Installing a monitoring agent on Linux or UNIX computers

Before you begin

Keep the following guidelines in mind:

You can install only one Application Management Console (which is also the robotic script file depot) in the IBM Tivoli Monitoring environment.
You can install more than one Web Response Time agent on the same UNIX or Linux host, as long as the agents are installed in different file systems (for example, /opt/IBM/ITM and /var/IBM/ITM2). Some configuration customization is also required. For more information about installing agents, see Installing multiple Web Response Time agents on the same Linux or UNIX host.
The Application Management Console must be on its own server-class box if monitoring more than 50 agents or if using Web Response Time to monitor a high volume web site generating thousands of transactions an hour. In test environments, it can be on the same computer as the Tivoli Enterprise Monitoring Server and the Tivoli Enterprise Portal.
The Tivoli Enterprise Portal has a 26 character limitation for the host and application names in the Navigator. When you have a combination of names that are longer than 26 characters, the application name is truncated. For example, if you have application names such as Websphere Plants and Websphere Petstore on a computer with a long host name, the names might be displayed in the Navigator as WebspherePl and WebspherePe, whereas if the host name are only four characters shorter they are displayed as WebspherePlants and WebspherePetsto. To avoid truncating the host and application names in the Navigator, install the Application Management Console on a computer with a short host name.
If you have a previously installed Response Time Tracking, version 6.1, you must uninstall it and verify that the following dll's were deleted before installing Robotic Response Time or Client Response Time.
- /usr/lib/libarm4.*
- /usr/lib/libarmjni4.*
- /usr/lib/libarmjni.*
- /usr/lib/libarm32.*
- /usr/lib/libarmcli.*
Install Rational Performance Tester software if you want to record and play back RPT test scripts on Robotic Response Time.
If your environment has installed a GSKit version newer than 7.0.3.18, the installation fails. Uninstall the GSKit and try installing the agent again.
If you are not logged in as root:
- You cannot install or run Web Response Time. If you install the Web Response Time agent and attempt to start the kfcm120 process as a non-root user, the process will fail because it cannot open the NIC. Root is required to access the Network Interface.
- You must have write access to the following:
  - ITM_HOME or the installation directory
  - TEMP directory for the system
  - The JLOG directory for the Tivoli Enterprise Monitoring agent, typically called /var/ibm/tivoli/common/BWM/logs/
- Do not install two monitoring agents in the same ITM_HOME or installation directory using different user IDs. You can install two agents in the same ITM_HOME or installation direction if you use the same user ID for the installations.
When installing Client Response Time (t4) or Robotic Response Time (t6), you must manually stop any applications that use ARM before starting the installation.

The following table provides an overview of what you need to for installation:

Table 1. Overview of installing a monitoring agent on Linux or UNIX computers
Steps	Where to find information
Install the monitoring agent.	Procedure: Installing the agent
Change the file permissions for files on the computer where you installed the agent, if you used a non-root user to install a monitoring agent	Procedure: Changing file permissions
Configure the monitoring agent.	Configuring Client Response Time from the GUI (UNIX and Linux) Configuring Application Management Console from the GUI (UNIX and Linux) Configuring Robotic Response Time from the GUI (UNIX and Linux) Configuring Web Response Time from the GUI (UNIX and Linux) Some agents require additional, agent-specific configuration parameters. See the agent documentation for the specific agents that you are configuring.
Start the monitoring agent.	Starting and stopping monitoring agents

Procedure: Installing the agent

In the directory where you extracted the installation files, run the following command: ./install.sh.
When prompted for the IBM Tivoli Monitoring home directory, press Enter to accept the default (/opt/IBM/ITM) or type the full path to a different directory.
You are given a list of processes currently running. You are warned that these processes will be stopped and restarted during the installation. Type 1 and press Enter to continue the installation.
A prompt similar to the following is displayed:
```
Select one of the following: 
1) Install products to the local host. 
2) Install products to depot for remote deployment (requires TEMS).
3) Exit install. 

Please enter a valid number:
```
Type 1 to start the installation and press Enter.
Type the number that corresponds to the language in which you want to display the software license agreement in and press Enter. (In some cases, you might not be prompted for this step. so you can skip it.)
Press Enter to display the agreement.
Type 1 to accept the agreement and press Enter.
Note: Step 8 applies to agents installed from the IBM Tivoli Monitoring installation image, if the agent is installed from the agent installation, you can skip the step.
Type a 32 character encryption key and press Enter. This key was specified during the installation of the monitoring server to which this monitoring agent connects.
Type the number for the operating system on which you are installing and press Enter. The default value is your current operating system.
Type 1 to validate your selection and press Enter.
A list of available products for installation is provided. Type the number that corresponds to the monitoring agent or agents that you want to install and press Enter. To install more than one agent, use a comma (,) or a space to separate the numbers for each agent.
Type 1 to validate your selection and press Enter to start the installation.
When prompted to install additional products or product support packages, type 2 for no and press Enter.

