HP Network Node Manager 9 : Discovering and Monitoring Your Network - Examining discovery results (part 1)

Discovery is a continuous process. After the initial discovery is completed, the devices are periodically discovered for configuration changes so that NNMi can display the most accurate map. However, you may want to check the discovery status. There is no straight green or red indicator telling you if everything was discovered by NNMi, as it's not only about the amount of discovered nodes, but it is also about the device configuration and their connections with each other, including Layer 2 and Layer 3 connections.

The following table describes the main issues that can be caused by discovery and their possible solutions:

Result	Solution
Too many devices have been discovered	See seeds and auto-discovery filters.
Missing device or devices	See auto-discovery filters. Check connection to device (devices). Add into seed.
Discovered device/devices is/are not accurate	Check if SNMP traffic is not limited to device (devices). Rediscover the node.

There are several places where you can examine your discovery results:

• The overall situation can be found by selecting Help | About HP NNMi. Here you will find general information about the management server state, including state poller data. As you will see later on, this window may help you recognize some primary issues on NNMi performance.
  
• Node discovery state: Every node shows their discovery state in the inventory tab. It has three status states:
  • Newly created: This state indicates that the node is in the NNMi's database, but more information needs to be collected to determine the node's state and status.
  • Discovery completed: This state shows that node discovery is completed and all the information needed is already collected.
  • Rediscovery in process: This state shows that NNMi is updating the information about the node.

  To see in which state node discovery is, make the following changes (here is one of possible ways):

  Inventory | Nodes | select a node | Open (it is a small icon on the top left window corner)

• Seed success: There is an option to check whether seeds were discovered. To verify, navigate to Configuration | Discovery Configuration | Discovery Seeds. Discovery Seed Result presents the result of the seed discovery. If you find Node created, that indicates that the seed was discovered. The following table presents all possible values of Discovery Seed Results.

  The following table provides a list of Discovery Seed Result status values:

  Result	Description
  New seed	This state is shown when a new seed is entered. As soon as the discovery begins, this is changed to In progress. Troubleshooting: If this state doesn't change, please make sure that the discovery process is running. Try restarting the discovery process.
  In progress	This state is shown while discovery is in progress.
  Node created	Discovery seed is discovered.
  Node created (non-SNMP device)	Discovered device doesn't support SNMP. The device is added into the database, but doesn't provide any SNMP information.
  Node not created (duplicate seed)	The address or hostname provided in the seed is already in the database.
  Node not created (DNS resolution failed)	The hostname provided in the seed cannot be resolved by DNS. Please add records into DNS or check if any mistakes are in setting the hostname.
  Node not created (license expired)	NNMi rejects new discovered nodes after the license limit is reached. Please expand your license or review existing discovery, if any node can be deleted to make space in the license pool.
  Failed	Some NNMi internal error has happened. For more details, please check discovery log file. Unix: $OV_LOG/nnm/disco.0.0.log Windows: %OV_LOG%\nnm\disco.0.0.log

• Overall inventory: A list of discovered devices. Check discovered devices in NNMi's discovery inventory table. To verify, navigate to Inventory | IP Addresses. Verify that nodes that you expected to be discovered are listed in this inventory table. The following is a screenshot of the inventory window:
  
• Layer 2 and Layer 3 results: Layer 2 and Layer 3 views are separated. Layer 2 represents traffic across physical links in the network. It also provides information such as MAC address, VLANs, and its related information; while Layer 3 represents traffic routed by address and its related information, such as IP address. To be generic, Layer 2 is switch-related information and Layer 3 is router or switch router related information.

Layer 2 shows the physical network's connection inventory. This can be checked by following these steps:

1. Click on the Inventory workspace.
2. Click on Nodes and select the node in which you are interested.
3. Go to Actions | Layer2 Neighbor View and change the Number of hops accordingly, to expand or reduce number of hops that you want to be shown on a map.

