PDP-D Engine


The PDP-D Core Engine provides an infrastructure and services for drools based applications in the context of Policies and ONAP.

A PDP-D supports applications by means of controllers. A controller is a named grouping of resources. These typically include references to communication endpoints, maven artifact coordinates, and coders for message mapping.

Controllers use communication endpoints to interact with remote networked entities typically using messaging (dmaap or ueb), or http.

PDP-D Engine capabilities can be extended via features. Integration with other Policy Framework components (API, PAP, and PDP-X) is through one of them (feature-lifecycle).

The PDP-D Engine infrastructure provides mechanisms for data migration, diagnostics, and application management.


Source Code repositories

The PDP-D software is mainly located in the policy/drools repository with the communication endpoints software residing in the policy/common repository and Tosca policy models in the policy/models repository.

Docker Image

Check the drools-pdp released versions page for the latest versions. At the time of this writing 1.8.2 is the latest version.

docker pull onap/policy-drools:1.8.2

A container instantiated from this image will run under the non-priviledged policy account.

The PDP-D root directory is located at the /opt/app/policy directory (or $POLICY_HOME), with the exception of the $HOME/.m2 which contains the local maven repository. The PDP-D configuration resides in the following directories:

  • /opt/app/policy/config: ($POLICY_HOME/config or $POLICY_CONFIG) contains engine, controllers, and endpoint configuration.

  • /home/policy/.m2: ($HOME/.m2) maven repository configuration.

  • /opt/app/policy/etc/: ($POLICY_HOME/etc) miscellaneous configuration such as certificate stores.

The following command can be used to explore the directory layout.

docker run --rm -it nexus3.onap.org:10001/onap/policy-drools:1.8.2 -- bash

Communication Endpoints

PDP-D supports the following networked infrastructures. This is also referred to as communication infrastructures in the source code.

  • DMaaP

  • UEB

  • NOOP

  • Http Servers

  • Http Clients

The source code is located at the policy-endpoints module in the policy/commons repository.

These network resources are named and typically have a global scope, therefore typically visible to the PDP-D engine (for administration purposes), application controllers, and features.

DMaaP, UEB, and NOOP are message-based communication infrastructures, hence the terminology of source and sinks, to denote their directionality into or out of the controller, respectively.

An endpoint can either be managed or unmanaged. The default for an endpoint is to be managed, meaning that they are globally accessible by name, and managed by the PDP-D engine. Unmanaged topics are used when neither global visibility, or centralized PDP-D management is desired. The software that uses unmanaged topics is responsible for their lifecycle management.

DMaaP Endpoints

These are messaging enpoints that use DMaaP as the communication infrastructure.

Typically, a managed endpoint configuration is stored in the <topic-name>-topic.properties files.

For example, the DCAE_TOPIC-topic.properties is defined as



In this example, the generic name of the source endpoint is DCAE_TOPIC. This is known as the canonical name. The actual topic used in communication exchanges in a physical lab is contained in the $DCAE_TOPIC environment variable. This environment variable is usually set up by devops on a per installation basis to meet the needs of each lab spec.

In the previous example, DCAE_TOPIC is a source-only topic.

Sink topics are similarly specified but indicating that are sink endpoints from the perspective of the controller. For example, the APPC-CL topic is configured as




Although not shown in these examples, additional configuration options are available such as user name, password, security keys, consumer group and consumer instance.

UEB Endpoints

Similary, UEB endpoints are messaging endpoints, similar to the DMaaP ones.

For example, the DCAE_TOPIC-topic.properties can be converted to an UEB one, by replacing the dmaap prefix with ueb. For example:



NOOP Endpoints

NOOP (no-operation) endpoints are messaging endpoints that don’t have any network attachments. They are used for testing convenience. To convert the DCAE_TOPIC-topic.properties to a NOOP endpoint, simply replace the dmaap prefix with noop:


HTTP Clients

HTTP Clients are typically stored in files following the naming convention: <name>-http-client.properties convention. One such example is the AAI HTTP Client:



HTTP Servers

HTTP Servers are stored in files that follow a similar naming convention <name>-http-server.properties. The following is an example of a server named CONFIG, getting most of its configuration from environment variables.



