_table_handles | 100000 |
| performance_schema_max_table_instances | 50000 |
| performance_schema_max_thread_classes | 50 |
| performance_schema_max_thread_instances | 1000 |
+---------------------------------------------------+---------+
Performance Schema system variables can be set at server startup on the command line or in option files, and many can be set at runtime. See Section 22.8, “Performance Schema Option and Variable Reference”.
Performance Schema system variables have the following meanings:
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema=# | ||
System Variable | Name | performance_schema | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | boolean | |
Default | OFF |
The value of this variable is ON
or
OFF
to indicate whether the Performance
Schema is enabled. By default, the value is
OFF
. At server startup, you can specify
this variable with no value or a value of 1 to enable it, or
with a value of 0 to disable it.
performance_schema_events_waits_history_long_size
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_events_waits_history_long_size=# | ||
System Variable | Name | performance_schema_events_waits_history_long_size | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 10000 |
The number of rows in the
events_waits_history_long
table.
performance_schema_events_waits_history_size
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_events_waits_history_size=# | ||
System Variable | Name | performance_schema_events_waits_history_size | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 10 |
The number of rows per thread in the
events_waits_history
table.
performance_schema_max_cond_classes
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_cond_classes=# | ||
System Variable | Name | performance_schema_max_cond_classes | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 80 |
The maximum number of condition instruments.
performance_schema_max_cond_instances
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_cond_instances=# | ||
System Variable | Name | performance_schema_max_cond_instances | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 1000 |
The maximum number of instrumented condition objects.
performance_schema_max_file_classes
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_file_classes=# | ||
System Variable | Name | performance_schema_max_file_classes | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 50 |
The maximum number of file instruments.
performance_schema_max_file_handles
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_file_handles=# | ||
System Variable | Name | performance_schema_max_file_handles | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 32768 |
The maximum number of opened file objects.
The value of
performance_schema_max_file_handles
should be greater than the value of
open_files_limit
:
open_files_limit
affects the
maximum number of open file handles the server can support and
performance_schema_max_file_handles
affects how many of these file handles can be instrumented.
performance_schema_max_file_instances
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_file_instances=# | ||
System Variable | Name | performance_schema_max_file_instances | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 10000 |
The maximum number of instrumented file objects.
performance_schema_max_mutex_classes
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_mutex_classes=# | ||
System Variable | Name | performance_schema_max_mutex_classes | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 200 |
The maximum number of mutex instruments.
performance_schema_max_mutex_instances
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_mutex_instances=# | ||
System Variable | Name | performance_schema_max_mutex_instances | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 1000 |
The maximum number of instrumented mutex objects.
performance_schema_max_rwlock_classes
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_rwlock_classes=# | ||
System Variable | Name | performance_schema_max_rwlock_classes | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values (<= 5.5.6) | Type | integer | |
Default | 20 | ||
Permitted Values (>= 5.5.7) | Type | integer | |
Default | 30 |
The maximum number of rwlock instruments.
performance_schema_max_rwlock_instances
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_rwlock_instances=# | ||
System Variable | Name | performance_schema_max_rwlock_instances | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 1000 |
The maximum number of instrumented rwlock objects.
performance_schema_max_table_handles
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_table_handles=# | ||
System Variable | Name | performance_schema_max_table_handles | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 100000 |
The maximum number of opened table objects.
performance_schema_max_table_instances
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_table_instances=# | ||
System Variable | Name | performance_schema_max_table_instances | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 50000 |
The maximum number of instrumented table objects.
performance_schema_max_thread_classes
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_thread_classes=# | ||
System Variable | Name | performance_schema_max_thread_classes | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 50 |
The maximum number of thread instruments.
performance_schema_max_thread_instances
Introduced | 5.5.3 | ||
Command-Line Format | --performance_schema_max_thread_instances=# | ||
System Variable | Name | performance_schema_max_thread_instances | |
Variable Scope | Global | ||
Dynamic Variable | No | ||
Permitted Values | Type | integer | |
Default | 1000 |
The maximum number of instrumented thread objects. The value
controls the size of the threads
table. If this maximum is exceeded such that a thread cannot
be instrumented, the Performance Schema increments the
Performance_schema_thread_instances_lost
status variable.
The max_connections
and
max_delayed_threads
system
variables affect how many threads are run in the server.
performance_schema_max_thread_instances
affects how many of these running threads can be instrumented.
If you increase
max_connections
or
max_delayed_threads
, you
should consider increasing
performance_schema_max_thread_instances
so that
performance_schema_max_thread_instances
is greater than the sum of
max_connections
and
max_delayed_threads
.
The Performance Schema implements several status variables that provide information about instrumentation that could not be loaded or created due to memory constraints:
mysql> SHOW STATUS LIKE 'perf%';
+------------------------------------------+-------+
| Variable_name | Value |
+------------------------------------------+-------+
| Performance_schema_cond_classes_lost | 0 |
| Performance_schema_cond_instances_lost | 0 |
| Performance_schema_file_classes_lost | 0 |
| Performance_schema_file_handles_lost | 0 |
| Performance_schema_file_instances_lost | 0 |
| Performance_schema_locker_lost | 0 |
| Performance_schema_mutex_classes_lost | 0 |
| Performance_schema_mutex_instances_lost | 0 |
| Performance_schema_rwlock_classes_lost | 0 |
| Performance_schema_rwlock_instances_lost | 0 |
| Performance_schema_table_handles_lost | 0 |
| Performance_schema_table_instances_lost | 0 |
| Performance_schema_thread_classes_lost | 0 |
| Performance_schema_thread_instances_lost | 0 |
+------------------------------------------+-------+
Performance Schema status variables have the following meanings:
Performance_schema_cond_classes_lost
How many condition instruments could not be loaded.
