Query Optimization with MySQL 5.6: Old and New Tricks

Query Optimization with MySQL 5.6:
Old and New Tricks
Jaime Crespo
Senior MySQL Instructor
Percona Live London 2013
November 11, 2013

About This Presentation
• Objective #1:
- Get faster queries
• Reminder of some well-known query
optimization techniques
- Focus on new 5.6 query features
• For other performance tricks with MySQL &
InnoDB:
- InnoDB Architecture and Performance
Optimization with Peter Zaitsev at 2pm
2 www.percona.com

Agenda
1. Introduction and Environment Setup
2. Server Update and Configuration
3. Tools for Query Profiling
4. Indexing: Old and New Strategies
5. New Query Planner Features
6. Results and wrap-up
3 www.percona.com

Example Platform
• Sample LAMP Application
- 'My Movies' PHP Script
- Example movie database
• Distributed as a VirtualBox VM
- Import percona_live.ova
- User: percona / pass: percona
- Make sure httpd and PS are running
- Optionally: startxfce4
• Currently no design optimization, no indexes
(except PK), no proper query design
4 www.percona.com

Query Optimization with MySQL 5.6: Old and New Tricks
SERVER UPDATE AND
CONFIGURATION
5 www.percona.com 5

Basic Configuration Review
• InnoDB as the default engine (default in 5.5)
• InnoDB File Per Table (default in 5.6)
• Larger Buffer Pool and Transaction Log
(changed in 5.6)
• Row Based Replication and tx_isolation
• Performance Schema overhead (enabled by
default in 5.6)
• InnoDB page checksum (changed in 5.6)
• InnoDB Old Blocks time (changed in 5.6)
6 www.percona.com

TOOLS FOR QUERY PROFILING
7 www.percona.com 5

EXPLAIN
• EXPLAIN is most important tool for the
MySQL DBA and Developer
• It provides:
- The query plan strategy
- The index(es) to be used
- The order of table access
- An estimation of the number of rows to be
examined!
8 www.percona.com 5

EXPLAIN on DML
• EXPLAIN is now available also for
INSERT, UPDATE and DELETE
mysql> EXPLAIN DELETE FROM title WHERE title = 'Pilot'G!
********************* 1. row *********************!
id: 1!
select_type: SIMPLE!
table: title!
type: range!
possible_keys: PRIMARY,title!
key: title!
key_len: 77!
ref: NULL!
rows: 1380!
Extra: Using where!
1 row in set (0.00 sec)!
9 www.percona.com 5

Structured EXPLAIN
mysql> EXPLAIN FORMAT=JSON SELECT COUNT(*) FROM title
WHERE (title = 'Pilot' OR production_year > 2010)
AND kind_id < 4G
*************************** 1. row ***************************!
EXPLAIN: {!
"query_block": {!
"select_id": 1,!
"table": {!
"table_name": "title",!
"access_type": "index_merge",!
"possible_keys": [!
"title",!
"production_year"!
],!
"key": "sort_union(title,production_year)",!
"key_length": "77,5",!
"rows": 4434,!
"filtered": 100,!
"attached_condition": "(((ìmdb`.`title`.`title` = 'Pilot') or
(ìmdb`.`title`.`production_year` > 2010)) and
(ìmdb`.`title`.`kind_id` < 4))"!
}!
}!
}!
10 www.percona.com 5

Optimizer Trace
• Allows profiling of MySQL query planner
• It shows not only information about the
final query plan (EXPLAIN), but also about
other discarded strategies, and its
“execution cost”
• It can be accessed via the
INFORMATION_SCHEMA database
• It is off by default

Checking the Optimizer Trace
• mysql> SET optimizer_trace="enabled=on";
• mysql> SELECT COUNT(*) FROM title WHERE
(title = 'Pilot' OR production_year >
2010) AND kind_id < 4;
+----------+
| COUNT(*) |
+----------+
| 3050 |
+----------+
1 row in set (0.01 sec)
• mysql> SELECT trace FROM
information_schema.optimizer_trace;