Endpoints configuration resides in the $POLICY_HOME/config (or $POLICY_CONFIG) directory in a container.


Controllers are the means for the PDP-D to run applications. Controllers are defined in <name>-controller.properties files.

For example, see the usecases controller configuration.

This configuration file has two sections: a) application maven coordinates, and b) endpoint references and coders.

Maven Coordinates

The coordinates section (rules) points to the controller-usecases kjar artifact. It is the brain of the control loop application.



This kjar contains the usecases DRL file (there may be more than one DRL file included).

        $policy : ToscaPolicy()


    ControlLoopParams params = ControlLoopUtils.toControlLoopParams($policy);
    if (params != null) {

The DRL in conjuction with the dependent java libraries in the kjar pom realizes the application’s function. For intance, it realizes the vFirewall, vCPE, and vDNS use cases in ONAP.


Endpoints References and Coders

The usecases-controller.properties configuration also contains a mix of source (of incoming controller traffic) and sink (of outgoing controller traffic) configuration. This configuration also contains specific filtering and mapping rules for incoming and outgoing dmaap messages known as coders.


dmaap.source.topics.APPC-LCM-WRITE.events.org.onap.policy.appclcm.AppcLcmDmaapWrapper.filter=[?($.type == 'response')]


In this example, the coders specify that incoming messages over the DMaaP endpoint reference APPC-LCM-WRITE, that have a field called type under the root JSON object with value response are allowed into the controller application. In this case, the incoming message is converted into an object (fact) of type org.onap.policy.appclcm.AppcLcmDmaapWrapper. The coder has attached a custom implementation provided by the application with class org.onap.policy.appclcm.util.Serialization. Note that the coder filter is expressed in JSONPath notation.

Note that not all the communication endpoint references need to be explicitly referenced within the controller configuration file. For example, Http clients do not. The reasons are historical, as the PDP-D was initially intended to only communicate through messaging-based protocols such as UEB or DMaaP in asynchronous unidirectional mode. The introduction of Http with synchronous bi-directional communication with remote endpoints made it more convenient for the application to manage each network exchange.

Controllers configuration resides in the $POLICY_HOME/config (or $POLICY_CONFIG) directory in a container.

Other Configuration Files

There are other types of configuration files that controllers can use, for example .environment files that provides a means to share data across applications. The controlloop.properties.environment is one such example.

Tosca Policies

PDP-D supports Tosca Policies through the feature-lifecycle. The PDP-D receives its policy set from the PAP. A policy conforms to its Policy Type specification. Policy Types and policy creation is done by the API component. Policy deployments are orchestrated by the PAP.

All communication between PAP and PDP-D is over the DMaaP POLICY-PDP-PAP topic.

Native Policy Types

The PDP-D Engine supports two (native) Tosca policy types by means of the lifecycle feature:

  • onap.policies.native.drools.Controller

  • onap.policies.native.drools.Artifact

These types can be used to dynamically deploy or undeploy application controllers, assign policy types, and upgrade or downgrade their attached maven artifact versions.

For instance, an example native controller policy is shown below.

    "tosca_definitions_version": "tosca_simple_yaml_1_0_0",
    "topology_template": {
        "policies": [
                "example.controller": {
                    "type": "onap.policies.native.drools.Controller",
                    "type_version": "1.0.0",
                    "version": "1.0.0",
                    "name": "example.controller",
                    "metadata": {
                        "policy-id": "example.controller"
                    "properties": {
                        "controllerName": "lifecycle",
                        "sourceTopics": [
                                "topicName": "DCAE_TOPIC",
                                "events": [
                                        "eventClass": "java.util.HashMap",
                                        "eventFilter": "[?($.closedLoopEventStatus == 'ONSET')]"
                                        "eventClass": "java.util.HashMap",
                                        "eventFilter": "[?($.closedLoopEventStatus == 'ABATED')]"
                        "sinkTopics": [
                                "topicName": "APPC-CL",
                                "events": [
                                        "eventClass": "java.util.HashMap",
                                        "eventFilter": "[?($.CommonHeader && $.Status)]"
                        "customConfig": {
                            "field1" : "value1"

