Agents (HOST)

Agents are programs that run on the target system, which can be either an Operating System or an application such as SAP, PeopleSoft, Siebel, etc. They establish the connection between the Automation Engine and those target systems, start the execution of tasks and make both their monitoring and the corresponding reporting possible. Agents create log files that record what occurs. As an administrator user, you install and configure them.

There are two possibilities for installing the Agents:

During the installation of the Automation Engine
Later on via the Automic Web Interface in the Process Assembly perspective of Client 0

As an administrator user, you create Agents manually in Client 0 in the Process Assembly perspective. You maintain and work with them in the Administration perspective. Agents are configured using their INI files. You can change some of their properties using the Automic Web Interface.

Important!

Once the Agents are installed, it is not possible to start working with them immediately. First you must authenticate them. There are three Authentication Methods, namely NO, LOCAL and LOCAL_REMOTE. Depending on the method you are applying, the steps to authenticate the Agents are slightly different. For more information, see Agent Authentication and Authenticating the Agents.
Rapid Automation Agents are an exception. They are not created automatically upon installation, you must create them manually after you have loaded the Rapid Automation solution in the database via the Utility DB Load.

Tip: When working with the Automation Engine system, you may need to work with many objects, lists and monitors simultaneously. To make your work easier, you can open them in different browser windows and arrange them side by side on your screen. For more information see Showing Windows Side by Side.

Object Definition

Object class: System object
Object type/Short form: HOST

This page includes the following:

Prerequisites

Do the following to be able to start working with an Agent:

Install it according to the target system requirements.

See Installing the Agents.
Set the Authentication Method.

See Agent Authentication and Authenticating the Agents.
Configure the INI file.

See Agents.
Authenticate the Agents in Client 0.
Assign the Agent to the Clients it should have access to and define the rights it should have.

See Assigning Clients to Agents.
Define user and user group authorizations on the Agent. This grants users and user groups specific rights on the Agent. You do it in each Client.

See Authorizations Page.

An Agent logs in only once and, once connected, it remains always connected. The following duplicate login validation takes place:

Upon installation, the Agent starts and tries to log in. If there is no Agent connection with the same name, the Automation Engine accepts the login attempt and stores the Agent name, port number, start time etc. It deletes this information when the particular Agent connection ends.
If the Automation Engine detects that an Agent with the same name has already logged in, it tries to establish contact. The time span in which it waits for a response can be specified in the duplicateTimeout= parameter in the INI file of the Agent (section [TCP/IP]). If the Agent responds within the specified time span, the login attempt is rejected. If the Agent does not respond, the double attempt to login is accepted and the login information (such as the start time) is updated.
Agents connect via TCP/IP to the communication process (CP). The communication between CP and Agent is encrypted . It is the Agents that are in charge of keeping the connection alive, as well. When it is lost, they try to reconnect; the interval of reconnect attempts is configurable.

For more information see Network Communication and Encryption and Encrypting the Communication.

Agents List

You define the number of Agents to be displayed on the list in the AGENT_LIST_LIMIT variable (see UC_SYSTEM_SETTINGS - Systemwide Settings). The default is 5000, the maximum is 20,000.

The number of agents currently displayed is shown on the list toolbar. A plus next to the number indicates that there are more Agents available than those displayed. For example, if the AGENT_LIST_LIMIT has been set to "200" and more than 200 Agents have been assigned to the Client, Agents (200+) is displayed on the toolbar.

In addition to the basic Agent properties, the list provides information on the following:

Status

Grayed out Agents are inactive.
Name
Authenticated

Some Agents support authenticated communication with the Automation Engine, some do not. If an Agent has been authenticated since it was installed, a tic is displayed.
Version

Version of the Agent.
Platform

Platform type of the Agent. This is not the OS.
Hardware

Information about the computer on which the Agent is running, for example, the system type, number of processors, or processor type.
Software

Type of platform or operating system of the computer on which the Agent is installed.
Software Version

Version of the platform or operating system.
Last check *

The time at which the status and other Agent-related information was last checked.
Communication Process *

Communication process with which the Agent is connected.
Time difference *

Time difference between the time of the agent and the time of the Automation Engine. This is helpful for interpreting Job reports when the Automation Engine and the Agent run on different hosts with different system times.
IP address

In systems that run with encryption, IP addresses of non-encrypted connections that will be accepted by the Agent.
Port no.

Port number of the Agent.
MIB *

MIB information of the Agent that is used in connection with SNMP. If it is available, this column displays the unique ID that identifies each monitored net element.