Checking the Optimizer Trace
(cont.)
{
"range_scan_alternatives": [
{
"index": "production_year",
"ranges": [
"2010 < production_year"
],
"index_dives_for_eq_ranges": true,
"rowid_ordered": false,
"using_mrr": false,
"index_only": true,
"rows": 3054,
"cost": 615.16,
"chosen": true
}
],
"index_to_merge": "production_year",
"cumulated_cost": 905.69
}
],
"cost_of_reading_ranges": 905.69,
"cost_sort_rowid_and_read_disk": 3672.8,
"cost_duplicate_removal": 9467.6,
"total_cost": 14046
}
],
"chosen_range_access_summary": {
"range_access_plan": {
"type": "index_merge",
"index_merge_of": [

Limitations of EXPLAIN
• It does not really execute the query
- Not even subqueries any more
• The row statistics can be sometimes very
off
• It does not give us information about
temporary tables (on disk or memory?),
subtask timing, query cache hit, etc.

Handler Statistics
• They provide time & memory status-independency:
mysql> SHOW SESSION STATUS LIKE 'Hand%';
+----------------------------+-------+
| Variable_name | Value |
+----------------------------+-------+
| Handler_commit | 1 |
| Handler_delete | 0 |
| Handler_discover | 0 |
| Handler_external_lock | 2 |
| Handler_mrr_init | 0 |
| Handler_prepare | 0 |
| Handler_read_first | 0 |
| Handler_read_key | 1 |
| Handler_read_last | 0 |
| Handler_read_next | 266 |
| Handler_read_prev | 0 |
| Handler_read_rnd | 0 |
| Handler_read_rnd_next | 0 |
| Handler_rollback | 0 |
| Handler_savepoint | 0 |
| Handler_savepoint_rollback | 0 |
| Handler_update | 0 |
| Handler_write | 0 |
+----------------------------+-------+
18 rows in set (0.00 sec)

Slow Query Log and pt-query-digest
• Works on all MySQL versions
- Specially useful in combination with the
extended log statistics in Percona Server
SET GLOBAL log_slow_verbosity = 'full';
!
# Time: 130601 8:01:06.058915
# User@Host: root[root] @ localhost [] Id: 42
# Schema: imdb Last_errno: 0 Killed: 0
# Query_time: 7.725616 Lock_time: 0.000328 Rows_sent: 4 Rows_examined: 1543720
Rows_affected: 0
# Bytes_sent: 272 Tmp_tables: 0 Tmp_disk_tables: 0 Tmp_table_sizes: 0
# QC_Hit: No Full_scan: Yes Full_join: No Tmp_table: No Tmp_table_on_disk: No
# Filesort: No Filesort_on_disk: No Merge_passes: 0
SET timestamp=1370073666;
SELECT id,title,production_year FROM title WHERE title = 'Bambi';!
16 www.percona.com

pt-query-digest
slow query log
general query log
tcpdump pt-query-digest report
binary log
process list
input tool output
17 www.percona.com

