Many projects that use data require connections to a relational database.
The main way of obtaining connections to a database is to use a datasource and configure a JDBC driver.
In Quarkus, the preferred datasource and connection pooling implementation is Agroal.
Agroal is a modern, light weight connection pool implementation designed for very high performance and scalability, and features first class integration with the other components in Quarkus, such as security, transaction management components, health metrics.
This guide will explain how to:
-
configure a datasource, or multiple datasources
-
how to obtain a reference to those datasources in code
-
which pool tuning configuration properties are available
Prerequisites
To complete this guide, you will need:
-
less than 10 minutes
-
an IDE
-
JDK 1.8+ installed with
JAVA_HOME
configured appropriately -
Apache Maven 3.5.3+
Creating the Maven project
First, we need a new project. Create a new project with the following command:
mvn io.quarkus:quarkus-maven-plugin:1.0.0.CR2:create \
-DprojectGroupId=org.acme \
-DprojectArtifactId=agroal-quickstart \
-DclassName="org.acme.datasource.GreetingResource" \
-Dpath="/hello"
cd agroal-quickstart
This will generate:
-
the Maven structure
-
a landing page accessible on
http://localhost:8080
-
an example
Dockerfile
files for bothnative
andjvm
modes -
the application configuration file
-
an
org.acme.datasource.GreetingResource
resource -
an example integration test
Adding maven dependencies
Next, you will need to add the quarkus-agroal
dependency to your project.
You can add it using a simple Maven command:
./mvnw quarkus:add-extension -Dextensions="agroal"
Agroal comes as a transitive dependency of the Hibernate ORM extension so if you are using Hibernate ORM, you don’t need to add the Agroal extension dependency explicitly. |
You will also need to choose, and add, the Quarkus extension for your relational database driver.
Quarkus provides driver extensions for:
-
H2 - jdbc-h2
-
PostgreSQL - jdbc-postgresql
-
MariaDB - jdbc-mariadb
-
MySQL - jdbc-mysql
-
Microsoft SQL Server - jdbc-mssql
-
Derby - jdbc-derby
The H2 and Derby databases can normally be configured to run in "embedded mode"; the extension does not support compiling the embedded database engine into native images. Read Testing with in-memory databases (below) for suggestions regarding integration testing. |
As usual, you can install the extension using add-extension
.
To install the PostgreSQL driver dependency for instance, just run the following command:
./mvnw quarkus:add-extension -Dextensions="jdbc-postgresql"
Configuring the datasource
Once the dependencies are added to your pom.xml file, you’ll need to configure Agroal.
This is done in the src/main/resources/application.properties
file.
A viable configuration file would be:
quarkus.datasource.url=jdbc:h2:tcp://localhost/mem:default
quarkus.datasource.driver=org.h2.Driver
quarkus.datasource.username=username-default
quarkus.datasource.min-size=3
quarkus.datasource.max-size=13
There are other configuration options, detailed below.
For more information about the Agroal extension configuration please refer to the Configuration Reference. |
JDBC URL configurations
Each of the supported databases contains different JDBC URL configuration options. Going into each of those options is beyond the scope of this document, but it gives an overview of each database URL and link to the official documentation.
H2
jdbc:h2:{ {.|mem:}[name] | [file:]fileName | {tcp|ssl}:[//]server[:port][,server2[:port]]/name }[;key=value…]
- Example
-
jdbc:h2:tcp://localhost/~/test
,jdbc:h2:mem:myDB
H2 is an embedded database. It can run as a server, based on a file, or live completely in memory. All of these options are available as listed above. You can find more information at the official documentation.
PostgreSQL
PostgreSQL only runs as a server, as do the rest of the databases below. As such, you must specify connection details, or use the defaults.
jdbc:postgresql:[//][host][:port][/database][?key=value…]
- Example
-
jdbc:postgresql://localhost/test
Defaults for the different parts are as follows:
host
-
localhost
port
-
5432
database
-
same name as the username
The official documentation go into more detail and list optional parameters as well.
MariaDB
jdbc:mariadb:[replication:|failover:|sequential:|aurora:]//<hostDescription>[,<hostDescription>…]/[database][?<key1>=<value1>[&<key2>=<value2>]]
hostDescription:: <host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
- Example
-
jdbc:mariadb://localhost:3306/test
You can find more information about this feature and others detailed in the official documentation.
