| There are really two issues here so I will try to address both of them.
Yes it is possible for an HSZ to present all physical devices as one
large logical device. This does not necessarily mean that it is
desirable to do so. You may want to follow-up in the HSZ conference
but I have been told that HSZ performance is degraded if requests are
not distributed amongst multiple targets and LUNs (front end).
The other issue is the question of whether or not a single device LSM
disk group is a good thing. One issue that comes to mind is the number
of copies of the configuration database. A single device can only hold
two copies of a configuration database and the recommended policy is
that each diskgroup have between four and eight copies. It is also
possible that both copies of the configuration database will end up on
the same physical disk on the HSZ. This means that the loss of single
disk will result in the loss of the entire diskgroup although I guess
this would be true anyway since the loss a single physical disk would
make the entire logical device unavailable on the HSZ.
Lastly, creating one large logical device limits your reconfiguration
options. If you decide that you want to change the mirroring or
striping characteristics for one LSM volume you are pretty much out of
luck. You need to reconfigure the entire HSZ logical device because it
is all or nothing.
|