pt-query-digest Report
# Profile!
# Rank Query ID Response time Calls R/Call Apdx V/M Item!
# ==== ================== =============== ===== ======= ==== ===== =======!
# 1 0xA8D2BBDE7EBE7822 4932.2992 28.8% 78 63.2346 0.00 5.22 SELECT person_info!
# 2 0xFE25DAF5DBB71F49 4205.2160 24.6% 130 32.3478 0.00 3.47 SELECT title!
# 3 0x70DAC639802CA233 1299.6269 7.6% 14 92.8305 0.00 0.17 SELECT cast_info!
# 4 0xE336B880F4FEC4B8 1184.5101 6.9% 294 4.0289 0.36 2.29 SELECT cast_info!
# 5 0x60550B93960F1837 905.1648 5.3% 60 15.0861 0.05 1.33 SELECT name!
# 6 0xF46D5C09B4E0CA2F 777.2446 4.5% 16340 0.0476 1.00 0.17 SELECT char_name!
# 7 0x09FCFFF0E5BC929F 747.4346 4.4% 130 5.7495 0.53 7.69 SELECT name!
# 8 0x9433950BE12B9470 744.1755 4.4% 14368 0.0518 1.00 0.18 SELECT name!
# 9 0x4DC0E044996DA715 448.5637 2.6% 130 3.4505 0.65 8.31 SELECT title!
# 10 0x09FB72D72ED18E93 361.1904 2.1% 78 4.6306 0.28 1.89 SELECT cast_info title
#!
# Query 1: 0.02 QPS, 1.40x concurrency, ID 0xA8D2BBDE7EBE7822 at byte 25744265!
# Attribute pct total min max avg 95% stddev median!
# ============ === ======= ======= ======= ======= ======= ======= =======!
# Count 0 78!
# Exec time 28 4932s 11s 109s 63s 88s 18s 63s!
# Lock time 0 724ms 36us 642ms 9ms 11ms 72ms 93us!
# Rows sent 7 12.95k 86 334 170.05 329.68 89.24 136.99!
# Rows examine 14 168.99M 2.17M 2.17M 2.17M 2.17M 0 2.17M!
# Rows affecte 0 0 0 0 0 0 0 0!
# Rows read 7 12.95k 86 334 170.05 329.68 89.24 136.99!
# Bytes sent 5 2.39M 16.18k 54.14k 31.43k 54.03k 13.47k 25.99k!
18 www.percona.com

performance_schema
• Integrated profiling tool
• Not new, but improved significately for 5.6
• Substitutes the old PROFILING interface
• Can have a 5-10% overhead by default
• Can be used to monitor several runtime
statistics, among them, query execution
statistics of the latest queries:
SELECT * FROM
performance_schema.events_statements_history_long!
19 www.percona.com

INDEXING: OLD AND NEW
STRATEGIES

Advantages of Indexing
• Faster Filtering
• Faster Column Return (Covering Index)
• Faster Ordering
• Certain Faster Functions (max, min)
21 www.percona.com

Faster Filtering
• Indexes allow to return results by reading
a fewer number of records
- We should minimize the rows value in
EXPLAIN/Handlers
- const > ref > range > index > ALL!
• Usually, only a single index can be used
per table access
- Compound indexes may provide higher
filtering rate
22 www.percona.com

Compound Indexes
• Column Order in Index Does Matter:
- Order by decreasing selectivity
- Only one range access per index*
23 www.percona.com

Covering Index Technique
• Additional columns can be indexed in
order to speed-up column retrieval
• Specially interesting in n-to-n relationship
tables
• Not possible if the number or size of
columns is too large (e.g. SELECT *,
BLOBs)
24 www.percona.com

Index Condition Pushdown
• Let’s prepare a use case:
UPDATE cast_info SET note = left(note,
250);
ALTER TABLE cast_info MODIFY note
varchar(250), ADD INDEX role_id_note
(role_id, note); !
• We want to execute:
SELECT * FROM cast_info
WHERE role_id = 1
and note like '%Jaime%‘;!

Without ICP (5.5)
mysql> EXPLAIN SELECT *
FROM cast_info
WHERE role_id = 1
and note like '%Jaime%'G
********** 1. row **********
id: 1
select_type: SIMPLE
table: cast_info
type: ref
possible_keys: role_id_note
key: role_id_note
key_len: 4
ref: const
rows: 11553718
Extra: Using where
mysql> SHOW STATUS like 'Hand
%';!
+----------------------------+---------+!
| Variable_name | Value |!
+----------------------------+---------+!
| Handler_commit | 1 |!
| Handler_delete | 0 |!
| Handler_discover | 0 |!
| Handler_prepare | 0 |!
| Handler_read_first | 0 |!
| Handler_read_key | 1 |!
| Handler_read_last | 0 |!
| Handler_read_next | 8346769 |!
| Handler_read_prev | 0 |!
| Handler_read_rnd | 0 |!
| Handler_read_rnd_next | 0 |!
| Handler_rollback | 0 |!
| Handler_savepoint | 0 |!
| Handler_savepoint_rollback | 0 |!
| Handler_update | 0 |!
| Handler_write | 0 |!
+----------------------------+---------+!
16 rows in set (0.00 sec)!

