Mysql tech day_paris_ps_and_sys

Copyright © 2014, Oracle and/or its affiliates. All rights reserved.Copyright © 2014, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 12!1

Copyright © 2014, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 12!2

MySQL Tech Day, Paris Performance Schema & MySQL sys

Mark Leith Senior Software Development Manager MySQL @ Oracle

Copyright © 2014, Oracle and/or its affiliates. All rights reserved.!3

Program Agenda

▪What is Performance Schema anyway? ▪ Performance Schema Configuration ▪ Profiling General Instance Activity ▪ Profiling Statement Activity ▪ Improvements made to date in MySQL 5.7 ▪ The MySQL SYS Schema ▪ Things still to come


Know Your Audience

▪ Using 5.1 (yes, still)? ▪ Using 5.5 (but I’ll upgrade soon, honest)? ▪ Using 5.6? ▪ Using … 5.7? ▪ Used Performance Schema? ▪ #monitoringlove ▪ #monitoringhate


What is Performance Schema anyway?



▪ Introduced in MySQL 5.5 ▪ A storage engine, built for recording instrumentation

– PERFORMANCE_SCHEMA – Manages fixed buffers in memory with lock free hashes – Real engine, unlike INFORMATION_SCHEMA

▪ A database schema to expose the instrumentation – performance_schema

▪ A set of interfaces in the MySQL Server wrapping calls to trace – i.e. pthread_mutex_lock() -> mysql_mutex_lock()



▪ Records latency of events that happen within the server ▪ All latency exposed to picosecond (a trillionth of a second) precision

▪ Though may be tracked at different precisions internally ▪ Also tracks other information as appropriate

▪ Bytes, source position, object metadata, etc.


Performance Schema in MySQL 5.5

▪ 17 Tables ▪ 222 Instruments

Instrument Event ClassFile IO wait/io/file/%Mutexes wait/synch/mutex/%Read/Write Locks

wait/synch/rwlock/%Conditions wait/synch/cond/%



▪ Laid out the framework for instrumentation ▪ Focused early on low level instruments

– Prove it could handle load, build from ground up ▪ Not turned on by default, had some overhead issues


Wait Eventsmysql> select * from events_waits_history_long where event_name like 'wait/io/table/%'\G!*************************** 1. row ***************************! THREAD_ID: 137746! EVENT_ID: 3604! END_EVENT_ID: 3604! EVENT_NAME: wait/io/table/sql/handler! SOURCE: handler.cc:2568! TIMER_START: 17101711846046490! TIMER_END: 17101711857643010! TIMER_WAIT: 11596520! SPINS: NULL! OBJECT_SCHEMA: mem__events! OBJECT_NAME: events! INDEX_NAME: PRIMARY! OBJECT_TYPE: TABLE!OBJECT_INSTANCE_BEGIN: 140246283168288! NESTING_EVENT_ID: 3603! NESTING_EVENT_TYPE: STAGE! OPERATION: fetch! NUMBER_OF_BYTES: NULL! FLAGS: NULL

Event Context

Event type and origin

Timing (~11.5 microseconds)

Database object info

Object type

Whether the event was nested

Further info as appropriate



▪ Fixed lots of performance issues ▪ Focused on bringing the instrumentation to the DBA / Developer

▪ Statements / Statement Digests ▪ Execution Stages ▪ Object tracking (table / index IO, table locks) ▪ Network IO

▪ Turned on by default, with a subset of instrumentation disabled



▪ 52 Tables (+35) ▪ 545 Instruments (+323)

Instrument Type

Event ClassStatements statement/%Stages stage/%Table IO wait/io/table/%Table Locks wait/lock/table/%Network IO wait/io/socket/%Idle Timing idle


“I really think Performance Schema overhead is reasonable for most workloads. … On my old server I got some 20.2K QPS with Performance Schema Disabled and 19.4 QPS with Performance Schema enabled which is overhead of less than 5%. For most workloads paying 5% to have insight about what is happening with the system is a very fair trade.”

▪ Peter Zaitsev, CEO, Percona

http://www.mysqlperformanceblog.com/2014/02/11/performance_schema-vs-slow-query-log/


Statement & Stage Events (5.6)

!▪ Statements have two types of event

▪ SQL Statements statement/sql/% ▪ Protocol Commands statement/com/%

▪ Stages are the thread states. Like SHOW PROFILE output, but available across connections ▪ Mostly one form stage/sql/% ▪ Apart from .. stage/mysys/Waiting for table level lock


Stage Eventsmysql> select * from events_stages_history limit 2\G!*************************** 1. row ***************************! THREAD_ID: 23! EVENT_ID: 18898! END_EVENT_ID: 18898! EVENT_NAME: stage/sql/Master has sent all binlog to slave; waiting for more updates! SOURCE: rpl_binlog_sender.cc:420! TIMER_START: 87465041640516000! TIMER_END: 87465045562629000! TIMER_WAIT: 3922113000! NESTING_EVENT_ID: 10!NESTING_EVENT_TYPE: STATEMENT!*************************** 2. row ***************************! THREAD_ID: 23! EVENT_ID: 18899! END_EVENT_ID: 18900! EVENT_NAME: stage/sql/Sending binlog event to slave! SOURCE: rpl_binlog_sender.cc:432! TIMER_START: 87465045562629000! TIMER_END: 87465045593812000! TIMER_WAIT: 31183000! NESTING_EVENT_ID: 10!NESTING_EVENT_TYPE: STATEMENT


Statement Eventsmysql> select * from events_statements_history_long limit 1\G!*************************** 1. row ***************************! THREAD_ID: 138683! EVENT_ID: 11268! END_EVENT_ID: 11295! EVENT_NAME: statement/sql/update! SOURCE: socket_connection.cc:94! TIMER_START: 87644080265364000! TIMER_END: 87644080504810000! TIMER_WAIT: 239446000! LOCK_TIME: 63000000! SQL_TEXT: /* mem dbpool.default */ update `mem__inventory`.`Agent` set `hasHostname`=1400665852927, `hasReachable`=1400665852927, `timestamp`=1400665852927, `hasVersion`=1400665852927 where hid=x'FA8FDC4C1BC344A0899DAB320757CDF2'! DIGEST: cc389abfcb093ae95cacfe42ed085191! DIGEST_TEXT: UPDATE `mem__inventory` . `Agent` SET `hasHostname` = ? , `hasReachable` = ? , `timestamp` = ? , `hasVersion` = ? WHERE `hid` = ?! CURRENT_SCHEMA: mem! OBJECT_TYPE: NULL! OBJECT_SCHEMA: NULL! OBJECT_NAME: NULL! OBJECT_INSTANCE_BEGIN: NULL!…………


Statement Events (cont.)mysql> select * from events_statements_history_long limit 1\G!*************************** 1. row ***************************!……! MYSQL_ERRNO: 0! RETURNED_SQLSTATE: 00000! MESSAGE_TEXT: Rows matched: 1 Changed: 1 Warnings: 0! ERRORS: 0! WARNINGS: 0! ROWS_AFFECTED: 1! ROWS_SENT: 0! ROWS_EXAMINED: 1!CREATED_TMP_DISK_TABLES: 0! CREATED_TMP_TABLES: 0! SELECT_FULL_JOIN: 0! SELECT_FULL_RANGE_JOIN: 0! SELECT_RANGE: 0! SELECT_RANGE_CHECK: 0! SELECT_SCAN: 0! SORT_MERGE_PASSES: 0! SORT_RANGE: 0! SORT_ROWS: 0! SORT_SCAN: 0

NO_INDEX_USED: 0! NO_GOOD_INDEX_USED: 0! NESTING_EVENT_ID: 11248! NESTING_EVENT_TYPE: TRANSACTION! NESTING_EVENT_LEVEL: 0


Performance Schema Table Types

Setup Tables

▪ Used to define certain configuration dynamically

▪ Can perform DML against these tables

mysql> select table_name! -> from information_schema.tables! -> where table_schema like 'perf%'! -> and table_name like 'setup%';!+-------------------+!| table_name |!+-------------------+!| setup_actors |!| setup_consumers |!| setup_instruments |!| setup_objects |!| setup_timers |!+-------------------+



Raw Data Tables

▪ Expose events, objects, or instances of instruments in a raw manner

▪ Allow seeing a (brief) history of raw event metrics as well

+-------------------------------------------+!| table_name |!+-------------------------------------------+!| accounts |!| cond_instances |!| events_stages_current |!| events_stages_history |!| events_stages_history_long |!| events_statements_current |!| events_statements_history |!| events_statements_history_long |!| events_waits_current |!| events_waits_history |!| events_waits_history_long |!| file_instances |!| host_cache |!| hosts |!| mutex_instances |!| performance_timers |!| rwlock_instances |!| session_account_connect_attrs |!| session_connect_attrs |!| socket_instances |!| threads |!| users |!+-------------------------------------------+



Summary Tables

▪ Summarise event information over multiple dimensions

▪ Useful for longer term monitoring of activity

+------------------------------------------------------+!| table_name |!+------------------------------------------------------+!| events_stages_summary_by_account_by_event_name |!| events_stages_summary_by_host_by_event_name |!| events_stages_summary_by_thread_by_event_name |!| events_stages_summary_by_user_by_event_name |!| events_stages_summary_global_by_event_name |!| events_statements_summary_by_account_by_event_name |!| events_statements_summary_by_digest |!| events_statements_summary_by_host_by_event_name |!| events_statements_summary_by_program |!| events_statements_summary_by_thread_by_event_name |!| events_statements_summary_by_user_by_event_name |!| events_statements_summary_global_by_event_name |!| events_waits_summary_by_account_by_event_name |!| events_waits_summary_by_host_by_event_name |!| events_waits_summary_by_instance |!| events_waits_summary_by_thread_by_event_name |!| events_waits_summary_by_user_by_event_name |!| events_waits_summary_global_by_event_name |!| file_summary_by_event_name |!| file_summary_by_instance |!| objects_summary_global_by_type |!| socket_summary_by_event_name |!| socket_summary_by_instance |!| table_io_waits_summary_by_index_usage |!| table_io_waits_summary_by_table |!| table_lock_waits_summary_by_table |!+------------------------------------------------------+


Performance Schema Configuration


Performance Schema Configuration

▪ There are three distinct types of configuration ▪ A way to configure how much memory to allocate for instruments

▪ Set with various system variables in options files, not dynamic ▪ A way to enable/disable instrumentation at startup

▪ Again set with options files, only available as of 5.6 ▪ A way to enable/disable instrumentation dynamically

▪ Set by updating the various setup_% tables dynamically


Performance Configuration

History / Summary Size

▪ Define total rows per type ▪ The %_history_size are per-thread ▪ The %_history_long_size are total

rows ▪ Only settable in my.[cnf|ini]

mysql> select variable_name,! -> variable_value! -> from information_schema.global_variables! -> where variable_name like 'perf%'! -> and variable_name not like '%instances'! -> and variable_name not like '%classes'! -> order by variable_name;!+--------------------------------------------------------+----------------+!| variable_name | variable_value |!+--------------------------------------------------------+----------------+!| PERFORMANCE_SCHEMA | ON |!| PERFORMANCE_SCHEMA_ACCOUNTS_SIZE | 100 |!| PERFORMANCE_SCHEMA_DIGESTS_SIZE | 5000 |!| PERFORMANCE_SCHEMA_EVENTS_STAGES_HISTORY_LONG_SIZE | 1000 |!| PERFORMANCE_SCHEMA_EVENTS_STAGES_HISTORY_SIZE | 10 |!| PERFORMANCE_SCHEMA_EVENTS_STATEMENTS_HISTORY_LONG_SIZE | 1000 |!| PERFORMANCE_SCHEMA_EVENTS_STATEMENTS_HISTORY_SIZE | 10 |!| PERFORMANCE_SCHEMA_EVENTS_WAITS_HISTORY_LONG_SIZE | 1000 |!| PERFORMANCE_SCHEMA_EVENTS_WAITS_HISTORY_SIZE | 10 |!| PERFORMANCE_SCHEMA_HOSTS_SIZE | 100 |!| PERFORMANCE_SCHEMA_MAX_FILE_HANDLES | 32768 |!| PERFORMANCE_SCHEMA_MAX_TABLE_HANDLES | 732 |!| PERFORMANCE_SCHEMA_SESSION_CONNECT_ATTRS_SIZE | 512 |!| PERFORMANCE_SCHEMA_SETUP_ACTORS_SIZE | 100 |!| PERFORMANCE_SCHEMA_SETUP_OBJECTS_SIZE | 100 |!| PERFORMANCE_SCHEMA_USERS_SIZE | 100 |!+--------------------------------------------------------+----------------+



Max Classes / Instances

▪ Classes count instrument implementations, i.e: ▪ “wait/io/file/sql/binlog”

