The rollup threshold is the number of devices that can become unavailable without affecting the agent’s status. The rollup threshold defaults to 1. For example, if you had a network test or non-production environment where you expected devices to go up and down regularly, you could increase the rollup threshold high enough to ignore the availability of these devices. This would keep the Agent group rollup status as Good (green). Rollup status thresholds are set per agent within the Agent Manager.

The Test Agent Credentials function is used to ensure the PCM server and Agent have the other's credentials (passwords) configured correctly. This function is available only for Agents configured for server-initiated connections.

Use the Agents tab to view additional information about Agents. Select the Agent Groups node or a specific agent group in the PCM server's navigation tree to display the Agents tab, which lists all Agents assigned to or detected by the selected Agent Group. Here you can quickly determine the operational status of the agent.

• The Custom Groups node is now located at the top of the navigation tree.
• Only devices in the region you have permission to view are displayed in custom group screens.
• The rules for creating custom groups and device/port assignments have changed, as explained in “Rules of Custom Groups”.
• Rules of Custom Groups.
  • Devices and ports/radios can belong to more than one group tree. A group tree is a top-level folder under the Custom Groups node, including all groups, subfolders, devices, and ports/radios beneath it.
  • Devices and ports/radios can only be assigned to leaf nodes. A leaf node is an end node in a branch of a group tree, which means that a leaf node has no child nodes.
  • Devices and ports/radios cannot belong to more than one leaf node in a group tree.

• Microsoft Windows XP Professional (SP1 or greater), Microsoft Windows 2000 (SP4 or later), Microsoft Windows 2003 Server.
• Windows Vista Business and Ultimate (32-bit) versions are supported on the remote client in PCM version 3.0.

	<installdirectory> PNB/server/config/ 

This file holds a list of all IP addresses that are "authorized" to access the PCM management server. Each line should contain 1 IP address - which can include wildcards. For example, the access.txt file might look something like this:

	192.168.*.*
	192.169.*.* 

If you are using PCM+2.3 and earlier, this requires a call to ProCurve Support for the engineer to release the Installation Identifier from the original hardware, which will allow new hardware to be attached to the Installation Identifier.

Networks with more than 750 managed devices should purchase an unlimited device license (J9293A).

HP ProCurve Network Immunity Manager 2.0 requires the purchase and installation of HP ProCurve Manager Plus 3.0 license.

HP ProCurve Identity Driven Manager 3.0 requires the purchase and installation of HP ProCurve Manager Plus 3.0 license.

When using PCM 2.x, the starting point used for Discovery is any network device that is reachable from the management server. For fastest discovery, a Core switch is recommended, rather than an Edge switch connected to primarily end-nodes.

Here are two common problems to be aware:

1. Not configuring an Action and an Alert. Both must be configured or the Policy will fail to execute.
2. In PCM 2.2 and later, the Policy Management configuration changes are disabled by default. In order for a configuration change to take place, the administrator must go to Preferences > Policy Management and in section Configuration Changes, select the radio button to “Log actions that would be taken by policies and allow device configuration changes.” Forgetting to select this radio button when creating new policies will cause the policy to fail to make any changes.

• Data Not Current—If the data is not current, the graphs and LED displays will not have information.
• Too Little Traffic on Network—If your network is carrying very little traffic at this time, the graphs may not indicate any traffic for sFlow data. You will get statistical polling on devices no matter how little traffic exists on the port. If there is no traffic, the reported values will be "0".
• One port (Port) is Gray—There may be a problem with this particular port. The data sampler may not be working, there may not be enough traffic on that port, or a device may have been disconnected from that port.
• Machine is Very Busy—The CPU may not be able to process the data because it is too busy.
• Switch is Very Busy—When an interconnect device becomes over-loaded, it may stop responding to traffic monitoring requests in order to execute its primary function of handling network traffic.

1. Make sure that the management station is set as a trap receiver on the devices.
2. Make sure there are no other processes bound to port 162. This includes the Windows SNMP Trap Receiver Service.
3. In the PCM preferences pane under the events node, uncheck "Ignore Unknown traps" just to see if the customer can receiver traps.
4. Check the server log files, PCM/server/logs/CS-err.log and PCM/server/SvrJCoreMain0.log to see if any errors are reported.