System Administration Guide: Network Interfaces and Network Virtualization

Recovering an IPMP Configuration With Dynamic Reconfiguration

This section contains procedures that relate to administering systems that support dynamic reconfiguration (DR).

ProcedureHow to Replace a Physical Card That Has Failed

This procedure explains how to replace a physical card on a system that supports DR. The procedure assumes the following conditions:

Before You Begin

The procedures for performing DR vary with the type of system. Therefore, make sure that you complete the following:

  1. On the system with the IPMP group configuration, assume the Primary Administrator role or become superuser.

    The Primary Administrator role includes the Primary Administrator profile. To create the role and assign the role to a user, see Chapter 2, Working With the Solaris Management Console (Tasks), in System Administration Guide: Basic Administration.

  2. Perform the appropriate DR steps to remove the failed NIC from the system.

    • If you are removing the card without intending to insert a replacement, then skip the rest of the steps after you remove the card.

    • If you are replacing a card, then proceed to the subsequent steps .

  3. Make sure that the replacement NIC is not being referenced by other configurations in the system.

    For example, the replacement NIC you install is bge0. If a /etc/hostname.bge0 file exists on the system, remove that file.


    # rm /etc/hostname.bge0
    
  4. Replace the default link name of the replacement NIC with the link name of the failed card.

    By default, the link name of the bge card that replaces the failed ce card is bgen, where n is the instance number, such as bge0.


    # dladm rename-link bge0 subitops0
    

    This step transfers the network configuration of subitops0 to bge0.

  5. Attach the replacement NIC to the system.

  6. Complete the DR process by enabling the new NIC's resources to become available for use.

    For example, you use the cfgadm command to perform this step. For more information, see the cfgadm(1M) man page.

    After this step, the new interface is configured with the test address, added as an underlying interface of the IPMP group, and deployed either as an active or a standby interface, all depending on the configurations that are specified in /etc/hostname.subitops0. The kernel can then allocate data addresses to this new interface according to the contents of the /etc/hostname.ipmp-interface configuration file.

About Missing Interfaces at System Boot

Certain systems might have the following configurations:

With the new IPMP implementation where data addresses belong to the IPMP interface, recovering the missing interface becomes automatic. During system boot, the boot script constructs a list of failed interfaces, including interfaces that are missing. Based on the /etc/hostname file of the IPMP interface as well as the hostname files of the underlying IP interfaces, the boot script can determine to which IPMP group an interface belongs. When the missing interface is subsequently dynamically reconfigured on the system, the script then automatically adds that interface to the appropriate IPMP group and the interface becomes immediately available for use.