▪ Instances are the different runtime instances of those things, i.e: ▪ “/data/logs/binlog.000123”

mysql> select variable_name, variable_value! -> from information_schema.global_variables! -> where variable_name like 'perf%classes'! -> or variable_name like 'perf%instances'! -> order by variable_name;!+------------------------------------------------------+----------------+!| variable_name | variable_value |!+------------------------------------------------------+----------------+!| PERFORMANCE_SCHEMA_MAX_COND_CLASSES | 80 |!| PERFORMANCE_SCHEMA_MAX_COND_INSTANCES | 946 |!| PERFORMANCE_SCHEMA_MAX_FILE_CLASSES | 50 |!| PERFORMANCE_SCHEMA_MAX_FILE_INSTANCES | 2170 |!| PERFORMANCE_SCHEMA_MAX_MUTEX_CLASSES | 200 |!| PERFORMANCE_SCHEMA_MAX_MUTEX_INSTANCES | 4586 |!| PERFORMANCE_SCHEMA_MAX_RWLOCK_CLASSES | 40 |!| PERFORMANCE_SCHEMA_MAX_RWLOCK_INSTANCES | 3283 |!| PERFORMANCE_SCHEMA_MAX_SOCKET_CLASSES | 10 |!| PERFORMANCE_SCHEMA_MAX_SOCKET_INSTANCES | 86 |!| PERFORMANCE_SCHEMA_MAX_STAGE_CLASSES | 150 |!| PERFORMANCE_SCHEMA_MAX_STATEMENT_CLASSES | 189 |!| PERFORMANCE_SCHEMA_MAX_TABLE_INSTANCES | 569 |!| PERFORMANCE_SCHEMA_MAX_THREAD_CLASSES | 50 |!| PERFORMANCE_SCHEMA_MAX_THREAD_INSTANCES | 143 |!+------------------------------------------------------+----------------+



Max Classes / Instances

▪ Monitor the %_lost status variables opposite to see whether these need tweaking

▪ You will generally not need to modify classes, only when loading new plugins

mysql> select variable_name, variable_value! -> from information_schema.global_status! -> where variable_name like 'perf%classes_lost'! -> or variable_name like 'perf%instances_lost'! -> order by variable_name;!+-------------------------------------------+----------------+!| variable_name | variable_value |!+-------------------------------------------+----------------+!| PERFORMANCE_SCHEMA_COND_CLASSES_LOST | 0 |!| PERFORMANCE_SCHEMA_COND_INSTANCES_LOST | 0 |!| PERFORMANCE_SCHEMA_FILE_CLASSES_LOST | 0 |!| PERFORMANCE_SCHEMA_FILE_INSTANCES_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_SOCKET_CLASSES_LOST | 0 |!| PERFORMANCE_SCHEMA_SOCKET_INSTANCES_LOST | 0 |!| PERFORMANCE_SCHEMA_STAGE_CLASSES_LOST | 0 |!| PERFORMANCE_SCHEMA_STATEMENT_CLASSES_LOST | 0 |!| PERFORMANCE_SCHEMA_TABLE_INSTANCES_LOST | 0 |!| PERFORMANCE_SCHEMA_THREAD_CLASSES_LOST | 0 |!| PERFORMANCE_SCHEMA_THREAD_INSTANCES_LOST | 0 |!+-------------------------------------------+----------------+



setup_instruments

▪ Turn on/off individual instruments, or whether to just count and not time events

▪ Use UPDATE to modify dynamically

mysql> select * from setup_instruments limit 5;!+-------------------------------------------------------+---------+-------+!| NAME | ENABLED | TIMED |!+-------------------------------------------------------+---------+-------+!| wait/synch/mutex/sql/TC_LOG_MMAP::LOCK_tc | NO | NO |!| wait/synch/mutex/sql/LOCK_des_key_file | NO | NO |!| wait/synch/mutex/sql/MYSQL_BIN_LOG::LOCK_commit | NO | NO |!| wait/synch/mutex/sql/MYSQL_BIN_LOG::LOCK_commit_queue | NO | NO |!| wait/synch/mutex/sql/MYSQL_BIN_LOG::LOCK_done | NO | NO |!+-------------------------------------------------------+---------+-------+!5 rows in set (0.00 sec)!!mysql> update setup_instruments! -> set enabled = 'yes', timed = 'yes'! -> where name like '%MYSQL_BIN_LOG%';!Query OK, 16 rows affected (0.00 sec)!Rows matched: 16 Changed: 16 Warnings: 0!!mysql> select * from setup_instruments limit 5;!+-------------------------------------------------------+---------+-------+!| NAME | ENABLED | TIMED |!+-------------------------------------------------------+---------+-------+!| wait/synch/mutex/sql/TC_LOG_MMAP::LOCK_tc | NO | NO |!| wait/synch/mutex/sql/LOCK_des_key_file | NO | NO |!| wait/synch/mutex/sql/MYSQL_BIN_LOG::LOCK_commit | YES | YES |!| wait/synch/mutex/sql/MYSQL_BIN_LOG::LOCK_commit_queue | YES | YES |!| wait/synch/mutex/sql/MYSQL_BIN_LOG::LOCK_done | YES | YES |!+-------------------------------------------------------+---------+-------+



setup_consumers

▪ Define how much to record, either in history, or summaries (by enabling the %_current for the class of event)

mysql> select * from setup_consumers;!+----------------------------------+---------+!| NAME | ENABLED |!+----------------------------------+---------+!| events_stages_current | NO |!| events_stages_history | NO |!| events_stages_history_long | NO |!| events_statements_current | YES |!| events_statements_history | NO |!| events_statements_history_long | NO |!| events_transactions_current | NO |!| events_transactions_history | NO |!| events_transactions_history_long | NO |!| events_waits_current | NO |!| events_waits_history | NO |!| events_waits_history_long | NO |!| global_instrumentation | YES |!| thread_instrumentation | YES |!| statements_digest | YES |!+----------------------------------+---------+


Performance Schema Consumer Hierarchy



setup_objects (5.6)

▪ Define exactly which database objects to monitor

▪ Filters standard databases by default

▪ Allows explicit enable/disable, matches on most specific

mysql> select * from setup_objects;!+-------------+--------------------+-------------+---------+-------+!| OBJECT_TYPE | OBJECT_SCHEMA | OBJECT_NAME | ENABLED | TIMED |!+-------------+--------------------+-------------+---------+-------+!| TABLE | mysql | % | NO | NO |!| TABLE | performance_schema | % | NO | NO |!| TABLE | information_schema | % | NO | NO |!| TABLE | % | % | YES | YES |!+-------------+--------------------+-------------+---------+-------+!4 rows in set (0.00 sec)!!mysql> insert into setup_objects! -> values ('TABLE', 'foo', 'bar', 'no', 'no');!Query OK, 1 row affected (0.01 sec)!!mysql> select * from setup_objects;!+-------------+--------------------+-------------+---------+-------+!| OBJECT_TYPE | OBJECT_SCHEMA | OBJECT_NAME | ENABLED | TIMED |!+-------------+--------------------+-------------+---------+-------+!| TABLE | mysql | % | NO | NO |!| TABLE | performance_schema | % | NO | NO |!| TABLE | information_schema | % | NO | NO |!| TABLE | % | % | YES | YES |!| TABLE | foo | bar | NO | NO |!+-------------+--------------------+-------------+---------+-------+



setup_actors (5.6)

▪ Define exactly which users to monitor

▪ By default monitors all connections (% is wildcard)

▪ Modifications only apply to new connections

mysql> select * from setup_actors;!+------+------+------+!| HOST | USER | ROLE |!+------+------+------+!| % | % | % |!+------+------+------+!1 row in set (0.00 sec)!!mysql> delete from setup_actors;!Query OK, 1 row affected (0.00 sec)!!mysql> insert into setup_actors! -> values ('%', 'mark', '%');!Query OK, 1 row affected (0.00 sec)!!mysql> select * from setup_actors;!+------+------+------+!| HOST | USER | ROLE |!+------+------+------+!| % | mark | % |!+------+------+------+!1 row in set (0.00 sec)


Performance Schema Options Files Config

▪ As of 5.6, allows altering configuration at start up ▪ Enabling / disabling instrumentation

▪ performance_schema_instrument = ‘instrument_name=value’ ▪ Value = [on|1|true] | [off|0|false] | counted ▪ instrument_name can have wildcards (‘wait/synch/mutex/%=off’)

▪ Enabling / disabling consumers ▪ performance_schema_consumer_consumer_name=value

▪ Value = [on|1|true] | [off|0|false]


Profiling General Instance Activity


Profiling Considerations

▪ It is easy to just enable everything, but there are overhead concerns ▪ On busy systems mutexes can be locked millions of times per second

▪ It is best to just pick higher latency event types ▪ Statements, Stages, Table IO, File IO, maybe Network IO ▪ It is good to enable all %_current consumers, and statement history

▪ For concurrency/profiling debugging toggle other instruments on an as needed basis


Profiling Types

▪ Once you’ve narrowed down what you’re interested in, there are two ways to start monitoring

▪ View raw data in the summary views ▪ Gives you an overall picture of usage on the instance

▪ Snapshot data, and compute deltas over time ▪ Gives you an idea of the rates of change for the events


Analyzing Global Waits

▪ Some waits can include other waits ▪ Table IO latency may also include some mutex and file IO latency

▪ Wait times include concurrency across all threads ▪ Do not assume you can sum all events in global tables and compare

to wall clock times


Top Waits By Latency

mysql> select event_name,! -> count_star as count,! -> sys.format_time(sum_timer_wait) as total_latency,! -> sys.format_time(avg_timer_wait) as avg_latency,! -> sys.format_time(max_timer_wait) as max_latency! -> from events_waits_summary_global_by_event_name! -> where event_name != 'idle'! -> order by sum_timer_wait desc limit 5;!+--------------------------------------+----------+---------------+-------------+-------------+!| event_name | count | total_latency | avg_latency | max_latency |!+--------------------------------------+----------+---------------+-------------+-------------+!| wait/io/table/sql/handler | 21888502 | 1.27h | 208.06 us | 2.21 s |!| wait/io/file/innodb/innodb_data_file | 4276800 | 00:48:12.49 | 676.32 us | 1.49 s |!| wait/io/file/innodb/innodb_log_file | 1948199 | 00:25:24.36 | 782.45 us | 1.30 s |!| wait/io/file/myisam/kfile | 4566406 | 00:13:45.92 | 180.87 us | 1.17 s |!| wait/io/file/myisam/dfile | 1277589 | 00:05:46.23 | 271.01 us | 1.18 s |!+--------------------------------------+----------+---------------+-------------+-------------+



▪ Some mutex events that can affect global concurrency (if high in list): ▪ wait/synch/mutex/innodb/buf_pool_mutex

▪ Increase innodb_buffer_pool_instances ▪ wait/synch/mutex/sql/Query_cache::structure_guard_mutex

▪ Look in to disabling the Query Cache ▪ wait/synch/mutex/myisam/MYISAM_SHARE::intern_lock

▪ Use Innodb …



▪ Some File IO events to watch for (if high in list): ▪ wait/io/file/sql/FRM

▪ Tune table_open_cache / table_definition_cache ▪ wait/io/file/sql/file_parser (view definition parsing)

▪ If high on 5.5, upgrade to 5.6, (which can cache these like tables) ▪ wait/io/file/sql/query_log and wait/io/file/sql/slow_log

▪ Disable the general ▪ Disable or tune what is logged to the slow log (decent long_query_time)


Top Files By Total IO mysql> select sys.format_path(file_name) as file,! -> count_read,! -> sys.format_bytes(sum_number_of_bytes_read) as total_read,! -> sys.format_bytes(IFNULL(sum_number_of_bytes_read / count_read, 0)) as avg_read,! -> count_write,! -> sys.format_bytes(sum_number_of_bytes_write) as total_written,! -> sys.format_bytes(IFNULL(sum_number_of_bytes_write / count_write, 0)) as avg_write,! -> sys.format_bytes(sum_number_of_bytes_read + sum_number_of_bytes_write) as total,! -> IFNULL(ROUND(100-((sum_number_of_bytes_read/(sum_number_of_bytes_read + sum_number_of_bytes_write))*100), 2), 0.00) as write_pct! -> from file_summary_by_instance! -> order by (sum_number_of_bytes_read + sum_number_of_bytes_write) desc limit 5;!

+----------------------------------+------------+------------+-----------+-------------+---------------+-----------+-------------+-----------+!| file | count_read | total_read | avg_read | count_write | total_written | avg_write | total | write_pct |!+----------------------------------+------------+------------+-----------+-------------+---------------+-----------+-------------+-----------+!| @@datadir/ibdata1 | 888 | 15.84 MiB | 18.27 KiB | 1089824 | 61.99 GiB | 59.64 KiB | 62.00 GiB | 99.98 |!| @@datadir/mem__events/events.ibd | 114 | 1.80 MiB | 16.14 KiB | 117370 | 2.14 GiB | 19.14 KiB | 2.14 GiB | 99.92 |!| @@datadir/cerberus-bin.000010 | 296103 | 1.01 GiB | 3.59 KiB | 362852 | 1.00 GiB | 2.89 KiB | 2.01 GiB | 49.63 |!| @@datadir/ib_logfile0 | 6 | 68.00 KiB | 11.33 KiB | 506837 | 1.46 GiB | 3.03 KiB | 1.46 GiB | 100.00 |!| @@datadir/ib_logfile1 | 0 | 0 bytes | 0 bytes | 476961 | 1.45 GiB | 3.19 KiB | 1.45 GiB | 100.00 |!+----------------------------------+------------+------------+-----------+-------------+---------------+-----------+-------------+-----------+

High IO on InnoDB per-tablespace tables can show candidates for a separate mountpoint/disk using “DATA DIRECTORY” in CREATE TABLE


Analyzing User Activity

▪ All event_% summaries are exposed with a number of dimensions ▪ To analyze connection activity you can do this in 4 ways

▪ By User ▪ By Host ▪ By Account (User@Host) ▪ By Thread

▪ The follow examples are by user, but could be replaced with the host, account or thread summary views


Top Users By Statement Latency

mysql> select user,! -> sum(count_star) as statements,! -> sys.format_time(sum(sum_timer_wait)) as total_latency,! -> sys.format_time(sum(sum_timer_wait) / sum(count_star)) as avg_latency! -> from events_statements_summary_by_user_by_event_name! -> where user is not null! -> group by user! -> order by sum(sum_timer_wait) desc;!+------+------------+---------------+-------------+!| user | statements | total_latency | avg_latency |!+------+------------+---------------+-------------+!| root | 7229032 | 15.17h | 7.55 ms |!| mark | 3072 | 1.77 s | 575.29 us |!+------+------------+---------------+-------------+!


