Enabling and Using Memory Acceleration

This section describes how to enable memory acceleration. The process is as follows:

Selecting a GeminiDB Instance

Creating a Mapping Rule

Using the Memory Acceleration Module

Usage Notes

After memory acceleration is enabled, commands such as RESET MASTER and FLUSH LOGS used to delete binlogs on MySQL instances are not allowed.
Currently, only hash data from MySQL can be converted to GeminiDB Redis API.
A Redis key prefix and a delimiter in a new rule can neither include those nor be included in those specified for an existing rule. For example, if the key prefix in a new rule is pre1: and is separated by a comma (,) and the key prefix in an existing rule is pre1 and is separated by a colon (:), the new rule cannot be created.
Currently, the ENUM, SET, and JSON data cannot be synchronized.
Currently, only single-table queries are supported during lightweight incremental synchronization. Joint queries are not supported.
Only GeminiDB Redis instances are charged. There are no other fees for this function.
If you delete an RDS instance, the GeminiDB Redis instance with DB Cache enabled will not be deleted. If you do not need the GeminiDB Redis instance, delete it in a timely manner to avoid extra fees.
When you purchase an RDS instance, if you select Buy Now for memory acceleration, a GeminiDB instance is automatically provisioned with DB Cache enabled. You can skip instance creation and start from Creating a Mapping Rule. This function is now in OBT. To use it, choose Service Tickets > Create Service Ticket in the upper right corner of the console and contact the customer service.

Procedure

Selecting a GeminiDB Instance

Log in to the Huawei Cloud console.
Click in the upper left corner and select a region and project.
Click in the upper left corner of the page and choose Databases > Relational Database Service.
On the Instances page, click the target instance name to go to the Overview page.
In the navigation pane, choose Memory Acceleration. On the Memory Acceleration page, click Create GeminiDB Instance or Use Existing GeminiDB Instance.
- Click Create GeminiDB Instance and perform 6.
- Click Use Existing GeminiDB Instance and select an existing GeminiDB Redis instance.
  
  When you select Use Existing GeminiDB Instance, only primary/standby instances are supported. The region, VPC, subnet, and security group of the GeminiDB and RDS instances must be the same.

Set parameters listed in Table 1 and click Submit.

**Table 1** Basic information
Parameter	Description
Instance Specifications	CPU and memory of the instance. For details, see Table 2.
Database Port	Port number for accessing the instance. You can specify a port number based on your requirements. The port number ranges from 1024 to 65535 except 2180, 2887, 3887, 6377, 6378, 6380, 8018, 8079, 8091, 8479, 8484, 8999, 12017, 12333, and 50069. If you do not specify a port number, port 6379 is used by default. NOTE: You cannot change the database port after an instance is created.
DB Instance Name	The instance name: Can be the same as an existing instance name. Can include 4 to 64 bytes and must start with a letter. It is case-sensitive and allows only letters, digits, hyphens (-), and underscores (_).
Database Password	Database password set by a user: Can contain 8 to 32 characters. Can contain at least two types of the following characters: uppercase letters, lowercase letters, digits, and special characters ~!@#%^*-_=+? For security reasons, set a strong password. The system will verify the password strength. Keep your password secure. The system cannot retrieve it if it is lost.
Confirm Password	Enter the administrator password again.

By default, the region, AZ, VPC, and subnet of the GeminiDB and RDS instances are the same.

**Table 2** GeminiDB Redis instance specifications
Storage (GB)	Nodes	vCPUs	QPS	Maximum Connections per Single-node Instance	Databases
16	2	1	10,000	10,000	1,000
24	2	2	20,000	10,000	1,000
32	2	2	20,000	10,000	1,000
48	2	4	40,000	20,000	1,000
64	2	4	40,000	20,000	1,000
96	2	8	80,000	20,000	1,000
128	2	16	160,000	20,000	1,000

Creating a Mapping Rule

Log in to the Huawei Cloud console.
Click in the upper left corner and select a region and project.
Click in the upper left corner of the page and choose Databases > Relational Database Service.
On the Instances page, click the target instance name to go to the Overview page.
In the navigation pane, choose Memory Acceleration. In the Mapping Rule area, click Create Mapping Rule.

Figure 1 Mapping rule
On the displayed page, configure parameters.
1. Enter a rule name.
  Rule Name: Enter a mapping rule name. The rule name must be unique within a GeminiDB instance and cannot exceed 256 characters or include number signs (#).
  Figure 2 Rule name
2. Configure source instance information.
  - Database Name: Select a database of the acceleration instance.
  - Table Name: Select a table of the acceleration instance.
  Figure 3 Configuring source instance information
3. Configure acceleration instance information.
  - Redis Key Prefix: This parameter is optional. The default value is in the format of Database name:Table name:Field name 1:Field name 2... and can contain a maximum of 1,024 characters. If you have created a custom prefix, it will be used instead of the default one.
  - Value Storage Type: Data type of the cache. Currently, only hash data is supported.
  - Database No. (0-999): ID of a database that stores cached data in the acceleration instance. The default value is 0.
  - TTL (s) Default value: 30 days: Validity period of cached data in the acceleration instance. The default value is 30 days (2,592,000 seconds). If you enter -1, the cached data will never expire.
  - Key Delimiter: Separator among the Redis key prefix, key, and key fields. It is a single character in length.
  Figure 4 Configuring acceleration instance information
4. Click Set Key, select a key field of the acceleration instance, and click OK.
  