The actual application coordinates are provided with a policy of type onap.policies.native.drools.Artifact, see the example native artifact

    "tosca_definitions_version": "tosca_simple_yaml_1_0_0",
    "topology_template": {
        "policies": [
                "example.artifact": {
                    "type": "onap.policies.native.drools.Artifact",
                    "type_version": "1.0.0",
                    "version": "1.0.0",
                    "name": "example.artifact",
                    "metadata": {
                        "policy-id": "example.artifact"
                    "properties": {
                        "rulesArtifact": {
                            "groupId": "org.onap.policy.drools.test",
                            "artifactId": "lifecycle",
                            "version": "1.0.0"
                        "controller": {
                            "name": "lifecycle"

Operational Policy Types

The PDP-D also recognizes Tosca Operational Policies, although it needs an application controller that understands them to execute them. These are:

  • onap.policies.controlloop.operational.common.Drools

A minimum of one application controller that supports these capabilities must be installed in order to honor the operational policy types. One such controller is the usecases controller residing in the policy/drools-applications repository.

Controller Policy Type Support

Note that a controller may support other policy types. A controller may declare them explicitly in a native onap.policies.native.drools.Controller policy.

"customConfig": {
    "controller.policy.types" : "policy.type.A"

The controller application could declare its supported policy types in the kjar. For example, the usecases controller packages this information in the kmodule.xml. One advantage of this approach is that the PDP-D would only commit to execute policies against these policy types if a supporting controller is up and running.

<kmodule xmlns="http://jboss.org/kie/6.0.0/kmodule">
    <kbase name="onap.policies.controlloop.operational.common.Drools" default="false" equalsBehavior="equality"/>
    <kbase name="onap.policies.controlloop.Operational" equalsBehavior="equality"
           packages="org.onap.policy.controlloop" includes="onap.policies.controlloop.operational.common.Drools">
        <ksession name="usecases"/>

Software Architecture

PDP-D is divided into 2 layers:

Core Layer

The core layer directly interfaces with the drools libraries with 2 main abstractions:

Policy Container and Sessions

The PolicyContainer abstracts the drools KieContainer, while a PolicySession abstracts a drools KieSession. PDP-D uses stateful sessions in active mode (fireUntilHalt) (please visit the drools website for additional documentation).

Management Layer

The management layer manages the PDP-D and builds on top of the core capabilities.


The PDP-D PolicyEngine is the top abstraction and abstracts away the PDP-D and all the resources it holds. The reader looking at the source code can start looking at this component in a top-down fashion. Note that the PolicyEngine abstraction should not be confused with the sofware in the policy/engine repository, there is no relationship whatsoever other than in the naming.

The PolicyEngine represents the PDP-D, holds all PDP-D resources, and orchestrates activities among those.

The PolicyEngine manages applications via the PolicyController abstractions in the base code. The relationship between the PolicyEngine and PolicyController is one to many.

The PolicyEngine holds other global resources such as a thread pool, policies validator, telemetry server, and unmanaged topics for administration purposes.

The PolicyEngine has interception points that allow *features* to observe and alter the default PolicyEngine behavior.

The PolicyEngine implements the *Startable* and *Lockable* interfaces. These operations have a cascading effect on the resources the PolicyEngine holds, as it is the top level entity, thus affecting controllers and endpoints. These capabilities are intended to be used for extensions, for example active/standby multi-node capabilities. This programmability is exposed via the telemetry API, and feature hooks.


PolicyEngine related configuration is located in the engine.properties, and engine-system.properties.

