Faulty Device Retirement Feature

Solaris Express Developer Edition 1/08: This release introduces a new device retirement mechanism to isolate a device as faulty by the fault management framework (FMA). This feature allows faulty devices to be safely and automatically inactivated to avoid data loss, data corruption, or panics and system down time. The retirement process is done safely, taking into account the stability of the system after the device has been retired.

Critical devices are never retired. If you need to manually replace a retired device, use the fmadm repair command after the device replacement so that system knows that the device is replaced, in addition to the manual replacement steps.

The fmadm repair process is as follows:

• Identify the faulted device with the fmadm faulty -a command.

  # fmadm faulty STATE RESOURCE / UUID -------- --------------------------------------------------------------------- faulty <fmri>

• Clear the fault by using the fmadm repair command.

  # fmadm repair <fmri>

• Run the fmadm faulty command again to be sure the fault is cleared.

  # fmadm faulty -a STATE RESOURCE / UUID

For more information, see fmadm(1M).

A general message regarding device retirement is displayed on the console and written to the /var/adm/messages file so that you aware of a retired device. For example:

Aug 9 18:14 starbug genunix: [ID 751201 kern.notice] NOTICE: One or more I/O devices have been retired

You can use the prtconf command to identify specific retired devices. For example:

# prtconf
.
.
.
pci, instance #2
        scsi, instance #0
            disk (driver not attached)
            tape (driver not attached)
            sd, instance #3
            sd, instance #0 (retired)
        scsi, instance #1 (retired)
            disk (retired)
            tape (retired)
    pci, instance #3
        network, instance #2 (driver not attached)
        network, instance #3 (driver not attached)
    os-io (driver not attached)
    iscsi, instance #0
    pseudo, instance #0 
.
.
.