Copy on write

Immediately after the FlashCopy® operation, the original fileset files (the source data) and the FlashCopy images (the copy data) of the files in the fileset share the same data blocks; that is, nothing is actually copied, which makes the operation space efficient.

As soon as any updates are made to the actual fileset contents (for example, a client adds or deletes files, or updates the contents of files), the fileset is updated by an operation called copy on write, in which only the changed blocks in the fileset are written to a new location on disk. The FlashCopy image continues to point to the old blocks, whereas the source fileset will be updated over time to point to the new blocks.

For example, when you create a FlashCopy image, the image is a set of pointers back to the original data in the fileset.
Figure 1. FlashCopy image
This illustration shows that a FlashCopy image is a set f pointers back to the original data.
If two blocks were changed (S and E), one block was deleted (T) and a new block was written (P) in the actual fileset, the new blocks would be written, and the FlashCopy image continues to point to the original blocks, preserving the point-in-time copy.
Figure 2. Copy on write
This illustration shows that the FlashCopy image still points to the original data even after the source data has been modified.

Therefore, any access to the FlashCopy image accesses the data blocks as they existed when the FlashCopy image was created, and any access to the fileset itself accesses the new data blocks.

Parent topic: FlashCopy images

Library | Support | Terms of use | Feedback
(C) Copyright IBM Corporation 2003, 2004. All Rights Reserved.
IBM TotalStorage SAN File System v2.2