The engine configuration files reside in the $POLICY_CONFIG directory.


A PolicyController represents an application. Each PolicyController has an instance of a DroolsController. The PolicyController provides the means to group application specific resources into a single unit. Such resources include the application’s maven coordinates, endpoint references, and coders.

A PolicyController uses a DroolsController to interface with the core layer (PolicyContainer and PolicySession).

The relationship between the PolicyController and the DroolsController is one-to-one. The DroolsController currently supports 2 implementations, the MavenDroolsController, and the NullDroolsController. The DroolsController’s polymorphic behavior depends on whether a maven artifact is attached to the controller or not.


The controllers configuration resides in the $POLICY_CONFIG directory.


PDP-D is programmable through:

  • Features and Event Listeners.

  • Maven-Drools applications.

Using Features and Listeners

Features hook into the interception points provided by the the PDP-D main entities.

Endpoint Listeners, see here and here, can be used in conjuction with features for additional capabilities.

Using Maven-Drools applications

Maven-based drools applications can run any arbitrary functionality structured with rules and java logic.

Telemetry Extensions

It is recommended to features (extensions) to offer a diagnostics REST API to integrate with the telemetry API. This is done by placing JAX-RS files under the package org.onap.policy.drools.server.restful. The root context path for all the telemetry services is /policy/pdp/engine.


Features is an extension mechanism for the PDP-D functionality. Features can be toggled on and off. A feature is composed of:

  • Java libraries.

  • Scripts and configuration files.

Java Extensions

Additional functionality can be provided in the form of java libraries that hook into the PolicyEngine, PolicyController, DroolsController, and PolicySession interception points to observe or alter the PDP-D logic.

See the Feature APIs available in the management and core layers.

The convention used for naming these extension modules are api-<name> for interfaces, and feature-<name> for the actual java extensions.

Configuration Items

Installation items such as scripts, SQL, maven artifacts, and configuration files.

The reader can refer to the policy/drools-pdp repository and the <https://git.onap.org/policy/drools-applications>`__ repository for miscellaneous feature implementations.


A feature is packaged in a feature-<name>.zip and has this internal layout:

# #######################################################################################
# Features Directory Layout:
#   L─ features/
#        L─ <feature-name>*/
#            L─ [config]/
#            |   L─ <config-file>+
#            L─ [bin]/
#            |   L─ <bin-file>+
#            L─ lib/
#            |   L─ [dependencies]/
#            |   |   L─ <dependent-jar>+
#            │   L─ feature/
#            │       L─ <feature-jar>
#            L─ [db]/
#            │   L─ <db-name>/+
#            │       L─ sql/
#            │           L─ <sql-scripts>*
#            L─ [artifacts]/
#                L─ <artifact>+
#            L─ [install]
#                L─ [enable]
#                L─ [disable]
#                L─ [other-directories-or-files]
# notes:  [] = optional , * = 0 or more , + = 1 or more
#   <feature-name> directory without "feature-" prefix.
#   [config]       feature configuration directory that contains all configuration
#                  needed for this features
#   [config]/<config-file>  preferably named with "feature-<feature-name>" prefix to
#                  precisely match it against the exact features, source code, and
#                  associated wiki page for configuration details.
#   [bin]       feature bin directory that contains helper scripts for this feature
#   [bin]/<executable-file>  preferably named with "feature-<feature-name>" prefix.
#   lib            jar libraries needed by this features
#   lib/[dependencies]  3rd party jar dependencies not provided by base installation
#                  of pdp-d that are necessary for <feature-name> to operate
#                  correctly.
#   lib/feature    the single feature jar that implements the feature.
#   [db]           database directory, if the feature contains sql.
#   [db]/<db-name> database to which underlying sql scripts should be applied.
#                  ideally, <db-name> = <feature-name> so it is easily to associate
#                  the db data with a feature itself.   In addition, since a feature is
#                  a somewhat independent isolated unit of functionality,the <db-name>
#                  database ideally isolates all its data.
#   [db]/<db-name>/sql  directory with all the sql scripts.
#   [db]/<db-name>/sql/<sql-scripts>  for this feature, sql
#                  upgrade scripts should be suffixed with ".upgrade.sql"
#                  and downgrade scripts should be suffixed with ".downgrade.sql"
#   [artifacts]    maven artifacts to be deployed in a maven repository.
#   [artifacts]/<artifact>  maven artifact with identifiable maven coordinates embedded
#                  in the artifact.
#   [install]      custom installation directory where custom enable or disable scripts
#                  and other free form data is included to be used for the enable and
#                  and disable scripts.
#   [install]/[enable] enable script executed when the enable operation is invoked in
#                  the feature.
#   [install]/[disable] disable script executed when the disable operation is invoked in
#                  the feature.
#   [install]/[other-directories-or-files] other executables, or data that can be used
#                  by the feature for any of its operations.   The content is determined
#                  by the feature designer.
# ########################################################################################

