HP OpenView System Administration Handbook [Electronic resources] : Network Node Manager, Customer Views, Service Information Portal, HP OpenView Operations نسخه متنی

14.3 AGENTS ALIVE AND WELL AT ALL TIMES

The server periodically checks the health of the agents. The frequency of the check is configured in the node configuration. The control agent sends status replies to the message receiver on the server. If the network interface is alive (via ping), the server sends an alive-check packet to the control agent. The DCE control agent is the parent process of all the other agents running on the node. When the control agent is not running, the server reports that the control agent is down. If the network interface is down, the server reports that the node is down. The control daemon (ovcd) is the parent process on HTTPS agents.

The agent can send alive checks to the server. This may be more efficient from a network traffic perspective. To configure the "agent sends alive check," modify the node configuration from the Node Add or Node Modify Window. See Figure 14-1.

14.3.1 RPC Service

HTTPS-based node communications is the default (for 8.x managed nodes). You can select the RPC-based node and communication type when you add the node to the OVO configuration. There are prerequisites that must be met prior to installation of the RPC-based agents. Refer to the system vendor's documentation for operating system RPC patch information. Refer to the DCE Agent Concepts and Configuration Guide for platform-specific node communication details.

14.3.2 HTTPS Communication Service

HTTPS-based communications provides information exchanges based on open standard HTTP 1.1 and SOAP. The HTTPS-based process (ovbbccb) runs on the managed node and listens for request on port 383. The HTTPS-based agent communications eliminate the need for DCE/NCS RPC technology and offer more efficient management through firewalls. Refer to the

HP OVO HTTPS Agent Concepts and Configuration Guide for further details.

14.3.3 Functional Tests for Primary Communications

There are some cases when the node and the server may not communicate. OpenView makes extensive use of the configured name resolution services (DNS, NIS or a host file). Here are questions to ask when you think communication is an issue:

• Are all of the agents running?

• Can the node ping the server?

• Can the server ping the node?

• Can you verify the control agent with opcragt?

• Is the communication service running on the node?

• Are the communication processes running on the server?

• Do the server and the node resolve the same names and addresses for each other?

• Does the node have more than one network interface card? If so,

• Which interface IP address is configured to communicate with the server? And also,

• Which interface IP address for the node is configured on the server?

• Did the node communication type change (i.e. UDP to TCP)?

• Can you verify the path for the network end-to-end connections?

• Can you ping the node's HTTPS process?

If you have tried most of the functional tests and the communication still does not work,it is possible that the queue files have become unusable. As a last resort, it may be necessary to stop the agents, remove the queue files, and restart the agents. Refer to Chapter 22, "Troubleshooting Tips and Techniques," for more information on troubleshooting.