Process Instance Migration Service

This service is intended to be deployed as a standalone application. It must be integrated with one or more KIE Process Server instances to be able to execute migrations remotely.

The two entities that will be used are Migration Plans and Migrations.

Migration Plan

Is a definition of how the migration will be performed. Includes the following information:

• Plan ID (Generated)
• Plan name
• Plan description
• Source and target container IDs
• Node mappings (if any)

Migration

Is the execution of a defined plan, applied to a set of process instances. These are the attributes that define a Migration:

• Migration ID (Generated)
• Plan ID
• Instances to migrate.
• Execution type
  • Sync or Async
  • Scheduled start time
  • Callback URL

Requirements

• JRE 11+
• Running KIE Process server

Requirements for testing

PIM Integration Tests requires an external Kie Server instance, for that you can provide your own instance or use TestContainers. TestContainers on Podman requires to connect through the external API so make sure you can connect remotely to Podman.

# Use a different time but bear in mind the default is 5 seconds
$ podman system service --time=0 &

# Confirm it's working 
$ podman-remote info
...
  os: linux
  remoteSocket:
    exists: true
    path: /run/user/1000/podman/podman.sock
...

Build

It is a Quarkus 2.x application.

JVM

mvn clean package

Native

Note: Native compilation is currently not supported.

mvn clean package -Pnative -Dquarkus.native.container-build=true

Run application

JVM

java -jar target/quarkus-app/quarkus-run.jar

Native

./target/process-migration-runner

You can provide your custom configuration file. Check application.yaml to see an example. The provided configuration will be added or override the one existing in application.yaml

Container image

If you want to create a container image using your JVM or Native build use the -Dquarkus.container-image.build=true flag when packaging your application and pass in any other configuration property you want to override. See the Container Image Guide.

$ mvn clean package -Dquarkus.container-image.build=true \
 -Dquarkus.container-image.push=true \
-Dquarkus.container-image.image=quay.io/kiegroup/process-migration-service:1.0
...
[INFO] [io.quarkus.container.image.jib.deployment.JibProcessor] Created container image quay.io/kiegroup/process-migration-service:1.0 (sha256:dc8028963923081941529ef0ea515bd1e970b8fc96d5ad6a9346eb4ad61028f6)

Or you can just use your favourite container builder tool and refer to any of the existing Dockerfiles

$ podman build -t quay.io/kiegroup/process-migration-service:1.0 -f ./src/main/docker/Dockerfile.jvm .

Configuration

Default configuration is as follows:

quarkus:
  class-loading:
    removed-artifacts: com.oracle.database.jdbc:ojdbc8,com.ibm.db2:jcc,com.microsoft.sqlserver:mssql-jdbc
  package:
    type: mutable-jar
    user-providers-directory: providers (1)
  http:
    auth:
      basic: true
      policy:
        main-policy:
          roles-allowed: admin
      permission:
        main:
          paths: /*
          policy: main-policy
        public:
          paths: /q/health/*
          policy: permit
          methods: GET
  security:
    users:
      file:
        realm-name: pim_file
        enabled: true
        plain-text: true
        users: users.properties
        roles: roles.properties
    jdbc:
      realm-name: pim_jdbc
      enabled: true
      principal-query:
        sql: SELECT u.password, u.role FROM users u WHERE u.username=?
    ldap:
      realm-name: pim_ldap
      dir-context:
        url: ldap://override-when-needed
      identity-mapping:
        search-base-dn: ou=users,o=YourCompany,c=ES
# Flyway to create PIM schema
  flyway:
    connect-retries: 10
    table: flyway_pim_history
    migrate-at-start: true (2)
    baseline-on-migrate: true
    baseline-version: 1.0
    baseline-description: PimDB      
# Quartz configuration
  quartz:
    store-type: jdbc-cmt
    start-mode: forced
  resteasy:
    path: /rest (3)
  datasource:
    db-kind: h2 (4)
    jdbc:
      url: jdbc:h2:mem:test;DB_CLOSE_DELAY=-1;DB_CLOSE_ON_EXIT=FALSE
    username: sa
    password: sa
  hibernate-orm:
    database:
      generation: update
pim:
  auth-method: file (5)

