Control-M/Server Cluster Configuration

You can configure Control-M/Server in a cluster environment on UNIX and Window:

Control-M/Server Windows Cluster Configuration

The Control-M/Server installation on Windows automatically recognizes the cluster environment and prompts you to define the Cluster group (cluster aware) or Local install (cluster non-aware). To install, see Control-M/Server Installation.

Automatic installation and automatic upgrade of Control-M/Server is not supported for Microsoft Windows cluster environments.

The Cluster Group installation defines the following components and as a resource in the cluster manager in online status:

  • Control-M/Server

  • Control-M PostgreSQL (If installed with PostgreSQL)

  • Control-M/Server Configuration Agent

By default, the Control-M/Server components are automatically defined as resources in the cluster manager. If a component shuts down, you must start it up from the cluster manager.

Configuring Control-M/Server as a Cluster Resource on Windows

This procedure describes how to configure Control-M/Server in a cluster-aware environment on Windows.

Begin

  1. Install Control-M/Server on the primary host, as described in Installing Control-M/Server on Windows.

    The Control-M/Server, Control-M Server Configuration Agent, and the PostgreSQL for Server900 (if installed) cluster resource components are installed and online.

  2. From the cluster manager on the primary node, run the Move Group command to move the Control-M/Server cluster group to the secondary node.

  3. Verify that the Disk, IP address, and Network Name resources are online on the secondary node.

  4. Do one of the following:

    • Verify that the DVD drive is available to both nodes and run Setup_files\3RD\setup_ctm.bat.

    • From the DVD, copy the Setup_files\3RD\setup_ctm.bat file to the secondary node and run setup_ctm.bat.

  5. Verify that the secondary node is online in the Windows Cluster Administrator window.

    You have now completed the configuration of Control-M/Server with clusters on Windows.

Control-M/Server UNIX Cluster Configuration

The Control-M/Server UNIX installation is not cluster aware. Therefore, you need to configure the UNIX cluster manager to manage Control-M/Server components.

Before you can configure Control-M/Server in a cluster environment on UNIX, you must do the following:

Configuring Control-M/Server as Cluster Resource on UNIX

This procedure describes how to configure Control-M/Server as a cluster resource on UNIX, which enables the cluster manager to start, check, and stop Control-M/Server processes.

Begin

    1. From the CCM, set the Control-M/Server component to Ignore, as described in Ignoring a component:

    2. On the active host create the following scripts in a directory that is not on the shared disk:

      • start_all: startdb && start_ca && start_ctm && start-ag -u `whoami` -p ALL -s

      • show_all: show_ca && shctm && shagent

      • shut_all: shut_ca && shut_ctm && shut-ag -u `whoami` -p ALL && shutdb

    3. Copy the scripts to the other host and allow access to the Control-M/Server user.

    4. On the active host, create the following scripts in a directory that is not on the shared disk:

      • ctms_start_all: sudo -u <ctmuser> -i start_all

      • ctms_show_all: sudo -u <ctmuser> -i show_all

      ret=$?

      if [ $ret -eq 0 ]

      then

      exit 0

      fi

      exit 100

      • ctms_shut_all: sudo -u <ctmuser> -i shut_all

    5. Copy the scripts to the other host and allow access to the cluster user.

    6. Verify that resource group and logical hostname resource are already defined on the cluster.

    7. From the cluster manager, run the following command to create a Control-M/Server resource, register it with required resource group, and allocate it to a virtual host:

      sudo /usr/cluster/bin/clrs create -g <resource_group> -t SUNW.gds:6 -p Scalable=false -p Start_timeout=120 -p Stop_timeout=300 -p Probe_timeout=20 -p Start_command="/<host_private_directory>/ctms_start_all.sh" -p Stop_command="/<host_private_directory>/ctms_shut_all.sh" -p Probe_command="/<host_private_directory>/ctms_show_all.sh" -p Child_mon_level=-1 -p Port_list="2369/tcp" -p Resource_dependencies=<logical_hostname> -p Failover_enabled=TRUE -p Stop_signal=9 <Control-M/Server_resource>