Top Users By IO Latency

mysql> select user, sum(count_star) as count,! -> sys.format_time(sum(sum_timer_wait)) as total_latency! -> from events_waits_summary_by_user_by_event_name! -> where event_name like 'wait/io/file/%'! -> and user is not null! -> group by user! -> order by sum(sum_timer_wait) desc;!+------+----------+---------------+!| user | count | total_latency |!+------+----------+---------------+!| root | 10892980 | 00:46:19.67 |!| mark | 20043 | 346.79 ms |!+------+----------+---------------+!

Replace with some other pattern here for “Top user by

wait class”


Top Users By Connectionsmysql> select * from users where user is not null order by current_connections desc;!+------+---------------------+-------------------+!| USER | CURRENT_CONNECTIONS | TOTAL_CONNECTIONS |!+------+---------------------+-------------------+!| root | 8 | 151655 |!| mark | 1 | 1 |!+------+---------------------+-------------------+!2 rows in set (0.00 sec)!!mysql> select * from accounts where user is not null order by current_connections desc;!+------+-----------+---------------------+-------------------+!| USER | HOST | CURRENT_CONNECTIONS | TOTAL_CONNECTIONS |!+------+-----------+---------------------+-------------------+!| root | localhost | 39 | 151698 |!| mark | localhost | 1 | 1 |!+------+-----------+---------------------+-------------------+!2 rows in set (0.00 sec)


User Statement Activitymysql> select *! -> from events_statements_summary_by_user_by_event_name! -> where user is not null! -> order by user, sum_timer_wait desc limit 20\G!*************************** 1. row ***************************! USER: mark! EVENT_NAME: statement/sql/select! COUNT_STAR: 4! SUM_TIMER_WAIT: 1692883853000! MIN_TIMER_WAIT: 88330000! AVG_TIMER_WAIT: 423220963000! MAX_TIMER_WAIT: 1692583043000! SUM_LOCK_TIME: 408000000! SUM_ERRORS: 0! SUM_WARNINGS: 0! SUM_ROWS_AFFECTED: 0! SUM_ROWS_SENT: 212! SUM_ROWS_EXAMINED: 15067!SUM_CREATED_TMP_DISK_TABLES: 3! SUM_CREATED_TMP_TABLES: 4

SUM_SELECT_FULL_JOIN: 0! SUM_SELECT_FULL_RANGE_JOIN: 0! SUM_SELECT_RANGE: 0! SUM_SELECT_RANGE_CHECK: 0! SUM_SELECT_SCAN: 2! SUM_SORT_MERGE_PASSES: 6! SUM_SORT_RANGE: 0! SUM_SORT_ROWS: 6458! SUM_SORT_SCAN: 2! SUM_NO_INDEX_USED: 1! SUM_NO_GOOD_INDEX_USED: 0

Summarized by statement type, e.g:

!statement/sql/select statement/sql/update


User Top Stagesmysql> select user, event_name as stage,! -> count_star as total,! -> sys.format_time(sum_timer_wait) as total_latency,! -> sys.format_time(avg_timer_wait) as avg_latency,! -> sys.format_time(max_timer_wait) as max_latency! -> from events_stages_summary_by_user_by_event_name! -> where sum_timer_wait > 0! -> order by user, sum_timer_wait desc;!+------+--------------------------------+--------+---------------+-------------+-------------+!| user | stage | total | total_latency | avg_latency | max_latency |!+------+--------------------------------+--------+---------------+-------------+-------------+!| root | stage/sql/starting | 591184 | 00:47:52.47 | 4.86 ms | 3.96 s |!| root | stage/sql/update | 193940 | 00:15:00.84 | 4.64 ms | 4.02 s |!| root | stage/sql/updating | 50334 | 00:01:41.38 | 2.01 ms | 2.35 s |!| root | stage/sql/statistics | 73869 | 00:01:41.38 | 1.37 ms | 4.54 s |!| root | stage/sql/Sending data | 73211 | 00:01:38.31 | 1.34 ms | 1.04 s |!| root | stage/sql/Opening tables | 404230 | 00:01:23.48 | 206.51 us | 473.72 ms |!| root | stage/sql/closing tables | 567351 | 00:01:13.40 | 129.38 us | 79.52 ms |!| root | stage/sql/removing tmp table | 11603 | 56.87 s | 4.90 ms | 964.17 ms |!| root | stage/sql/freeing items | 567520 | 19.75 s | 34.80 us | 219.19 ms |!| root | stage/sql/init | 316805 | 8.26 s | 26.06 us | 217.54 ms |!………


Analyzing Table Activity

▪ 3 summary views have been added to 5.6 on top of Table IO ▪ objects_summary_global_by_type

▪ A high level aggregate view of database object latency ▪ table_io_waits_summary_by_table

▪ A detailed aggregate by table, breaking down reads, writes, etc. ▪ table_io_waits_summary_by_index_usage

▪ Further breaking down usage per index


Top Tables By Latencymysql> select object_schema,! -> object_name,! -> count_star,! -> sys.format_time(sum_timer_wait) as total_latency,! -> sys.format_time(sum_timer_wait / count_star) as avg_latency,! -> sys.format_time(max_timer_wait) as max_latency! -> from objects_summary_global_by_type! -> order by sum_timer_wait desc limit 5;!+------------------+-------------------+------------+---------------+-------------+-------------+!| object_schema | object_name | count_star | total_latency | avg_latency | max_latency |!+------------------+-------------------+------------+---------------+-------------+-------------+!| mem__inventory | mysqlserver | 1049496 | 00:04:48.32 | 274.72 us | 1.50 s |!| mem__events | events | 227939 | 00:01:00.50 | 265.41 us | 2.15 s |!| mem__events | event_chain_heads | 328664 | 50.80 s | 154.57 us | 2.21 s |!| mem__inventory | agent | 442751 | 44.82 s | 101.23 us | 1.69 s |!| mem__instruments | fsstatistics | 2233484 | 16.53 s | 7.40 us | 269.27 ms |!+------------------+-------------------+------------+---------------+-------------+-------------+

Less events, yet higher latency, is a sign that there could be

concurrency issues


Table Usage Detailed Breakdown!mysql> select object_schema, object_name,! -> count_fetch as selects, sys.format_time(sum_timer_fetch) as select_latency,! -> count_insert as inserts, sys.format_time(sum_timer_insert) as insert_latency,! -> count_update as updates, sys.format_time(sum_timer_update) as update_latency,! -> count_delete as deletes, sys.format_time(sum_timer_delete) as delete_latency! -> from table_io_waits_summary_by_table! -> order by sum_timer_wait desc limit 10;!+------------------+-------------------+---------+----------------+---------+----------------+---------+----------------+---------+----------------+!| object_schema | object_name | selects | select_latency | inserts | insert_latency | updates | update_latency | deletes | delete_latency |!+------------------+-------------------+---------+----------------+---------+----------------+---------+----------------+---------+----------------+!| mem__inventory | mysqlserver | 418518 | 00:04:32.21 | 0 | 0 ps | 48534 | 15.51 s | 0 | 0 ps |!| mem__events | events | 58512 | 5.54 s | 2701 | 4.01 s | 52414 | 50.83 s | 0 | 0 ps |!| mem__events | event_chain_heads | 156194 | 6.93 s | 78 | 59.75 ms | 52366 | 43.60 s | 0 | 0 ps |!| mem__inventory | agent | 188113 | 40.06 s | 0 | 0 ps | 6998 | 4.54 s | 0 | 0 ps |!| mem__instruments | fsstatistics | 2222058 | 15.12 s | 1517 | 1.37 s | 1 | 72.69 us | 0 | 0 ps |!| mem__instruments | qrtidata | 2771546 | 2.57 s | 3030 | 12.52 s | 0 | 0 ps | 0 | 0 ps |!| mem__inventory | network | 91187 | 2.24 s | 0 | 0 ps | 11864 | 8.42 s | 0 | 0 ps |!| mem__inventory | os | 79736 | 6.08 s | 0 | 0 ps | 8827 | 4.07 s | 0 | 0 ps |!| mem__inventory | networking | 99114 | 4.24 s | 0 | 0 ps | 5928 | 5.54 s | 0 | 0 ps |!| mem__inventory | environment | 100445 | 5.29 s | 0 | 0 ps | 6540 | 2.53 s | 0 | 0 ps |!+------------------+-------------------+---------+----------------+---------+----------------+---------+----------------+---------+----------------+



Tables with full scans

▪ Search in table_io_waits_summary_by_index_usage where index_name is null

▪ This is a catch all row for rows that are read without indexes

mysql> select concat(object_schema, '.', object_name) as object,! -> count_read as rows_scanned,! -> sys.format_time(sum_timer_wait) as latency! -> from table_io_waits_summary_by_index_usage! -> where index_name is null! -> and count_read > 0! -> order by sum_timer_wait desc limit 20;!+---------------------------------------+--------------+-----------+!| object | rows_scanned | latency |!+---------------------------------------+--------------+-----------+!| mem__instruments.qrtidata | 2735434 | 14.56 s |!| mem__instruments.databaseactivitydata | 2735633 | 4.68 s |!| mem__instruments.diskiototaldata | 31813 | 3.26 s |!| mem__instruments.connectionsdata | 2530816 | 3.17 s |!| mem__inventory.agent | 117544 | 2.89 s |!| mem__instruments.fsstatistics | 738626 | 2.69 s |!| mem__instruments.diskioopstotaldata | 25955 | 2.43 s |!| mem__inventory.environment | 42549 | 1.55 s |!| mem__enterprise.whats_new_entries | 1453 | 1.16 s |!| mem__inventory.networking | 42549 | 559.14 ms |!+---------------------------------------+--------------+-----------+


Host Cache Usage and Errors

▪ New host_cache view within 5.6, exposing hosts in the host cache ▪ Counts errors, by the type of error that can occur ▪ Can show when errors started happening ▪ Compare the sum_connect_errors counter to the

max_connect_errors system variable


host_cachemysql> desc host_cache;!+--------------------------------------------+------------------+------+-----+---------------------+-------+!| Field | Type | Null | Key | Default | Extra |!+--------------------------------------------+------------------+------+-----+---------------------+-------+!| IP | varchar(64) | NO | | NULL | |!| HOST | varchar(255) | YES | | NULL | |!| HOST_VALIDATED | enum('YES','NO') | NO | | NULL | |!| SUM_CONNECT_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_HOST_BLOCKED_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_NAMEINFO_TRANSIENT_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_NAMEINFO_PERMANENT_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_FORMAT_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_ADDRINFO_TRANSIENT_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_ADDRINFO_PERMANENT_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_FCRDNS_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_HOST_ACL_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_NO_AUTH_PLUGIN_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_AUTH_PLUGIN_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_HANDSHAKE_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_PROXY_USER_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_PROXY_USER_ACL_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_AUTHENTICATION_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_SSL_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_MAX_USER_CONNECTIONS_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_MAX_USER_CONNECTIONS_PER_HOUR_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_DEFAULT_DATABASE_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_INIT_CONNECT_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_LOCAL_ERRORS | bigint(20) | NO | | NULL | |!| COUNT_UNKNOWN_ERRORS | bigint(20) | NO | | NULL | |!| FIRST_SEEN | timestamp | NO | | 0000-00-00 00:00:00 | |!| LAST_SEEN | timestamp | NO | | 0000-00-00 00:00:00 | |!| FIRST_ERROR_SEEN | timestamp | YES | | 0000-00-00 00:00:00 | |!| LAST_ERROR_SEEN | timestamp | YES | | 0000-00-00 00:00:00 | |


Computing Rates Of Change

▪ To look at how things change over time, you have a couple of options ▪ Record the events you are interested in within persistent history tables

▪ Get current stats from performance_schema, the last rows of history, compute the time delta and event count deltas, store new deltas along with the raw values

▪ Or use some tool like MySQL Enterprise Monitor ▪ Read the tables regularly, and run TRUNCATE TABLE on them

immediately afterwards ▪ Some users have done this by dumping local traces to files on

disk, and resetting stats every 15 minutes, for example



▪ For an example, the following event tracks the time spent waiting in the replication SQL thread for new events to be written to the relay log by the IO thread ▪ wait/synch/cond/sql/MYSQL_RELAY_LOG::update_cond

▪ This is essentially the SQL thread idle time, for a single SQL thread, if we snapshot this over time, we can find overall busy time


Slave SQL Load Average

▪ An example of monitoring this event over time !

http://www.markleith.co.uk/2012/07/24/a-mysql-replication-load-average-with-performance-schema/


Slave SQL Load Average

http://www.markleith.co.uk/2012/07/24/a-mysql-replication-load-average-with-performance-schema/



▪ If interested in calculating percentages, copy the approach shown in the replication load average – you have to take in to account waiting 100% on other events, or just this event, etc.