1. In case we wanted to enable a different database requiring a license agreement, the JDBC driver will have to be located in this providers folder manually.
2. Out of the box, Flyway will automatically create PIM database schema for H2 in-memory database. Disable it (-Dquarkus.flyway.migrate-at-start=false) when you do not want PIM to handle the schema creation.
3. Deploy the application on /rest
4. H2 in-memory datasource. Override it by one of your choice
5. Authentication method. Defaults to file but jdbc or ldap are also valid options

Configuration overrides

Check Quarkus Configuration Reference Guide for further details.

It is possible to override or extend the provided configuration. You can provide one or more additional configuration files that will allow you to customize the application. Several examples are provided in the examples folder.

As an example, if you want to replace the H2 default persistence configuration by MariaDB and the authentication mechanism to use LDAP, see sample config files below.

• MariaDB database configuration file
• LDAP configuration file

Note: These files will override or extend the already defined properties in the application.yaml file

Using Keystore Vault

This feature uses the Quarkiverse File Vault extension. By default, the Credentials Provider interface supports some credential consumer extensions like agroal, used for database configuration. In Process Instance Migration we have added support for KIE Server basic authentication and client certificate definition.

Usage

Add passwords to the keystore that will be used as Vault

keytool -importpass -alias pimdb -keystore pimvault.p12 -storepass password -storetype PKCS12 
keytool -importpass -alias kieserver -keystore pimvault.p12 -storepass password -storetype PKCS12
keytool -importpass -alias cert -keystore pimvault.p12 -storepass password -storetype PKCS12
keytool -importpass -alias keystore -keystore pimvault.p12 -storepass password -storetype PKCS12
keytool -importpass -alias truststore -keystore pimvault.p12 -storepass password -storetype PKCS12

Configure the vault provider to use the keystore we just created

quarkus:
  file:
    vault:
      provider:
        pim:
          path: pimvault.p12
          secret: ${vault.storepassword} # This will be provided as a property

Configure your application to use the credentials from the vault

quarkus:
  datasource:
    credentials-provider: quarkus.file.vault.provider.pim.pimdb
kieservers:
  - host: http://localhost:18080/kie-server/services/rest/server
    credentials-provider: quarkus.file.vault.provider.pim.kieserver

Defining KIE Servers

The right way to configure the connection to one or more KIE Servers in order to perform the migrations, a list of KIE servers should exist in the configuration file.

Basic Authentication

Example

kieservers:
  - host: http://kieserver1.example.com:8080/kie-server/services/rest/server
    username: joe
    password: secret
  - host: http://kieserver2.example.com:8080/kie-server/services/rest/server
    username: jim
    password: secret

Token Based Authentication

Example

kieservers:
- host: http://localhost:18080/kie-server/services/rest/server
  token: __REDACTED__

Credentials Provider

Don't forget to configure the Vault. See Configuring a File Vault. If the credentials-provider property is used the other properties will be ignored. The vault key will be the username and the value will be the password.

Example

kieservers:
  - host: http://localhost:18080/kie-server/services/rest/server
    credentials-provider: quarkus.file.vault.provider.pim.kieserver

Mutual TLS Authentication

Example

kieservers:
  - host: http://localhost:18080/kie-server/services/rest/server
# HTTP Basic or Token Authentication can still be used, but usually a client identity is sufficient.
#   username: foo
#   password: bar
client-cert:
  cert-name: pim
  cert-password: pim-secret
  keystore-path: /path/to/keystore.jks
  keystore-password: ks-secret
  truststore-path: /path/to/truststore.jks
  truststore-password: ts-secret

