HP 418800-B21 - StorageWorks Modular Smart Array 70 Storage Enclosure, User Manual

Page: 142 / 462

Each logical volume in a volume group is considered to be a component with which hosts can perform
I/O. Logical volumes can contain fi le systems or be raw storage. See raw disks .
LUN
In each storage system, logical unit numbers (LUNs) are assigned to its virtual disks. When a virtual disk
is presented to hosts, the storage system and the hosts perform I/O by referencing the LUN.
At a low level, a host OS typically reports each storage device that it detects in the format of C# T#
D#, where:
C #
Identi fi es a host I/O c ontroller
T # Identi fi es the t arget storage system on the controller
D #
Identi fi es the virtual
d isk (LUN) on the storage system
Automatic LUN assignment
• When presenting a virtual disk to a host, enter zero (0) to allow the storage controller software to
automatically assign the LUN.
Mirrorclones - fractured
Mirrorclone replication establishes and maintains a copy of a original virtual disk, via a local replication
link. See virtual disk types .
source
=====||=====
mirrorclone
local link (fractured)
When the local replication between a synchronized mirrorclone and its source is stopped by an action or
command, the mirrorclone is said to be fractured. In a fractured state, the mirrorclone is not updated
when the source virtual disk is updated. At the instant replication is stopped, the mirrorclone is a
point-in-time copy of its source. See also mirrorclone states and synchronized mirrorclones .
Task summary for fractured mirrorclones
Fractured mirrorclones
Deleting No. The disk must fi rst be detached, then deleted.
Detaching Yes.
Fracturing
Not applicable.
Presenting
Yes. The disk can immediately be presented to hosts for I/O.
Replicating - snapclones No.
Replicating - snapshots Yes. Multiple snapshots are allowed.
Restoring
Yes. The disk must be unpresented fi rst.
Resynchronizing Yes. The disk must be unpresented fi rst.
See also general mirrorclone guidelines and mirrorclone FAQ .
Mirrorclone write-cache fl ush
The source virtual disk write cache must be fl ushed before a fracture is started. (See cache policies
write cache .) This ensures that the source virtual disk and its mirrorclone contain identical data when the
fracture occurs. The following table shows how a write cache fl ush is implemented.
142
Host volumes