Installing or replacing an adapter with the system power turned on in a Virtual I/O Server

Find information about how to install or replace an adapter in the Virtual I/O Server logical partition.

Before you begin

The Virtual I/O Server includes a Hot Plug Manager that is similar to the Hot Plug Manager in the AIX® operating system. The Hot Plug Manager allows you to hot plug adapters into the system and then activate them for the logical partition without having to reboot the system. Use the Hot Plug Manager for adding, identifying, or replacing adapters in the system that are currently assigned to the Virtual I/O Server.

Prerequisites:
  • If you are installing a new adapter, an empty system slot must be assigned to the Virtual I/O Server logical partition. This task can be done through dynamic logical partitioning (DLPAR) operations.
  • If you are using a Hardware Management Console (HMC), you must also update the logical partition profile of the Virtual I/O Server so that the new adapter is configured to the Virtual I/O Server after you restart the system.
  • If you are installing a new adapter, ensure that you have the software required to support the new adapter and determine whether there are any existing PTF prerequisites to install. For information about software prerequisites, see the IBM® Prerequisite website.

About this task

Choose from the following tasks:

Installing an adapter

About this task

To install an adapter with the system power on in Virtual I/O Server, complete the following steps:

Procedure

  1. From the Hot Plug Manager, select Add a PCIe Hot Plug Adapter, then press Enter.
    The Add a Hot-Plug Adapter window is displayed.
  2. Select the appropriate empty slot from those listed, and press Enter.
    A fast-blinking amber LED located at the back of the server near the adapter indicates that the slot has been identified.
  3. Follow the instructions on the screen to install the adapter until the LED for the specified slot is set to the Action state.
    1. Set the adapter LED to the action state so that the indicator light for the adapter slot flashes
    2. Physically install the adapter
    3. Finish the adapter installation task in diagmenu.
  4. Enter cfgdev to configure the device for the Virtual I/O Server.

Results

If you are installing a PCIe, Fibre Channel adapter, it is now ready to be attached to a SAN and have LUNs assigned to the Virtual I/O Server for virtualization.

Replacing an adapter

Before you begin

Prerequisite: Before you can remove or replace a storage adapter, you must unconfigure that adapter. See Unconfiguring storage adapters for instructions.

About this task

To replace an adapter with the system power turned on in Virtual I/O Server, complete the following steps:

Procedure

  1. From the PCIe Hot Plug Manager, select Unconfigure a Device, then press Enter.
  2. Press F4 (or Esc +4) to display the Device Names menu.
  3. Select the adapter you are removing in the Device Names menu.
  4. In the Keep Definition field, use the Tab key to answer Yes. In the Unconfigure Child Devices field, use the Tab key again to answer YES, then press Enter.
  5. Press Enter to verify the information on the ARE YOU SURE screen. Successful unconfiguration is indicated by the OK message displayed next to the Command field at the top of the screen.
  6. Press F4 (or Esc +4) twice to return to the Hot Plug Manager.
  7. Select replace/remove PCIe Hot Plug adapter.
  8. Select the slot that has the device to be removed from the system.
  9. Select replace.
    A fast-blinking amber LED located at the back of the machine near the adapter indicates that the slot has been identified.
  10. Press Enter which places the adapter in the action state, meaning it is ready to be removed from the system.

Unconfiguring storage adapters

About this task

Before you can remove or replace a storage adapter, you must unconfigure that adapter. Storage adapters are generally parent devices to media devices, such as disk drives or tape drives. Removing the parent requires that all attached child devices either be removed or placed in the define state.

Unconfiguring a storage adapter involves the following tasks:
  • Closing all applications that are using the adapter you are removing, replacing, or moving
  • Unmounting file systems
  • Ensuring that all devices connected to the adapter are identified and stopped
  • Listing all slots that are currently in use or a slot that is occupied by a specific adapter
  • Identifying the adapter's slot location
  • Making parent and child devices unavailable
  • Making the adapter unavailable

If the adapter supports physical volumes that are in use by a client logical partition, then You can perform steps on the client logical partition before unconfiguring the storage adapter. For instructions, see Preparing the client logical partitions. For example, the adapter might be in use because the physical volume was used to create a virtual target device, or it might be part of a volume group used to create a virtual target device.

To unconfigure SCSI, SSA, and Fibre Channel storage adapters, complete the following steps:

Procedure

  1. Connect to the Virtual I/O Server command-line interface.
  2. Enter oem_setup_env to close all applications that are using the adapter you are unconfiguring.
  3. Type lsslot-c PCI to list all the hot plug slots in the system unit and display their characteristics.
  4. Type lsdev -C to list the current state of all the devices in the system unit.
  5. Type unmount to unmount previously mounted file systems, directories, or files using this adapter.
  6. Type rmdev -l adapter -R to make the adapter unavailable.
    Attention: Do not use the -d flag with the rmdev command for hot plug operations because this action removes your configuration.

Preparing the client logical partitions

About this task

If the virtual target devices of the client logical partitions are not available, the client logical partitions can fail or they might be unable to perform I/O operations for a particular application. If you use the HMC to manage the system, you might have redundant Virtual I/O Server logical partitions, which allow for Virtual I/O Server maintenance and avoid downtime for client logical partitions. If you are replacing an adapter on the Virtual I/O Server and your client logical partition is dependent on one or more of the physical volumes accessed by that adapter, then You can take action on the client before you unconfigure the adapter.

The virtual target devices must be in the define state before the Virtual I/O Server adapter can be replaced. Do not remove the virtual devices permanently.

Procedure

To prepare the client logical partitions so that you can unconfigure an adapter, complete the following steps depending on your situation.
Table 1. Situations and steps for preparing the client logical partitions
Situation Steps
You have redundant hardware on the Virtual I/O Server for the adapter. No action is required on the client logical partition.
HMC-managed systems only: You have redundant Virtual I/O Server logical partitions that, in conjunction with virtual client adapters, provide multiple paths to the physical volume on the client logical partition. No action is required on the client logical partition. However, path errors might be logged on the client logical partition.
HMC-managed systems only: You have redundant Virtual I/O Server logical partitions that, in conjunction with virtual client adapters, provide multiple physical volumes that are used to mirror a volume group. See the procedures for your client operating system. For example, for AIX, see Replacing a disk on the Virtual I/O Server in the Advanced POWER® Virtualization Best Practices Redpaper. The procedure for Linux® is similar to this procedure for AIX.For example, for AIX, see Replacing a disk on the Virtual I/O Server in the Advanced POWER Virtualization Best Practices Redpaper. The procedure for Linux is similar to this procedure for AIX.
You do not have redundant Virtual I/O Server logical partitions. Shut down the client logical partition.

For systems that are managed by the HMC, see Stopping a system by using the HMC.