Planning for a high availability environment

Starting in IBM® Control Center Monitor V6.1, you can choose to have a high availability environment. To help you successfully incorporate IBM Control Center Monitor in a high availability environment, you must plan your implementation.

Before you create a high availability environment in IBM Control Center Monitor, consider the following planning questions:
  • Which operating system is your high availability environment going to be installed on? For more information, see Installing IBM Control Center Monitor in a high availability environment.
  • What database type are you going to use in your high availability environment? For more information, see Creating and setting up databases and Databases.
  • What is your CPU allocation for each event processor (EP)? If you are running at 80 percent or more CPU capacity with IBM Control Center Monitor V6.0, then in V6.1 you need to add more CPU. Adding more CPUs is necessary to properly scale an IBM Control Center Monitor environment with multiple EPs. For more information, see Detailed System Requirements.
  • What is your CPU allocation for each database?
  • What is your planned heap memory allocation? The max heap default allocation for an EP is 4 GB and web console is 2 GB, but to increase performance in a high availability environment, the allocation can be set to 8 GB for an EP and 4 GB for the web console.
  • How many EPs are you going to install? For more information, see Event processors.
  • Which EP are you going to start first? The EP that is started first becomes the controller EP. For more information, see Controller event processor.
  • Are there any EPs that are going to be carrying a heavy load? If an EP with a heavy load becomes the controller EP, then your performance of IBM Control Center Monitor can become slower.
  • Do you want to configure the Jetty Web Application server to not automatically shut down when an EP shuts down?
  • Do you use WS_FTP, Sterling Connect:Enterprise® for z/OS®, FTP for z/OS, or Sterling Connect:Direct® File Agent that send traps to IBM Control Center Monitor? In a high availability environment, you need to plan to configure WS_FTP, Sterling Connect:Enterprise for z/OS, and FTP for z/OS to send traps to the EP they are assigned to. Sterling Connect:Direct File Agent must be configured to send their traps to all EPs. Another option is to set up a Trap Forwarder, external to Control Center, and to configure those servers to send their traps to the trap forwarder, and to configure the trap forwarder to send all the traps it receives to all the EPs.
  • When you add new servers for IBM Control Center Monitor to monitor, are any of the EPs going to accept newly added servers, accept servers from failed EPs, or allow other EPs to monitor servers upon failure? For more information, see Setting a server assignment, failover, and monitoring policy for an event processor.
  • Are you going to use a load balancer in your high availability environment? Configuring a load balancer evenly distributes your web console sessions, or events of a dynamically discovered server that are posted to the IBM Control Center Monitor event repository. For more information, see Load balancer configuration for IBM Control Center.
  • In your high availability environment, are you going to rebalance your servers? You can rebalance your servers, including servers that are not running, to other EPs for a more even server distribution according to their reassignment policy. For more information, see Rebalancing servers across event processors.
    Important: By default, if you have Sterling Connect:Enterprise for z/OS, FTP z/OS, or WS_FTP server types, then those server types are pinned and not included in a rebalance. If you want to unpin these server types and include them in the rebalance, then go to the advanced properties for each server. The system administrator of those servers must reconfigure the SNMP host to send traps to the new EP.

After you plan for a high availability environment, you can begin installing EPs.