Manually Faulting a Disk Shelf Drive

For the user's role or exceptions, add the disk-fault BUI authorization or the allow_disk-fault=true CLI authorization under the hardware scope. For information about user configuration, including roles, authorizations, and exceptions, see Configuring Users in Oracle ZFS Storage Appliance Administration Guide, Release OS8.8.x.

Use the following procedure to manually fault a disk shelf drive, which is especially useful when a drive is causing a performance impact, but has not reached its fault threshold. This command-line-interface procedure, which is also available in the RESTful API but not in the BUI, allows a drive to be gracefully removed from the system. If the drive is in a pool with a configured spare drive, the spare drive will take over for the manually faulted drive.

To remove the manual fault, repair the associated drive problem or replace the faulted drive. The fault cannot be removed manually by setting the faulted parameter to false. To repair the faulted drive, go to the Maintenance Problems area of the software and follow procedures described in Repairing Active Problems. To replace a faulty drive, see Replacing a Faulty Disk Shelf Drive.

Note:

Do not perform a pool configuration operation while a disk firmware upgrade is occurring. To check if an upgrade is in progress, from the Maintenance BUI menu, select System, or navigate to maintenance system updates in the CLI.

1. Identify the drive to manually fault by running, from the BUI, such analytics as "Disk: Disks: Broken down by percent utilization" and drilling down to the individual disks, and "Disks: I/O bytes: Broken down by disk".

   For details, refer to the Oracle ZFS Storage Appliance Analytics Guide, Release OS8.8.x.

2. In the CLI, go to maintenance hardware and enter ls to list the system hardware components.

   hostname:> maintenance hardware
   hostname:maintenance hardware> ls
   hostname:maintenance hardware>
   Chassis:
                NAME        STATE  MANUFACTURER  MODEL
   chassis-000  hostname    ok     Oracle        Oracle ZFS Storage ZS7-2
   ...
   chassis-001  1235FM401W  ok     Oracle        Oracle Storage DE2-24P
   disk-000     HDD 0       ok     HGST          H7210A520SUN010T
   disk-001     HDD 1       ok     HGST          H7210A520SUN010T
   disk-002     HDD 2       ok     HITACHI       H109090SESUN900G
   ...

3. Select the chassis with the drive to be faulted and enter ls to list its properties and components.

   hostname:maintenance hardware> select chassis-001
   hostname:maintenance chassis-001> ls
   Properties:
              name = 1227FMT00E
           faulted = false
      manufacturer = Oracle
             model = Oracle Storage DE2-24C
              part = 7046842
            serial = 1227FMT00E
          revision = 001E
              type = hdd
               rpm = 7200
              path = 2
            locate = false
   
   Components:
             LABEL  STATE  MANUFACTURER  MODEL             SERIAL  RPM   TYPE
   disk-000  HDD 0  ok     HITACHI       H7230AS60SUN3.0T  001...  7200  data
   disk-001  HDD 1  ok     HITACHI       H7230AS60SUN3.0T  001...  7200  data
   disk-002  HDD 2  ok     HITACHI       H7230AS60SUN3.0T  001...  7200  data
   ...

4. Enter select disk.

   hostname:maintenance chassis-001> select disk

5. Enter the disk to be manually faulted, and enter ls to list its properties.

   hostname:maintenance chassis-001 disk> select disk-002
   hostname:maintenance chassis-001 disk-002> ls
   Properties:
             label = HDD 2
           present = true
           faulted = false
      manufacturer = HITACHI
             model = H7230AS60SUN3.0T
            serial = 001218RRV9GD--------YHJRV9GD
          revision = A6C0
              size = 2.73T
              type = data
               rpm = 7200
            device = c0t5000CCA01A9A47ACd0
         pathcount = 2
         interface = SAS
            locate = false
           offline = false

6. Set the faulted property to true.

   hostname:maintenance chassis-001 disk-002> set faulted=true
        faulted = true (uncommitted)

7. Enter commit.

   hostname:maintenance chassis-001 disk-002> commit