The features is the tool used for administration purposes:

Usage:  features status
            Get enabled/disabled status on all features
        features enable <feature> ...
            Enable the specified feature
        features disable <feature> ...
            Disable the specified feature
        features install [ <feature> | <file-name> ] ...
            Install the specified feature
        features uninstall <feature> ...
            Uninstall the specified feature

Features available in the Docker image

The only enabled feature in the onap/policy-drools image is:

  • lifecycle: enables the lifecycle capability to integrate with the Policy Framework components.

The following features are included in the image but disabled.

  • distributed locking: distributed resource locking.

  • healthcheck: basic PDP-D Engine healthcheck.


The Healthcheck feature provides reports used to verify the health of PolicyEngine.manager in addition to the construction, operation, and deconstruction of HTTP server/client objects.

When enabled, the feature takes as input a properties file named “feature-healtcheck.properties. This file should contain configuration properties necessary for the construction of HTTP client and server objects.

Upon initialization, the feature first constructs HTTP server and client objects using the properties from its properties file. A healthCheck operation is then triggered. The logic of the healthCheck verifies that PolicyEngine.manager is alive, and iteratively tests each HTTP server object by sending HTTP GET requests using its respective client object. If a server returns a “200 OK” message, it is marked as “healthy” in its individual report. Any other return code results in an “unhealthy” report.

After the testing of the server objects has completed, the feature returns a single consolidated report.


The “lifecycle” feature enables a PDP-D to work with the architectural framework introduced in the Dublin release.

The lifecycle feature maintains three states: TERMINATED, PASSIVE, and ACTIVE. The PAP interacts with the lifecycle feature to put a PDP-D in PASSIVE or ACTIVE states. The PASSIVE state allows for Tosca Operational policies to be deployed. Policy execution is enabled when the PDP-D transitions to the ACTIVE state.

This feature can coexist side by side with the legacy mode of operation that pre-dates the Dublin release.

Distributed Locking

The Distributed Locking Feature provides locking of resources across a pool of PDP-D hosts. The list of locks is maintained in a database, where each record includes a resource identifier, an owner identifier, and an expiration time. Typically, a drools application will unlock the resource when it’s operation completes. However, if it fails to do so, then the resource will be automatically released when the lock expires, thus preventing a resource from becoming permanently locked.

Other features

The following features have been contributed to the policy/drools-pdp but are either unnecessary or have not been thoroughly tested:

Data Migration

PDP-D data is migrated across releases with the db-migrator.

