This documentation is for WSO2 Stream Processor 4.3.0. View documentation for the latest release.

All docs This doc
Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 2 Next »

Introduction

The most common deployment pattern for WSO2 SP is the Minimum High Availability Deployment that offers high availability with the minimum amount of resources. However, there are a few user scenarios where the HA (High Availability) deployment is not sufficient to handle the throughput. 

The Distributed Deployment pattern is supported so that a high volume of data can be distributed among multiple SP instances instead of having them accumulated at a single point. It is suitable to be used in scenarios where the volume of data handled is too high to be managed in a single SP instance or a minimum high availability deployment.

Distributed Siddhi applications

A Siddhi Application is a combination of multiple Siddhi executional elements. A Siddhi executional element can be a Siddhi Query or a Siddhi Partition. In distributed processing perspective, a collection of these execution elements is called an execution group. Execution group is the smallest unit of execution.

Distributed processing of a Siddhi application allows users to execute multiple instances of each execution group in-parallel in multiple SP instances.

Users can specify execution groups and the parallelism to execute them by annotating existing Siddhi applications. Following sample application is annotated in that manner.


@App:name('wso2-app')
 
@info(name = ‘query-1')
@dist(execGroup='group-1')
from TempStream#window.time(2 min)
select avg(temp) as avgTemp, roomNo, deviceID
insert all events into AvgTempStream;
  
@info(name = ‘query-2')
@dist(execGroup='group-1')
from every( e1=TempStream ) ->
e2=TempStream[e1.roomNo==roomNo and (e1.temp + 5) <= temp ] within 10 min
select e1.roomNo, e1.temp as initialTemp, e2.temp as finalTemp
insert into AlertStream;
  
@info(name = ‘query-3')
@dist(execGroup='group-2' ,parallel ='2')
from TempStream [(roomNo >= 100 and roomNo < 110) and temp > 40 ]
select roomNo, temp
insert into HighTempStream;

This sample disributed Siddhi application contains two execution groups named group-1 and group-2 (defined via execGroup='<GROUP_ID>' e.g., execGroup='group-1').   group-1 contains two queries named query-1 and query-2. group-2 contains query-3. No specific number of parallel instances are specified for group-1. Therefore, only one instance is created for it at runtime by default. Two parallel instances are specified for group-2.


The following is an illustration of how each parallel instance is created as a separate Siddhi application.



Each Siddhi application is deployed in the available resource nodes of the distributed cluster. All these Siddhi applications communicate with each other using Kafka topics. The system creates Kafka topics representing each stream and configures the Siddhi applications to use these topics as required.

For detailed information, see Converting to a Distributed Streaming Application.

Deployment architecture

WSO2 Stream Processor has a component named Dashboard in the User Interface and Dashboard layer. The Dashboard allows users to view the output of analytics in an interactive manner. It also conveys observability information the cluster, the status of the list of the applications (i.e., Siddhi applications) currently submitted, and the status of each Stream Processor node. The JVM metrics, as well as Siddhi application level metrics, can be viewed through this dashboard.


Job Manager nodes handle all the Management layer related functionalities. This layer contains two WSO2 SP Manager instances configured to run in high availability mode. Here, the Manager parses the distributed Siddhi application provided by the user, partitions it into multiple Siddhi applications, wires them using Kafka topics, and deploys them in the available worker nodes. Management layer also handles the effects of the worker nodes joining/leaving the distributed cluster by re-distributing the Siddhi applications accordingly.


The processing layer (also known as the resource cluster) is represented by multiple WSO2 SP Worker instances that are configured as workers. Each WSO2 SP worker instance in this layer registers itself to the Manager Cluster when it starts. These workers periodically send their heartbeats to the Manager Cluster. This allows the Managers to identify the active worker nodes and the inactive ones. The worker nodes (resource nodes) run the Siddhi applications assigned to them by their Manager nodes. In addition, they are also capable of handling network partitions in a graceful manner as depicted in the following diagram. 


As depicted above, a worker node periodically synchronizes its configurations and the Siddhi applications with the manager Node. If the network gets partitioned or if the manager becomes unreachable, it undeploys the applications deployed in it. By doing so, it allows the Siddhi applications to be rescheduled in other work nodes that are maintaining their connections with the manager nodes.

It is required to use Apache Kafka and Apache Zookeeper to configure a fully distributed SP cluster. As explained above, a Kafka cluster which is a part of the Persistence layer and the Transport layer holds all the topics that are used for communication by distributed Siddhi applications. Persistence stores of the Persistence layer can be RDBMS databases that store both configuration and system state data. Identity and access management of all the WSO2 Stream Processor nodes are handled by any OAuth Identity provider such as of the WSO2 Identity and Access Management(WSO2 IAM).

