Shulou(Shulou.com)06/01 Report--

What is Mycat?

As for what Mycat is, there may be different understandings under different roles. If you have any knowledge of MySQL architecture, you all know that MySQL is actually made up of Server layer and storage engine layer. So for DBA, Mycat is the Server layer of MySQL. The MySQL Server connected behind the Mycat is like the storage engine of MySQL. Therefore, Mycat itself does not store data, the data is stored on the back-end MySQL, so the reliability of the data

And transactions are still guaranteed by MySQL.

For developers, Mycat is basically equivalent to MySQL. Because you can connect to the Mycat by connecting to the MySQL, there is no difference in the connection method except for the default port. So Mycat is transparent to developers, but there are some restrictions on the use of SQL.

For architects, Mycat is a database middleware that can be used for read-write separation, table partition, and disaster recovery backup. For applications using Mycat, it hides the storage logic of the database. Let us not need to control the storage logic such as read-write separation in the code, so as to expand the database architecture without modifying the application code.

As Mycat is a domestic open source project, there are very perfect and rich Chinese documents. Therefore, this article will not repeat the basic concepts, it is recommended to refer to the online version of the authoritative guide to Mycat or to the official website of Mycat to learn more about Mycat.

What is database middleware

Database middleware is an intermediate service between database and application for data processing and interaction. After slicing the data, it is divided into multiple sliced databases from the original database, and all the sliced database clusters constitute the whole complete database storage.

As shown in the figure above, after the data is divided into multiple sharded databases, if the application needs to read the data, it needs to process the data from multiple data sources.

If there is no database middleware, then the application will directly face the sharding cluster. Then the problems such as data source switching, transaction processing and data aggregation all need to be dealt with directly. This will lead to applications that are supposed to focus on the business, will spend a lot of work to deal with the problems after sharding, and most importantly, each application processing will be a complete repetition of the wheel.

Therefore, with database middleware, applications only need to focus on business processing. A large number of general data aggregation, transactions and data source switching are handled by middleware, and the performance and processing capacity of middleware will directly determine the read and write performance of the application, so a good database middleware is very important.

The main role of Mycat

1. Used as distributed database middleware:

Mycat can connect not only to mysql but also to other databases such as oracle and sql server, because Mycat is written in Java. Theoretically, Mycat can connect to any database that Java can connect to. This allows the application to connect to the Mycat instead of directly connecting to the database, and the back-end database is transparent to the application. In this way, the complexity of the back-end database can be shielded from the application, and the connection mode can be maintained.

2. Implement read-write separation and load balancing of database cluster:

When the system develops to the later stage and the amount of data comes up, the database architecture usually evolves to the cluster. In the Replication cluster scheme, there is a master-slave role in the node, that is, the master-slave architecture. In this kind of cluster, the master node is usually responsible for writing and the slave node is responsible for reading. On the other hand, Mycat can be used as an intermediate layer to achieve read-write separation, without the need for each application to write a set of read-write separation logic. And when there are multiple slave nodes, Mycat can also balance the load of read requests to each slave node.

3. Realize the high availability of the database:

When the database architecture is one master and one slave, Mycat can be used to achieve high availability of the database. When the master node dies, the Mycat can switch to the slave node for reading and writing. However, it is not applicable to the case of one master and multiple slaves, because after the master node dies, Mycat can only switch to one of the slave nodes, but cannot change the role of the node. This will cause the remaining slave nodes to be unable to synchronize them. Therefore, one master and multiple slaves need to use MMM or MHA to achieve high availability.

4. Split the business database vertically:

When the write load in the system comes up, a single master node begins to be unable to bear the pressure of write requests, so it is necessary to increase the master node to share the load. When you have multiple master nodes, you usually split the undependent tables in a library into different libraries, or split some large tables into small tables to spread the load. This is the vertical segmentation of the database.

The existence of multiple master nodes means that the application may need to operate on multiple databases, while Mycat can abstract multiple databases into a logical library. In this way, there is no need to change the code to operate multiple data sources for the application, only the logic library in Mycat can be operated.

5. Split the business database horizontally:

When the database is vertically split, there may still be some libraries with a high write load, or the amount of data reaches the limit of a single table. Then you need to consider horizontal segmentation, the so-called horizontal segmentation is the sub-database table. The difference between vertical sharding and horizontal sharding is that after horizontal sharding, the table structure in each library is the same, and the data is split. On the other hand, the vertical segmentation is that the structure of the library table is different, and the segmentation is the table.