For more information see SNMP Support.
Services *

Whether the Agent performs a special service:
Compression *

Compression method used as standard in File Transfers.
Authorizations *

Privileges granted to this Agent in the current Client.
Job res. max. *

Maximum Agent resources available for Jobs.
FT. res. max. *

Maximum Agent resources available for File Transfers.
Keep alive *

Interval in seconds in which the connection to the communication process is checked. If the connection is lost, the Agent tries to reconnect automatically to the first available communication process (CP), which could be a different Agent to the previous one.
Roles *

Administrators can define a role (free text) for the agent in its INI file. If so, it is displayed here.
Last logoff

Last time in which the Agent logged off the Automation Engine system.
Number of clients

Number of user-defined clients in which the Agent has authorizations.
Linked

Whether the Agent is linked to a ServiceManager or not. This depends on the settings made in the Service Manager section.
Process ID

Process identification number of the Agent.
Net Area

For details see Net Areas in the Automation Engine.

By default, this is the name of the Automation Engine system.
ZDU Version

Either base (B) or target (T) version of the Agent in connection with a Zero Downtime Upgrade (ZDU).

* These columns are not displayed by default. To be able to see them, you must expand the list of additional columns by clicking the arrow at the right end of the header row and select them.

Displaying the Agent Variables

To see the variables that have been defined for a particular Agent, select it and click the Details button on the toolbar. The Details pane displays the variables and the values that have been defined for the Agent .

Filtering the Content of the List

It is possible to use different filters to customize the content of the Agents list. Click the Filter button on the toolbar to open the Filter pane. Agent Name, Status and IP Address are always displayed but you can add additional criteria.

Editing the Agent Properties

Once an Agent has been added to your Automation Engine system, you can edit its properties as follows:

Right-click the Agent and select Open or Open in New Window.
Edit the properties on the Agent pages described below.

Authorizations Section

Grant or remove the rights of the Agent on Clients.

Attributes Section

Trusted IPs

If your system runs with encryption, you can specify the IP addresses from which the Agent accepts non-encrypted connections. They must be separated by semicolons. Communication to components such as the Job Messenger, Event Monitors (BS2000, z/OS) or the Output Collector (NSK) takes place without encryption.

By default, the Agent accepts non-encrypted messages from the local host (127.0.0.1 on IPv4 and ::1 on IPv6) and its own IP address. If Event Monitors on z/OS are used in LPARs on different computers, you must specify the relevant IP addresses here.

Releasing individual IP addresses is only reasonable if login attempts of former Agent versions should be rejected.

Service Manager Section

These properties are necessary for the connection between the Agent and the Service Manager, an external monitoring tool that allows you to start and stop the Automation Engine.

The Service Manager comes with a proprietary user interface. If you want to use the AWI instead, you can do so provided you enter the necessary information. This way, Agents can start directly via the User Interface or via the MODIFY_SYSTEM script element.

All values will automatically be replaced if the connection to the corresponding Service Manager service was successful.

To refresh the connection, right-click the Agent and select Update Service Manager Link from the context menu. The default port number 8871 is used, unless a different specification is provided in the SMGR_PORT_RANGE key in the UC_SYSTEM_SETTINGS variable (UC_SYSTEM_SETTINGS - Systemwide Settings).

Phrase is the name of the Service Manager environment.
Name refers to the service in the Service Manager.
Port is the port number of the Service Manager.
Activate Link Agent to Service Manager, otherwise you can no longer start the Agent either through the Administration perspective or using the MODIFY_SYSTEM script element. This specification is automatically set if the Service Manager scan for the Agent was successful.

Job Execution

If a Job is started on an OS Agent, it will be initiated in the OS independently of the Agent. The Agent monitors it and reports its status to the Automation Engine. When the Job finishes, follow-up Jobs can be triggered.

Job Reports

Job reports can be configured in the Job object. When the Job is executed they are transferred to the Automation Engine database. The OS Agent provides the report as a file. The Automation Engine loads the report as individual blocks into the database.
Should the report exceed a configurable limit, only data with the configured size will be transferred into the database, where the last block transferred always contains the file's end. Thus the beginning and end of a Job report are available in the database.

File Transfer Objects

A File Transfer is an executable object that is always executed between two OS Agents. A TCP/IP connection is created between both. The data transfer is encrypted.

The Agent sending files reports the Job status and the end of the Job execution to the Automation Engine system. The connection between the two Agents is closed at the end of the File Transfer. For more information, see File Transfers (JOBF).

Monitoring and Reporting

As soon as a Job is started, the Agent starts monitoring and reporting its status to the Automation Engine. Likewise, when the Job finishes, it reports it. See Reports for Servers and Agents.