Migration Guide 3.2

Table of Contents

• @InjectMock
• Native Compilation - Native executables and .so files
• Native Compilation - Work around missing CPU features
• Transaction manager (Narayana)
• Kubernetes
  • Improve the logic to generate TLS-based container ports
  • Management interface port is not bound as a Service any longer
• Hibernate Search
  • Defaults for projectable/sortable on geo-point fields
  • Different DB schema for outbox-polling
  • Deprecated/renamed configuration properties

@InjectMock

io.quarkus.test.junit.mockito.InjectMock has been deprecated in favor of io.quarkus.test.InjectMock.

Native Compilation - Native executables and .so files

Due to changes made to GraalVM/Mandrel, if you are using an extension that is relying on .so files (such as the AWT extension), you need to add these .so files to the native container similarly to:

...
# Copy of the .so files to the container
COPY --chown=1001:root target/*.so /work/
COPY --chown=1001:root target/*-runner /work/application
...

Native Compilation - Work around missing CPU features

When building on recent machines and running your native executable on older machines, you may see the following failure when starting the application:

The current machine does not support all of the following CPU features that are required by the image: [CX8, CMOV, FXSR, MMX, SSE, SSE2, SSE3, SSSE3, SSE4_1, SSE4_2, POPCNT, LZCNT, AVX, AVX2, BMI1, BMI2, FMA].
Please rebuild the executable with an appropriate setting of the -march option.

This error message means that the native compilation used more advanced instruction sets not supported by the CPU running the application. To work around that issue, add the following line to the application.properties:

quarkus.native.additional-build-args=-march=compatibility

Then, rebuild your native executable. This setting forces the native compilation to use an older instruction set, increasing the chance of compatibility.

To explicitly define the target architecture, run native-image -march=list to get the supported configurations and then set -march to one of them, e.g., quarkus.native.additional-build-args=-march=x86-64-v4. If you are targeting an AMD64 host, -march=x86-64-v2 would work in most cases.

Note	The march parameter is only available on GraalVM 23+.

Transaction manager (Narayana)

The configuration property quarkus.transaction-manager.object-store-directory has been renamed to quarkus.transaction-manager.object-store.directory.

Kubernetes

Improve the logic to generate TLS-based container ports

At the moment, the Kubernetes extension is always adding a container port called https to the generated Deployment resources. This approach is problematic because the port won’t work unless SSL is configured.

After 3.2, the container port https will not be added to the generated Deployment resources by default, unless: - Either users explicitly provide some of the quarkus.http.ssl.* properties at build time. - Or users add this property quarkus.kubernetes.ports.https.tls=true at build time.

More information about these changes in here.

Management interface port is not bound as a Service any longer

When the management port is enabled (see more information here), the Kubernetes extension will not generate the Kubernetes Service that binds the management port by default for security reasons.

Still, if you want to expose it via Service and Ingress resources, you can add the following properties:

quarkus.kubernetes.ingress.expose=true
quarkus.kubernetes.ingress.target-port=management

For OpenShift, you would need:

quarkus.openshift.route.expose=true
quarkus.openshift.route.target-port=management

And now, a new Kubernetes Service will be generated to bind the Management interface port.

Hibernate Search

Defaults for projectable/sortable on geo-point fields

Note	The following applies to Quarkus 3.2.3+ only, as Quarkus 3.2.0, 3.2.1, 3.2.2 were still using Hibernate Search 6.1.

Hibernate Search 6.2 changed how defaults are handled for geo-point fields.

If your Hibernate Search mapping includes geo-point fields that are using the default value for the projectable option, and are using either the default value or Sortable.NO for the sortable option, Elasticsearch schema validation will fail on startup because of missing docvalues on those fields.

To address that, either:

• Revert to the previous defaults by adding projectable = Projectable.NO to the mapping annotation of relevant geo-point fields.

• Or recreate your Elasticsearch indexes and reindex your database. The easiest way to do so is to use the MassIndexer with dropAndCreateSchemaOnStart(true).

See also https://docs.jboss.org/hibernate/search/6.2/migration/html_single/#data-format

Different DB schema for outbox-polling

Note	The following applies to Quarkus 3.2.3+ only, as Quarkus 3.2.0, 3.2.1, 3.2.2 were still using Hibernate Search 6.1.

If you use the quarkus-hibernate-search-orm-outbox-polling extension, be aware that the database schema for the tables used by that extension changed.

See https://docs.jboss.org/hibernate/search/6.2/migration/html_single/#data-format for help on how to migrate to the new schema.

Deprecated/renamed configuration properties

Note	The following applies to Quarkus 3.2.3+ only, as Quarkus 3.2.0, 3.2.1, 3.2.2 were still using Hibernate Search 6.1.

• quarkus.hibernate-search-orm.automatic-indexing.synchronization.strategy is now deprecated in favor of quarkus.hibernate-search-orm.indexing.plan.synchronization.strategy.

• quarkus.hibernate-search-orm.automatic-indexing.enable-dirty-check is now deprecated with no alternative to replace it. After its removal in a future version, a dirty check will always be performed when considering whether to trigger reindexing.

See also https://docs.jboss.org/hibernate/search/6.2/migration/html_single/#configuration