Note these instruction refers to current master branch. The release-1.0.0 branch have development issues that have only been fixed in master branch.
DCAE Controller Overview
Make sure your settings.xml includes the onap public repo for some snapshot jar downloads during the build.
DCAE Controller LF GIT repositories
- dcae/controller
- dcae/operation/utils
- dcae/controller/analytics
- dcae/demo
SOMF - Sirius Operational Management Framework
SOMF is a model-driven framework for building controllers. SOMF Is using the Eclipse Modeling Framework (EMF) for data modeling.
LF SOMF repositories
- ncomp/maven - contains a few maven related artifacts
- ncomp/utils - contains Java utilities and libraries that are used across DCAE Controller projects
- ncomp/core - contains core models
- ncomp/sirius/manager - contains the core SOMF implementation
- ncomp/docker - contains SOMF adaptor used to communicate with Docker Engine API
- ncomp/openstack- contains SOMF adaptor used to communicate with OpenStack API
- ncomp/cdap- contains SOMF adaptor used to communicate with CDAP API+
Development Setup
GIT repositories and build order
- ncomp/maven
- ncomp/utils
- dcae/operation/utils
- ncomp/core
- ncomp/sirius/manager
- ncomp/cdap
- ncomp/docker
- ncomp/openstack
- dcae/controller
- dcae/controller/analytics
- dcae/demo
Eclipse Setup
Since DCAE Controller/SOMF is build on top of EMF developement is tightly tied to using Eclipse IDE with specific plugin installed. This section describe the process to setup a developement Eclipse.
- Install Neon.3 Java Developer Version http://www.eclipse.org/downloads/packages/eclipse-ide-java-developers/neon3
- Install software from Neon update site
- Eclipse Plug-in Development Environment 3.12.2.v20161124-1400
- EMF - Eclipse Modeling Framework Xcore SDK 1.4.0.v20160526-0606
- Install software from Eclipse Market Place
- YEdit Feature 1.0.20.201509041456-RELEASE
- Install Groovy from http://dist.springsource.org/snapshot/GRECLIPSE/e4.6/
- Groovy-Eclipse Feature 2.9.2.xx-201703131833-e46
- Groovy-Eclipse M2E integration 2.9.2.xx-201703131833-e46
- Groovy Compiler 2.4 Feature 2.9.2.xx-201703131833-e46
Import Eclipse Projects
The GIT repositories include the Eclipse Project Metadata (e.g., .project, .classpath, and etc files) and thus can be importing into a workspace using the "Import Projects" → "Import existing Eclipse projects". Only the projects that need to be worked on need to be in the workspace.
Working on DCAE Controller projects
There are 2 projects for most features.
- A model project which contains the data model using XCORE file. These projects always have a name (and artifactId) that ends in
-model
.This file is located on thesrc/main/xcore
directory. And the project is setup so Eclipse will update automatically the generated Java code in thesrc/main/xcore-gen
directory. This project is a straight EMF type project and contain very little SOMF specifics. - A SOMF project which contains SOMF generated code and implementation of the APIs that the model defines.
- SOMF generator class
Generator.java
under thesrc/main/java
folder. It contains the defined customization for creating the SOMF implementation. Running this class will generate new contents in the two generated folders. src/main/sirius-gen
folder contains generated Java code for implementing Client and Server side functions.src/main/server-gen
folder contains generated Bash and Groovy scripts for implementing Client and Server side functions.- SOMF provider
*Provider.java
Java classes under thesrc/main/java
folder. These implements the operation/methods that are defined in the model.
- SOMF generator class
DCAE Configuration Setup
The DCAE configuration setup is trying to carefully handle Environmental and Non-Environmental configuration.
The ONAP Demo setup is creating another level of complexity to the setup since the overall ONAP Heat templates will deliver environmental information that need to be include in the Environmental configuration.
This is the various configuration entities.
- DCAE Non Environmental ONAP Demo configuration. https://gerrit.onap.org/r/gitweb?p=dcae/demo.git;a=tree;f=OPENECOMP-DEMO;h=38b2f51a1be5c12f357e228b7a655c0ee97606b8;hb=HEAD with the main files
- location-types.yaml define the various DCAE entities (e.g., docker-host VMs, cdap cluster VMs, VES collector etc) that are part of the demo deployment
- vm-templates/vm-docker-host.yaml, vm-templates/vm-cdap-cluster.yaml, vm-templates/vm-postgresql.yaml . Define the deployment configuration for various VM deployments.
- docker-templates/docker-XX.yaml. Similar for docker deployments.
- cdap-templates/cdap-YY.yaml. Similar for docker deployments.
- steams.yaml. Define DCAE DMaaP setup.
- HEAT provided information. When the ONAP demo is getting deploy the vm1-dcae-controller VM which will come up with the DCAE controller Docker container. This container will use configuration attributes setup in the file /opt/app/dcae-controller/config.yaml to apply to DCAE controller environment file, which will have expressions like @{XXX} replaced with the value of XXX in the config.yaml file.
- DCAE Environmental ONAP Demo configuration. Currently 3 deployment scenarios are supported (See
-
DCAE-7Getting issue details...
STATUS
for details about current status.)
- RACKSPACE. This scenario only work when the Cloud provider is Rackspace. In this each DCAE VM will be setup with a predefined fixed IP on a private network and it will get a random public IP on the public network. This scenario depends on the following values from the HEAT provided environment file: DCAE-VERSION, DOCKER-REGISTRY, DOCKER-VERSION, GIT-MR-REPO, HORIZON-URL, KEYSTONE-URL, NEXUS-PASSWORD, NEXUS-RAWURL, NEXUS-USER, OPENSTACK-KEYNAME, OPENSTACK-PASSWORD, OPENSTACK-PRIVATE-NETWORK, OPENSTACK-PUBKEY, OPENSTACK-REGION, OPENSTACK-TENANT-ID, OPENSTACK-TENANT-NAME, OPENSTACK-USER, POLICY-IP, STATE, ZONE.