VLAN information is also part of the Layer 2 inventory. You can check VLAN results by doing the following:

1. Click on the Inventory workspace.
2. Click on VLANs and select the VLAN in which you are interested.
3. Click on Open (icon on top left corner) to open the VLAN form and check the VLAN discovery results.

Layer 3 inventory results can be checked by following these steps:

1. Click on the Inventory workspace.
2. Click on Nodes and select the router.
3. Go to Actions | Layer3 Neighbor View and change the number of hops to expand or reduce map area.

If the results are not exactly what you expected, try one or more of the following:

• Check ordering of your Auto-Discovery rules.
• Check poll node for rediscovering inventory by clicking on Actions | Configuration Poll.
• Verify whether IP addresses are not listed in the Excluded IP Address filter.
• Manually add or remove the connection.

If the discovery results are not something you have expected and you need some adjustments, there are several ways to modify discovery accuracy. An example in real life could be if you are missing some devices, which have to be discovered, or a device that was reconfigured, displays the old configuration information. The following is a list of suggested solutions:

• Spiral discovery: As NNMi also uses information collected from neighbor devices, some information may be not accurate until the neighbor devices are polled and configuration data is collected. So, if you see some data that is inaccurate, which was recently discovered, be patient and allow NNMi to collect it. Depending on network size, 10-30 minutes would be a reasonable time.
• Scheduled discovery: NNMi periodically makes discovery in scheduled time intervals. If changes in your network appeared after your last scheduled discovery, you may want to wait for the next scheduled discovery. To check what intervals are configured on your system, go to Discovery Configuration and see the Global Control window for Rediscovery Interval and Sweep Interval (if any other option than None is selected in Ping Sweep configuration).
• Delete node: If you delete a node, NNMi forces node rediscovery. One of the ways in which we can use this method is that the IP address, which was assigned to one of the discovered devices, should be reassigned to some other device (no matter whether or not it was discovered before). If you have both devices discovered, it is a good idea to delete both of the devices and let NNMi rediscover them with updated information.
• Add or delete discovery seeds: Whenever a seed is added or deleted, NNMi forces the rediscovery of nodes in the seed list.
• Accurately detect interface changes: If your device doesn't show all interfaces or interfaces are not accurate, this can be because the node has interface renumbering and NNMi does not have accurate information. To fix this issue, follow the steps that need to be done are listed as follows:
• Add or delete connections: If you are using Frame Relay, ATM, or MPLS links between your WAN, your map may need a manual edition of connections. Connections can be added or deleted by the following sequence:
  • nnmconnedit.ovpl command with option -t should be initiated to generate an XML template.
  • The XML template is modified according to what connection needs to be modified.
  • nnmconnedit.ovpl command with option -f is initiated to load configuration into the NNMi database.

Example 1: If the connection between interface Customer_ABC, which is in Router_1 and interface WAN_LINK, which is in Customer_ABC_Router needs to be added, we do it as follows:

1. Run this command to create add.xml file:

   nnmconnedit.ovpl -t add
   

2. Open add.xml file, which was created after completion of this command, and modify the file accordingly:

   <connectionedits>
   <connection>
   <operation>add</operation>
   <node>Router_1</node>
   <interface>Customer_ABC</interface>
   <node>Customer_ABC_Router</node>
   <interface>WAN_Link</interface>
   </connection>
   </connectionedits>
   nnmconnedit.ovpl -f add.xml
   

Example 2: If the connection between interface Customer_ABC, which is in Router_1 and interface WAN_LINK, which is in Customer_ABC_Router needs to be deleted, we do following:

1. Run this command to generate the delete.xml file:

   nnmconnedit.ovpl -t delete
   

2. Open the delete.xml file, which was created after completion of this command, and modify the file accordingly:

   <connectionedits>
   <connection>
   <operation>delete</operation>
   <node>Router_1</node>
   <interface>Customer_ABC</interface>
   <node>Customer_ABC_Router</node>
   <interface>WAN_Link</interface>
   </connection>
   </connectionedits>
   nnmconnedit.ovpl -f delete.xml