SonarQube Platform Overview
The SonarQube platform is made of 3 components:
- the configuration of the SonarQube instance (security, plugins settings, etc.)
- the quality snapshots of projects, views, etc.
- One for users to browse quality snapshots and configure the SonarQube instance
- One or more
- The SonarQube platform cannot have more than a single web server and a single database.
- For optimal performance, each component (Database, Web Server, Analyzers) should be installed on a dedicated machine. However, the Server and the Database can be installed on the same machine.
- Analyzers will scale by adding machines.
- The Database and the Analyzers must be located in the same network.
- The Database and the Analyzers must not be installed on the same machine.
- All the machines must be time synchronized.
Installing the Database
Several database engines are supported.
Create an empty schema and a sonarqube user. Grant this sonarqube user permissions to create, update and delete objects for this schema. The charset of the database has to be set to "UTF-8" and the language to "English". Here is a sample script for MySQL.
Microsoft SQL Server
Note that collation must be case-sensitive (CS) and accent-sensitive (AS).
There are two well-known engines that can be used in MySQL: MyISAM and InnoDB. MyISAM is the oldest of the two engines and is being progressively replaced by InnoDB. InnoDB is clearly faster and scales better with SonarQube as the number of projects under quality control increases. If you were an early adopter of SonarQube, you probably have a series of table that are still using MyISAM. To improve performances, you should change the engine for all tables to InnoDB.
Once all SonarQube tables are using the InnoDB engine, the first thing to do is allocate a maximum amount of RAM to your MySQL instance with the
innodb_buffer_pool_size parameter and give at least 15Mb to the
query_cache_size parameter. Read this article about InnoDB Performance Optimization Basics for more information.
When having two SonarQube schemas on the same Oracle instance, especially if they are of two different versions, SonarQube gets confused and picks the first it finds. To fix this issue:
- Either privileges associated to the SonarQube Oracle users should be decreased
- Or a trigger should be defined on Oracle side to automatically alter the SonarQube Oracle user session when establishing a new connection:
If you want to use a custom schema and not the default "public" one, the PostgreSQL
search_path property must be set:
Installing the Web Server
First, check the requirements.
<install_directory> (below) refers to the path to the directory where the SonarQube distribution has been unzipped.
Setting the access to the Database
Edit <install_directory>/conf/sonar.properties to configure the database settings. Templates are available for every supported database. Just uncomment and configure the template you need and comment out the lines dedicated to H2:
Adding the JDBC Driver
Drivers for the supported databases (except Oracle) are already provided. Do not replace the provided drivers; they are the only ones supported.
For Oracle, copy the JDBC driver into <install_directory>/extensions/jdbc-driver/oracle.
Starting the Web Server
The default port is "9000" and the context path is "/". These values can be changed in <install_directory>/conf/sonar.properties:
Execute the following script to start the server:
- On Linux/Mac OS: bin/<YOUR OS>/sonar.sh start
- On Windows: bin/windows-x86-XX/StartSonar.bat
- Read Installing SonarQube on AIX, HPUX or Solaris for other platforms
Tuning the Web Server
By default, SonarQube is configured to run on any computer with a simple Java JRE. For better performance, the first thing to do when installing a production instance is to use a Java JDK and activate the server mode by uncommenting the following line in <install_directory>/conf/wrapper.conf:
To change the Java JVM used by SonarQube, simply edit <install_directory>/conf/wrapper.conf and update the following line:
Advanced Installation Features
- Running SonarQube as a Service on Windows or Linux
- Running SonarQube behind a Proxy
- Running SonarQube over HTTPS
Installing the Analyzers
Multiple analyzers are available:
- SonarQube Runner: recommended for all non-Maven projects
- Maven: recommended for all projects built with Maven
- SonarQube Ant Task: to integrate with projects built with Ant
- Gradle: to integrate with projects built with Gradle
- CI Engine: see Jenkins plugin or Hudson plugin or Bamboo plugin or AnthillPro plugin. Note that other CI engines can be used even if there is no SonarQube plugin available.