There are no restrictions to run WSO2 Stream Processor in the distributed mode on any environment. It can run in the distributed mode on bare metal, VMs, and containers. Here the manager nodes are grouped in a single cluster backed by a database for correlation. Similarly, dashboard nodes can also be deployed in a separate cluster. The worker nodes, on the other hand, are not aware of each other. They are synchronized with manager nodes from which they receive instructions. 

Manager cluster

The manager cluster contains two or more WSO2 SP instances configured to run in the high availability mode. The manager cluster is responsible for parsing a user-defined distributed Siddhi application, dividing it to multiple Siddhi applications, creating the required topics and then deploying them in the available resource nodes. The manager cluster also handles resource nodes that join/leave the distributed cluster, and re-schedules the Siddhi applications accordingly. Since manager nodes are deployed in a high availability mode, if and when the active manager node goes down, another node in the manager will be elected as the cluster to handle the resource cluster.

Resource cluster

A resource cluster contains multiple WSO2 SP instances. Each instance sends a periodic heartbeat to the manager cluster so that the managers at any given time can identify the resource nodes that are active in the cluster. The resource nodes are responsible for running Siddhi applications assigned to them by the manager nodes. A resource node continues to run its Siddhi applications until a manager node undeploys them, or until it is no longer able to reach a manager node to send its heartbeat. If a manager node is unreachable for a specified amount of time, the resource node stops operating, removes its deployed Siddhi applications and waits until it can reach a manager node again.

The resource cluster can include both receiver workers and resource workers. You can specify the minimum number of receiver worker nodes to be included. However, you need to ensure that the minimum number specified is greater than one. This is because, if one or more distributed Siddhi applications contain a user-defined source such as HTTP or Thrift, then that Siddhi application cannot be deployed in a resource worker node. Therefore, at least one receiver worker node needs to be available in the resource cluster to ensure that distributed Siddhi applications are successfully deployed.

Deployed Siddhi applications communicate among themselves via Kafka topics.

Kafka cluster

It is required to insrtall Kafka and Zookeeper to configure a fully distributed deployment.


A Kafka cluster holds all the topics used by distributed Siddhi applications. All communications between execution groups take place via Kafka.

Publishing and receiving data from distributed Siddhi applications can be done via Kafka or other Siddhi sources as follows:

  • Via Kafka
    To use Kafka for publishing and receiving data, you can either define a Kafka source in the initial distributed Siddhi application or use the Kafka source created by the distributed implementation. 
  • Via Other Siddhi Sources
    This invoves definingh the source in the initial distributed Siddhi application.

    At least one receiver worker node should be configured in the cluster.

Configuring a distributed cluster

This section explains how to configure a distributed WSO2 SP cluster.

Prerequisites

In order to configure a fully distributed HA cluster, the following prerequisites must be completed:

  • A WSO2 SP binary pack must be available for each node in the cluster.
  • Each SP node should have an ID under wso2.carbon in the <SP_HOME>/conf/manager/deployment.yaml or <SP_HOME>/conf/worker/deployment.yaml  file depending on the cluster node being configured.
  • A working RDBMS instance to be used for clustering of the manager nodes. 

    We currently support only MySQL. Support for other databases will be added soon.

  • The datasource to be shared by the nodes in the manager cluster must be already defined in the <SP_HOME>/conf/manager/deployment.yaml file.
  • For MySQL to work with the Stream Processor, download the MySQL connector from  here, extract and find the  mysql-connector-java-5.*.*-bin.jar.  Drop the jar to the  <SP_HOME>/lib  directory in both manager nodes.
  • In order to retrieve the state of the Siddhi Applications deployed in the system in case of a scenario where both the nodes fail, state persistence must be enabled for all worker nodes. For detailed instructions, see Configuring Database and File System State Persistence.
  • A Zookeeper cluster and Kafka cluster should already be started and hosts and ports should be known.
    • Supported Zookeeper version - 3.4.6
    • Supported Kafka version - 2.11-0.10.0.0
  • For all manager and resource nodes to communicate with the Kafka broker the following kafka libs found in <KAFKA_HOME>/libs should be converted to OSGI and added to <SP_HOME>/libs directory:
    • kafka_2.11-0.10.0.0.jar

    • kafka-clients-0.10.0.0.jar

    • metrics-core-2.2.0.jar

    • scala-parser-combinators_2.11-1.0.4.jar
    • scala-library-2.11.8.jar

    • zkclient-0.8.jar

    • zookeeper-3.4.6.jar

  • To convert Kafka libs to OSGI,

    • Create the source folder (eg: kafka) and copy the respective Kafka libs to the created folder.

    • Create another folder (eg: kafka-osgi) as the destination folder to which the converted OSGI bundles will be added.

