Advance MySQL Training by Pratyush Majumdar

Advance MySQL training
Pratyush Majumdar
VP - Technology
91Mobiles

Topics
1. Database Optimization and Performance Tuning
2. Stored Procedures
3. Triggers
4. Partitioning
5. Advanced Indexing Techniques

Topics
6. User Management and Security
7. Replication
8. Backup and Recovery
9. Show processlist
10. Monitoring

Tuning
2. Add an Index and check the performance again
mysql> create index idx_first_name on employees(first_name);
Query OK, 0 rows affected (1.52 sec)
Records: 0 Duplicates: 0 Warnings: 0
mysql> explain select * from employees where first_name='John';
+----+-------------+-----------+------+----------------+----------------+---------+-------+------+-------------+
+----+-------------+-----------+------+----------------+----------------+---------+-------+------+-------------+
| 1 | SIMPLE | employees | ref | idx_first_name | idx_first_name | 16 | const | 1 | Using where |
+----+-------------+-----------+------+----------------+----------------+---------+-------+------+-------------+

Tuning
query_cache_type = 1
query_cache_size = 128M
innodb_buffer_pool_size = 3 GB (80% of RAM)
innodb_file_per_table = 1
max_connections = 1000
datadir = /var/lib/mysql

Tuning
Enabling slow query in MySQL conﬁguration.
slow-query-log = 1
long-query-time = 5
log-queries-not-using-indexes = 1 (‘0’ for Production)
mysql> show variables where Variable_name in ('log_slow_queries','long_query_time','slow_query_log_file');
+---------------------+--------------------------------------+
| Variable_name | Value |
+---------------------+--------------------------------------+
| log_slow_queries | ON |
| long_query_time | 5.000000 |
| slow_query_log_file | /var/lib/mysql/00e805bd8a3f-slow.log |
+---------------------+--------------------------------------+

Get top 10 Slow queries
mysqldumpslow -a -t 10 /var/lib/mysql/00e805bd8a3f-slow.log
Reading mysql slow query log from /var/lib/mysql/00e805bd8a3f-slow.log
Count: 1 Time=6.08s (6s) Lock=0.00s (0s) Rows=1.0 (1), root[root]@localhost
select count(*) from salaries_wo_partition where from_date between '1998-01-01' and '1998-12-31'
Tuning

MySQL Memory calculator
Sum of global buffers + max_connections x (sum of local buffers)
SELECT ( @@key_buffer_size
+ @@query_cache_size
+ @@innodb_buffer_pool_size
+ @@innodb_log_buffer_size
+ @@innodb_additional_mem_pool_size
+ @@max_connections * (
@@read_buffer_size
+ @@read_rnd_buffer_size
+ @@sort_buffer_size
+ @@join_buffer_size
+ @@binlog_cache_size
+ @@thread_stack
+ @@max_heap_table_size )
) / (1024 * 1024 * 1024) AS MAX_MEMORY_GB;
Tool: https://www.mysqlcalculator.com/
Tuning

2.1 Stored Procedures
Stored procedures allows us to encapsulate SQL code and reuse it later.
DELIMITER //
CREATE PROCEDURE CalculateSalaryCredit(IN employee_no INT, IN from_year VARCHAR(4), IN
to_year VARCHAR(4))
BEGIN
SELECT SUM(salary) AS SalaryCredit
FROM salaries
WHERE emp_no = employee_no AND year(from_date)>from_year and year(to_date)<to_year;
END //
DELIMITER ;

2.2 Stored Procedures
Calling the Stored Procedure
mysql> call CalculateSalaryCredit(10009, '1985', '1990');
+--------------+
| SalaryCredit |
+--------------+
| 195686 |
+--------------+

3. Triggers
Triggers are database objects that are automatically ﬁred when certain events
occur such as inserting, updating or deleting data.
CREATE TRIGGER customer_update_trigger
AFTER UPDATE ON customers
FOR EACH ROW
BEGIN
IF NEW.address <> OLD.address THEN
INSERT INTO customer_address_history (customer_id, old_address, new_address, updated_at)
VALUES (NEW.customer_id, OLD.address, NEW.address, NOW());
END IF;
END;

4.1 Partitioning
Partition the salaries table by date range.
ALTER TABLE salaries partition by range COLUMNS (from_date)
(
partition p01 values less than ('1985-12-31'),
…………
partition p19 values less than (MAXVALUE)
);

4.2 Partitioning
Performance difference on tables with and without partitions
mysql> select count(*) from salaries where from_date between '1998-01-01' and '1998-12-31';
+----------+
| count(*) |
+----------+
| 247489 |
+----------+
mysql> select count(*) from salaries_wo_partition where from_date between '1998-01-01' and '1998-12-31';
+----------+
| count(*) |
+----------+
| 247489 |
+----------+