Or using vault

Example

kieservers:
  - host: http://localhost:18080/kie-server/services/rest/server
client-cert:
  cert-name: pim
  cert-credentials-provider: quarkus.file.vault.provider.pim.cert
  keystore-path: /path/to/keystore.jks
  keystore-credentials-provider: quarkus.file.vault.provider.pim.keystore
  truststore-path: /path/to/truststore.jks
  truststore-credentials-provider: quarkus.file.vault.provider.pim.truststore

MariaDB Datasource

See Using other JDBC extensions for details on how to include additional JDBC drivers to the runtime.

quarkus:
  datasource:
    db-kind: mariadb
    jdbc:
      url: jdbc:mariadb://localhost:3306/pimdb
    username: pim
    password: pim123

Refer to the Quarkus Datasource configuration for further details

Basic authentication and authorization

Authorization example available here. Shows how to define basic authorization.

Note that this is a build time configuration (See Quarkus Authorization)

  http:
    auth:
      basic: true
      policy:
        main-policy:
          roles-allowed: admin
      permission:
        main:
          paths: /*
          policy: main-policy
        public:
          paths: /q/health/*
          policy: permit
          methods: GET

1. Authentication type BASIC
2. Every resource under root path requires the admin role
3. Health checks are not secured

By default, the user properties IdentityProvider is provided in plain text:

Note that this is a build time configuration (See Quarkus Security)

  security:
    users:
      file:
        enabled: true
        plain-text: true
        users: users.properties
        roles: roles.properties

Configuring Quartz

By default, Quartz jobs are stored in-memory database through connections managed by the container, but it is possible to configure Quartz to persist the jobs in a different way either using a user transaction based configuration or the in-memory database. This can be achieved by setting the quarkus.quartz.store-type property to ram or jdbc-tx. See the examples and the Quarkus Quartz documentation.

Using other JDBC extensions

The H2 JDBC extension is set by default. However, users will be able to use different JDBC extensions to connect to any supported database. For that purpose you will have to re-augment the base build to include the right build properties.

For instance, the below command will add a PostgreSQL database extension by re-augmenting the application.

java -jar -Dquarkus.launch.rebuild=true -Dquarkus.datasource.db-kind=postgresql target/quarkus-app/quarkus-run.jar

Note that in case of using a database requiring a JDBC driver license agreement (such as Oracle, DB2 and MS SQL Server), we will need to drop the JDBC drivers in the providers folder manually before re-augmenting the Quarkus application.

Afterwards, you can start up the application normally.

java -jar -Dquarkus.flyway.migrate-at-start=false -Dquarkus.datasource.username=jbpm -Dquarkus.datasource.password=jbpm -Dquarkus.datasource.jdbc.url=jdbc:postgresql://localhost:5432/jbpm?pinGlobalTxToPhysicalConnection=true target/quarkus-app/quarkus-run.jar

Reference:

• Quarkus Datasources
• Quarkus Re-augmentation
• Quarkus Flyway

Disabling database schema auto-creation

By default, Process Migration service generates an in-memory database schema automatically. You should be able to disable it and manage the schema creation by setting and passing the quarkus.flyway.migrate-at-start=false system property.

java -jar -Dquarkus.flyway.migrate-at-start=false target/quarkus-app/quarkus-run.jar

How to change build time properties in a mutable-jar

By default, the Process Migration Service is built as a mutable-jar and configured to fail if any built-time property is changed at runtime. If you want to change a build-time property it is required that you re-augment the build.

java -jar -Dquarkus.launch.rebuild=true -Dquarkus.datasource.db-kind=mariadb target/quarkus-app/quarkus-run.jar

Usage

Define the plan (without node mappings)

Request:

URL: http://localhost:8080/rest/plans
Method: POST
HTTP Headers:
  Content-Type: application/json
  Authorization: Basic a2VybWl0OnRoZWZyb2c=
Body:
{
    "name": "Test plan",
    "description": "Evaluation Process Test Plan",
    "source": {
      "containerId": "evaluation_1.0",
      "processId": "evaluation"
    },
    "target": {
      "containerId": "evaluation_1.1",
      "processId": "evaluation"
    }
}