    • Navigate to <SP_HOME>/bin and issue the following command
      - For Linux: ./jartobundle.sh <path_to_source_folder> <path_to_destination_folder>
      - For Windows: ./jartobundle.bat <path_to_source_folder> <path_to_destination_folder>

    • If converted successfully then for each lib, following messages would be shown on the terminal

    - INFO: Created the OSGi bundle <kafka-lib-name>.jar for JAR file <absolute_path>/kafka/<kafka-lib-name>.jar
    
    • You can find the osgi converted libs in the destination folder. Copy them to the <SP_HOME>/lib directory.


Configuring the cluster

To configure a fully distributed HA cluster, follow the procedure below:

Configure manager nodes

To configure a node as a manager node, update the <SP_HOME>/conf/manager/deployment.yaml file as follows. The fully distributed cluster can have one or more manager nodes. For more information on how to set up cluster coordination see Configuring Cluster Coordination.

  1. In the cluster.config section, make the following changes.
    1. To enable the cluster mode, set the enabled property to true.
    2. In order to cluster all the manager nodes together, enter the same cluster ID as the group ID for all the nodes (e.g., groupId: group-1).
    3. Enter the ID of the class that defines the coordination strategy for the cluster as shown in the example below.
      e.g., coordinationStrategyClass: org.wso2.carbon.cluster.coordinator.rdbms.RDBMSCoordinationStrategy
  2. In the strategyConfig section of cluster.config, enter information for the required parameters as follows.
    1. Enter the ID of the datasource shared by the nodes in the cluster as shown in the example below. Data handled by the cluster are persisted here.
      datasource: SP_MGT_DB

      The SP_MGT_DB datasource is configured to an h2 database by default. You must create a MySQL database and then configure this datasource in the <SP_HOME>/conf/manager/deployment.yaml file of the required manager. The following is a sample configuration.

        - name: SP_MGT_DB
          description: The MySQL datasource used for Cluster Coordination
          # JNDI mapping of a data source
          jndiConfig:
             name: jdbc/WSO2ClusterDB
          # data source definition
          definition:
             # data source type
             type: RDBMS
             # data source configuration
             configuration:
               jdbcUrl: 'jdbc:mysql://<host>:<port>/<database_name>?useSSL=false'
               username: <Username_Here>
               password: '<Password_Here>'
               driverClassName: com.mysql.jdbc.Driver
               maxPoolSize: 50
               idleTimeout: 60000
               connectionTestQuery: SELECT 1
               validationTimeout: 30000
               isAutoCommit: false
    2. Specify the time interval (in milliseconds) at which heartbeat pulse should occur within the manager cluster to indicate that a manager is in an active state as shown in the example below.
      heartbeatInterval: 500
    3. Specify the number of times the heartbeat pulse must be unavailable at the specified time interval in order to consider a manager node as inactive as shown in the example below. A value of four means that if a manager node fails to send four consecutive heart beat pulses, it will be identified as unresponsive and another manager node will act as the active node in the manager cluster.
      heartbeatMaxRetry: 4
    4. Specify the time interval (in milliseconds) at which each node should listen for changes that occur in the cluster as shown in the example below.
      eventPollingInterval: 1000
  3. In the deployment.config section, enter information as follows:
    1. In the type field, enter the type of the cluster as distributed
      type: distributed    
    2. For the httpsInterface parameter, specify the host and the port of the node.

      Host should be the IP of the network interface though which nodes are connected. (i.e LAN IP). Each node should have a separate port if deployed in same physical machine.

      e.g., host:localhost, port:9543

    3. Specify the time interval (in milliseconds) at which resource nodes connected to this manager should send heartbeat pulses to indicate that they are in a working state as shown in the example below.
      e.g., heartbeatInterval: 2000
    4. Specify the number of times a resource node's heartbeat should be unavailable for the manager node to identify that the resource node as unresponsive. i.e. according to the below example, if the resource node fails to send 4 consecutive heartbeat pulses it will be recognized as unresponsive and the siddhi applications deployed in that node will be rescheduled to another available resource node.
      e.g., heartbeatMaxRetry: 4

    5. In the minResourceCount parameter, specify the minimum number of resource nodes required to operate the distributed setup. Siddhi applications are not deployed when the number of available resource nodes is less than the number specified here. The default value is 1.