▪ Note, this method can not be used generally, you have to be able to correlate it to specific thread(s)

▪ As latency is per thread, and you have to compare to a known time delta, summary views do not give this to you (you don’t get thread counts per interval)

▪ Should still compute the deltas for global events however, for event throughput / latency statistics over time !


Profiling Statement Activity


Statement Profiling Options

▪ The new instrumentation in 5.6 gives many options ▪ See statements running currently with events_statements_current ▪ Summary views by user, host, account ▪ Statement histories within events_statements_history% tables ▪ A normalized view within events_statements_summary_by_digest


Currently Executing Statementsmysql> select * from events_statements_current where timer_end is null\G!*************************** 8. row ***************************! THREAD_ID: 156945! EVENT_ID: 15312! END_EVENT_ID: NULL! EVENT_NAME: statement/sql/select! SOURCE: socket_connection.cc:94! TIMER_START: 99343994725205000! TIMER_END: NULL! TIMER_WAIT: NULL! LOCK_TIME: 145000000! SQL_TEXT: /* mem dbpool.ui */ select normalized0_.round_robin_bin as round1_1256_, /* ..snip .. */! DIGEST: NULL! DIGEST_TEXT: NULL! CURRENT_SCHEMA: mem! OBJECT_TYPE: NULL! OBJECT_SCHEMA: NULL! OBJECT_NAME: NULL! OBJECT_INSTANCE_BEGIN: NULL! MYSQL_ERRNO: 0! RETURNED_SQLSTATE: NULL! MESSAGE_TEXT: NULL! ERRORS: 0! WARNINGS: 0! ROWS_AFFECTED: 0! ROWS_SENT: 27! ROWS_EXAMINED: 0!CREATED_TMP_DISK_TABLES: 0! CREATED_TMP_TABLES: 0

SELECT_FULL_JOIN: 0! SELECT_FULL_RANGE_JOIN: 0! SELECT_RANGE: 1! SELECT_RANGE_CHECK: 0! SELECT_SCAN: 0! SORT_MERGE_PASSES: 0! SORT_RANGE: 0! SORT_ROWS: 0! SORT_SCAN: 0! NO_INDEX_USED: 0! NO_GOOD_INDEX_USED: 0! NESTING_EVENT_ID: NULL! NESTING_EVENT_TYPE: NULL! NESTING_EVENT_LEVEL: 0

All counters are live, and increment whilst the statements

execute


Per Thread Statement Historymysql> select thread_id,! -> sys.format_statement(sql_text) as stmt,! -> sys.format_time(timer_wait) as latency! -> from events_statements_history! -> order by thread_id, event_id;!+-----------+-------------------------------------------------------------------+-----------+!| thread_id | stmt | latency |!+-----------+-------------------------------------------------------------------+-----------+!...!| 158162 | /* mem dbpool.default */ commit | 5.39 ms |!| 158162 | /* mem dbpool.default */ inser ... currentlyOpen), currentlyOpen) | 1.25 ms |!| 158162 | /* mem dbpool.default */ commit | 9.69 ms |!| 158162 | /* mem dbpool.default */ inser ... , VALUES(lastSeen)), lastSeen) | 259.75 us |!| 158162 | /* mem dbpool.default */ inser ... (bytesTotal), bytesTotal), col | 383.05 us |!| 158162 | /* mem dbpool.default */ commit | 51.41 ms |!| 158162 | /* mem dbpool.default */ inser ... ullScan), rowsReadviaFullScan) | 147.04 ms |!| 158162 | /* mem dbpool.default */ commit | 33.87 ms |!| 158162 | /* mem dbpool.default */ inser ... ALUES(indexUsage), indexUsage) | 1.10 ms |!| 158162 | /* mem dbpool.default */ commit | 14.78 ms |!...!+-----------+-------------------------------------------------------------------+-----------+!370 rows in set (0.17 sec)

Also has same columns as events_statements_current


Statement Digests

▪ A new summary view introduced in 5.6 ▪ Aggregates statistics based on normalised statements ▪ Can be used to drill in to your problem statements instead of using the

slow query log (with less contention, and no need for post-processing) ▪ Each statement gets an MD5 ‘DIGEST’, that is also available to link in

to events_statements_history% etc. for raw per-query statistics


events_statements_summary_by_digestmysql> desc events_statements_summary_by_digest;!+-----------------------------+---------------------+------+-----+---------------------+-------+!| Field | Type | Null | Key | Default | Extra |!+-----------------------------+---------------------+------+-----+---------------------+-------+!| SCHEMA_NAME | varchar(64) | YES | | NULL | |!| DIGEST | varchar(32) | YES | | NULL | |!| DIGEST_TEXT | longtext | YES | | NULL | |!| COUNT_STAR | bigint(20) unsigned | NO | | NULL | |!| SUM_TIMER_WAIT | bigint(20) unsigned | NO | | NULL | |!| MIN_TIMER_WAIT | bigint(20) unsigned | NO | | NULL | |!| AVG_TIMER_WAIT | bigint(20) unsigned | NO | | NULL | |!| MAX_TIMER_WAIT | bigint(20) unsigned | NO | | NULL | |!| SUM_LOCK_TIME | bigint(20) unsigned | NO | | NULL | |!| SUM_ERRORS | bigint(20) unsigned | NO | | NULL | |!| SUM_WARNINGS | bigint(20) unsigned | NO | | NULL | |!| SUM_ROWS_AFFECTED | bigint(20) unsigned | NO | | NULL | |!| SUM_ROWS_SENT | bigint(20) unsigned | NO | | NULL | |!| SUM_ROWS_EXAMINED | bigint(20) unsigned | NO | | NULL | |!| SUM_CREATED_TMP_DISK_TABLES | bigint(20) unsigned | NO | | NULL | |!| SUM_CREATED_TMP_TABLES | bigint(20) unsigned | NO | | NULL | |!| SUM_SELECT_FULL_JOIN | bigint(20) unsigned | NO | | NULL | |!| SUM_SELECT_FULL_RANGE_JOIN | bigint(20) unsigned | NO | | NULL | |!| SUM_SELECT_RANGE | bigint(20) unsigned | NO | | NULL | |!| SUM_SELECT_RANGE_CHECK | bigint(20) unsigned | NO | | NULL | |!| SUM_SELECT_SCAN | bigint(20) unsigned | NO | | NULL | |!| SUM_SORT_MERGE_PASSES | bigint(20) unsigned | NO | | NULL | |!| SUM_SORT_RANGE | bigint(20) unsigned | NO | | NULL | |!| SUM_SORT_ROWS | bigint(20) unsigned | NO | | NULL | |!| SUM_SORT_SCAN | bigint(20) unsigned | NO | | NULL | |!| SUM_NO_INDEX_USED | bigint(20) unsigned | NO | | NULL | |!| SUM_NO_GOOD_INDEX_USED | bigint(20) unsigned | NO | | NULL | |!| FIRST_SEEN | timestamp | NO | | 0000-00-00 00:00:00 | |!| LAST_SEEN | timestamp | NO | | 0000-00-00 00:00:00 | |


Statement Normalization

▪ Normalization folds certain constructs of statements ▪ Strip whitespace / comments ▪ Replace literals with ?

▪ “WHERE foo = 1” becomes “WHERE foo = ?” ▪ Fold lists of things

▪ “IN (1,2,3)” becomes “IN (…)” ▪ Fold multi-row inserts

▪ “VALUES (1), (2), (3)” becomes “VALUES (?) /*, … */” !


Statement Digest Output

mysql> select * from events_statements_summary_by_digest order by sum_timer_wait desc limit 5\G!...!*************************** 4. row ***************************! SCHEMA_NAME: mem! DIGEST: 5f41a0036bae4fa1c79339c1c7da3c9e! DIGEST_TEXT: SELECT DISTINCTROW `agent0_` . `hid` AS `hid1239_` ...! COUNT_STAR: 64136! SUM_TIMER_WAIT: 1835031591954000! MIN_TIMER_WAIT: 209280000! AVG_TIMER_WAIT: 28611569000! MAX_TIMER_WAIT: 4050621865000! SUM_LOCK_TIME: 31388434000000! SUM_ERRORS: 0! SUM_WARNINGS: 0! SUM_ROWS_AFFECTED: 0! SUM_ROWS_SENT: 64133! SUM_ROWS_EXAMINED: 187363!SUM_CREATED_TMP_DISK_TABLES: 59624! SUM_CREATED_TMP_TABLES: 59624!

SUM_SELECT_FULL_JOIN: 0! SUM_SELECT_FULL_RANGE_JOIN: 0! SUM_SELECT_RANGE: 0! SUM_SELECT_RANGE_CHECK: 0! SUM_SELECT_SCAN: 59625! SUM_SORT_MERGE_PASSES: 0! SUM_SORT_RANGE: 0! SUM_SORT_ROWS: 0! SUM_SORT_SCAN: 0! SUM_NO_INDEX_USED: 59625! SUM_NO_GOOD_INDEX_USED: 0! FIRST_SEEN: 2014-05-20 11:26:54! LAST_SEEN: 2014-05-21 14:50:54


Statements With Temporary Tables

mysql> select sys.format_statement(digest_text) as stmt,! -> count_star as total,! -> sum_created_tmp_tables as in_memory,! -> sum_created_tmp_disk_tables as on_disk,! -> round(sum_created_tmp_tables / count_star) as avg_per_stmt,! -> round((sum_created_tmp_disk_tables/sum_created_tmp_tables) * 100) as to_disk_pct! -> from events_statements_summary_by_digest! -> where sum_created_tmp_tables > 0! -> order by sum_created_tmp_disk_tables desc, sum_created_tmp_tables desc limit 5;!+-------------------------------------------------------------------+-------+-----------+---------+--------------+-------------+!| stmt | total | in_memory | on_disk | avg_per_stmt | to_disk_pct |!+-------------------------------------------------------------------+-------+-----------+---------+--------------+-------------+!| SELECT DISTINCTROW àgent0_` . ... gent` àgent0_` INNER JOIN ... | 64668 | 60116 | 60116 | 1 | 100 |!| SELECT DISTINCTROW `mysqlconne ... conne0_` . `socketPath` AS ... | 5605 | 5604 | 5604 | 1 | 100 |!| SELECT * FROM ( SELECT digest ... ed AS `noIndexUsedCount` , ... | 1431 | 10017 | 4293 | 7 | 43 |!| SELECT `s` . ìdentityId` , `s ... `subject_id` = `s` . ìd` ... | 1406 | 4218 | 2812 | 3 | 67 |!| SELECT plugin_name FROM inform ... atus = ? ORDER BY plugin_name | 1726 | 1726 | 1726 | 1 | 100 |!+-------------------------------------------------------------------+-------+-----------+---------+--------------+-------------+!5 rows in set (0.01 sec)


Statements With Full Table Scans

mysql> select sys.format_statement(digest_text) as stmt,! -> count_star as total,! -> sum_no_index_used as scan_count,! -> round((sum_no_index_used/count_star) * 100) as scan_pct,! -> sum_rows_examined as rows_scanned! -> from events_statements_summary_by_digest! -> where digest_text like 'select%'! -> and (sum_no_index_used > 0 OR sum_no_good_index_used > 0)! -> order by rows_scanned desc limit 5;!+-------------------------------------------------------------------+-------+------------+----------+--------------+!| stmt | total | scan_count | scan_pct | rows_scanned |!+-------------------------------------------------------------------+-------+------------+----------+--------------+!| SELECT `hid` , TIMESTAMP , `en ... DIV ? AS `slice` , `hid` , ... | 4258 | 4257 | 100 | 31644768 |!| SELECT TIMESTAMP , SUM ( `sele ... by_bucket` GROUP BY TIMESTAMP | 807 | 807 | 100 | 4657959 |!| SELECT TIMESTAMP , SUM ( `opti ... by_bucket` GROUP BY TIMESTAMP | 807 | 807 | 100 | 4656726 |!| SELECT TIMESTAMP , SUM ( `tota ... by_bucket` GROUP BY TIMESTAMP | 807 | 807 | 100 | 4454821 |!| SELECT * FROM ( SELECT digest ... ed AS `noIndexUsedCount` , ... | 1434 | 1434 | 100 | 3874018 |!+-------------------------------------------------------------------+-------+------------+----------+--------------+


Tracing Individual Statement Activity

▪ By taking all available data within the %_history_long tables, we can build a complete picture of where latency lines within statements

▪ We can link the histories of Statements, Stages and Waits using the event_id and nesting_event_id columns, which define hierarchy !

