Configure Multiple UMP Servers

This article describes how to set up an initial multiple-UMP configuration. If you are upgrading a multiple-UMP configuration, see .
uim902-9-0-2
This article describes how to set up an initial multiple-UMP configuration. If you are upgrading a multiple-UMP configuration, see Upgrade a Multiple UMP Configuration.
 
 
2
 
 
 Ensure that you deploy the wasp_alarmviewer_api package to the secondary UMP.
Configure a Secondary UMP Server
Use the steps in this section to configure a secondary UMP server. Repeat these steps to configure additional UMP servers.
 No steps are required for configuring the primary UMP server.
Review the following points:
  • A secondary UMP server can only be installed on a robot. Do not attempt to install a secondary UMP server on a hub.
  • All probe addresses, for example, 
    /domain/hub/robot/
    ace, are case sensitive.
  • Before wasp is started on the secondary UMP server, import the .pem file to the secondary UMP server and add the cryptkey path for that .pem file in the robot.cfg file. For more information, see Configure the robot.cfg File.
  • Verify that robot 7.96 is running on the secondary UMP robot (without this the wasp will not start).
 
Follow these steps:
 
  1. Deploy a robot to the host that you plan to use as the secondary UMP server.
    • The robot you choose must report directly to the primary hub, and must be on the same network subnet.
    • Installing a secondary UMP instance on a robot that reports to a secondary hub is not supported.
    • Do not run the UMP installer on the secondary UMP server.
  2. In Admin Console or Infrastructure Manager, distribute the following UMP server packages from the Archive to the secondary UMP server. Ensure that you drag each of the ump_portlet_name packages that are required for your environment from the archive.
    • java_jre
    • wasp
    • nisapi_wasp
    • ump
    • ump_
      portlet_name
       
  3. Configure the wasp probe to use the correct data_engine probe address. For example, 
    /domain/hub/robot
    /data_engine
  4. Configure the wasp probe 
    relationshipviewer
     webapp to have the custom properties 
    GraphServiceHost
     and 
    GraphServicePort
    GraphServiceHost
     is the IP address of the robot running the relationship_services probe. 
    GraphServicePort
     is the HTTP port that relationship_services is using. 
    Default
    : 8182
  5. Use Raw Configure to modify the 
    <ump_common>
     section of the wasp configuration to specify the following probe addresses:
    • ace: /
      domain/hub/robot
      /ace
    • automated_deployment_engine: /
      domain/hub/robot
      /automated_deployment_engine
    • discovery_server: /
      domain/hub/robot
      /discovery_server
    • nas: /
      domain/hub/robot
      /nas
    • service_host: /service_host Functionality for the service_host probe has been moved to wasp, and the probe is no longer installed as part of CA UIM v8.47 or later. The address key for the probe still exists in wasp but is not functional.
    • sla_engine: /
      domain/hub/robot
      /sla_engine
  6. Add the following keys with addresses as follows:
    • maintenance_mode = /
      domain/hub/robot
      /maintenance_mode
    • udm_manager = /
      domain/hub/robot
      /udm_manager
    • mpse = /
      domain/hub/robot
      /mpse
  7. Activate the wasp probe on the secondary UMP server. When the wasp probe is activated, the probes create database tables.
  8. Repeat Steps 1 - 7 on other robots to implement more UMP servers.
Configure the Portal for Multiple UMP Servers
Use the steps in this section after configuring the secondary UMP server.
 Be sure that you have permissions to change directories for your operating system before following these steps. Minimum permissions include Read, Write, and Execute.
 
Follow these steps:
 
  1. Deactivate the wasp probe on each UMP server.
  2. On each UMP server, add or uncomment the following two lines in <UIM_installation>/probes/service/wasp/webapps/ROOT/WEB-INF/classes/portal-ext.properties:
    • cluster.link.enabled=true
    • lucene.replicate.write=true
  3. Create and share a shared document_library directory on the primary UMP server.
    • For Windows, share the directory <UIM_installation>/probes/service/data/document_library with full read/write access.
    • For Linux, share the directory <UIM_installation>/probes/service/data/document_library via NFS, and ensure that no_root_squash is enabled.
  4. On the secondary UMP server, replace the directory <UIM_installation>/probes/service/data/document_library with a link to the primary document_library share.
    • For Windows, use the 
      mklink 
      command as follows:
      mklink /d "c:\<UIM_installation>\probes\service\data\document_library" "\\<primary>\<document_library_share>"
    • For Linux, use the 
      mount 
      command as follows:
      mount -t nfs <primary>:/<UIM_installation>/probes/service/data/document_library <UIM_installation>/probes/service/data/document_library
  5. (Only required for UMP servers that have network connections configured with 
    both IPv4 and IPv6 addresses
    ). Add the following Java VM parameter to the wasp configuration:
    -Djava.net.preferIPv4Stack=true
  6. Activate the wasp probe on the primary UMP server.
  7. When the wasp is running on the primary UMP server, activate the wasp on the secondary UMP server.
The portal is now configured for multiple UMP servers.
Configure a Load Balancer
This section provides the high-level steps for configuring a load balancer for a multiple UMP server configuration. Use the steps in this section after you install and configure the primary and secondary UMP server.
 Load balancers and the terms that vendors use to describe them vary. Refer to the documentation for your load balancer for specific configuration details. You may need to perform the high-level steps in this section in a different order than shown.
 
Follow these steps:
 
  1. Deploy the load balancer:
    1. Configure the load balancer with IP addresses for each UMP server.
  2. Create a node/device for each UMP server.
    1. Enter the name and IP address of each UMP server in your configuration.
  3. Create a pool/server farm:
    1. Provide a unique name for the pool/server farm.
    2. Add one or more health monitors, such as the gateway_icmp and http_head.
    3. Select a load balancing algorithm. The most common load balancing algorithm is round robin, where one connection is sent to each server on the list in turn.
    4. Add the nodes you created previously to the pool for port 80. Use port 443 for an HTTPS connection.
  4. Create a virtual server/context:
    1. Provide a unique name for the virtual server/context.
    2. Provide an IP address for the virtual server/context.
    3. Configure additional settings for the virtual server as follows:
      • Protocol = tcp
      • HTTP Profile = https 
        or 
        http
      • Source Port = preserve strict
      • Default Persistent Profile = cookie.
       Regardless of the load balancer, setting the Default Persistent Profile setting to 
      cookie
       is required.
  5. Verify the load balancer was successfully configured by entering the IP of the virtual server/context in a web browser.
The load balancer is now configured for a multiple UMP server configuration.