    6. In the bootstrapURLs parameter, specify the Kafka server URLs used by the cluster as a comma separated list. 
      e.g., It should be given in the format <host_1>:<port_1><host_2>:<port_2>

    7. In the zooKeeperURLs parameter, specify the server URL of the zookeeper of the cluster in the format given below:
      <ZOOKEEPER_HOST>:<ZOOKEEPER_PORT>

      The following is an example of a deployment.config section configured as described above.

      deployment.config:
        type: distributed
        httpsInterface:
          host: 192.168.1.1
          port: 9543
        heartbeatInterval: 2000
        heartbeatMaxRetry: 2
        datasource: SP_MGT_DB    # define a mysql datasource in datasources and refer it from here.
        minResourceCount: 1
        bootstrapURLs: 192.168.1.10:9092, 192.168.1.11:9092 # only required in manager (kafka urls)
        zooKeeperURLs: 192.168.1.10:2181
Configure resource nodes

To configure the resource nodes for a fully distributed HA cluster, edit the <SP_HOME>/conf/worker/deployment.yaml file as follows. You have to uncomment (remove the # in front of each line) the section under # Sample of deployment.config for Distributed deployment. Now start performing following steps under deployment.config section.

  1. In the type field, enter the type of the cluster as distributed.
    type: distributed    
  2. For the httpsInterface parameter, specify the host, port and the user credentials of the configuring resource node.

    The host must be the IP of the network interface through which the nodes are connected (i.e., LAN IP). If all the nodes are deployed in the same physical machine, each node must have a separate port.

    e.g., host:localhost, port:9090, username:admin, password:admin

  3. In the leaderRetryInterval parameter, enter the number of milliseconds for which the resource node must keep retrying to connect with a manager node. If the time specified for this parameter elapses without the resource node connecting to a manager node, the resource node is shut down.
    e.g., leaderRetryInterval: 5000 
  4. In the resourceManagers parameter, specify the hosts, ports and user credentials of the manager nodes to which the resource node must try to connect. If there are multiple managers, a sequence must be specified. 

    Following is a sample deployment configuration for a resource node.

    deployment.config:
      type: distributed           # required in both manager / resource
      httpsInterface:              # required in both manager / resource
        host: 192.168.1.3
        port: 9090
        username: admin			  # username of current resource node
        password: admin			  # password of current resource node
      leaderRetryInterval: 10000  # only required in worker
      resourceManagers:           # only required in worker
        - host: 192.168.1.1
          port: 9543
          username: admin		  # username of manager node
          password: admin		  # password of manager node
        - host: 192.168.1.2
          port: 9543
          username: admin		  # username of manager node
          password: admin		  # password of manager node

    If you want to configure this noded as a receiver node instead of a resource node, you need to add the following parameter below the type parameter.

    isReceiverNode : true

    This is the only difference between a resource node and a receiver node. The following is the sample configuration of a receiver node.

    deployment.config:
     type: distributed       
     isReceiverNode : true
     httpsInterface:          
       host: 192.168.1.3
       port: 9090
       username: admin     
       password: admin     
     leaderRetryInterval: 10000
     resourceManagers:          - host: 192.168.1.1
         port: 9543
         username: admin      # username of manager node
         password: admin      # password of manager node
       - host: 192.168.1.2
         port: 9543
         username: admin      # username of manager node
         password: admin      # password of manager node

    Important

    State persistence must be enabled for all worker nodes using a common database. For detailed instructions, see Configuring Database and File System State Persistence

Starting the cluster

To start the distributed SP cluster, follow the procedure below:

    1. Start each manager by navigating to the <SP_HOME>/bin directory and issuing the following command:
      For Windows: manager.bat
      For Linux : ./manager.sh
    2. Start each worker by navigating to the <SP_HOME>/bin directory and issuing the following command:
      For Windows: worker.bat
      For Linux : ./worker.sh
    3. When both manager and resource nodes are successfully started, the following is printed in the log entry.

      INFO {org.wso2.carbon.kernel.internal.CarbonStartupHandler} - WSO2 Stream Processor started in x sec

    4. Siddhi applications should be deployed to the manager cluster using one of the following methods.

      1. Dropping the .siddhi file in to the <SP_HOME>/wso2/manager/deployment/siddhi-files/ directory before or after starting the manager node.

      2. Sending a "POST" request to  http://<host>:<port>/siddhi-apps, with the Siddhi App attached as a file in the request as shown in the example below. Refer Stream Processor REST API Guide for more information on using WSO2 Strean Processor APIs.

        Sample CURL request to deploy Siddhi application
        curl -X POST "https://localhost:9543/siddhi-apps" -H "accept: application/json" -H "Content-Type: text/plain" -d @TestSiddhiApp.siddhi -u admin:admin -k

      Important

      To deploy Siddhi applications in Distributed deployment it is recommended to use a content synchronization mechanism since Siddhi applications must be deployed to both manager nodes. You can use a common shared file system such as Network File System (NFS) or any other shared file system that is available. You need to mount the <SP_HOME>/wso2/manager/deployment/siddhi-files/ directory of the two nodes to the shared file system.
  • No labels