Response:

Status: 201 CREATED
HTTP Headers:
  Content-Type: application/json
Body:
{
    "id": 1,
    "name": "Test plan",
    "description": "Evaluation Process Test Plan",
    "source": {
      "containerId": "evaluation_1.0",
      "processId": "evaluation"
    },
    "target": {
      "containerId": "evaluation_1.1",
      "processId": "evaluation"
    }
}

Deploy some processes to test the migration

1. Start a KIE Server
2. Deploy two versions of the evaluation project (evaluation_1.0 and evaluation_1.1)
3. Start one instance of the evaluation process (evaluation_1.0)

Create a sync migration

URL: http://localhost:8080/rest/migrations
Method: POST
HTTP Headers:
  Content-Type: application/json
  Authorization: Basic a2VybWl0OnRoZWZyb2c=
Body:
{
    "planId": 1,
    "processInstanceIds": [1],
    "kieserverId": "sample-server",
    "execution": {
      "type": "SYNC"
    }
}

Response:

Status: 201 CREATED
HTTP Headers:
  Content-Type: application/json
Body:
{
    "id": 1,
    "definition": {
        "planId": 1,
        "processInstanceIds": [1],
        "kieserverId": "sample-server",
        "requester": "kermit",
        "execution": {
            "type": "SYNC"
        }
    },
    "createdAt": "2018-11-29T13:47:07.839Z",
    "startedAt": "2018-11-29T13:47:07.839Z",
    "finishedAt": "2018-11-29T13:47:07.874Z",
    "status": "COMPLETED"
}

As it is a Synchronous migration, the result of the migration will be returned once it has finished.

Check the migration output

The following request will fetch the overall result of the migration

Request:

URL: http://localhost:8080/rest/migrations/1
Method: GET
HTTP Headers:
  Content-Type: application/json
  Authorization: Basic a2VybWl0OnRoZWZyb2c=

Response:

Status: 200 OK
HTTP Headers:
  Content-Type: application/json
Body:
{
    "id": 1,
    "definition": {
        "planId": 1,
        "processInstanceIds": [],
        "kieserverId": "sample-server",
        "requester": "kermit",
        "execution": {
            "type": "SYNC"
        }
    },
    "createdAt": "2018-11-27T14:28:58.918Z",
    "startedAt": "2018-11-27T14:28:59.861Z",
    "finishedAt": "2018-11-27T14:29:00.167Z",
    "status": "COMPLETED"
}

To retrieve the individual results of the migration of each process instance

Request:

URL: http://localhost:8080/rest/migrations/1/results
Method: GET
HTTP Headers:
  Content-Type: application/json
  Authorization: Basic a2VybWl0OnRoZWZyb2c=

Response:

Status: 200 OK
HTTP Headers:
  Content-Type: application/json
Body:
[
    {
        "id": 1,
        "migrationId": 3,
        "processInstanceId": 5,
        "startDate": "2018-12-18T11:16:26.779Z",
        "endDate": "2018-12-18T11:16:26.906Z",
        "successful": true,
        "logs": [
            "INFO Tue Dec 18 12:16:26 CET 2018 Variable instances updated = 1 for process instance id 5",
            "INFO Tue Dec 18 12:16:26 CET 2018 Node instances updated = 3 for process instance id 5",
            "INFO Tue Dec 18 12:16:26 CET 2018 Process instances updated = 1 for process instance id 5",
            "INFO Tue Dec 18 12:16:26 CET 2018 Task variables updated = 1 for process instance id 5",
            "INFO Tue Dec 18 12:16:26 CET 2018 Task audit updated = 1 for process instance id 5",
            "INFO Tue Dec 18 12:16:26 CET 2018 Tasks updated = 1 for process instance id 5",
            "INFO Tue Dec 18 12:16:26 CET 2018 Context info updated = 0 for process instance id 5",
            "INFO Tue Dec 18 12:16:26 CET 2018 Executor Jobs updated = 0 for process instance id 5",
            "WARN Tue Dec 18 12:16:26 CET 2018 Source and target process id is exactly the same (test.myprocess) it's recommended to use unique process ids",
            "INFO Tue Dec 18 12:16:26 CET 2018 Mapping: Node instance logs to be updated  = [1]",
            "INFO Tue Dec 18 12:16:26 CET 2018 Mapping: Node instance logs updated = 1 for node instance id 1",
            "INFO Tue Dec 18 12:16:26 CET 2018 Mapping: Task audit updated = 1 for task id 5",
            "INFO Tue Dec 18 12:16:26 CET 2018 Mapping: Task updated = 1 for task id 5",
            "INFO Tue Dec 18 12:16:26 CET 2018 Migration of process instance (5) completed successfully to process test.myprocess"
        ]
    },
    {
        "id": 2,
        "migrationId": 3,
        "processInstanceId": 6,
        "startDate": "2018-12-18T11:16:26.992Z",
        "endDate": "2018-12-18T11:16:27.039Z",
        "successful": true,
        "logs": [
            "INFO Tue Dec 18 12:16:27 CET 2018 Variable instances updated = 1 for process instance id 6",
            "INFO Tue Dec 18 12:16:27 CET 2018 Node instances updated = 3 for process instance id 6",
            "INFO Tue Dec 18 12:16:27 CET 2018 Process instances updated = 1 for process instance id 6",
            "INFO Tue Dec 18 12:16:27 CET 2018 Task variables updated = 1 for process instance id 6",
            "INFO Tue Dec 18 12:16:27 CET 2018 Task audit updated = 1 for process instance id 6",
            "INFO Tue Dec 18 12:16:27 CET 2018 Tasks updated = 1 for process instance id 6",
            "INFO Tue Dec 18 12:16:27 CET 2018 Context info updated = 0 for process instance id 6",
            "INFO Tue Dec 18 12:16:27 CET 2018 Executor Jobs updated = 0 for process instance id 6",
            "WARN Tue Dec 18 12:16:27 CET 2018 Source and target process id is exactly the same (test.myprocess) it's recommended to use unique process ids",
            "INFO Tue Dec 18 12:16:27 CET 2018 Mapping: Node instance logs to be updated  = [1]",
            "INFO Tue Dec 18 12:16:27 CET 2018 Mapping: Node instance logs updated = 1 for node instance id 1",
            "INFO Tue Dec 18 12:16:27 CET 2018 Mapping: Task audit updated = 1 for task id 6",
            "INFO Tue Dec 18 12:16:27 CET 2018 Mapping: Task updated = 1 for task id 6",
            "INFO Tue Dec 18 12:16:27 CET 2018 Migration of process instance (6) completed successfully to process test.myprocess"
        ]
    }
]

Create an Async migration

1. Start two more processes
2. Trigger the migration of all the existing active processes

Request:

URL: http://localhost:8080/rest/migrations
Method: POST
HTTP Headers:
  Content-Type: application/json
  Authorization: Basic a2VybWl0OnRoZWZyb2c=
Body:
{
    "planId": 1,
    "processInstanceIds": [],
    "kieserverId": "sample-server",
    "execution": {
      "type": "ASYNC",
      "scheduledStartTime": "2018-12-11T12:35:00.000Z"
    }
}

Response:

Status: 202 Accepted
HTTP Headers:
  Content-Type: application/json
Body:
{
    "id": 2,
    "definition": {
        "execution": {
            "type": "ASYNC",
            "scheduledStartTime": "2018-12-11T12:35:00.000Z"
        },
        "planId": 1,
        "processInstanceIds": [],
        "kieServerId": "sample-server",
    },
    "status": "SCHEDULED",
    "createdAt": "2018-11-07T11:28:43.828Z",
    "startedAt": null,
    "finishedAt": null
}

The migration status can be checked using the migrations api with the id returned as done before

User Interface

The Process Instance Migration User Interface can be accessed in the following URL

http://localhost:8080/