When the database is split horizontally, when the application is writing data, you need to consider which library the data should be written to. On the other hand, Mycat can distribute write requests to a database through a specific strategy, so that applications only need to send write requests to Mycat, regardless of which library should be written.

6. Control the number of database connections:

There are many applications in the system, and each application has its own connection pool, so when the number of visits increases, it may exceed the maximum number of connections to the database. At this point, a common connection pool can be established through Mycat, and all applications can obtain database connections through this connection pool. In order to control the number of applications connected to the database at the same time, play a current-limiting role, can ensure that the system is more stable.

Mycat basic elements

Logical Library:

Usually for practical applications, it is not necessary to know the existence of Mycat, business developers only need to know the concept of database, so database middleware can be regarded as a logical library composed of one or more database clusters, it should be noted that the logical library does not store data. So no matter what kind of database architecture is behind the middleware, it is operating a database for the application, and this database is the logical library abstracted by Mycat.

Logical tables:

Since there is a logical library, then there will be logical tables. In distributed databases, tables that read and write data are logical tables for applications. Logical tables, which can be divided into data and distributed in one or more sharding libraries. There can also be no data segmentation, no slicing, only one table. Similarly, logical tables do not store data.

Logical table classification:

Sharded tables: tables whose data is split into multiple databases, non-sharded tables: tables whose data has not been split: tables that are similar to dictionaries and exist in all shards, and are designed to solve the problem of cross-sharding data joinER tables: tables that are sharded according to ER relationships, so that the records of child tables and the associated parent table records are stored on the same data shard, ensuring that data join will not be installed across library Mycat

Environment:

System: CentOS 7JDK:11Mycat:1.6.7.4

First, go to the official download address to get the download link, and then go to linux to download and decompress it using the following command:

[root@txy-server ~] # cd / usr/local/src [root@txy-server / usr/local/src] # wget http://dl.mycat.io/1.6.7.4/Mycat-server-1.6.7.4-release/Mycat-server-1.6.7.4-release-20200105164103-linux.tar.gz[root@txy-server / usr/local/src] # tar-zxvf Mycat-server-1.6.7.4-release-20200105164103-linux.tar.gz-C / usr/local/

After decompression, the directory structure of Mycat is as follows:

[root@txy-server / usr/local/mycat] # ll-rhtotal 24K-rwxrwxrwx 1 root root 227 Jan 5 16:41 version.txtdrwxrwxrwx 2 root root 4.0K Jan 5 16:40 logs # log file directory drwxr-xr-x 2 root root 4.0K Jan 6 20:40 lib # library file directory drwxrwxrwx 4 root root 4.0K Jan 6 20:40 conf # configuration file directory drwxrwxrwx 2 root root 4.0K Oct 22 21:26 catletdrwxr-xr-x 2 root root 4. 0K Jan 6 20:40 bin # executable directory [root@txy-server / usr/local/mycat] #

Create a system account for Mycat and change the user of the mycat directory, because it is not possible for the program to have root account permissions in the real world. The command is as follows:

[root@txy-server / usr/local] # useradd mycat-s / sbin/nologin [root@txy-server / usr/local] # chown-R mycat:mycat mycat/

Configure environment variables:

[root@txy-server ~] # vim / etc/profileexport MYCAT_HOME=/usr/local/mycatexport PATH=$PATH:$MYCAT_HOME/ [root @ txy-server ~] # source / etc/profile

Since I have only 2 gigabytes of memory here, I need to modify the Mycat startup parameter to change the maximum memory to 1 gigabyte, which defaults to 2 gigabytes. Of course, if your machine has enough memory, you don't need to change it:

[root@txy-server / usr/local] # vim mycat/conf/wrapper.confwrapper.java.additional.4=-XX:MaxDirectMemorySize=1G

Start the Mycat service:

[root@txy-server ~] # mycat start

Check the log. The output of successfully indicates that the startup is successful:

[root@txy-server ~] # more / usr/local/mycat/logs/wrapper.log | grep successfullyINFO | jvm 1 | 21:16:51 on 2020-01-06 | MyCAT Server startup successfully. See logs in logs/mycat.log [root@txy-server ~] #

After Mycat starts, ports such as 8066 and 9066 will be listened by default. If the firewall is enabled, the port needs to be released. The command to open the port is as follows:

[root@txy-server] # firewall-cmd-- zone=public-- add-port=$ {port number} / tcp-- permanent [root@txy-server ~] # firewall-cmd-- reload

At this point, we have successfully built the Mycat service.

After setting up the Mycat, how to use it? The first step is to understand its core configuration file, which will be introduced in the next article because of the limited space:

Detailed explanation of Mycat Core configuration

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.