This chapter describes the steps to integrate the Agent Connection into your management system. Using the Agent Connection with your management system requires the following steps:

1. Load the SNMP MIB for TUXEDO and M3 into the management system.

   The MIB defines the data types and access permissions for the various managed objects that can be accessed through the Agent Connection. It also defines the event notifications that can be generated by the Agent Connection. The MIB thus provides the management system with information it requires to manage TUXEDO and M3 resources.

   By default, this file is installed in the installation_directory/etc directory. This MIB must be imported into the management database of your management platform. Some management platforms refer to this process as loading a MIB. Refer to the BEA Manager Release Notes for a list of management platforms tested with the TUXEDO Agent Connection.

2. Decide what kind of information you need in order to meet your system management goals.

   For example, are there particular attributes of the resources you are managing that you want to monitor? Do you want to be notified when certain TUXEDO system events occur?

3. Configure the management system response to incoming TUXEDO system events.

   This is described in the next section, "Integrating M3 and TUXEDO Event Notifications."

4. Configure polling or data collection rules on the manager for performance and fault management.

   Periodic collection of values of pertinent objects is valuable for analysis of trends. This is valuable for capacity-planning and load-balancing. Polling can also be used to generate alarms which is useful for fault management.

5. Define Agent Integrator polling rules (optional)

   If you are using the Agent Connection as a subagent with the Agent Integrator, you may want to offload some threshold checking to the Agent Integrator. The Agent Integrator generates enterprise-specific traps when the user-defined threshold is crossed. Offloading checking of selective thresholds to the Agent Integrator reduces the network bandwidth consumed by the management station's polling activities.

6. Configure and start the agents.

   The procedure for configuring and starting the SNMP agents is described in Chapter 2, "Setting Up the Agents."

To integrate the M3 and TUXEDO system event traps with your management platform, perform the following actions:

1. Make sure the TUXEDO Event Broker server (TMSYSEVT) is running for the domain being managed.

   The Agent Connection will not receive event notifications unless the Event Broker server (TMSYSEVT) is running. Information on the TUXEDO Event Broker can be found in Section 5 of the BEA TUXEDO Reference Manual.

2. Modify the TRAP_HOST entry in the BEA Manager configuration file (beamgr.conf) to specify the location of the management machine that is to be the destination for traps generated by the agent.

3. Load the BEA Manager SNMP MIB file (bea.asn1) into your management platform (if you have not already done so).

   For example, on HP OpenView Network Node Manager, do the following:

   1. Select Options'Load/Unload MIBs: SNMP (Figure 3-1).

   2. Select Load.

   3. Specify the path to the BEA Manager configuration file (bea.asn1). By default, this file is installed in:

      install_directory\etc\bea.asn1

   4. Select OK.

      Figure 3-1 Selecting Load/Unload MIBs in OpenView

      

4. Configure the management system to take the appropriate action in response to incoming TUXEDO SNMP traps.

   You may want to change the way in which TUXEDO SNMP traps are displayed on your management console, or the actions that the management system takes in response to specified events. For example, you might choose to ignore some routine informational notifications. For example, to view the event configuration on HP OpenView, do the following:

   1. Select Options'Event Configuration.

   2. Select the enterprise tuxedo.

   3. Select an event type

      In Figure 3-2, networkFlowTrap is the selected event type.

      Figure 3-2 OpenView Event Configuration

      

   4. Customize the management system response to incoming events of that type. Select Edit'Modify Event in OpenView. This invokes the Event Configuration window (Figure 3-2). You can modify the event configuration to ignore an event, or generate a pop-up notification or run a program or script when the event is received. You might also want to create a separate category for TUXEDO events, as shown in Figure 3-3.

      Figure 3-3 OpenView Event Categories Window

      

Monitoring of multiple domains is done by running a separate TUXEDO or M3 agent for each domain being monitored. These agents must be run as subagents under the Agent Integrator.

When more than one M3 or TUXEDO SMP agent is running on a node, then SNMP manager GET or SET requests to an agent must be addressed using a community of the form:

community@logical_agent_name

For example:

public@payrollagent

In this example payrollagent is a logical agent name that identifies the agent to which the request is to be forwarded by the Agent Connection.

Integrating Events Generated by Agent Integrator Polling

Agent Integrator can be used to poll TUXEDO or M3 objects, or other managed resources. To integrate the Agent Integrator threshold-checking activity with the management system, do the following:

1. Set up the Agent Integrator polling rules.

   A polling rule is defined by a RULE_ACTION entry in the BEA Manager configuration file (beamgr.conf). This is described in the Agent Integrator Reference Manual.

2. Configure the management system to recognize the events generated by Agent Integrator.

   For example, in HP OpenView, you can add a new event type by doing the following:

   1. Select Options'Event Configuration

   2. Select beaSystemDescr

   3. Select Edit'Add Event.

   In the window that is invoked you would use the following as the event number:

   .1.3.6.1.4.1.140.1.1.0.specific_trap_number

3. Configure the management system to respond appropriately to incoming Agent Integrator events.

   This involves essentially the same process as described in Step 4 in the previous section.

The Agent Connection for TUXEDO and M3 has the ability to translate TUXEDO system event notifications into SNMP trap notifications. These traps can be sent to the Tivoli Event Console (TEC) using the TEC SNMP Adapter. The TEC SNMP Adapter must be installed on the TUXEDO master node along with the Agent Connection, as illustrated in Figure 3-4.

