FlashCopy is an IBM feature supported on ESS (Enterprise Storage Servers) that allows you to make nearly instantaneous Point in Time copies of entire logical volumes or data sets. The HDS (Hitachi Data Systems) implementation providing similar function is branded as TrueCopy. Using either implementation, the copies are immediately available for both read and write access. Often times, the FlashCopied version of the data is used as input to a backup process allowing a time-consistent copy of the data to be written to tape.
There are three distinct stages to FlashCopy operations: Establish, Copy and Withdraw.
The FlashCopy Establish occurs very rapidly. It involves just setting up an additional set of pointers within the ESS that describe the FlashCopy relationship. Once the new pointers have been created, both the source and target can be referenced independently. For the application, this equates to the copy being immediately available (via the new pointers) even though no physical data has yet been copied.
There are two variants to the copy operation. The first is “COPY” which will begin a physical copy of the source data to the target once the pointers have been established. During the copy operation, if data is required from the target that has not yet been copied from the source, the data will automatically be provided from the source via the pointers without impacting either the application or the copy operation.
The second variant is “NOCOPY”. The NOCOPY operation does not copy the entire source data to the target, but rather, only copies source data that is about to be changed (on the source volume). Applications accessing the target data automatically access the source data for any information that has not changed since the FlashCopy Establish. For any source data that has changed since the FlashCopy Establish, the original data is read from the target. Both GDPS and FDRinstant invoke FlashCopy using the “NOCOPY” mode of operations.
The first implementation of FlashCopy, Version 1 allowed entire volumes to be instantaneously “copied” to another volume by using the facilities of the newer Enterprise Storage Subsystems (ESS).
Version 1 of FlashCopy had its limitations however. Although the copy (or “flash” of a volume occurred instantaneously, the FlashCopy commands were issued sequentially and the ESS required a brief moment to establish the new pointers. Because of this minute processing delay, the data residing on two volumes that were FlashCopied are not exactly time consistent. When FlashCopy is established for several hundred volumes, there is a finite amount of time - a minute or more - between the first and last established, so that copies will not be created at a consistent point-in-time.
FlashCopy Version 2 introduced the ability to flash individual data sets and more recently added support for “consistency groups”. FlashCopy consistency groups can be used to help create a consistent point-in-time copy across multiple volumes, and even across multiple ESSs, thus managing the consistency of dependent writes.
By using the Freeze FlashCopy Consistency Group option, the disk subsystem will hold off I/O activity to a volume for a brief time period by putting the source volume in an extended long busy state. In this way, a window is created during which dependent write updates will not occur and FlashCopy will use that window to obtain a consistent point-in-time copy of the related volumes. I/O activity resumes when a FlashCopy consistency group is created.
FlashCopy consistency groups are used in a single-site scenario in order to create a time-consistent copy of data that can then be backed-up and sent offsite, or in a multi-site Global Mirror for ESS implementation to force time consistency at the remote site.
The implementation of consistency groups is not limited to FlashCopy. Global Mirror for z/Series (formerly known as XRC or eXtended Remote Copy) also creates consistency groups to asynchronously mirror disk data from one site to another over any distance. In this implementation, changed data from the production site is time-stamped and asynchronously transmitted to the recovery site. Once at the recovery site, the records are journaled until a time-consistent copy of updates for all volumes are present. Once present, these changes are written to copies of the production volumes, and the process repeats.
This document was printed from http://recoveryspecialties.com/