Troubleshooting Discovery

This section describes potential problems with server discovery and explains how to deal with these problems.

Discovery and Routers

Discovery of manageable servers using N1 System Manager works across routers if the network services used by the discovery process are not blocked by a firewall. Network services used by the discovery process can include SSH and SNMP.

RSC Server Discovery Problems

Manageable servers based on the Remote System Control (RSC) technology, such as Sun Fire V800 series servers, must be powered off before they can be discovered by the N1 System Manager. See Discovery of RSC Servers in Sun N1 System Manager 1.3 Troubleshooting Guide for details.

Discovering and Identifying Duplicate Servers

For managed servers that were discovered manually or by OS-based discovery, the N1 System Manager could discover the same server more than once. This duplication can happen in the following conditions:

• You discover a server manually. You then use SP-based discovery to discover another server but one of the platform MAC addresses of the second server matches a MAC address in the manual discovery file. A duplicate server has been discovered.

• You discover a server using OS-based discovery but its service processor is on a different subnet from its OS. You then use SP-based discovery to discover another server, but due to the conflict with subnets, the SP-based discovery command discovers the same server. A duplicate server has been discovered.

• Both of the above cases happen. Two duplicate servers have been discovered, and the same server appears three times.

Discovery of duplicate servers can lead to confusion and is not recommended. In addition, there is a risk that multiple attempts to provision an OS on the same managed server might occur simultaneously, or simultaneous attempts might be made to provision an OS on a server and power off the server.

If you use OS-based discovery or manual discovery to discover servers, use the detectduplicates utility to identify duplicate servers:

N1-ok> /opt/sun/n1gc/bin/detectduplicates
Name           Hardware       Discovered At  Network
manual1        V20z           -              File
manual2        V20z           -              File
192.168.79.2   V20z           192.168.79.2   Management
 
192.168.79.67  SF-T2000       192.168.79.67  Management
 manual3        T2000          -             File

In the output of the detectduplicates utility, duplicates are organized into groups, separated by a blank line. In this example, the detectduplicates utility has detected two groups of duplicates.

The output of the detectduplicates utility displays the following information:

• Name – The name of the server, as reported by the show server command.

• Hardware – The model of the server as reported by the show server command.

• Discovered At – The IP address that was used to discover the server. The IP address is reported as '-' for manually discovered servers.

• Network – The network that was used to discover the server. Possible values are:

  • Management – The management (service processor) network

  • Data – The provisioning network

  • File – Manually asserted

Server Details Information Is Missing

The N1 System Manager provides a limited set of features for managed server that were discovered manually or using OS-based discovery. Some server details might not be displayed in the Server details page for the server, or using the show server command. See Capability of Managed Servers Based on Discovery for more information about what capabilities are provided for servers depending on how they were discovered.

Identifying How a Managed Server Was Discovered

You can identify whether a managed server was discovered by OS-based discovery if in the server details section of the N1 System Manager browser interface or by using the show server command, the following states should be true:

• The connection section does not show a mgmtEth interface

• The management IP address is identical to the IP address used in the discover command when the server was discovered

• Power control capability is listed as unavailable

Some functionality might be absent, based on how the server was discovered. Table 4–1 lists supported and unsupported operations for managed servers based on how they were discovered. For managed servers discovered manually or by OS-based discovery, attempts to execute operations that are unsupported because of how the server was discovered are flagged by unsupported operation error messages.

Some servers support Remote System Control (RSC). For some models of RSC servers, the model number displayed by the N1 System Manager can depend on how the server was discovered.

• For Sun Fire V490 servers discovered manually or through OS-based discovery, the model name returned is SF-V490. If the Sun Fire V490 server was discovered through its service processor, the model name is returned as SF-RSC.

• For Sun Fire V890 servers discovered manually or through OS-based discovery, the model name returned is SF-V890. If the server was discovered automatically through its service processor, the model name is returned as SF-RSC.

See Discovering and Identifying Servers by Their Model Numbers for details.

Reprovisioning Servers That Were Discovered Manually or Using OS-based Discovery

There are several issues to be aware of when attempting to reprovision managed servers that were discovered manually or using OS-based discovery. Using the load server command, reset the SSH and management IP address if they have changed. For more information, see Sun N1 System Manager 1.3 Operating System Provisioning Guide.

When using the N1 System Manager to load an OS on managed servers that were discovered manually or using OS-based discovery, the manualnetboot feature must be turned on. For more information, see Sun N1 System Manager 1.3 Operating System Provisioning Guide.