Failed I/O Host Card

What Caused the Problem?

An I/O host card in one of the controllers is not functioning properly. The Recovery Guru Details area provides specific information you will need as you follow the recovery steps.

 Caution: Electrostatic discharge can damage sensitive components. Always use proper antistatic protection when handling components. Touching components without using a proper ground may damage the equipment.

Important Note

You will need to replace the controller that has the failed host card. The failed controller is listed in the Component Requiring Service field in the Details area.

Recovery Steps

If... Then...
Your storage subsystem has one controller Go to "Procedure for Storage Subsystems with One Controller."
Your storage subsystem has two controllers If there are any hosts connected to this storage subsystem that are NOT running a host-based, multi-path failover driver, stop I/O to the storage subsystem from each of these hosts.

Go to "Procedure for Storage Subsystems with Two Controllers."

Procedure for Storage Subsystems with One Controller

1 Select the affected controller in the physical view and then select the Controller >> Properties menu option. Check the replacement part number of the affected controller and ensure that you have a new controller with the same replacement part number.

If you do not have a controller with the appropriate replacement part number, contact your technical support representative.

2 Stop all I/O to this storage subsystem.

3

Turn off power to all power-fan canisters in the enclosure containing the failed controller.

4 Remove the affected controller.

Refer to the Enterprise Management Window (EMW) to view which management method you are using to manage this storage subsystem.

If... Then...
You are using In-Band management for ALL hosts attached to this storage subsystem Go to step 5.
You are using Out-of-Band management for ANY host attached to this storage subsystem Before you insert a new controller canister into the storage subsystem, you must update the DHCP/BOOTP server for each Out-of-Band managed host so that it will associate the new controller's hardware Ethernet (MAC) address with the DNS/network name and IP address previously assigned to the removed controller.

To update the DHCP/BOOTP server, find the entry associated with the removed controller and replace its Ethernet (MAC) address with the new controller's Ethernet (MAC) address. The controller's Ethernet (MAC) address is located on an Ethernet ID label on the controller canister in the form xx.xx.xx.xx.xx.xx.

When you are finished, go to step 5.

5 If necessary, insert the battery from the old controller canister into the new replacement controller canister. Consult your hardware documentation for details.

Make sure at least 1 minute has elapsed. Then, insert the new (compatible) controller canister firmly into place.

6 Turn on power to all power-fan canisters in the enclosure.

Wait until all drives have completed the spin-up process, and then go to step 7.

7 Select the affected controller in the physical view and then select the Controller >> Properties menu option.

Check the status of the controller. The controller should have an Optimal status.

8 Click the Recheck button to rerun the Recovery Guru. The failure should no longer appear in the Summary area. If the failure appears again, contact your technical support representative.

Procedure for Storage Subsystems with Two Controllers

1 Place the affected controller offline.
a Select the controller in the physical view of the Subsystem Management Window.
b Select Advanced >> Recovery >> Place Controller >> Offline.
c Complete the instructions in the dialog, then select Yes.
2 Read all of the following steps before taking any action.
a Click the Recheck button to rerun the Recovery Guru.
b Select the Offline Controller problem that is being reported in the Summary area.
c Complete the Recovery Steps in the Offline Controller problem to replace the controller.