!


This means we can graph their relationships!



▪ This is only really effective on development / test systems, production systems with high concurrency / throughput age history very quickly, and turning the %_history_long tables on with all instrumentation on is not recommend generally

▪ Good to disable all other threads, or all actors, and just enable a single thread in development / test environments, to get a full trace

▪ Procedure to dump the graph data is available within the MySQL SYS schema !

!


Improvements made to date in MySQL 5.7

Copyright © 2014, Oracle and/or its affiliates. All rights reserved.Copyright © 2014, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 12

Insert Picture Here

!76

▪Memory Usage ▪Metadata Locking ▪ Replication Configuration & Status ▪ Prepared Statements ▪ Transactions ▪ Stored Programs



▪ 75 Tables (+23) ▪ 784 Instruments (+239)

Instrument Type Event Class

Transactions transaction

Memory memory/%

Metadata Locks wait/lock/metadata/%


WL#3249!PERFORMANCE SCHEMA, Instrument memory usage


“Understanding where MySQL can allocate memory can help us to find the cause in most cases. It is not as straightforward as it should be and I’m very hopeful future releases of MySQL, MariaDB or Drizzle bring improvements in this space allowing us to see directly for what purpose memory is allocated and so detect all kinds of memory usage problems easier.”

▪ Peter Zaitsev, CEO, Percona

http://www.mysqlperformanceblog.com/2012/03/21/troubleshooting-mysql-memory-usage/


Instrument memory usage

▪ Added 212 different memory instrumentation types so far ▪ Records current, high and low water marks of allocations ▪ Does not track latency of memory allocation ▪ 5 new tables

▪ memory_summary_by_account_by_event_name ▪ memory_summary_by_host_by_event_name ▪ memory_summary_by_thread_by_event_name ▪ memory_summary_by_user_by_event_name ▪ memory_summary_global_by_event_name


Instrument memory usage - global summarymysql> SELECT * FROM sys.memory_global_by_current_allocated\G!*************************** 1. row ***************************! event_name: memory/performance_schema/internal_buffers! current_count: 60! current_alloc: 497.00 MiB!current_avg_alloc: 8.28 MiB! high_count: 60! high_alloc: 497.00 MiB! high_avg_alloc: 8.28 MiB!*************************** 2. row ***************************! event_name: memory/mysys/KEY_CACHE! current_count: 3! current_alloc: 8.00 MiB!current_avg_alloc: 2.67 MiB! high_count: 3! high_alloc: 8.00 MiB! high_avg_alloc: 2.67 MiB


Instrument memory usage - thread summariesmysql> select * from memory_by_thread_by_current_allocated\G!*************************** 1. row ***************************! user: sql/main! current_count: 2407! current_alloc: 10.89 MiB!current_avg_alloc: 4.63 KiB!current_max_alloc: 8.00 MiB! total_allocated: 30.55 MiB! thread_id: 1!*************************** 2. row ***************************! user: mem@localhost! current_count: 1914! current_alloc: 1.50 MiB!current_avg_alloc: 824 bytes!current_max_alloc: 816.67 KiB! total_allocated: 9.25 GiB! thread_id: 4336


Instrument memory usage - thread detailsmysql> SELECT event_name,! -> sys.format_bytes(current_number_of_bytes_used) AS current_used! -> FROM performance_schema.memory_summary_by_thread_by_event_name! -> WHERE thread_id = 24! -> ORDER BY current_number_of_bytes_used DESC;!+-----------------------------------------------------+--------------+!| event_name | current_used |!+-----------------------------------------------------+--------------+!| memory/sql/Filesort_buffer::sort_keys | 255.94 KiB |!| memory/sql/sp_head::main_mem_root | 103.64 KiB |!| memory/mysys/IO_CACHE | 64.05 KiB |!| memory/mysys/lf_dynarray | 46.17 KiB |!| memory/mysys/array_buffer | 24.20 KiB |!| memory/sql/thd::main_mem_root | 23.95 KiB |!| memory/sql/String::value | 16.13 KiB |!| memory/sql/TABLE | 9.44 KiB |!| memory/sql/TABLE_SHARE::mem_root | 8.70 KiB |!| memory/myisam/MI_INFO | 7.07 KiB |!| memory/sql/THD::transactions::mem_root | 4.02 KiB |!| memory/myisam/MYISAM_SHARE | 3.29 KiB |!…


Instrument memory usage - other details

!▪ Disabled by default !!!!

▪ Still requires InnoDB instrumentation

UPDATE setup_instruments SET enabled = ‘YES’ WHERE name LIKE ‘memory/%’;


WL#5879!PERFORMANCE SCHEMA, MDL lock instrumentation


“Unfortunately, it’s unlikely that I’ll be able to create a reproducible test case, because there’s no way to actually see what is happening. I hope that a future version of MySQL will include a more comprehensive set of tables for inspecting locks, requests, and waits at all layers of the server.”

▪ Baron Schwartz, CEO, VividCortex

http://www.xaprb.com/blog/2012/08/28/debugging-metadata-locking-in-mysql-5-5/


MDL Lock Instrumentation

▪ Added the wait/lock/metadata/sql/mdl instrument ▪ 2 new tables

▪ metadata_locks ▪ table_handles


MDL Lock Instrumentation▪ OBJECT_TYPE

▪ GLOBAL, SCHEMA, TABLE, FUNCTION, PROCEDURE, TRIGGER, EVENT, COMMIT

▪ LOCK_TYPE ▪ INTENTION_EXCLUSIVE, SHARED, SHARED_HIGH_PRIO,

SHARED_READ, SHARED_WRITE, SHARED_UPGRADABLE, SHARED_NO_WRITE, SHARED_NO_READ_WRITE, EXCLUSIVE

▪ LOCK_DURATION ▪ STATEMENT, TRANSACTION, EXPLICIT

▪ LOCK_STATUS ▪ PENDING, GRANTED, VICTIM, TIMEOUT, KILLED


WL#3656!PERFORMANCE SCHEMA table for!SHOW SLAVE STATUS


Replication Instrumentation

▪ 6 new tables ▪ replication_connection_configuration ▪ replication_connection_status ▪ replication_execute_configuration ▪ replication_execute_status ▪ replication_execute_status_by_coordinator ▪ replication_execute_status_by_worker


Replication Instrumentation


Replication Connection Configurationmysql> select * from performance_schema.replication_connection_configuration\G *************************** 1. row ***************************! HOST: 127.0.0.1! PORT: 3306! USER: rpl! NETWORK_INTERFACE:! AUTO_POSITION: 1! SSL_ALLOWED: YES! SSL_CA_FILE:! SSL_CA_PATH:! SSL_CERTIFICATE:! SSL_CIPHER:! SSL_KEY:!SSL_VERIFY_SERVER_CERTIFICATE: YES! SSL_CRL_FILE:! SSL_CRL_PATH:! CONNECTION_RETRY_INTERVAL: 40! CONNECTION_RETRY_COUNT: 10


Replication Connection Status

mysql> select * from replication_connection_status\G!*************************** 1. row ***************************! SOURCE_UUID:! THREAD_ID: NULL! SERVICE_STATE: CONNECTING!RECEIVED_TRANSACTION_SET:! LAST_ERROR_NUMBER: 1045! LAST_ERROR_MESSAGE: error connecting to master 'repl@localhost:5613' - retry-time: 60 retries: 1! LAST_ERROR_TIMESTAMP: 2014-04-02 05:41:20


Replication Coordinator Status

mysql> select * from replication_execute_status_by_coordinator\G!*************************** 1. row ***************************! THREAD_ID: 1281! SERVICE_STATE: ON! LAST_ERROR_NUMBER: 0! LAST_ERROR_MESSAGE:!LAST_ERROR_TIMESTAMP: 0000-00-00 00:00:00

▪ Status of Coordinator thread in multi-threaded replication


Replication Worker Status

mysql> select * from replication_execute_status_by_worker;!+-----------+-----------+---------------+-----------------------------------------+-------------------+--------------------+----------------------+!| WORKER_ID | THREAD_ID | SERVICE_STATE | LAST_SEEN_TRANSACTION | LAST_ERROR_NUMBER | LAST_ERROR_MESSAGE | LAST_ERROR_TIMESTAMP |!+-----------+-----------+---------------+-----------------------------------------+-------------------+--------------------+----------------------+!| 1 | 1282 | ON | | 0 | | 0000-00-00 00:00:00 |!| 2 | 1283 | ON | | 0 | | 0000-00-00 00:00:00 |!| 3 | 1284 | ON | | 0 | | 0000-00-00 00:00:00 |!| 4 | 1285 | ON | | 0 | | 0000-00-00 00:00:00 |!| 5 | 1286 | ON | | 0 | | 0000-00-00 00:00:00 |!| 6 | 1287 | ON | | 0 | | 0000-00-00 00:00:00 |!| 7 | 1288 | ON | | 0 | | 0000-00-00 00:00:00 |!| 8 | 1289 | ON | 181f2b36-a0b4-11e3-9ac7-1025863574a7:16 | 0 | | 0000-00-00 00:00:00 |!+-----------+-----------+---------------+-----------------------------------------+-------------------+--------------------+----------------------+

▪ Status of worker threads within multi-threaded replication


WL#5768!PERFORMANCE SCHEMA, prepared statements instrumentation


Prepared Statement Instrumentation

▪ Instrumentation depends on already existing statement instruments ▪ statement/com/prepare, statement/com/execute ▪ statement/sql/prepare_sql, statement/sql/execute_sql

▪ 1 new table ▪ prepared_statements_instances


Prepared Statement Instancesmysql> select * from performance_schema.prepared_statements_instances\G!*************************** 1. row ***************************! OBJECT_INSTANCE_BEGIN: 140198306602144! STATEMENT_ID: 1! STATEMENT_NAME: stmt1! SQL_TEXT: select * from test.t1! OWNER_THREAD_ID: 54003! OWNER_EVENT_ID: 935! OWNER_OBJECT_TYPE: NULL! OWNER_OBJECT_SCHEMA: NULL! OWNER_OBJECT_NAME: NULL! TIMER_PREPARE: 15386274000! COUNT_REPREPARE: 0! COUNT_EXECUTE: 1! SUM_TIMER_EXECUTE: 217699000! MIN_TIMER_EXECUTE: 217699000! AVG_TIMER_EXECUTE: 217699000! MAX_TIMER_EXECUTE: 217699000

SUM_LOCK_TIME: 0! SUM_ERRORS: 0! SUM_WARNINGS: 0! SUM_ROWS_AFFECTED: 0! SUM_ROWS_SENT: 0! SUM_ROWS_EXAMINED: 0!SUM_CREATED_TMP_DISK_TABLES: 0! SUM_CREATED_TMP_TABLES: 0! SUM_SELECT_FULL_JOIN: 0! SUM_SELECT_FULL_RANGE_JOIN: 0! SUM_SELECT_RANGE: 0! SUM_SELECT_RANGE_CHECK: 0! SUM_SELECT_SCAN: 0! SUM_SORT_MERGE_PASSES: 0! SUM_SORT_RANGE: 0! SUM_SORT_ROWS: 0! SUM_SORT_SCAN: 0! SUM_NO_INDEX_USED: 0! SUM_NO_GOOD_INDEX_USED: 0


WL#5864!PERFORMANCE SCHEMA, instrument TRANSACTIONS


Transaction Instrumentation▪ Added the transaction instrument ▪ 8 new tables

▪ events_transactions_current ▪ events_transactions_history ▪ events_transactions_history_long ▪ events_transactions_summary_by_account_by_event_name ▪ events_transactions_summary_by_host_by_event_name ▪ events_transactions_summary_by_thread_by_event_name ▪ events_transactions_summary_by_user_by_event_name ▪ events_transactions_summary_global_by_event_name


Transaction Instrumentation

▪ Exposes details of transactions in raw or summary views ▪ Show details such as

▪ Transaction latency ▪ Isolation levels ▪ Auto commit ▪ Savepoint info ▪ GTID or transaction IDs

▪ Link transactions to the statements within the transactions


Current Transaction Detailsmysql> select * from events_transactions_current\G!*************************** 1. row ***************************! THREAD_ID: 1! EVENT_ID: 23733! END_EVENT_ID: 23742! EVENT_NAME: transaction! STATE: COMMITTED! TRX_ID: 281479898269256! GTID: NULL! XID: NULL! XA_STATE: NULL! SOURCE: handler.cc:1246! TIMER_START: 31140612726000! TIMER_END: 31140647445000! TIMER_WAIT: 34719000

ACCESS_MODE: READ WRITE! ISOLATION_LEVEL: REPEATABLE READ! AUTOCOMMIT: YES! NUMBER_OF_SAVEPOINTS: 0!NUMBER_OF_ROLLBACK_TO_SAVEPOINT: 0! NUMBER_OF_RELEASE_SAVEPOINT: 0! OBJECT_INSTANCE_BEGIN: NULL! NESTING_EVENT_ID: NULL! NESTING_EVENT_TYPE: NULL


Transaction Summary - per usermysql> select * from events_transactions_summary_by_user_by_event_name\G!*************************** 1. row ***************************! USER: mem! EVENT_NAME: transaction! COUNT_STAR: 400044! SUM_TIMER_WAIT: 21208048458267000! MIN_TIMER_WAIT: 160744000! AVG_TIMER_WAIT: 53014289000! MAX_TIMER_WAIT: 24028520397000! COUNT_READ_WRITE: 400044!SUM_TIMER_READ_WRITE: 21208048458267000!MIN_TIMER_READ_WRITE: 160744000!AVG_TIMER_READ_WRITE: 53014289000!MAX_TIMER_READ_WRITE: 24028520397000! COUNT_READ_ONLY: 0! SUM_TIMER_READ_ONLY: 0! MIN_TIMER_READ_ONLY: 0! AVG_TIMER_READ_ONLY: 0! MAX_TIMER_READ_ONLY: 0


WL#5766!PERFORMANCE SCHEMA, stored programs instrumentation


Stored Program Instrumentation▪ Tracks Stored Procedures, Stored Functions, Triggers and Events ▪ Added 16 new statement/sp/% instruments

▪ Expose the different work flows that stored programs use, such as cursor operations, workflow controls etc.