The migration occurs when different release data is detected. db-migrator will look under the $POLICY_HOME/etc/db/migration for databases and SQL scripts to migrate.


where <sql-file> is of the form:


The db-migrator tool syntax is

syntax: db-migrator
     -s <schema-name>
     [-b <migration-dir>]
     [-f <from-version>]
     [-t <target-version>]
     -o <operations>

     where <operations>=upgrade|downgrade|auto|version|erase|report

Configuration Options:
     -s|--schema|--database:  schema to operate on ('ALL' to apply on all)
     -b|--basedir: overrides base DB migration directory
     -f|--from: overrides current release version for operations
     -t|--target: overrides target release to upgrade/downgrade

     upgrade: upgrade operation
     downgrade: performs a downgrade operation
     auto: autonomous operation, determines upgrade or downgrade
     version: returns current version, and in conjunction if '-f' sets the current version
     erase: erase all data related <schema> (use with care)
     report: migration detailed report on an schema
     ok: is the migration status valid

See the feature-distributed-locking sql directory for an example of upgrade/downgrade scripts.

The following command will provide a report on the upgrade or downgrade activies:

db-migrator -s ALL -o report

For example in the official guilin delivery:

policy@dev-drools-0:/tmp/policy-install$ db-migrator -s ALL -o report
| name    | version |
| pooling | 1811    |
| script                              | operation | success | atTime              |
| 1804-distributedlocking.upgrade.sql | upgrade   | 1       | 2020-05-22 19:33:09 |
| 1811-distributedlocking.upgrade.sql | upgrade   | 1       | 2020-05-22 19:33:09 |

In order to use the db-migrator tool, the system must be configured with a database.


Maven Repositories

The drools libraries in the PDP-D uses maven to fetch rules artifacts and software dependencies.

The default settings.xml file specifies the repositories to search. This configuration can be overriden with a custom copy that would sit in a mounted configuration directory. See an example of the OOM override settings.xml.

The default ONAP installation of the control loop child image onap/policy-pdpd-cl:1.6.4 is OFFLINE. In this configuration, the rules artifact and the dependencies retrieves all the artifacts from the local maven repository. Of course, this requires that the maven dependencies are preloaded in the local repository for it to work.

An offline configuration requires two items:

  • OFFLINE environment variable set to true.

  • override settings.xml customization, see settings.xml.

The default mode in the onap/policy-drools:1.6.3 is ONLINE instead.

In ONLINE mode, the controller initialization can take a significant amount of time.

The Policy ONAP installation includes a nexus repository component that can be used to host any arbitrary artifacts that an PDP-D application may require. The following environment variables configure its location:


The deploy-artifact tool is used to deploy artifacts to the local or remote maven repositories. It also allows for dependencies to be installed locally. The features tool invokes it when artifacts are to be deployed as part of a feature. The tool can be useful for developers to test a new application in a container.

syntax: deploy-artifact
     -s <custom-settings>
     -a <artifact>

     -f|--file-repo: deploy in the file repository
     -l|--local-repo: install in the local repository
     -d|--dependencies: install dependencies in the local repository
     -s|--settings: custom settings.xml
     -a|--artifact: file artifact (jar or pom) to deploy and/or install


Policy can talk to AAF for authorization requests. To enable AAF set the following environment variables:


By default AAF is disabled.

Policy Tool

The policy tool can be used to stop, start, and provide status on the PDP-D.

syntax: policy [--debug] status|start|stop

The status option provides generic status of the system.

 L []: Policy Management (pid 408) is running
    0 cron jobs installed.

name                   version         status
----                   -------         ------
healthcheck            1.6.3           enabled
distributed-locking    1.6.3           enabled
lifecycle              1.6.3           enabled
controlloop-management 1.6.4           enabled
controlloop-utils      1.6.4           enabled
controlloop-trans      1.6.4           enabled
controlloop-usecases   1.6.4           enabled

pooling: OK @ 1811

It contains 3 sections:

  • PDP-D running status

  • features applied

  • Data migration status on a per database basis.

The start and stop commands are useful for developers testing functionality on a docker container instance.

Telemetry Shell

PDP-D offers an ample set of REST APIs to debug, introspect, and change state on a running PDP-D. This is known as the telemetry API. The telemetry shell wraps these APIs for shell-like access using http-prompt.

policy@dev-drools-0:~$ telemetry
Version: 1.0.0
https://localhost:9696/policy/pdp/engine> get controllers
HTTP/1.1 200 OK
Content-Length: 13
Content-Type: application/json
Date: Thu, 04 Jun 2020 01:07:38 GMT
Server: Jetty(9.4.24.v20191120)


https://localhost:9696/policy/pdp/engine> exit

Other tools

Refer to the $POLICY_HOME/bin/ directory for additional tooling.