MySQL
jdbc:mysql:[replication:|failover:|sequential:|aurora:]//<hostDescription>[,<hostDescription>…]/[database][?<key1>=<value1>[&<key2>=<value2>]]
hostDescription:: <host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
- Example
-
jdbc:mysql://localhost:3306/test
You can find more information about this feature and others detailed in the official documentation.
Microsoft SQL Server
Microsoft SQL Server takes a connection URL in the following form:
jdbc:sqlserver://[serverName[\instanceName][:portNumber]][;property=value[;property=value]]
- Example
-
jdbc:sqlserver://localhost:1433;databaseName=AdventureWorks
The Microsoft SQL Server JDBC driver works essentially the same as the others. More details can be found in the official documentation.
Derby
jdbc:derby:[//serverName[:portNumber]/][memory:]databaseName[;property=value[;property=value]]
- Example
-
jdbc:derby://localhost:1527/myDB
,jdbc:derby:memory:myDB;create=true
Derby is an embedded database. It can run as a server, based on a file, or live completely in memory. All of these options are available as listed above. You can find more information at the official documentation.
Injecting a Datasource
Because Quarkus uses CDI, injecting a datasource is very simple:
@Inject
AgroalDataSource defaultDataSource;
In the above example, the type is AgroalDataSource
which is a subtype of javax.sql.DataSource
.
Because of this, you can also use javax.sql.DataSource
.
Multiple Datasources
Agroal allows you to configure multiple datasources. It works exactly the same way as a single datasource, with one important change: a name.
quarkus.datasource.driver=org.h2.Driver
quarkus.datasource.url=jdbc:h2:tcp://localhost/mem:default
quarkus.datasource.username=username-default
quarkus.datasource.min-size=3
quarkus.datasource.max-size=13
quarkus.datasource.users.driver=org.h2.Driver
quarkus.datasource.users.url=jdbc:h2:tcp://localhost/mem:users
quarkus.datasource.users.username=username1
quarkus.datasource.users.min-size=1
quarkus.datasource.users.max-size=11
quarkus.datasource.inventory.driver=org.h2.Driver
quarkus.datasource.inventory.url=jdbc:h2:tcp://localhost/mem:inventory
quarkus.datasource.inventory.username=username2
quarkus.datasource.inventory.min-size=2
quarkus.datasource.inventory.max-size=12
Notice there’s an extra bit in the key.
The syntax is as follows: quarkus.datasource.[optional name.][datasource property]
.
Named Datasource Injection
When using multiple datasources, each DataSource
also has the io.quarkus.agroal.DataSource
qualifier with the name of the datasource in the property as the value.
Using the above properties to configure three different datasources, you can also inject each one as follows:
@Inject
AgroalDataSource defaultDataSource;
@Inject
@DataSource("users")
AgroalDataSource dataSource1;
@Inject
@DataSource("inventory")
AgroalDataSource dataSource2;
Datasource Health Check
If you are using the quarkus-smallrye-health
extension, quarkus-agroal
will automatically add a readiness health check
to validate the datasource.
So when you access the /health/ready
endpoint of your application you will have information about the datasource validation status.
If you have multiple datasources, all datasources will be checked and the status will be DOWN
as soon as there is one datasource validation failure.
This behavior can be disabled via the property quarkus.datasource.health.enabled
.
Narayana Transaction Manager integration
If the Narayana JTA extension is also available, integration is automatic.
You can override this by setting the transactions
configuration property - see the Configuration Reference below.
Testing with in-memory databases
Some databases like H2 and Derby are commonly used in "embedded mode" as a facility to run quick integration tests.
Our suggestion is to use the real database you intend to use in production; container technologies made this simple enough so you no longer have an excuse. Still, there are sometimes good reasons to also want the ability to run quick integration tests using the JVM powered databases, so this is possible as well.
It is important to remember that when configuring H2 (or Derby) to use the embedded engine, this will work as usual in JVM mode but such an application will not compile into a native image, as the Quarkus extensions only cover for making the JDBC client code compatible with the native compilation step: embedding the whole database engine into a native image is currently not implemented.