▪ Integrated in to the normal statement instrumentation ▪ 1 new tables

▪ events_statements_summary_by_program


Stored Program Summarymysql> select * from events_statements_summary_by_program\G!*************************** 1. row ***************************! OBJECT_TYPE: PROCEDURE! OBJECT_SCHEMA: ps_demo! OBJECT_NAME: ps_demo_proc! COUNT_STAR: 1! SUM_TIMER_WAIT: 6970931000! MIN_TIMER_WAIT: 6970931000! AVG_TIMER_WAIT: 6970931000! MAX_TIMER_WAIT: 6970931000! COUNT_STATEMENTS: 5! SUM_STATEMENTS_WAIT: 6802181000! MIN_STATEMENTS_WAIT: 16372000! AVG_STATEMENTS_WAIT: 1360436000! MAX_STATEMENTS_WAIT: 6484366000! SUM_LOCK_TIME: 176401000000! SUM_ERRORS: 0! SUM_WARNINGS: 0! SUM_ROWS_AFFECTED: 1! SUM_ROWS_SENT: 0

SUM_ROWS_EXAMINED: 1!SUM_CREATED_TMP_DISK_TABLES: 0! SUM_CREATED_TMP_TABLES: 0! SUM_SELECT_FULL_JOIN: 0! SUM_SELECT_FULL_RANGE_JOIN: 0! SUM_SELECT_RANGE: 0! SUM_SELECT_RANGE_CHECK: 0! SUM_SELECT_SCAN: 0! SUM_SORT_MERGE_PASSES: 0! SUM_SORT_RANGE: 0! SUM_SORT_ROWS: 0! SUM_SORT_SCAN: 0! SUM_NO_INDEX_USED: 0! SUM_NO_GOOD_INDEX_USED: 0


Stored Program Instrumentation

▪ Like Transaction instrumentation, stored programs link to the statements within them, or even link to transactions within them, which in turn link to their statements


Full Hierarchy of New Instrumentation


The MySQL SYS Schema


MySQL SYS Schema Overview

▪ Originally called “ps_helper” ▪ Started as a collection of views, procedures and functions, designed

to make reading raw Performance Schema data easier ▪ Implements many of the common DBA and Developer use cases,

including many of those shown already ▪ Now bundled within MySQL Workbench 6.1 ▪ Available on GitHub

▪ https://github.com/MarkLeith/mysql-sys

https://github.com/MarkLeith/mysql-sys


Loading MySQL SYS

$ git clone https://github.com/MarkLeith/mysql-sys.git /tmp/mysql-sys $ cd /tmp/mysql-sys $ mysql -u user -p < sys_<version>.sql !<version> can be 56 or 57, for 5.6 and 5.7 respectively

https://github.com/MarkLeith/mysql-sys.git


Loading MySQL SYS

▪ Once loaded creates a new “sys” schema containing all objects

mysql> select object_type, count! -> from sys.schema_object_overview! -> where db = 'sys';!+-------------+-------+!| object_type | count |!+-------------+-------+!| PROCEDURE | 16 |!| FUNCTION | 8 |!| VIEW | 71 |!+-------------+-------+


MySQL SYS Functions

▪ Make the raw data more readable to a human ▪ format_time() / format_bytes()

▪ Compress data for CLI output ▪ format_statement() / format_path()

▪ Extract some data to assist with JOIN in certain cases ▪ extract_[schema|table]_from_file_name()

▪ Other functions ▪ ps_is_account_enabled() / ps_thread_stack()


MySQL SYS Views

▪ Reference set of views solving various admin use cases ▪ Build upon both Performance Schema and INFORMATION_SCHEMA ▪ Both formatted and raw views are available

▪ All raw views are prefixed with x$ ▪ Allows tooling to poll raw views, but humans to use normal ones


User Summary Views

▪ High level user overview ▪ Breakdowns of IO usage ▪ Drill in to stages per user ▪ Drill in to statement details per

user

mysql> show tables like 'user%';!+-----------------------------------+!| Tables_in_sys (user%) |!+-----------------------------------+!| user_summary |!| user_summary_by_file_io |!| user_summary_by_file_io_type |!| user_summary_by_stages |!| user_summary_by_statement_latency |!| user_summary_by_statement_type |!+-----------------------------------+


User Summary (5.7)mysql> select * from user_summary\G!*************************** 1. row ***************************! user: mark! statements: 3072! statement_latency: 1.77 s! statement_avg_latency: 575.29 us! table_scans: 7! file_ios: 20043! file_io_latency: 346.79 ms! current_connections: 1! total_connections: 1! unique_hosts: 1! current_memory: 515.81 KiB!total_memory_allocated: 30.69 MiB


User File IO Summarymysql> select * from user_summary_by_file_io_type where user != 'background';!+------+--------------------------------------+---------+-------------+-------------+!| user | event_name | total | latency | max_latency |!+------+--------------------------------------+---------+-------------+-------------+!| mark | wait/io/file/myisam/dfile | 19540 | 215.20 ms | 120.79 ms |!| mark | wait/io/file/myisam/kfile | 501 | 131.58 ms | 57.74 ms |!| mark | wait/io/file/sql/dbopt | 2 | 17.29 us | 9.17 us |!| root | wait/io/file/innodb/innodb_log_file | 2079432 | 00:28:27.62 | 1.30 s |!| root | wait/io/file/myisam/kfile | 5195927 | 00:17:07.33 | 1.17 s |!| root | wait/io/file/myisam/dfile | 1533813 | 00:07:27.59 | 1.18 s |!| root | wait/io/file/sql/binlog | 3517008 | 00:03:07.34 | 1.42 s |!| root | wait/io/file/innodb/innodb_data_file | 20720 | 00:02:37.81 | 492.23 ms |!| root | wait/io/file/sql/FRM | 41412 | 3.17 s | 80.43 ms |!| root | wait/io/file/sql/dbopt | 110993 | 1.23 s | 63.27 ms |!| root | wait/io/file/sql/binlog_index | 80 | 1.16 s | 439.79 ms |!| root | wait/io/file/csv/metadata | 16 | 175.97 ms | 79.53 ms |!| root | wait/io/file/archive/data | 3345 | 18.35 ms | 5.05 ms |!| root | wait/io/file/sql/file_parser | 142 | 5.80 ms | 331.82 us |!| root | wait/io/file/sql/misc | 99 | 760.67 us | 25.19 us |!| root | wait/io/file/csv/data | 6 | 596.99 us | 348.11 us |!+------+--------------------------------------+---------+-------------+-------------+


User Statement Latency Summary

mysql> select * from user_summary_by_statement_latency where user != 'background';!+------+---------+---------------+-------------+--------------+-----------+---------------+---------------+------------+!| user | total | total_latency | max_latency | lock_latency | rows_sent | rows_examined | rows_affected | full_scans |!+------+---------+---------------+-------------+--------------+-----------+---------------+---------------+------------+!| root | 8322401 | 19.37h | 00:07:08.11 | 00:40:26.25 | 10110692 | 68391445 | 4198113 | 189531 |!| mark | 3072 | 1.77 s | 1.73 s | 1.96 ms | 534 | 15389 | 0 | 7 |!+------+---------+---------------+-------------+--------------+-----------+---------------+---------------+------------+!


User Statement Type Summary

mysql> select * from user_summary_by_statement_type where user != 'background';!+------+-----------------------+---------+---------------+-------------+--------------+-----------+---------------+---------------+------------+!| user | statement | total | total_latency | max_latency | lock_latency | rows_sent | rows_examined | rows_affected | full_scans |!+------+-----------------------+---------+---------------+-------------+--------------+-----------+---------------+---------------+------------+!| mark | select | 4 | 1.69 s | 1.69 s | 408.00 us | 212 | 15067 | 0 | 1 |!| mark | Field List | 146 | 67.91 ms | 40.02 ms | 1.11 ms | 0 | 0 | 0 | 0 |!| mark | jump_if_not | 2496 | 2.68 ms | 7.07 us | 0 ps | 0 | 0 | 0 | 0 |!| mark | show_tables | 4 | 1.70 ms | 545.41 us | 296.00 us | 292 | 292 | 0 | 4 |!| mark | freturn | 418 | 1.15 ms | 9.29 us | 0 ps | 0 | 0 | 0 | 0 |!| mark | show_databases | 2 | 849.40 us | 431.05 us | 146.00 us | 30 | 30 | 0 | 2 |!| mark | Init DB | 2 | 114.95 us | 59.58 us | 0 ps | 0 | 0 | 0 | 0 |!| root | commit | 2463190 | 13.33h | 12.18 s | 0 ps | 0 | 0 | 0 | 0 |!| root | insert | 2501839 | 3.78h | 6.64 s | 00:31:18.63 | 192 | 0 | 3674716 | 0 |!| root | select | 1119999 | 1.74h | 16.34 s | 00:07:42.99 | 6187336 | 63933992 | 0 | 163913 |!| root | update | 494217 | 00:15:23.09 | 4.83 s | 00:01:05.05 | 126 | 496604 | 495334 | 0 |!| root | Binlog Dump GTID | 1 | 00:06:12.69 | 00:06:12.69 | 0 ps | 0 | 0 | 0 | 0 |!| root | delete | 206176 | 00:01:59.54 | 2.22 s | 6.69 s | 12 | 20201 | 20203 | 0 |!| root | set_option | 1068360 | 00:01:10.73 | 45.37 ms | 0 ps | 0 | 0 | 0 | 0 |!| root | show_engine_status | 2222 | 46.62 s | 4.96 s | 0 ps | 0 | 0 | 0 | 0 |!| root | rollback | 64527 | 26.21 s | 407.15 ms | 0 ps | 0 | 0 | 0 | 0 |!| root | show_variables | 11219 | 12.28 s | 255.00 ms | 1.19 s | 2542172 | 2542172 | 0 | 11219 |!| root | show_binlogs | 1850 | 8.96 s | 1.64 s | 0 ps | 0 | 0 | 0 | 0 |!| root | show_tables | 4250 | 6.86 s | 123.34 ms | 182.80 ms | 5277 | 5277 | 0 | 4250 |!| root | show_status | 5485 | 6.50 s | 242.60 ms | 581.02 ms | 1331264 | 1331264 | 0 | 5485 |!