Figure 3-4 Integrating Agent Connection with the Tivoli Event Console

To integrate the Agent Connection with the Tivoli Event Console, do the following:

1. Install the TEC SNMP Adapter.

   The Tivoli Event Console SNMP Adapter needs to be installed - preferably on the TUXEDO master node.

2. Configure the TEC SNMP Adapter.

   The following is a suggested mapping of TUXEDO to Tivoli events. You may need to modify this mapping to meet your own requirements.

   1. To provide a class definition for TUXEDO events, copy the following lines from the file bea2tiv.baroc, provided with Agent Connection, to the file tecad_snmp.baroc of the TEC SNMP Adapter.

      TEC_CLASS :
           Tux_Event ISA  Specific_SNMP_Trap
             DEFINES {
                     severity: default = WARNING;
                     class: STRING;
                     ulogcat: STRING;
                     ulogmsgnum: INTEGER;
                     };
      END

   2. To provide a class description for TUXEDO events, copy the following lines from the file bea2tiv.cds, provided with Agent Connection, to the file tecad_snmp.cds of the TEC SNMP Adapter:

      CLASS Tux_Event
         SELECT
            1: ATTR(=, "tuxEventsName" ) ;
            2: ATTR(=, "tuxEventsLmid" ) ;
            3: ATTR(=, "tuxEventsTime" ) ;
            4: ATTR(=, "tuxEventsDescription" ) ;
            5: ATTR(=, "tuxEventsClass" ) ;
            6: ATTR(=, "tuxEventsUlogCat" ) ;
            7: ATTR(=, "tuxEventsUlogMsgNum" ) ;
         MAP
            source = "TUXEDO";
            enterprise = "tuxedo";
            sub_source = $V1;
            hostname = $V2;
            date = $V3;
            msg = $V4;
            class = $V5;
            ulogcat = $V6
            ulogmsgnum = $V7;
      END

   3. To provide the OID definitions needed for TUXEDO events, copy the following lines from the file bea2tiv.oid, provided with Agent Connection, to the file tecad_snmp.oid of the TEC SNMP Adapter:

      "tuxedo"                 "1.3.6.1.4.1.140"
      "tuxEventsName"          "1.3.6.1.4.1.140.300.2.6.1"
      "tuxEventsSeverity"      "1.3.6.1.4.1.140.300.2.6.2"
      "tuxEventsLmid"          "1.3.6.1.4.1.140.300.2.6.3"
      "tuxEventsTime"          "1.3.6.1.4.1.140.300.2.6.4"
      "tuxEventsUsec"          "1.3.6.1.4.1.140.300.2.6.5"
      "tuxEventsDescription"   "1.3.6.1.4.1.140.300.2.6.6"
      "tuxEventsClass"         "1.3.6.1.4.1.140.300.2.6.7"
      "tuxEventsUlogCat"       "1.3.6.1.4.1.140.300.2.6.8"
      "tuxEventsUlogMsgNum"    "1.3.6.1.4.1.140.300.2.6.9"

3. Set up trap destination for BEA Manager.

   Ensure that the TRAP_HOST entry in the BEA Manager configuration file (beamgr.conf) points to the host where the Tivoli Event Console SNMP Adapter is running.

4. Start the TUXEDO or M3 SNMP agents.

   Refer to Chapter 2, "Setting Up the Agents," for more information.

5. Set up the Tivoli Event Server for TUXEDO events.

   Set up the rules on the Tivoli management system that are to be applied to incoming TUXEDO system events. This is essentially the same process as described in Step 4 under "Integrating M3 and TUXEDO Event Notifications."

6. Set up the Tivoli Event Console for display of TUXEDO events.

   You may need to configure the Tivoli management system to select which TUXEDO events to display and how they should be displayed.

Tivoli has the ability to distribute polling or data collecting to distributed Tivoli agents. The following steps are an example of how to implement distributed monitoring of TUXEDO or M3 MIB objects on the Tivoli platform:

1. Create a new monitor that specifies the object you want to monitor.

   The Profiles Property window lists the monitors which you have configured to do threshold-checking. Select Add Monitor from the Profile Properties window.

2. Select the UserSNMP monitoring collection, as shown in Figure 3-5.

   Figure 3-5 Adding a New Monitor to a Tivoli Profile

   

   You can define the frequency of polling, the SNMP community, and the object that you wish to monitor. The object must be specified using the absolute object identifier (OID). In the example we have used .1.3.6.1.4.1.140.305.1.6.0, which is the OID for beaDomainStatus. In this case we are setting up a monitor to check whether a TUXEDO domain is active.

   In the case of columnar objects, you need to know which instance to monitor. For information on how to specify an instance using an object identifier, refer to the "Polling" chapter in the Agent Integrator Reference Manual.

3. Define the polling threshold and the desired response.

   Select Add Empty to invoke the Monitor window. This allows you to specify the polling threshold, and other properties of the monitor, such as the severity of the Tivoli event that is generated when the threshold is crossed. In Figure 3-6, we specify a threshold of Not equal to 1 because we want to know when the domain is not active.

   Figure 3-6 Defining a Polling Threshold

   

4. Distribute the monitor to the Tivoli agents.

   Once you have defined a monitor that specifies the polling interval and condition being checked, you can deploy this monitoring profile to the Tivoli agents for distributed monitoring. Consult the Tivoli documentation for further information.