Performance_schema_cond_instances_lost
How many condition instrument instances could not be created.
Performance_schema_file_classes_lost
How many file instruments could not be loaded.
Performance_schema_file_handles_lost
How many file instrument instances could not be opened.
Performance_schema_file_instances_lost
How many file instrument instances could not be created.
Performance_schema_locker_lost
How many events are “lost” or not recorded, due to the following conditions:
Events are recursive (for example, waiting for A caused a wait on B, which caused a wait on C).
The depth of the nested events stack is greater than the limit imposed by the implementation.
Events recorded by the Performance Schema are not recursive, so this variable should always be 0.
Performance_schema_mutex_classes_lost
How many mutex instruments could not be loaded.
Performance_schema_mutex_instances_lost
How many mutex instrument instances could not be created.
Performance_schema_rwlock_classes_lost
How many rwlock instruments could not be loaded.
Performance_schema_rwlock_instances_lost
How many rwlock instrument instances could not be created.
Performance_schema_table_handles_lost
How many table instrument instances could not be opened.
Performance_schema_table_instances_lost
How many table instrument instances could not be created.
Performance_schema_thread_classes_lost
How many thread instruments could not be loaded.
Performance_schema_thread_instances_lost
The number of thread instances that could not be instrumented
in the threads
table. This can be
nonzero if the value of
performance_schema_max_thread_instances
is too small.
For information on using these variables to check Performance Schema status, see Section 22.5, “Performance Schema Status Monitoring”.
Removing a plugin with UNINSTALL
PLUGIN
does not affect information already collected for
code in that plugin. Time spent executing the code while the
plugin was loaded was still spent even if the plugin is unloaded
later. The associated event information, including aggregate
information, remains readable in
performance_schema
database tables. For
additional information about the effect of plugin installation and
removal, see
Section 22.5, “Performance Schema Status Monitoring”.
A plugin implementor who instruments plugin code should document its instrumentation characteristics to enable those who load the plugin to account for its requirements. For example, a third-party storage engine should include in its documentation how much memory the engine needs for mutex and other instruments.
The Performance Schema is a tool to help a DBA do performance tuning by taking real measurements instead of “wild guesses.” This section demonstrates some ways to use the Performance Schema for this purpose. The discussion here relies on the use of event filtering, which is described in Section 22.2.3.2, “Performance Schema Event Filtering”.
The following example provides one methodology that you can use to analyze a repeatable problem, such as investigating a performance bottleneck. To begin, you should have a repeatable use case where performance is deemed “too slow” and needs optimization, and you should enable all instrumentation (no pre-filtering at all).
Run the use case.
Using the Performance Schema tables, analyze the root cause of the performance problem. This analysis will rely heavily on post-filtering.
For problem areas that are ruled out, disable the corresponding instruments. For example, if analysis shows that the issue is not related to file I/O in a particular storage engine, disable the file I/O instruments for that engine. Then truncate the history and summary tables to remove previously collected events.
Repeat the process at step 1.
At each iteration, the Performance Schema output, particularly
the events_waits_history_long
table, will contain less and less “noise” caused
by nonsignificant instruments, and given that this table has a
fixed size, will contain more and more data relevant to the
analysis of the problem at hand.
At each iteration, investigation should lead closer and closer to the root cause of the problem, as the “signal/noise” ratio will improve, making analysis easier.
Once a root cause of performance bottleneck is identified, take the appropriate corrective action, such as:
Tune the server parameters (cache sizes, memory, and so forth).
Tune a query by writing it differently,
Tune the database schema (tables, indexes, and so forth).
Tune the code (this applies to storage engine or server developers only).
Start again at step 1, to see the effects of the changes on performance.
The mutex_instances.LOCKED_BY_THREAD_ID
and
rwlock_instances.WRITE_LOCKED_BY_THREAD_ID
columns are extremely important for investigating performance
bottlenecks or deadlocks. This is made possible by Performance
Schema instrumentation as follows:
Suppose that thread 1 is stuck waiting for a mutex.
You can determine what the thread is waiting for:
SELECT * FROM events_waits_current WHERE THREAD_ID = thread_1
;
Say the query result identifies that the thread is waiting for
mutex A, found in
events_waits_current.OBJECT_INSTANCE_BEGIN
.
You can determine which thread is holding mutex A:
SELECT * FROM mutex_instances WHERE OBJECT_INSTANCE_BEGIN = mutex_A
;
Say the query result identifies that it is thread 2 holding
mutex A, as found in
mutex_instances.LOCKED_BY_THREAD_ID
.
You can see what thread 2 is doing:
SELECT * FROM events_waits_current WHERE THREAD_ID = thread_2
;