IO Summary Views

▪ IO breakdown by thread ▪ Global summaries by file and wait

class, by both bytes and latency ▪ Stream of last raw file IO stats

mysql> show tables like 'io_%';!+------------------------------+!| Tables_in_sys (io_%) |!+------------------------------+!| io_by_thread_by_latency |!| io_global_by_file_by_bytes |!| io_global_by_file_by_latency |!| io_global_by_wait_by_bytes |!| io_global_by_wait_by_latency |!| latest_file_io |!+------------------------------+


IO Per Thread Summarymysql> select * from io_by_thread_by_latency;!+---------------------+---------+---------------+-------------+-------------+-------------+-----------+----------------+!| user | total | total_latency | min_latency | avg_latency | max_latency | thread_id | processlist_id |!+---------------------+---------+---------------+-------------+-------------+-------------+-----------+----------------+!| io_write_thread | 1270599 | 00:28:51.42 | 1.46 us | 1.36 ms | 1.39 s | 9 | NULL |!| page_cleaner_thread | 3356577 | 00:21:01.53 | 410.93 ns | 2.08 ms | 1.49 s | 19 | NULL |!| io_write_thread | 338216 | 00:14:31.16 | 1.67 us | 2.58 ms | 1.17 s | 12 | NULL |!| io_log_thread | 47023 | 00:04:06.10 | 3.65 us | 5.23 ms | 538.77 ms | 4 | NULL |!| io_write_thread | 290066 | 00:02:20.99 | 2.78 us | 486.07 us | 703.20 ms | 10 | NULL |!| io_write_thread | 221417 | 00:02:12.64 | 1.85 us | 599.05 us | 773.64 ms | 11 | NULL |!| root@localhost | 587531 | 00:02:03.29 | 444.86 ns | 200.24 us | 372.09 ms | 2515 | 2495 |!| io_read_thread | 5102 | 00:01:09.66 | 12.24 us | 13.65 ms | 281.13 ms | 5 | NULL |!| srv_master_thread | 55199 | 58.11 s | 693.68 ns | 1.40 ms | 504.03 ms | 17 | NULL |!| root@localhost | 1622373 | 41.20 s | 486.33 ns | 25.40 us | 359.49 ms | 23 | 3 |!| io_read_thread | 2401 | 37.27 s | 11.86 us | 15.52 ms | 258.97 ms | 6 | NULL |!| io_read_thread | 2156 | 35.73 s | 12.02 us | 16.57 ms | 284.81 ms | 8 | NULL |!| io_read_thread | 1719 | 29.19 s | 11.37 us | 16.98 ms | 269.93 ms | 7 | NULL |!| srv_purge_thread | 3874 | 12.16 s | 433.55 ns | 3.14 ms | 246.79 ms | 18 | NULL |!| main | 11083 | 3.12 s | 588.12 ns | 547.21 us | 234.02 ms | 1 | NULL |!| root@localhost | 5985 | 1.48 s | 625.82 ns | 1.06 ms | 239.17 ms | 137759 | 137739 |!| root@localhost | 157 | 434.38 ms | 497.64 ns | 8.09 ms | 205.84 ms | 178521 | 178501 |!| mark@localhost | 20043 | 346.79 ms | 418.47 ns | 94.10 us | 120.79 ms | 150606 | 150586 |!| root@localhost | 138 | 204.72 ms | 486.33 ns | 2.68 ms | 86.54 ms | 178524 | 178504 |!| root@localhost | 415 | 126.43 ms | 471.25 ns | 195.15 us | 37.21 ms | 178269 | 178249 |!| root@localhost | 9 | 106.21 ms | 6.40 us | 8.85 ms | 103.88 ms | 178522 | 178502 |!+---------------------+---------+---------------+-------------+-------------+-------------+-----------+----------------+!21 rows in set (0.02 sec)


IO Per File Summariesmysql> select * from io_global_by_file_by_latency limit 1\G!*************************** 1. row ***************************! file: @@datadir/ibdata1! total: 1395714!total_latency: 00:28:53.06! count_read: 1006! read_latency: 4.51 s! count_write: 1326529!write_latency: 33.39 s! count_misc: 68179! misc_latency: 00:28:15.16!!mysql> select * from io_global_by_file_by_bytes limit 1\G!*************************** 1. row ***************************! file: @@datadir/ibdata1! count_read: 1006! total_read: 17.69 MiB! avg_read: 18.00 KiB! count_write: 1327221!total_written: 77.82 GiB! avg_write: 61.48 KiB! total: 77.84 GiB! write_pct: 99.98


Schema Analysis Views

▪ Object overview ▪ Table usage stats ▪ Index usage stats

mysql> show tables like 'schema%';!+-------------------------------------+!| Tables_in_sys (schema%) |!+-------------------------------------+!| schema_index_statistics |!| schema_object_overview |!| schema_table_statistics |!| schema_table_statistics_with_buffer |!| schema_tables_with_full_table_scans |!| schema_unused_indexes |!+-------------------------------------+


Schema Object Overviewmysql> select * from schema_object_overview;!+-------------------------+---------------+-------+!| db | object_type | count |!+-------------------------+---------------+-------+!| information_schema | SYSTEM VIEW | 60 |!| mem__advisors | BASE TABLE | 3 |!| mem__advisors | INDEX (BTREE) | 5 |!| mem__advisor_text | BASE TABLE | 2 |!| mem__advisor_text | INDEX (BTREE) | 5 |!| mem__bean_config | BASE TABLE | 4 |!| mem__bean_config | INDEX (BTREE) | 6 |!| mem__config | BASE TABLE | 12 |!| mem__config | INDEX (BTREE) | 21 |!| mem__enterprise | BASE TABLE | 2 |!| mem__enterprise | INDEX (BTREE) | 3 |!| mem__events | BASE TABLE | 32 |!| mem__events | INDEX (BTREE) | 69 |!| mem__instruments | BASE TABLE | 117 |!| mem__instruments | INDEX (BTREE) | 582 |


Table statisticsmysql> select * from schema_table_statistics limit 1\G!*************************** 1. row ***************************! table_schema: mem__quan! table_name: normalized_statements_by_server_by_schema_data! total_latency: 00:55:07.80! rows_fetched: 239685! fetch_latency: 17.10 s! rows_inserted: 1102895! insert_latency: 00:52:18.10! rows_updated: 55910! update_latency: 00:02:28.62! rows_deleted: 17899! delete_latency: 3.98 s! io_read_requests: 20639! io_read: 324.94 MiB! io_read_latency: 00:03:04.10!io_write_requests: 1965973! io_write: 40.36 GiB! io_write_latency: 25.46 s! io_misc_requests: 18217! io_misc_latency: 00:14:06.02


Index statistics

mysql> select * from schema_index_statistics limit 1\G!*************************** 1. row ***************************! table_schema: mem__inventory! table_name: mysqlconnectionconfiguration! index_name: PRIMARY! rows_selected: 131399!select_latency: 00:07:05.03! rows_inserted: 0!insert_latency: 0 ps! rows_updated: 41608!update_latency: 21.96 s! rows_deleted: 0!delete_latency: 0 ps


InnoDB Buffer Usage Views

▪ Buffer usage by schema ▪ Buffer usage by table

mysql> show tables like 'innodb%';!+-------------------------------+!| Tables_in_sys (innodb%) |!+-------------------------------+!| innodb_buffer_stats_by_schema |!| innodb_buffer_stats_by_table |!+-------------------------------+


Buffer Usage By Schema

mysql> select * from innodb_buffer_stats_by_schema;!+------------------+------------+------------+-------+--------------+-----------+-------------+!| object_schema | allocated | data | pages | pages_hashed | pages_old | rows_cached |!+------------------+------------+------------+-------+--------------+-----------+-------------+!| mem__quan | 69.94 MiB | 45.39 MiB | 4476 | 4476 | 4476 | 123226 |!| mem__instruments | 20.70 MiB | 10.75 MiB | 1325 | 1325 | 1325 | 82812 |!| mem__events | 3.08 MiB | 2.09 MiB | 197 | 197 | 197 | 3411 |!| mem__inventory | 1.34 MiB | 52.49 KiB | 86 | 86 | 86 | 18 |!| mysql | 432.00 KiB | 251.34 KiB | 27 | 27 | 27 | 2252 |!| mem__config | 144.00 KiB | 3.86 KiB | 9 | 9 | 9 | 29 |!+------------------+------------+------------+-------+--------------+-----------+-------------+


Buffer Usage By Tablemysql> select * from innodb_buffer_stats_by_table limit 20;!+------------------+------------------------------------------------+------------+------------+-------+--------------+-----------+-------------+!| object_schema | object_name | allocated | data | pages | pages_hashed | pages_old | rows_cached |!+------------------+------------------------------------------------+------------+------------+-------+--------------+-----------+-------------+!| mem__quan | normalized_statements_by_server_by_schema_data | 49.97 MiB | 33.12 MiB | 3198 | 3198 | 3198 | 238845 |!| mem__quan | example_statements | 15.95 MiB | 9.71 MiB | 1021 | 1021 | 1021 | 32435 |!| mem__events | events | 2.75 MiB | 1.97 MiB | 176 | 176 | 176 | 5044 |!| mem__quan | normalized_statements | 2.19 MiB | 1.42 MiB | 140 | 140 | 140 | 2297 |!| mem__quan | normalized_statements_by_server_by_schema | 1.06 MiB | 693.58 KiB | 68 | 68 | 68 | 3840 |!| mem__instruments | innodbundologsdata | 1.05 MiB | 754.07 KiB | 67 | 67 | 67 | 5529 |!| mem__instruments | fsstatistics | 736.00 KiB | 603.64 KiB | 46 | 46 | 46 | 2092 |!| mem__instruments | networktrafficadvisor_networktraffic | 496.00 KiB | 291.38 KiB | 31 | 31 | 31 | 1160 |!| mysql | innodb_index_stats | 384.00 KiB | 230.37 KiB | 24 | 24 | 24 | 1986 |!| mem__quan | explains | 368.00 KiB | 274.42 KiB | 23 | 23 | 23 | 333 |!| mem__instruments | innodbchangebufferusagedata | 272.00 KiB | 126.94 KiB | 17 | 17 | 17 | 1104 |!| mem__instruments | rowscanratiodata | 272.00 KiB | 141.18 KiB | 17 | 17 | 17 | 1207 |!| mem__instruments | connectionsmaxdata | 256.00 KiB | 150.98 KiB | 16 | 16 | 16 | 1262 |!| mem__instruments | innodbopenfilesdata | 256.00 KiB | 145.30 KiB | 16 | 16 | 16 | 1170 |!| mem__instruments | myisamkeybufferusagedata | 256.00 KiB | 155.60 KiB | 16 | 16 | 16 | 1192 |!| mem__instruments | connectionsdata | 240.00 KiB | 129.12 KiB | 15 | 15 | 15 | 1106 |!| mem__instruments | diskioopstotaldata | 240.00 KiB | 143.94 KiB | 15 | 15 | 15 | 533 |!| mem__instruments | diskiototaldata | 240.00 KiB | 143.94 KiB | 15 | 15 | 15 | 533 |!| mem__instruments | innodbchangebufferactivitydata | 240.00 KiB | 124.43 KiB | 15 | 15 | 15 | 1003 |!| mem__instruments | innodbcheckpointagedata | 240.00 KiB | 122.65 KiB | 15 | 15 | 15 | 1052 |!+------------------+------------------------------------------------+------------+------------+-------+--------------+-----------+-------------+


Statement Analysis Views

▪ Statement overview ▪ Find statements by errors, full

scans, sorting, temporary tables ▪ Find the statements with the

longest runtimes !

mysql> show tables like 'statement%';!+---------------------------------------------+!| Tables_in_sys (statement%) |!+---------------------------------------------+!| statement_analysis |!| statements_with_errors_or_warnings |!| statements_with_full_table_scans |!| statements_with_runtimes_in_95th_percentile |!| statements_with_sorting |!| statements_with_temp_tables |!+---------------------------------------------+


Statement Analysismysql> select * from statement_analysis limit 11\G!*************************** 1. row ***************************! query: INSERT INTO `mem__quan` . `nor ... nDuration` = IF ( VALUES ( ...! db: mem! full_scan:! exec_count: 1110067! err_count: 0! warn_count: 0! total_latency: 1.93h! max_latency: 5.03 s! avg_latency: 6.27 ms! lock_latency: 00:18:29.18! rows_sent: 0! rows_sent_avg: 0! rows_examined: 0!rows_examined_avg: 0! tmp_tables: 0! tmp_disk_tables: 0! rows_sorted: 0!sort_merge_passes: 0! digest: d48316a218e95b1b8b72db5e6b177788! first_seen: 2014-05-20 10:42:17


Statement With Errors And Warnings

mysql> select * from statements_with_errors_or_warnings limit 1\G!*************************** 1. row ***************************! query: CREATE TEMPORARY TABLE IF NOT ... _logs` ( `id` INT8 NOT NULL )! db: mem! exec_count: 1725! errors: 1725! error_pct: 100.0000! warnings: 0!warning_pct: 0.0000! first_seen: 2014-05-20 10:42:32! last_seen: 2014-05-21 18:39:22! digest: 51fb979dbc3910a6294c7cdabd7839c3


Statement With Sorting

mysql> select * from statements_with_sorting limit 1\G!*************************** 1. row ***************************! query: SELECT `s` . `identityId` , `s ... `subject_id` = `s` . `id` ...! db: mem! exec_count: 1406! total_latency: 00:03:01.22!sort_merge_passes: 0! avg_sort_merges: 0!sorts_using_scans: 1406! sort_using_range: 1406! rows_sorted: 39865! avg_rows_sorted: 28! first_seen: 2014-05-20 10:42:08! last_seen: 2014-05-21 14:04:02! digest: 9972d1d2cc12b92fe938ec5bd74d2f1d


Wait Analysis Views

▪ Wait summaries by class and per instrument globally

▪ Wait details per user !

mysql> show tables like 'wait%';!+------------------------------------+!| Tables_in_sys (wait%) |!+------------------------------------+!| wait_classes_global_by_avg_latency |!| wait_classes_global_by_latency |!| waits_by_user_by_latency |!| waits_global_by_latency |!+------------------------------------+


