trellis-deployment

This provides composite deployments for Trellis application modules.

trellis-compose: contains docker-compositions

docker images

trellis-app-triplestore - Provides the Trellis Application configured with an external Triplestore and Kafka Notification System.

• 1.8 (Stable 0.6.1 Release Build)

• 1.10 (JPMS)

kafka - Provides Kafka.

zookeeper - Provides Zookeeper.

fuseki - Provides a Fuseki Triplestore.

trellis-processing - Provides a Beam async processor.

Administration

• Trellis can be monitored at http://localhost:8502
• Fuseki SPARQL interface at http://localhost:3030/fuseki/dataset.html?tab=query&ds=/trellis
• Zookeeper can be monitored at http://localhost:8500/commands

Mount Points

To persist data, create these mount points before running docker-compose:

• /mnt/trellis-binaries
• /mnt/kafka-data
• /mnt/fuseki-data
• /mnt/fuseki-backup
• /mnt/trellis-data (only with trellis-rosid-app)

Configuration

• Trellis can be configured with config.yml
• Zookeeper can be configured with zoo.cfg

JDK

• the trellis-app is compiled for 1.8, 1.10.
• the 10 build use an experimental JPMS branch

External Triplestore

• Jena Fuseki 3.4.0 in Tomcat 9 with HTTP/2 over TLS

HTTP/2 over TLS

• a development keystore trellis.jks is included in trellis-app/etc
• https is available on port 8445
• see ldp-client for an HTTP/2 capable client

Scaling

• Kafka can be scaled by running docker-compose up --scale kafka=3 where 3 is the number of brokers.
• Scaling can be performed after trellis processing has been started.

File-Based Resource Service

• To run trellis-rosid-app run docker-compose -f rosid-app.yml up
• To start the async processor, run docker-compose up in the trellis-compose/trellis-processing directory.
• This should be done after all other containers have started.
• Trellis Processing is required for the creation of LDP containment triples with this resource service.