If you plan to run such integration tests in the JVM exclusively, it will of course work as usual.
If you want the ability to run such integration test in both JVM and/or native images, we have some cool helpers for you: just add either @QuarkusTestResource(H2DatabaseTestResource.class)
or @QuarkusTestResource(DerbyDatabaseTestResource.class)
on any class in your integration tests, this will make sure the testsuite starts (and stops) the embedded database into a separate process as necessary to run your tests.
These additional helpers are provided by the artifacts having Maven coordinates io.quarkus:quarkus-test-h2:1.0.0.CR2
and io.quarkus:quarkus-test-derby:1.0.0.CR2
, respectively for H2 and Derby.
Follows an example for H2:
package my.app.integrationtests.db;
import io.quarkus.test.common.QuarkusTestResource;
import io.quarkus.test.h2.H2DatabaseTestResource;
@QuarkusTestResource(H2DatabaseTestResource.class)
public class TestResources {
}
This will allow you to test your application even when it’s compiled into a native image, while the database will run in the JVM as usual.
Connect to it using:
quarkus.datasource.url=jdbc:h2:tcp://localhost/mem:test
quarkus.datasource.driver=org.h2.Driver
Agroal Configuration Reference
Configuration property fixed at build time - ️ Configuration property overridable at runtime
Type |
Default |
|
---|---|---|
The datasource driver class name |
string |
|
Whether we want to use regular JDBC transactions, XA, or disable all transactional capabilities.
When enabling XA you will need a driver implementing |
|
|
Whether or not an healtcheck is published in case the smallrye-health extension is present (default to true). |
boolean |
|
The datasource URL |
string |
|
The datasource username |
string |
|
The datasource password |
string |
|
The credentials provider name |
string |
|
The credentials provider type.
It is the |
string |
|
The initial size of the pool. Usually you will want to set the initial size to match at least the minimal size, but this is not enforced so to allow for architectures which prefer a lazy initialization of the connections on boot, while being able to sustain a minimal pool size after boot. |
int |
|
The datasource pool minimum size |
int |
|
The datasource pool maximum size |
int |
|
The interval at which we validate idle connections in the background.
Set to |
|
|
The timeout before cancelling the acquisition of a new connection |
|
|
The interval at which we check for connection leaks. |
||
The interval at which we try to remove idle connections. |
|
|
The max lifetime of a connection. |
||
The transaction isolation level. |
|
|
Enable datasource metrics collection. |
boolean |
|
When enabled Agroal will be able to produce a warning when a connection is returned to the pool without the application having closed all open statements. This is unrelated with tracking of open connections. Disable for peak performance, but only when there’s high confidence that no leaks are happening. |
boolean |
|
Query executed when first using a connection. |
string |
|
Type |
Default |
|
The datasource driver class name |
string |
|
Whether we want to use regular JDBC transactions, XA, or disable all transactional capabilities.
When enabling XA you will need a driver implementing |
|
|
The datasource URL |
string |
|
The datasource username |
string |
|
The datasource password |
string |
|
The credentials provider name |
string |
|
The credentials provider type.
It is the |
string |
|
The initial size of the pool. Usually you will want to set the initial size to match at least the minimal size, but this is not enforced so to allow for architectures which prefer a lazy initialization of the connections on boot, while being able to sustain a minimal pool size after boot. |
int |
|
The datasource pool minimum size |
int |
|
The datasource pool maximum size |
int |
|
The interval at which we validate idle connections in the background.
Set to |
|
|
The timeout before cancelling the acquisition of a new connection |
|
|
The interval at which we check for connection leaks. |
||
The interval at which we try to remove idle connections. |
|
|
The max lifetime of a connection. |
||
The transaction isolation level. |
|
|
Enable datasource metrics collection. |
boolean |
|
When enabled Agroal will be able to produce a warning when a connection is returned to the pool without the application having closed all open statements. This is unrelated with tracking of open connections. Disable for peak performance, but only when there’s high confidence that no leaks are happening. |
boolean |
|
Query executed when first using a connection. |
string |
About the Duration format
The format for durations uses the standard You can also provide duration values starting with a number.
In this case, if the value consists only of a number, the converter treats the value as seconds.
Otherwise, |