PDP-D Docker Container Configuration

Both the PDP-D onap/policy-drools and onap/policy-pdpd-cl images can be used without other components.

There are 2 types of configuration data provided to the container:

  1. environment variables.

  2. configuration files and shell scripts.

Environment variables

As it was shown in the controller and endpoint sections, PDP-D configuration can rely on environment variables. In a container environment, these variables are set up by the user in the host environment.

Configuration Files and Shell Scripts

PDP-D is very flexible in its configuration.

The following file types are recognized when mounted under /tmp/policy-install/config.

These are the configuration items that can reside externally and override the default configuration:

  • settings.xml if working with external nexus repositories.

  • standalone-settings.xml if an external policy nexus repository is not available.

  • *.conf files containing environment variables. This is an alternative to use environment variables, as these files will be sourced in before the PDP-D starts.

  • features*.zip to load any arbitrary feature not present in the image.

  • *.pre.sh scripts that will be executed before the PDP-D starts.

  • *.post.sh scripts that will be executed after the PDP-D starts.

  • policy-keystore to override the default PDP-D java keystore.

  • policy-truststore to override the default PDP-D java truststore.

  • aaf-cadi.keyfile to override the default AAF CADI Key generated by AAF.

  • *.properties to override or add any properties file for the PDP-D, this includes controller, endpoint, engine or system configurations.

  • logback*.xml to override the default logging configuration.

  • *.xml to override other .xml configuration that may be used for example by an application.

  • *.json json configuration that may be used by an application.

Running PDP-D with a single container

Environment File

First create an environment file (in this example env.conf) to configure the PDP-D.

# SYSTEM software configuration


# Telemetry credentials


# nexus repository


# Relational (SQL) DB access




# PDP-D DMaaP configuration channel


# PAP-PDP configuration channel


# DMaaP


Note that SQL_HOST, and REPOSITORY are empty, so the PDP-D does not attempt to integrate with those components.


In order to avoid the noise in the logs that relate to dmaap configuration, a startup script (noop.pre.sh) is added to convert dmaap endpoints to noop in the host directory to be mounted.


#!/bin/bash -x

sed -i "s/^dmaap/noop/g" $POLICY_HOME/config/*.properties


To put the controller directly in active mode at initialization, place an active.post.sh script under the mounted host directory:

#!/bin/bash -x

bash -c "http --verify=no -a ${TELEMETRY_USER}:${TELEMETRY_PASSWORD} PUT https://localhost:9696/policy/pdp/engine/lifecycle/state/ACTIVE"

Bring up the PDP-D

docker run --rm -p 9696:9696 -v ${PWD}/config:/tmp/policy-install/config --env-file ${PWD}/env/env.conf -it --name PDPD -h pdpd nexus3.onap.org:10001/onap/policy-drools:1.6.3

To run the container in detached mode, add the -d flag.

Note that in this command, we are opening the 9696 telemetry API port to the outside world, the config directory (where the noop.pre.sh customization script resides) is mounted as /tmp/policy-install/config, and the customization environment variables (env/env.conf) are passed into the container.