With ICP (5.6)
mysql> EXPLAIN SELECT *
FROM cast_info
WHERE role_id = 1 and note like
'%Jaime%'G
************ 1. row ************
id: 1
select_type: SIMPLE
table: cast_info
type: ref
possible_keys: role_id_note
key: role_id_note
key_len: 4
ref: const
rows: 10259274
Extra: Using index
condition
mysql> SHOW STATUS like 'Hand%';!
+----------------------------+-------+!
+----------------------------+-------+!
| Handler_external_lock | 2 |!
| Handler_mrr_init | 0 |!
+----------------------------+-------+!

Comparison of ICP Execution
• Execution time for this example:
• MySQL 5.5: 5.76 sec
• MySQL 5.6: 1.09 sec
• Over 5x faster
• In this example, it would not work with a
prefix index (e.g. TEXT/BLOB), as it must
search the whole field
• Fun fact: if covering technique is tried, it actually runs
slower than with ICP for this case/hardware (#68554)

ICP and Indexing
• ICP will change the way we index our tables
FORCE INDEX(person_role_id_role_id)
WHERE person_role_id > 0
and person_role_id < 150000
and role_id > 1 and role_id < 7;!
• For example, a multiple column index can be, in
some cases, efficiently used for a range
condition on several columns:
• Effectiveness is highly dependent on how selective
is the second part of the index (an “index scan” is
still done at engine level)
29 www.percona.com

Index for Sorts
• Additional columns can be indexed in
order to speed up ordering of rows
• Only effective in simple patterns
• GROUP BY operations are essentially
sorts
• They can also be greatly improved
with the right indexes in some cases
30 www.percona.com

Multi-Range Read (MRR)
• Reorders access to table data when using a
secondary index on disk for sequential I/O
- Like ICP, its efficiency is highly dependent on
data distribution and memory contents
• It also depends on hardware sequential
access (e.g., InnoDB’s read ahead)
- May not be useful on SSDs
• Not compatible with Using index
• Only for range access and equi-joins

MRR Example
• We want to execute:
WHERE person_role_id > 0
and person_role_id < 150000;
WHERE role_id = 3
and person_role_id > 0
and person_role_id < 500000;!
• We will use these indexes, respectively:
ALTER TABLE cast_info
ADD INDEX person_role_id (person_role_id);
ALTER TABLE cast_info
ADD INDEX person_role_id_role_id (person_role_id,
role_id);!

Without MRR (5.5)
+----------------------------+---------+!
+----------------------------+---------+!
+----------------------------+---------+!
mysql> EXPLAIN SELECT * FROM
cast_info FORCE
INDEX(person_role_id) WHERE
person_role_id > 0 and
person_role_id < 150000G
*********** 1. row ***********
id: 1
select_type: SIMPLE
table: cast_info
type: range
possible_keys: person_role_id
key: person_role_id
key_len: 5
ref: NULL
rows: 8966490
Extra: Using where

With MRR (5.6)
+----------------------------+---------+!
+----------------------------+---------+!
| Handler_external_lock | 4 |!
| Handler_mrr_init | 0 |!
+----------------------------+---------+!
mysql> EXPLAIN SELECT * FROM
cast_info FORCE
INDEX(person_role_id) WHERE
person_role_id > 0 and
person_role_id < 150000G
*********** 1. row ***********
id: 1
select_type: SIMPLE
table: cast_info
type: range
possible_keys: person_role_id
key: person_role_id
key_len: 5
ref: NULL
rows: 8966490
Extra: Using index
condition; Using MRR

