HP Application Recovery Manager software A.06.10 Administrator's guide (March 2008)

Replica redundancy level
EVA supports creations of snapshots and snapclones which have different storage
redundancy level (RAID) than their source storage volumes. Vraid1 consumes the
most storage space, followed by Vraid5 and then Vraid0.
Standard snapshots and vsnaps cannot be created with more space-consuming RAID
level than their source volumes have. For details, see the following table:
Table 4 Redundancy levels for snapshots
Standard snapshot
or vsnap - Vraid 0
Standard snapshot
or vsnap - Vraid 5
Standard snapshot
or vsnap - Vraid 1
AllowedAllowedAllowedSource volume - Vraid 1
AllowedAllowedNot allowedSource volume - Vraid 5
AllowedNot allowedNot allowedSource volume - Vraid 0
If the RAID level of a source volumes is such that the specified snapshot RAID level
is not allowed, the session does not fail, but the created snapshot has the same RAID
level as its source volume.
Advantages
The selection of RAID level has a significant impact on the amount of storage
space required.
Snapclone normalization is faster if Vraid5 is used instead of Vraid1.
ZDB in HP-UX LVM mirroring environments
Your HP-UX LVM mirroring environment should be configured as follows:
All logical volumes inside a volume group must be created with the PVG-strict
allocation policy. Consequently, the mirrors are created on different PVGs.
As a best practice, different PVGs should be located on separate arrays.
Consequently, mirrors are created on separate arrays.
At least one PVG must contain a consistent mirror copy for all logical volumes of
the volume group.
During a backup, Application Recovery Manager first checks the status of all mirror
copies (see Figure 2 on page 33). Out of all consistent mirror copies (mirrors without
stale extents), one is backed up, preferably the one residing on a different array than
Administrator's guide 31