To open a shell into the PDP-D:

docker exec -it pdp-d bash

Once in the container, run tools such as telemetry, db-migrator, policy to look at the system state:

To run the telemetry shell and other tools from the host:

docker exec -it PDPD bash -c "/opt/app/policy/bin/telemetry"
docker exec -it PDPD bash -c "/opt/app/policy/bin/policy status"
docker exec -it PDPD bash -c "/opt/app/policy/bin/db-migrator -s ALL -o report"

Controlled instantiation of the PDP-D

Sometimes a developer may want to start and stop the PDP-D manually:

# start a bash

docker run --rm -p 9696:9696 -v ${PWD}/config:/tmp/policy-install/config --env-file ${PWD}/env/env.conf -it --name PDPD -h pdpd nexus3.onap.org:10001/onap/policy-drools:1.6.3 bash

# use this command to start policy applying host customizations from /tmp/policy-install/config

pdpd-entrypoint.sh vmboot

# or use this command to start policy without host customization

policy start

# at any time use the following command to stop the PDP-D

policy stop

# and this command to start the PDP-D back again

policy start

Running PDP-D with nexus and mariadb

docker-compose can be used to test the PDP-D with other components. This is an example configuration that brings up nexus, mariadb and the PDP-D (docker-compose-pdp.yml)


version: '3'
      image: mariadb:10.2.25
      container_name: mariadb
      hostname: mariadb
      command: ['--lower-case-table-names=1', '--wait_timeout=28800']
         - ${PWD}/db/db.conf
         - ${PWD}/db:/docker-entrypoint-initdb.d
         - "3306:3306"
      image: sonatype/nexus:2.14.8-01
      container_name: nexus
      hostname: nexus
         - "8081:8081"
      image: nexus3.onap.org:10001/onap/policy-drools:1.6.3
      container_name: drools
         - mariadb
         - nexus
      hostname: drools
         - "9696:9696"
         - ${PWD}/config:/tmp/policy-install/config
         - ${PWD}/env/env.conf

with ${PWD}/db/db.conf:



and ${PWD}/db/db.sh:


for db in support onap_sdk log migration operationshistory10 pooling policyadmin operationshistory
    mysql -uroot -p"${MYSQL_ROOT_PASSWORD}" --execute "CREATE DATABASE IF NOT EXISTS ${db};"
    mysql -uroot -p"${MYSQL_ROOT_PASSWORD}" --execute "GRANT ALL PRIVILEGES ON \`${db}\`.* TO '${MYSQL_USER}'@'%' ;"

mysql -uroot -p"${MYSQL_ROOT_PASSWORD}" --execute "FLUSH PRIVILEGES;"


The environment file env/env.conf for PDP-D can be set up with appropriate variables to point to the nexus instance and the mariadb database:

# SYSTEM software configuration


# Telemetry credentials


# nexus repository



# Relational (SQL) DB access




# PDP-D DMaaP configuration channel


# PAP-PDP configuration channel


# DMaaP



A pre-start script config/prepare.pres.sh”can be added the custom config directory to prepare the PDP-D to activate the distributed-locking feature (using the database) and to use “noop” topics instead of *dmaap topics:


bash -c "/opt/app/policy/bin/features enable distributed-locking"
sed -i "s/^dmaap/noop/g" $POLICY_HOME/config/*.properties


A post-start script config/active.post.sh can place PDP-D in active mode at initialization:

bash -c “http –verify=no -a ${TELEMETRY_USER}:${TELEMETRY_PASSWORD} PUT <http|https>://localhost:9696/policy/pdp/engine/lifecycle/state/ACTIVE”

Bring up the PDP-D, nexus, and mariadb

To bring up the containers:

docker-compose -f docker-compose-pdpd.yaml up -d

To take it down:

docker-compose -f docker-compose-pdpd.yaml down -v

Other examples

The reader can also look at the policy/docker repository. More specifically, these directories have examples of other PDP-D configurations:

Configuring the PDP-D in an OOM Kubernetes installation

The PDP-D OOM chart can be customized at the following locations:

  • values.yaml: custom values for your installation.

  • configmaps: place in this directory any configuration extensions or overrides to customize the PDP-D that does not contain sensitive information.

  • secrets: place in this directory any configuration extensions or overrides to customize the PDP-D that does contain sensitive information.

The same customization techniques described in the docker sections for PDP-D, fully apply here, by placing the corresponding files or scripts in these two directories.

Additional information

For additional information, please see the Drools PDP Development and Testing (In Depth) page.