Comparison of MRR
Execution
• MySQL 5.5: 4654312 rows in set (1 min 4.79 sec)
• MySQL 5.6 (w/MRR, wo/ICP): 4654312 rows in set
(48.64 sec)
• Consistent 33% execution improvement
for this test case
• Difficult to see for smaller resultsets
• What has changed?
• 15% more “read_ahead”s, resulting in a 40% less data reads
• Reordering has an overhead, can be tuned with
read_rnd_buffer_size

What’s the Point for the
‘FORCE INDEX’?
• For demonstration purposes (full table scan
is better here)
• Some features are not properly detected yet:
SELECT *
FROM cast_info FORCE INDEX(person_role_id_role_id)
WHERE role_id = 3
and person_role_id > 0
and person_role_id < 500000;
• Full table scan (any version): Empty set
(8.08 sec)
• MySQL 5.5: Empty set (1 min 16.88 sec)
• MySQL 5.6 ICP+MRR*: Empty set (0.46 sec)
*ICP is responsible for the dramatic change in execution time, not MRR
36 www.percona.com

Batched Key Access (BKA)
• It retrieves keys in batches and allows
MRR usage for JOINs, as an alternative to
standard Nested Loop Join execution
• Not enabled by default
SET optimizer_switch=
'mrr=on,mrr_cost_based=off,batched_key_access=on';!
37 www.percona.com

Without BKA (5.5) - EXPLAIN
...!
************* 2. row ************* id: 1!
table: cast_info!
type: ref!
possible_keys: person_role_id!
key: person_role_id!
key_len: 5!
ref: imdb.char_name.id!
rows: 4!
Extra: NULL!
mysql> EXPLAIN
SELECT cast_info.note, char_name.name
FROM cast_info FORCE index(person_role_id)
JOIN char_name
ON cast_info.person_role_id = char_name.id
WHERE cast_info.person_role_id > 0 and
cast_info.person_role_id < 150000
************ 1. row ************!
id: 1!
table: char_name!
type: range!
possible_keys: PRIMARY!
key: PRIMARY!
key_len: 4!
ref: NULL!
rows: 313782!
Extra: Using where!
...!
38 www.percona.com

Without BKA (5.5) - Handlers
mysql> SHOW STATUS like ‘Hand%’;!
+----------------------------+---------+
+----------------------------+---------+
+----------------------------+---------+
39 www.percona.com

With BKA (5.6) - EXPLAIN
...!
************ 2. row ************!
id: 1!
table: cast_info!
type: ref!
possible_keys: person_role_id!
key: person_role_id!
key_len: 5!
ref: imdb.char_name.id!
rows: 4!
Extra: Using join buffer
(Batched Key Access)!
mysql> EXPLAIN
SELECT cast_info.note, char_name.name
FROM cast_info FORCE index(person_role_id)
JOIN char_name ON cast_info.person_role_id
= char_name.id
WHERE cast_info.person_role_id > 0 and
cast_info.person_role_id < 150000G!
************ 1. row ************!
id: 1!
table: char_name!
type: range!
possible_keys: PRIMARY!
key: PRIMARY!
key_len: 4!
ref: NULL!
rows: 313782!
Extra: Using where
...!
40 www.percona.com

Comparison of BKA
Execution
• MySQL 5.5: 4654312 rows in set (1 min 6.78 sec)
• MySQL 5.6 (w/MRR, w/BKA): 4654312 rows in set
(1 min 0.47 sec)
• The results are consistent between
executions, but the gain is not too big. But
if we change the join_buffer_size…
• MySQL 5.6 (w/MRR, w/BKA, join_buffer_size =
50M): 4654312 rows in set (19.54 sec)
(join_buffer_size does not affect execution time in the 5.5 version)