  If an acceleration instance key consists of multiple source instance fields, the key must be unique in a MySQL instance. You can click Up or Down to adjust the sequence of each field.
  
  Figure 5 Key settings
  
  After the parameters are set, the key is displayed.
  
  Figure 6 Key
5. Configure the acceleration instance fields.
  Move the required fields in the source instance to the acceleration instance.
  Figure 7 Configuring acceleration instance fields
6. After setting the parameters, click Submit.

Using the Memory Acceleration Module

Create database db1 in the source MySQL instance and create table students in db1.

mysql> CREATE DATABASE db1;
Query OK, 1 row affected (0.00 sec)

mysql> CREATE TABLE db1.students(
       sid INT UNSIGNED PRIMARY KEY AUTO_INCREMENT NOT NULL,
       sname VARCHAR(20),
       sclass INT,
       sgender VARCHAR(10),
       sbirthday DATE
       );
Query OK, 0 rows affected (0.00 sec)

mysql> DESC db1.students;
+-----------+--------------+------+-----+---------+----------------+
| Field     | Type         | Null | Key | Default | Extra          |
+-----------+--------------+------+-----+---------+----------------+
| sid       | int unsigned | NO   | PRI | NULL    | auto_increment |
| sname     | varchar(20)  | YES  |     | NULL    |                |
| sclass    | int          | YES  |     | NULL    |                |
| sgender   | varchar(10)  | YES  |     | NULL    |                |
| sbirthday | date         | YES  |     | NULL    |                |
+-----------+--------------+------+-----+---------+----------------+
5 rows in set (0.00 sec)

After the table is created, on the memory acceleration page, create a mapping rule to convert each row in the students table into a Redis hash. The key of a hash is in the format of Database name:Data table name:sid:<sid value>. The selected fields are sname, sclass, sgender, and sbirthday.
Figure 8 Configuring a mapping rule
After a mapping rule is created, check the mapping rule and information.
Figure 9 Mapping information

Insert a new data record to the students table in the MySQL instance.

mysql> INSERT INTO db1.students (sname, sclass, sgender, sbirthday) VALUES ('zhangsan', 1, 'male', '2015-05-20');
Query OK, 1 row affected (0.01 sec)

mysql> SELECT * FROM db1.students;
+-----+----------+--------+---------+------------+
| sid | sname    | sclass | sgender | sbirthday  |
+-----+----------+--------+---------+------------+
|   1 | zhangsan |      1 | male    | 2015-05-20 |
+-----+----------+--------+---------+------------+
1 row in set (0.00 sec)

After the mapping rule is created, the data is automatically synchronized to the GeminiDB instance. Run commands in the GeminiDB instance to query the data.

127.0.0.1:6379> KEYS *
1) "db1:students:sid:1"

127.0.0.1:6379> HGETALL db1:students:sid:1
1) "sbirthday"
2) "2015-05-20"
3) "sclass"
4) "1"
5) "sgender"
6) "male"
7) "sname"
8) "zhangsan"

Insert a new data record to the students table in the MySQL instance.

mysql> INSERT INTO db1.students (sname, sclass, sgender, sbirthday) VALUES ('lisi', 10, 'male', '2015-05-22');
Query OK, 1 row affected (0.00 sec)

mysql> SELECT * FROM db1.students;
+-----+----------+--------+---------+------------+
| sid | sname    | sclass | sgender | sbirthday  |
+-----+----------+--------+---------+------------+
|   1 | zhangsan |      1 | male    | 2015-05-20 |
|   2 | lisi     |     10 | male    | 2015-05-22 |
+-----+----------+--------+---------+------------+
2 rows in set (0.00 sec)

Check whether the new data is synchronized to the GeminiDB instance.

127.0.0.1:6379> KEYS *
1) "db1:students:sid:1"
2) "db1:students:sid:2"

127.0.0.1:6379> HGETALL db1:students:sid:2
1) "sbirthday"
2) "2015-05-22"
3) "sclass"
4) "10"
5) "sgender"
6) "male"
7) "sname"

Update data in the students table in the MySQL instance.

mysql> UPDATE db1.students SET sclass=12, sname='wangwu' WHERE sid = 1;
Query OK, 1 row affected (0.00 sec)
Rows matched: 1  Changed: 1  Warnings: 0

mysql> SELECT * FROM db1.students;
+-----+--------+--------+---------+------------+
| sid | sname  | sclass | sgender | sbirthday  |
+-----+--------+--------+---------+------------+
|   1 | wangwu |     12 | male    | 2015-05-20 |
|   2 | lisi   |     10 | male    | 2015-05-22 |
+-----+--------+--------+---------+------------+
2 rows in set (0.00 sec)

Check whether the data is updated in the GeminiDB instance.

127.0.0.1:6379> KEYS *
1) "db1:students:sid:1"
2) "db1:students:sid:2"

127.0.0.1:6379> HGETALL db1:students:sid:1
1) "sbirthday"
2) "2015-05-20"
3) "sclass"
4) "12"
5) "sgender"
6) "male"
7) "sname"
8) "wangwu"

Delete data from the students table in the MySQL instance.

mysql> DELETE FROM db1.students WHERE sid = 1;
Query OK, 1 row affected (0.00 sec)

mysql> SELECT * FROM db1.students;
+-----+-------+--------+---------+------------+
| sid | sname | sclass | sgender | sbirthday  |
+-----+-------+--------+---------+------------+
|   2 | lisi  |     10 | male    | 2015-05-22 |
+-----+-------+--------+---------+------------+
1 row in set (0.00 sec)