[EMAIL PROTECTED] wrote:
In a message dated 3/12/2006 12:19:13 P.M. Central Standard Time,
[EMAIL PROTECTED] writes:
It would have to be a very precise alien laser beam.
Murphy is an awfully good shot! Spent many hours recovering disk failures in
dual copy/PPRC where the primary failed during the process.
...
Dual copy and PPRC are both distinctly different from FlashCopy. Both
require physical movement of data for the copy to be valid and usable
and thus have a window that is orders of magnitude larger for Murphy to
hit.
FlashCopy is only within the same DASD Subsystem, and creating the
virtual copy (which may eventually become a physical copy) only requires
building some internal bit table in the subsystem controller to indicate
which tracks have been copied and must be read from the target device
(none initially) versus those that must be read from the source device
and flagging the devices as being in a FlashCopy relationship. Any
physical movement of data occurs later, if at all.
I would, however, concur that the copy is not "instantaneous".
Establishing the FlashCopy relationship for 300 - 400 drives takes us
somewhere on the order of a minute, so you must, for example, suspend
DB2 updates and some other update activities during the establishment
process to be sure that you have consistent volume backups that will
actually be usable for a Disaster Recovery. The establishment time is
short enough that this appears to affected users as a short-term period
of bad response rather than a system outage, and users that only need
read access to data shouldn't even see an interruption. If you
incorrectly assume that multi-volume FlashCopy is instantaneous and take
no precautionary actions to insure data consistency, you are likely to
end up with inconsistencies in your DB2 tables, catalogs out of sync
with datasets on other volumes, etc., etc.
Just initiating FlashCopy is obviously only part of a DR solution. To
have the data usable for an actual Disaster Recovery after loss of your
computer facility obviously requires that some physical movement of data
to an offsite location must have completed. Any disaster that
intervenes before this has occurred means you must recover from the
previous DR backups with potential loss of intervening transactions.
--
Joel C. Ewing, Fort Smith, AR [EMAIL PROTECTED]
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