Better Index Merge
• In this example, it avoids a full table scan:
mysql> EXPLAIN SELECT COUNT(*) FROM title WHERE (title =
'Pilot' OR production_year > 2010) AND kind_id < 4G
************************ 1. row ************************
id: 1
select_type: SIMPLE
table: title
type: index_merge
possible_keys: title,production_year
key: title,production_year
key_len: 77,5
ref: NULL
rows: 4434
Extra: Using sort_union(title,production_year);
Using where
(almost all titles have kind_id < 4)
• MySQL 5.5: 0.79s – MySQL 5.6: 0.01s

Extended Secondary Keys
• Implicit primary keys inside secondary keys
can be used for filtering (ref, range, etc), not
only for covering index or sorting.
• Requires use_index_extensions=on
(default)
ALTER TABLE title add index
(title(25));
SELECT COUNT(*) FROM title
WHERE title = 'Pilot'
AND id BETWEEN 1000 AND 1000000;

Extended Secondary Keys
mysql> EXPLAIN SELECT COUNT(*) FROM title WHERE
title = 'Pilot' AND id BETWEEN 1000 AND
1000000G
****************** 1. row *******************
id: 1
select_type: SIMPLE
table: title
type: range
possible_keys: PRIMARY,title
key: title
key_len: 81
ref: NULL
rows: 531
Extra: Using index condition; Using
where

Duplicate Key Check
mysql> alter table title add index (production_year);!
Query OK, 0 rows affected (10.08 sec)!
Records: 0 Duplicates: 0 Warnings: 0!
mysql> alter table title add index (production_year);
Query OK, 0 rows affected, 1 warning (5.11 sec)
Records: 0 Duplicates: 0 Warnings: 1!
!
mysql> show warningsG!
************************* 1. row *************************!
Level: Note!
Code: 1831!
Message: Duplicate index 'production_year_2' defined on
the table 'imdb.title'. This is deprecated and will be
disallowed in a future release.!

pt-duplicate-key-checker
• It detects duplicate and redundant
indexes
- Including primary keys at the end of
secondary keys on InnoDB tables
47 www.percona.com

pt-index-usage & user_stats
• Can detect unused indexes
- That slow down writes
- Can also affect reads!
- Wasted memory
- Query optimizer woes
48 www.percona.com

NEW QUERY PLANNER FEATURES

Filesort with Short LIMIT
• Queries that require a filesort but only the first
records are selected can benefit from this
optimization:
mysql> EXPLAIN select * from movie_info
ORDER BY info LIMIT 100G
********************* 1. row *********************
id: 1
select_type: SIMPLE
table: movie_info
type: ALL
possible_keys: NULL
key: NULL
key_len: NULL
ref: NULL
rows: 6927988
Extra: Using filesort
!
Note: Both EXPLAIN and the STATUS Handlers show the same outcome

Filesort with Short LIMIT
(cont.)
• SELECT * FROM movie_info ORDER
BY info LIMIT 100;!
- MySQL 5.5: 20.06 sec
- MySQL 5.6 (P_S on): 9.14 sec
- MySQL 5.6 (P_S off): 8.51 sec
• Over 2x faster.
• Exact speed-up may depend on the
original sort buffer size

Subquery Optimization
• Late Subquery Materialization
- Useful for FROM subqueries if further
filtering is done
- MySQL may also decide to create its own
index on a temporary table on memory
SELECT * FROM title JOIN (SELECT * FROM cast_info)
AS subselect ON subselect.movie_id = title.id WHERE
subselect.id = 1;
EXPLAIN: 0.00 sec (vs >2 min in 5.5)
EXECUTION: 1m 22 sec (vs > 4 min in 5.5)

JOINs and Subqueries
• Better detection of IN + non-dependent
subqueries
- Many queries do not need to be converted
to a JOIN anymore for a proper execution
plan
- MySQL decides what is the best strategy:
Materialization, Semi-Join, etc.

