HP StorageWorks 9100 Extreme Data Storage System administration guide V1.0.2 (AN540-96018, February 2010)

• The exds_stdiag reports that disks "have been ejected" from a logical disk – this indicates that
the disk drive is failed. The command also reports "SMART predicts failure" for disks in the predict-
fail state. The following show a small section of an exds_stdiagreport:
disk problems
* box 2 drive [12] missing or failed on storage block SGA90100CV
disk status problems
* SMART predicts failure on HP 1000.2GB SAS disk in box 3 bay 19 of SGA90100CV
lun problems
* box 2 drive [12] evicted, spares in box 1 drive [1] active on 7452GB ADG lun 06
on box 2 drive [11-20] of SGA90100CV
IMPORTANT:
Before replacing a failed hard disk drive, perform the following steps:
1. Check the global service indicator (GSI) light on the front panel of the hard drive drawer. If the
GSI light is lit amber, there is a problem with one of the enclosure components such as a power
supply, fan, or I/O module. Address that issue before replacing the hard disk drive. If the issue
is with the I/O module, and it is corrected, then the HDD LED light should turn green after the
HDD is replaced.
2. Before replacing a failed HDD with a new drive, try removing the drive, waiting two minutes,
and re-seating the drive. If it continues to report a failure, replace the drive. Do not attempt to
re-seat a disk drive in the predict-fail state -- simply replace it.
It is always safe to remove a drive in the failed state. However, if the drive is in the ok or
predict-fail state, do not remove it if any LUN on that array is not in the ok state (that is, if LUNs
are degraded or rebuilding). This is because the RAID logic might need the data on that drive to
reconstruct data for another drive that has failed. Removing a drive in this situation could break the
logical disk and set the LUN to the failed state.
To replace a drive:
NOTE:
For best results, HP recommends that you replace disk drives with both the HP ExDS9100c and HP
ExDS9100cx powered on.
1. Identify the drive. To ensure that you have the correct array, use the exdsmgr set array com-
mand with the locator=on box=<box-number> bay=<bay-number> option to turn on
the locator LEDs of the array or drive. For example, to turn on the locator LED of bay 1, box 2
of array 3, use the following command:
set array 3 locator=on box=2 bay=1
See sections HP ExDS9100c (array controller with 12 disk drives) and HP ExDS9100cx (dense
JBOD with 70 disk drives) for more information on box and bay numbering.
2. Remove the drive. Wait two minutes.
3. Insert a replacement drive.
4. Wait two minutes, and then use the exds_stdiag command to verify that LUNs are rebuilding
(that is, recovery is in progress). No action is needed to trigger the HP ExDS9100cc controllers
to rebuild LUNs; they do this automatically. The first 4 LUNs on an array share the same set of
disk drives. Rebuilding (recovery in progress) happens on them one by one. So immediately after
9100 Extreme Data Storage System administration guide 113