4.3 Partitioning
mysql> explain select count(*) from salaries where from_date between '1998-01-01' and '1998-12-31';
+----+-------------+----------+-------+---------------+---------+---------+------+--------+--------------------------+
+----+-------------+----------+-------+---------------+---------+---------+------+--------+--------------------------+
| 1 | SIMPLE | salaries | index | NULL | PRIMARY | 7 | NULL | 509532 | Using where; Using index |
+----+-------------+----------+-------+---------------+---------+---------+------+--------+--------------------------+
mysql> explain select count(*) from salaries_wo_partition where from_date between '1998-01-01' and '1998-12-31';
+----+-------------+-----------------------+-------+---------------+---------+---------+------+---------+--------------------------+
+----+-------------+-----------------------+-------+---------------+---------+---------+------+---------+--------------------------+
| 1 | SIMPLE | salaries_wo_partition | index | NULL | PRIMARY | 7 | NULL | 3003272 | Using where; Using index |
+----+-------------+-----------------------+-------+---------------+---------+---------+------+---------+--------------------------+

4.4 Partitioning
Conclusion of using MySQL Partitions
● Use for huge tables (index size > available RAM)
● Use where there are more write operations (indexes may slow down writes)
● Works best for time-series data that needs to be analyzed by date range.

5.1 Advanced Indexing Techniques
Covering Indexes
Covering indexes include all columns needed for a query within the index itself. This eliminates the
need to access the actual table data, signiﬁcantly boosting performance for queries that only need
those speciﬁc columns.
SELECT name, email FROM users WHERE id = 1;
CREATE INDEX idx_user_id_name_email ON users (id, name, email);

Composite Indexes
These indexes combine multiple columns, allowing eﬃcient searching based on multiple criteria
simultaneously. They're ideal for queries that ﬁlter or sort based on more than one column.
SELECT name, email FROM users WHERE country = 'US' AND age > 25;
CREATE INDEX idx_user_country_age ON users ( country, age);

Full-text Indexes
These enable eﬃcient searching for keywords within text columns. They're crucial for applications
involving text-based searches, signiﬁcantly improving search speed and relevance.
SELECT * FROM articles WHERE MATCH(title, content) AGAINST ('artificial
intelligence');
CREATE FULLTEXT INDEX idx_article_fulltext ON articles (title, content);

Function-based indexes
These enable eﬃcient searching for keywords within text columns. They're crucial for applications
involving text-based searches, signiﬁcantly improving search speed and relevance.
SELECT * FROM products WHERE UPPER(name) LIKE 'APPLE%';
CREATE INDEX idx_product_name_upper ON products ( UPPER(name));

6.1 User Management and Security
Create a user and assign privileges.
mysql> CREATE USER 'webapp'@'localhost' IDENTIFIED BY 'password';
mysql> GRANT SELECT, INSERT, UPDATE ON mydatabase.* TO 'webapp'@'localhost';
mysql> show grants for 'webapp'@'localhost';
+---------------------------------------------------------------------------------------------------------------+
| Grants for webapp@localhost |
+---------------------------------------------------------------------------------------------------------------+
| GRANT USAGE ON *.* TO 'webapp'@'localhost' IDENTIFIED BY PASSWORD '*2470C0C06DEE42FD1618BB99005ADCA2EC9D1E19' |
| GRANT SELECT, INSERT, UPDATE ON `mydatabase`.* TO 'webapp'@'localhost' |
+---------------------------------------------------------------------------------------------------------------+
2 rows in set (0.00 sec)

Best Practices
1. mysqld should be run as an ordinary, unprivileged user instead.
user = mysql
2. Remote connections to the MySQL server can be disabled
bind_address = 127.0.0.1 # comment this line to allow remote connections
3. There should not be any user with root, super privileges. Avoid using
wildcard(*) in case of hostnames
mysql> CREATE USER 'webapp'@'%' IDENTIFIED BY 'password';
4. Do not create single user for all databases
mysql> GRANT SELECT, INSERT, UPDATE, DELETE ON *.* TO 'webapp'@'localhost';
6.2 User Management and Security

7.2 Replication
Master-Slave replication settings
# On the master server
server-id=1
log-bin=mysql-bin
binlog-format=ROW
# On the slave server
server-id=2
replicate-do-db=mydatabase

Master-Master replication settings
# master-01 server
server-id = 1
log-bin = /var/lib/mysql/00e805bd8a3f-mysql-bin.log
binlog-do-db = test
auto-increment_increment = 2
auto-increment_offset = 1
# master-02 server
server-id = 2
log-bin = /var/lib/mysql/c1065b70d158-mysql-bin.log
binlog-do-db = test
auto-increment-increment = 2
auto-increment-offset = 2
7.3 Replication

8.1 Backup and Recovery
MySQL Backup
mysqldump -u username -p mydatabase > /home/user/backup.sql # dump whole database
mysqldump -u username -p mydatabase mytable > /home/user/backup.sql # single table
mysqldump -u username -p --routines mydatabase > /home/user/backup.sql # dump with
Stored Procedures and Functions
mysqldump -u username -p --no-data mydatabase > /home/user/backup.sql # structure only
dump
mysqldump -u username -p --skip-lock-tables mydatabase > /home/user/backup.sql # dump
the database without locking any table

8.2 Backup and Recovery
MySQL Restore
root@00e805bd8a3f:/# mysql -u root -p < /home/user/backup.sql
mysql> source /home/user/backup.sql

Advance MySQL Training by Pratyush Majumdar

More Related Content

Similar to Advance MySQL Training by Pratyush Majumdar (20)

More from Pratyush Majumdar (9)

Recently uploaded (20)

Advance MySQL Training by Pratyush Majumdar