That Subquery in 5.5
mysql> EXPLAIN SELECT * FROM title WHERE kind_id IN
(SELECT id FROM kind_type WHERE kind='video game')G
*************************** 1. row ***************************
id: 1
select_type: PRIMARY
table: title
type: ALL
possible_keys: NULL
key: NULL
key_len: NULL
ref: NULL
rows: 1567676
Extra: Using where
An index on title.kind_id
won't fix it!
*************************** 2. row ***************************
id: 2
select_type: DEPENDENT SUBQUERY
table: kind_type
type: const
possible_keys: PRIMARY,kind_id
key: kind_id
key_len: 47
ref: const
rows: 1
Extra: Using index

That Subquery in 5.6
mysql> EXPLAIN SELECT * FROM title WHERE kind_id IN (SELECT id FROM kind_type WHERE
kind='video game')G
*************************** 1. row ***************************
id: 1
select_type: SIMPLE
table: kind_type
type: const
possible_keys: PRIMARY,kind
key: kind
key_len: 47
ref: const
rows: 1
Extra: Using index
*************************** 2. row ***************************
id: 1
select_type: SIMPLE
table: title
type: ALL
possible_keys: NULL
key: NULL
key_len: NULL
ref: NULL
rows: 1477989
Extra: Using where
An index on title.kind_id
can and will be very useful!

Join Order
• Table order algorithm has been
optimized, which leads to better query
plans when joining many tables

Persistent Optimizer
Statistics
• InnoDB index statistics are no longer
discarded on server shutdown and
recomputed the next time a table is
accessed
• Controlled by variable:
innodb_stats_persistent = ON
(default)
• Remember that SHOW commands/accessing to
I_S do not automatically regenerate statistics by
default

READ ONLY Transactions
• START TRANSACTION READ ONLY;
- It avoids some overhead of R/W
transactions (MVCC)
- Automatically used for SELECT in
auto_commit
- Once started, a new transaction has to be
started to perform a write

Explicit Partition Selection
• Usually unnecessary, as MySQL does
automatically partition pruning
• Can be useful combined with some non-deterministic
functions or in some special
cases:
SELECT count(*)
FROM title PARTITION(p01, p02);

RESULTS AND WRAP-UP

5.6 is a Great Release
• Many optimizer changes that can
potentially improve query execution time
• Some of them are transparent, some
others require tuning and understanding
- Some old tricks and indexing strategies
become obsolete in 5.6
• pt-upgrade, from Percona Toolkit, and
Percona Playback can be great tools to
analyze improvements and regressions
61 www.percona.com

Where to Learn More ?
• More query and server optimization
techniques in our training courses (
http://www.percona.com/training):
- Scaling and Optimization for MySQL
- Operations and Troubleshooting with
MySQL
- Moving to MySQL 5.6
- Analysing Queries with Percona Toolkit
- Percona Server and Tools (NEW!)
• Ask me for a discount (only to PLUK
attendants)!
62 www.percona.com

Where to Learn More ? (cont.)
• Books:
• High Performance MySQL, 3rd Edition
• SQL Antipatterns
• Free MySQL Webinars:
• http://www.percona.com/resources/
mysql-webinars
63 www.percona.com

Percona Live MySQL
Conference and Expo 2014
• For more information and to register:
Visit: http://www.percona.com/live/mysql-conference-2014/

Thank You!
Jaime Crespo
jaime.crespo@percona.com
65 wwwww.wpe.prceorncao.cnoam.c om

Query Optimization with MySQL 5.6: Old and New Tricks

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to Query Optimization with MySQL 5.6: Old and New Tricks (20)

More from MYXPLAIN (10)

Recently uploaded (20)

Query Optimization with MySQL 5.6: Old and New Tricks