Procedure: Changing file permissions

If you used a non-root user to install a monitoring agent on a UNIX computer, the file permissions are initially set to a low level. Run the following procedure to change these file permissions:

Log in to the computer as root, or become the root user by running the su command.
Create a new group (such as itmusers) to own all of the files in the IBM Tivoli Monitoring installation directory.
For Linux, Solaris, and HPUX computers, run the following command: groupadd itmusers

For an AIX® computer, run the following command: mkgroup itmusers
Run the following command so that the CANDLEHOME environment variable correctly identifies IBM Tivoli Monitoring installation directory: export CANDLEHOME = <ITM_Install_Directory>
CAUTION:
Running the following steps in the wrong directory can change the permissions on every file in every file system on the computer.
Run the following command to change to the directory returned by the previous step: cd $CANDLEHOME
Run the following command to confirm that you are in the correct directory: pwd
Run the following commands:
```
chgrp -R itmusers 
chmod -R o-rwx 
```
Run the following command to change the ownership of additional agent files: bin/SetPerm
(Optional) To run the agent as a particular user, add the user to the itmusers group by editing the /etc/group file and adding the user to the itmusers group.
For example, to run the agent as user test1, add following line is in the /etc/group file: itmusers:x:504:test1
Run the su command to switch to the user that you want to run the agent as or log in as that user.
Start the monitoring agent.

Installing multiple Web Response Time agents on the same Linux or UNIX host

On supported Linux and UNIX systems, multiple Web Response Time agents can run on the same system, as long as the agents are installed in different file systems (for example, /opt/IBM/ITM and /var/IBM/ITM2). Some configuration customization is also required.

Complete the following steps to achieve this setup:

Install multiple instances of the Web Response Time (WRT) agent in different file systems. These path names must not be subsets of each other. For example, for an installation deploying two WRT agents on the same box:
- Correct:
```
/opt/IBM/ITM
/var/IBM/ITM2
```
- Incorrect:
```
/opt/candle
/opt/candle/itcam_wrt/ServPoint
```
In the <ITM_HOME>/tmaitm6/wrm directory, add the following line to the kfcmenv file for each WRT agent, where <port> is an open port for the machine:
```
KFC_API_MEDIASERVER_LISTEN_PORT=<port>;
export KFC_API_MEDIASERVER_LISTEN_PORT
```
Use port 12121 for the first agent, port 12122 for the second agent, port 12123 for the third agent, and so on.
In <ITM_HOME>/config, add this same line to the t5.ini and t5.config files for each WRT agent.
In <ITM_HOME>/config, add the following line to the t5.config and t5.ini files for each WRT agent. In <ITM_HOME>/tmaitm6/wrm, add the following line to the kfcmenv file for each WRT agent, where <port> is an open port on the system:
```
KT5ERRORMESSAGEPORT=<port>
```
Ensure that the port is unique for each agent.
In <ITM_HOME>/config, set the CTIRA_HOST variable in the t5.config and t5.ini files for each WRT agent to define the node name that the Tivoli Enterprise Portal (TEP) uses to identify each agent.

Avoid monitoring identical data: To avoid having each agent monitoring identical data, you must satisfy either of the following conditions:

Each agent monitors a different NIC.
In the <ITM_HOME>/tmaitm6/wrm/kfcmenv file for each WRT agent, use the following option to define which servers each WRT agent monitors:
```
KFC_RESTRICT_HOST=<ip1>:<port1>,<ip2>:<port2>,...
```

Follow-up or related tasks

Complete the following tasks:

Configure the agent you installed from either the command line or from the GUI:
Configuring the Eclipse help server