IBM AIX HACMP SG24-5131-00, Hardware User Manual

Page: 179 / 240

Cluster Management and Administration
161
• The new adapter must be of the same type or a compatible type as the
replaced adapter.
• When replacing or adding an SCSI adapter, remove the resistors for
shared buses. Furthermore, set the SCSI ID of the adapter to a value
different than 7.
8.3.3 Disks
Disk failures are handled differently according to the capabilities of the disk
type and the HACMP version you are running. Whether your data is still
available after a disk crash, and whether you will need down time to
exchange it, will depend on the following questions:
• Is all the data on the failed disk mirrored to another disk, or is the failed
disk part of a RAID array?
• Will the volume group stay online (Quorum)?
• Is the type of disk you are using hot-swappable?
8.3.3.1 SSA/SCSI Disk Replacement (RAID)
RAID arrays are typically designed for concurrent maintenance. No command
line intervention should be necessary to replace a failed disk in a RAID array.
Do the following steps in order to replace a disk that is a member of a RAID
array:
1. Remove the disk logically from the RAID array (for example with the
appropriate SMIT menu). Removing a disk from a RAID array is known as
reducing the RAID array. No more than one disk can be removed from an
array at one time.
2. Remove the failed disk and plug in the substitute disk.
3. Add the replacement disk logically to the RAID array. All information from
the original disk will be regenerated on the substitute disk. Once data
regeneration has completed on the new disk, the array will return to its
normal optimal mode of operation.
8.3.3.2 Disk Replacement (Non-RAID) before HACMP version 4.3
If LVM mirroring is used, some careful manual steps must be followed to
replace a failed SCSI or SSA disk:
1. Identify which disk has failed, using
errpt, lspv, lsvg, diags
.
2. Remove all LV copies from the failed disk (
rmlvcopy
).
3. Remove the disk from the VG (
reducevg
).