- 2-NIC. This scenario works in most OpenStack environments and allows higher level of flexibility in assigning IPs etc. Similar to the RACKSPACE setup each DCAE VM will be setup with a predefined fixed IP on a private network and it will get a random public IP on the public network.This scenario depends on the following values from the HEAT provided environment file: DCAE-VERSION, DNS-IP-ADDR, DOCKER-REGISTRY, DOCKER-VERSION, FLAVOR-LARGE, GIT-MR-REPO, HORIZON-URL, KEYSTONE-URL, NEXUS-PASSWORD, NEXUS-RAWURL, NEXUS-USER, OPENSTACK-AUTH-METHOD, OPENSTACK-KEYNAME, OPENSTACK-PASSWORD, OPENSTACK-PRIVATE-NETWORK, OPENSTACK-PUBKEY, OPENSTACK-REGION, OPENSTACK-TENANT-ID, OPENSTACK-TENANT-NAME, OPENSTACK-USER, POLICY-IP, STATE, UBUNTU-1404-IMAGE, UBUNTU-1604-IMAGE, ZONE, dcae_cdap00_ip_addr, dcae_cdap01_ip_addr, dcae_cdap02_ip_addr, dcae_coll00_ip_addr, dcae_ip_addr, dcae_pstg00_ip_addr, public_net_id. The HEAT template demo/heat/OpenECOMP/onap_openstack_nofloat.yaml provides these variables.
- 1-NIC-FLOATING-IPS. This scenario works in most OpenStack environments and allows higher level of flexibility in assigning IPs etc. In this setup each DCAE VM will only have one NIC with an IP on the private network. In addition each VM will have a floating IP from the public network. In this case it is the floating VMs that can get predefined values and the IPs from the private network that are assigned to the NIC are randomly setup. This makes a few issues with the rest of the demo setup but should be working in the 1.1 release. This scenario depends on the following values from the HEAT provided environment file:DCAE-VERSION, DNS-IP-ADDR, DOCKER-REGISTRY, DOCKER-VERSION, FLAVOR-LARGE, GIT-MR-REPO, HORIZON-URL, KEYSTONE-URL, NEXUS-PASSWORD, NEXUS-RAWURL, NEXUS-USER, OPENSTACK-KEYNAME, OPENSTACK-PASSWORD, OPENSTACK-PRIVATE-NETWORK, OPENSTACK-PUBKEY, OPENSTACK-REGION, OPENSTACK-TENANT-ID, OPENSTACK-TENANT-NAME, OPENSTACK-USER, POLICY-IP, STATE, UBUNTU-1404-IMAGE, UBUNTU-1604-IMAGE, ZONE, dcae_cdap00_float_ip_addr, dcae_cdap01_float_ip_addr, dcae_cdap02_float_ip_addr, dcae_coll00_float_ip_addr, dcae_float_ip_addr, dcae_pstg00_float_ip_addr. The HEAT template demo/heat/OpenECOMP/onap_openstack_float.yaml provides these variables.
The overall deployment flow is the following.
- HEAT template
- Create the vm1-dcae-controller VM.
- Setup /opt/app/dcae-controller/config.yaml
- Start the Docker Container for the DCAE Controller
- The DCAE controller will (running /opt/app/dcae-controller-platform-server/bin/controller-startup.sh)
- Determine the BASE based on the BASE attribute in config.yaml
- Determine the ZONE based on the ZONE attribute in config.yaml
- Substitute the HEAT delivered attributes from config.yaml into /opt/app/dcae-controller-platform-server/OPENECOMP-DEMO-$BASE and produce a /opt/app/dcae-controller-platform-server//opt/app/dcae-controller-platform-server/OPENECOMP-DEMO-$ZONE.
bin/dcae-controller.sh rackspace-substitute --from OPENECOMP-DEMO-$BASE --to OPENECOMP-DEMO-$ZONE --file /opt/app/dcae-controller/config.yaml
- Merge the Non-environmental configuration OPENECOMP-DEMO with the environmental configuration OPENECOMP-DEMO-$ZONE to create the complete configuration setup for this specific environment which will be placed in GITLINK/OPENECOMP-DEMO-$ZONE
java -cp 'lib/*' org.openecomp.dcae.controller.operation.utils.GenControllerConfiguration $ZONE . GITLINK OPENECOMP-DEMO
- The DCAE Controller will be started up and synced with the configuration
bin/dcae-controller.sh start
bin/dcae-controller.sh sync-configuration --environment OPENECOMP-DEMO-$ZONE
The DCAE Controller can then Deploy the various DCAE components define in the demo environment
bin/dcae-controller.sh deploy-service-instance -i $ZONE -s vm-docker-host-1
bin/dcae-controller.sh deploy-service-instance -i $ZONE -s vm-postgresql
bin/dcae-controller.sh deploy-service-instance -i $ZONE -s vm-cdap-cluster
bin/dcae-controller.sh deploy-service-instance -i $ZONE -s docker-databus-controller
bin/dcae-controller.sh deploy-service-instance -i $ZONE -s cdap-helloworld
bin/dcae-controller.sh deploy-service-instance -i $ZONE -s cdap-tca-hi-lo
bin/dcae-controller.sh deploy-service-instance -i $ZONE -s docker-common-event