Global Wait Summarymysql> select * from waits_global_by_latency;!+------------------------------------------------------+----------+---------------+-------------+-------------+!| events | total | total_latency | avg_latency | max_latency |!+------------------------------------------------------+----------+---------------+-------------+-------------+!| wait/io/table/sql/handler | 24196659 | 1.75h | 259.72 us | 2.29 s |!| wait/io/file/innodb/innodb_data_file | 5964255 | 1.36h | 822.01 us | 1.49 s |!| wait/io/file/innodb/innodb_log_file | 2272647 | 00:36:41.37 | 968.64 us | 1.30 s |!| wait/io/file/myisam/kfile | 5463587 | 00:18:14.16 | 200.26 us | 1.17 s |!| wait/io/file/myisam/dfile | 1715187 | 00:08:00.18 | 279.96 us | 1.18 s |!| wait/io/file/sql/binlog | 3722079 | 00:03:23.71 | 54.73 us | 1.42 s |!| wait/lock/table/sql/handler | 10074274 | 00:01:56.63 | 11.58 us | 186.90 ms |!| wait/io/socket/sql/client_connection | 3879827 | 25.36 s | 6.54 us | 219.25 ms |!| wait/io/file/sql/FRM | 43994 | 3.27 s | 74.35 us | 80.43 ms |!| wait/io/file/sql/dbopt | 116765 | 1.29 s | 11.03 us | 63.27 ms |!| wait/io/file/sql/binlog_index | 95 | 1.20 s | 12.67 ms | 439.79 ms |!| wait/io/file/csv/metadata | 32 | 577.39 ms | 18.04 ms | 245.23 ms |!| wait/synch/mutex/sql/THD::LOCK_thd_data | 10300455 | 221.84 ms | 21.49 ns | 12.62 ms |!| wait/synch/mutex/sql/THD::LOCK_query_plan | 5656929 | 168.50 ms | 29.78 ns | 11.89 ms |!| wait/io/file/archive/data | 3533 | 19.11 ms | 5.41 us | 5.05 ms |!| wait/io/file/mysys/cnf | 5 | 3.75 ms | 750.59 us | 3.60 ms |!| wait/synch/mutex/myisam/MYISAM_SHARE::intern_lock | 88257 | 2.20 ms | 24.88 ns | 361.49 us |!| wait/io/file/sql/ERRMSG | 5 | 1.14 ms | 227.99 us | 502.64 us |!| wait/io/file/csv/data | 12 | 1.03 ms | 85.56 us | 348.11 us |!……


User Wait Summarymysql> select * from waits_by_user_by_latency;!+------+------------------------------------------------------+----------+---------------+-------------+-------------+!| user | event | total | total_latency | avg_latency | max_latency |!+------+------------------------------------------------------+----------+---------------+-------------+-------------+!| mark | wait/io/file/myisam/dfile | 19540 | 215.20 ms | 11.01 us | 120.79 ms |!| mark | wait/io/file/myisam/kfile | 501 | 131.58 ms | 262.63 us | 57.74 ms |!| mark | wait/io/file/sql/dbopt | 2 | 17.29 us | 8.64 us | 9.17 us |!| root | wait/io/table/sql/handler | 24189319 | 1.74h | 258.77 us | 2.29 s |!| root | wait/io/file/innodb/innodb_log_file | 2162650 | 00:30:52.87 | 856.76 us | 1.30 s |!| root | wait/io/file/myisam/kfile | 5460992 | 00:18:12.83 | 200.11 us | 1.17 s |!| root | wait/io/file/myisam/dfile | 1694624 | 00:07:59.50 | 282.95 us | 1.18 s |!| root | wait/io/file/sql/binlog | 3711637 | 00:03:23.16 | 54.74 us | 1.42 s |!| root | wait/io/file/innodb/innodb_data_file | 22577 | 00:03:05.73 | 8.23 ms | 492.23 ms |!| root | wait/lock/table/sql/handler | 10069000 | 00:01:56.59 | 11.58 us | 186.90 ms |!| root | wait/io/socket/sql/client_connection | 3858476 | 25.36 s | 6.57 us | 219.25 ms |!| root | wait/io/file/sql/FRM | 42701 | 3.18 s | 74.52 us | 80.43 ms |!| root | wait/io/file/sql/dbopt | 116693 | 1.29 s | 11.03 us | 63.27 ms |!| root | wait/io/file/sql/binlog_index | 80 | 1.16 s | 14.47 ms | 439.79 ms |!| root | wait/io/file/csv/metadata | 32 | 577.39 ms | 18.04 ms | 245.23 ms |!| root | wait/synch/mutex/sql/THD::LOCK_thd_data | 10280287 | 221.84 ms | 21.49 ns | 12.62 ms |!| root | wait/synch/mutex/sql/THD::LOCK_query_plan | 5642052 | 168.50 ms | 29.78 ns | 11.89 ms |!| root | wait/io/file/archive/data | 3529 | 19.09 ms | 5.41 us | 5.05 ms |!……


MySQL SYS Procedures - P_S Analysis Helpers

▪ Dump the data to create a graph dot file for a thread trace ▪ ps_trace_thread()

▪ Look in the statement history table trying to capture more info ▪ ps_trace_statement_digest()

▪ Reset all summary data ▪ ps_truncate_all_tables()


Tracing Statement Digests

▪ ps_trace_statement_digest() analyzes live traffic looking for certain statement digest

▪ Captures statistics on each matching statement it finds ▪ Returns a report of the statistics

▪ An overall summary ▪ A break down for the longest running example ▪ An EXPLAIN (if the statement is not truncated)


Tracing Statement Digests

▪ ps_trace_statement_digest() parameters ▪ in_digest The statement digest to analyze ▪ in_runtime How long to run analysis for ▪ in_interval How often to snapshot for data ▪ in_start_fresh Whether to truncate p_s tables first ▪ in_auto_enable Whether to auto enable required config


Tracing Statement Digests Examplemysql> call analyze_statement_digest('6134e9d6f25eb8e6cddf11f6938f202a', 60, 1, true, true); +--------------------+ | SUMMARY STATISTICS | +--------------------+ | SUMMARY STATISTICS | +--------------------+ 1 row in set (59.93 sec) !+------------+-----------+-----------+-----------+---------------+------------+------------+ | executions | exec_time | lock_time | rows_sent | rows_examined | tmp_tables | full_scans | +------------+-----------+-----------+-----------+---------------+------------+------------+ | 360 | 1.41 s | 138.39 ms | 720 | 0 | 0 | 0 | +------------+-----------+-----------+-----------+---------------+------------+------------+ 1 row in set (59.93 sec) !+--------------------------------+-------+-----------+ | event_name | count | latency | +--------------------------------+-------+-----------+ | stage/sql/Sending data | 14 | 1.10 s | | stage/sql/update | 198 | 667.34 ms | | stage/sql/removing tmp table | 9 | 240.48 ms | ... !


Tracing Statement Digests Example+---------------------------+ | LONGEST RUNNING STATEMENT | +---------------------------+ | LONGEST RUNNING STATEMENT | +---------------------------+ 1 row in set (59.95 sec) !+-----------+-----------+-----------+-----------+---------------+------------+-----------+ | thread_id | exec_time | lock_time | rows_sent | rows_examined | tmp_tables | full_scan | +-----------+-----------+-----------+-----------+---------------+------------+-----------+ | 23277 | 77.64 ms | 97.00 us | 2 | 0 | 0 | 0 | +-----------+-----------+-----------+-----------+---------------+------------+-----------+ 1 row in set (59.95 sec) !+-------------------------------------------------------------------------------------------+ | sql_text | +-------------------------------------------------------------------------------------------+ | /* mem dbpool.default */ insert into `mem__quan`.`normalized_statements_by_server_by_schema` (firstSeen, lastSeen, normalized_statement_id, `schema`, server, id) values (1378983745000, 1378998600000, x'808ACEB88FC4B45BC47C4FE9D86C7F26', 'mem', '753c939e-1b99-11e3-b6d4-bc761a1f4f2f', x'7CFE81F98B1E3FE0895D1AA2C39B326D') ON DUPLICATE KEY UPDATE ... +-------------------------------------------------------------------------------------------+ ... !


Tracing Threads

▪ ps_trace_thread() monitors a specific thread for a period ▪ Captures as much information on the thread activity as possible ▪ Returns a “dot” formatted file, that can graph the event hierarchy as

previously seen ▪ http://en.wikipedia.org/wiki/DOT_(graph_description_language)

http://en.wikipedia.org/wiki/DOT_(graph_description_language)


Tracing Threads

▪ ps_trace_thread() parameters ▪ in_thread_id The thread to analyze ▪ in_outfile The file to dump the data to (INTO OUTFILE) ▪ in_max_runtime How long to run analysis for ▪ in_interval How often to snapshot for data ▪ in_start_fresh Whether to truncate p_s tables first ▪ in_auto_setup Whether to auto enable required config ▪ in_debug Whether to also print debug info (source info)


Tracing Threads Examplemysql> call dump_thread_stack(27768, '/tmp/stack_27768.dot', 60, 0.1, true, true, true); +------------------------------------------------+ | Info | +------------------------------------------------+ | Data collection starting for THREAD_ID = 27768 | +------------------------------------------------+ 1 row in set (4.82 sec) !+---------------------------------------------+ | Info | +---------------------------------------------+ | Stack trace written to /tmp/stack_27768.dot | +---------------------------------------------+ 1 row in set (60.90 sec) !+--------------------------------------------------------+ | Convert to PDF | +--------------------------------------------------------+ | dot -Tpdf -o /tmp/stack_27768.pdf /tmp/stack_27768.dot | +--------------------------------------------------------+ 1 row in set (60.90 sec) !+--------------------------------------------------------+ | Convert to PNG | +--------------------------------------------------------+ | dot -Tpng -o /tmp/stack_27768.png /tmp/stack_27768.dot | +--------------------------------------------------------+ 1 row in set (60.90 sec)+ !!


Tracing Threads Dot Format

cerberus:~ mark$ more Documents/statement_graphs/deletes.dot!digraph events {! graph [rankdir=LR, nodesep="0.10"];! node [label="\N"];! graph [bb="0,0,1860,17476"];! 1 [label="(0.20μ) mysys/THR_LOCK_threads\n", style=filled, color=lightskyblue, pos="209,18", width="3.19", height="0.50"];! 2 [label="(0.18μ) sql/THD::LOCK_thd_data\n", style=filled, color=lightskyblue, pos="209,61", width="3.06", height="0.50"];! 3 [label="(0.24μ) sql/hash_filo::lock\n", style=filled, color=lightskyblue, pos="209,104", width="2.42", height=“0.50"];! 4 [label="(0.31μ) sql/LOGGER::LOCK_logger - read_lock\n", style=filled, color=orchid, pos="209,147", width="4.22", height="0.50"];! 5 [label="(0.08μ) sql/LOG::LOCK_log\n", style=filled, color=lightskyblue, pos="209,190", width="2.64", height="0.50"];! 6 [label="(12.97μ) wait/io/file/sql/query_log - write 49 bytes\n/Users/mark/mysql/lp-mysql-trunk/mysql-test/var/mysqld.1/mysqld.log\n", style=filled, color=red, shape=box, pos="209,235", width="5.78", height="0.56"];! 7 [label="(176.14μ) sql/select\nselect @@version_comment limit 1\nerrors: 0\nwarnings: 0\nlock time: 0.00μ\nrows affected: 0\nrows sent:\!209,18", width="3.19", height="0.50"];mp disk tables: 0\nselect scan: 0\nselect full join: :! 2 [label="(0.18μ) sql/THD::LOCK_thd_data\n", style=filled, color=lightskyblue, pos="209,61", width="3.06", height="0.50"];! 3 [label="(0.24μ) sql/hash_filo::lock\n", style=filled, color=lightskyblue, pos="209,104", width="2.42", height="0.50"];! 4 [label="(0.31μ) sql/LOGGER::LOCK_logger - read_lock\n", style=filled, color=orchid, pos="209,147", width="4.22", height="0.50"];! 5 [label="(0.08μ) sql/LOG::LOCK_log\n", style=filled, color=lightskyblue, pos="209,190", width="2.64", height="0.50"];!……!


Tracing Statement Output


MySQL SYS Procedures - P_S Setup Helpers

▪ ps_setup_show_disabled() / ps_setup_show_enabled() ▪ ps_setup_disable_thread() / ps_setup_enable_thread() ▪ ps_setup_disable_background_threads() /

ps_setup_enable_background_threads() ▪ ps_setup_disable_instrument() / ps_setup_enable_instrument() ▪ ps_setup_disable_consumer() / ps_setup_enable_consumer() ▪ ps_setup_save() / ps_setup_reload_saved() ▪ ps_setup_reset_to_default() ▪ ps_truncate_all_tables()


Things still to come


Status / Variable Instrumentation

▪ Deprecate I_S.[GLOBAL|SESSION]_[STATUS|VARIABLES] ▪ Replace with tables that show true state / aggregates

▪ Per thread, user, host, account ▪ See session statistics or configuration individually

▪ Improve SHOW commands to only show appropriate variables ▪ SHOW SESSION STATUS to only show session specific items ▪ Backwards compatibility switch available

▪ Less internal mutex contention for polling status


Further Replication Statistics

▪ Redesigned lag monitoring ▪ Consistently check lag across a replication chain

▪ Further statistics on load of the replication threads ▪ Queue lengths ▪ Event counting ▪ Busy times


Stage Progress Tracking

▪ View the progress of longer stages of execution ▪ Extends stage tables with “estimate” and “completed” columns ▪ Bounded (“estimated”) or unbounded (only “completed” filled in) ▪ Units depend on stage

▪ stage/sql/copy to tmp table - 130 (rows) completed


Continued focus …

▪ On performance! ▪ On filling missing instrumentation points! ▪ On consolidating instrumentation within Performance schema!


The preceding is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. The development, release, and timing of any features or functionality described for Oracle’s products remains at the sole discretion of Oracle.

Mysql tech day_paris_ps_and_sys

Technology

Transcript of Mysql tech day_paris_ps_and_sys