SMISA ZDB for HP StorageWorks EVA in HP-UX LVM Mirroring Configuration

SMISA ZDB for EVA in HP-UX LVM Mirroring Configuration
Concepts
Chapter 1 7
HP-UX LVM and Mirroring
NOTE You need HP MirrorDisk/UX for mirroring.
Mirroring is the capability of storing identical copies of data in logical volumes, preferably on
separate disks. This redundancy has several advantages:
• The system can survive LVM disk crashes if you mirror the root file system and swap.
• Valuable data is available on more than one LVM disk thus providing high availability.
• If an I/O channel fails, LVM can recover the data from the duplicate source.
Mirror-write recovery mechanisms enable the system to synchronize data. Mirroring speeds
up read-intensive applications by enabling the hardware to read data from the most
convenient LVM disk, thus optimizing I/O. One copy of the data can be backed up while
another copy continues to run.
Single or Double Mirrors
• Single mirroring occurs when data is mapped from one logical extent to two sets of
physical extents on LVM disks.
• Double mirroring maps each logical extent to three sets of physical extents on LVM disks.
Sets of logical extents can map strictly to separate LVM disks or non-strictly to different areas
of the same disk, although this is not recommended.
Each copy of mirrored data maps to one logical volume. The number of logical extents remains
constant, but the number of physical extents used (and therefore occupied disk space)
changes, depending on the number of mirrored copies. Mirrored logical volumes must belong
to the same volume group; you cannot mirror across volume groups.
Physical Volume Group
Physical volume groups are subgroups of LVM disks (physical volumes) within a volume
group. An ASCII file, /etc/lvmpvg, contains all the mapping information for the physical
volume group, but the mapping is not recorded on disk. Physical volume groups have no fixed
naming convention; you might name them PVG0, PVG1, and so on. The /etc/lvmpvg file is
created and updated using the vgcreate, vgextend, and vgreduce commands.
I/O channel separation is particularly useful for databases because it heightens availability
(LVM has more flexibility in reading data on the most accessible logical extent), resulting in
better performance. If you define your physical volume groups to span I/O devices, you ensure