Introduction

The A&AI Model Loader Service is an application that facilitates the distribution and ingestion of new service and resource models and VNF catalogs from the SDC to the A&AI.

Features

The Model Loader:

• registers with the SDC to receive notification events
• polls the UEB/DMaap cluster for notification events
• downloads artifacts from SDC upon receipt of a distribution event
• pushes distribution components to A&AI

VNF Catalog loading

The Model Loader supports two methods for supplying VNF Catalog data for loading into A&AI:

• Embedded TOSCA image and vendor data
  VNF Catalog data can be embedded within the TOSCA yaml files contained in the CSAR.

• VNF Catalog XML files
  VNF Catalog data in the form of XML files can be supplied in the CSAR under the path Artifacts/Deployment/VNF_CATALOG

Note: Each CSAR should provide VNF Catalog information using only one of the above methods. If a CSAR contains both TOSCA and XML VNF Catalog information, a deploy failure will be logged and published to SDC, and no VNF Catalog data will be loaded into A&AI

Compiling Model Loader

Model Loader can be compiled by running mvn clean install A Model Loader docker image can be created by running docker build -t onap/model-loader target

Running Model Loader

Push the Docker image to your Docker repository. Pull this down to the host machine.

Create the following directories on the host machine:

./logs
./opt/app/model-loader/appconfig
./opt/app/model-loader/appconfig/auth

You will be mounting these as data volumes when you start the Docker container. For examples of the files required in these directories, see the aai/test/config repository (https://gerrit.onap.org/r/#/admin/projects/aai/test-config)

Populate these directories as follows:

Contents of /opt/app/model-loader/appconfig

The following file must be present in this directory on the host machine:

model-loader.properties

# Always false.  TLS Auth currently not supported 
ml.distribution.ACTIVE_SERVER_TLS_AUTH=false

# Address/port of the SDC
ml.distribution.ASDC_ADDRESS=<SDC-Hostname>:8443

# Kafka consumer group.  
ml.distribution.CONSUMER_GROUP=aai-ml-group

# Kafka consumer ID
ml.distribution.CONSUMER_ID=aai-ml

# SDC Environment Name.  This must match the environment name configured on the SDC
ml.distribution.ENVIRONMENT_NAME=<Environment Name>

# Currently not used
ml.distribution.KEYSTORE_PASSWORD=

# Currently not used
ml.distribution.KEYSTORE_FILE=

# Obfuscated password to connect to the SDC.  To obtain this value, use the following Jetty library to 
# obfuscate the cleartext password:  http://www.eclipse.org/jetty/documentation/9.4.x/configuring-security-secure-passwords.html
ml.distribution.PASSWORD=OBF:<password>

# How often (in seconds) to poll the Kafka topic for new model events
ml.distribution.POLLING_INTERVAL=<integer>

# Timeout value (in seconds) when polling the Kafka topic for new model events
ml.distribution.POLLING_TIMEOUT=<integer>

# Username to use when connecting to the SDC
ml.distribution.USER=<username>

# Artifact type we want to download from the SDC (the values below will typically suffice)
ml.distribution.ARTIFACT_TYPES=MODEL_QUERY_SPEC,TOSCA_CSAR

# URL of the A&AI
ml.aai.BASE_URL=https://<AAI-Hostname>:8443

# A&AI endpoint to post models
ml.aai.MODEL_URL=/aai/v*/service-design-and-creation/models/model/

# A&AI endpoint to post named queries
ml.aai.NAMED_QUERY_URL=/aai/v*/service-design-and-creation/named-queries/named-query/

# A&AI endpoint to post vnf images
ml.aai.VNF_IMAGE_URL=/aai/v*/service-design-and-creation/vnf-images

# Name of certificate to use in connecting to the A&AI
ml.aai.KEYSTORE_FILE=aai-os-cert.p12

# Obfuscated keystore password to connect to the A&AI.  This is only required if using 2-way SSL (not basic auth).
# To obtain this value, use the following Jetty library to obfuscate the cleartext password:
# http://www.eclipse.org/jetty/documentation/9.4.x/configuring-security-secure-passwords.html
ml.aai.KEYSTORE_PASSWORD=OBF:<password>

# Name of user to use when connecting to the A&AI.  This is only required if using basic auth (not 2-way SSL).
ml.aai.AUTH_USER=<username>

# Obfuscated password to connect to the A&AI.  This is only required if using basic auth (not 2-way SSL).
# To obtain this value, use the following Jetty library to obfuscate the cleartext password:
# http://www.eclipse.org/jetty/documentation/9.4.x/configuring-security-secure-passwords.html
ml.aai.AUTH_PASSWORD=OBF:<password>

Contents of the /opt/app/model-loader/app-config/auth Directory

The following files must be present in this directory on the host machine:

aai-os-cert.p12

The certificate used to connected to the A&AI

Start the service:

You can now start the Docker container for the Model Loader Service, e.g:

docker run -d \
	-e CONFIG_HOME=/opt/app/model-loader/config/ \
    -v /logs:/logs \
    -v /opt/app/model-loader/appconfig:/opt/app/model-loader/config \
    --name model-loader \
    {{your docker repo}}/model-loader

where

{{your docker repo}}

is the Docker repository you have published your image to.