Hello everyone,

I am wondering if it is possible to mount a disk image that has been cut in blocks (here used to refer to pieces stored in separate files). The easy answer would be to:

Code:
 cat block1.img block2.img ... blockN.img > whole.img
and then to mount it by (at least in a simplified case):

Code:
 mount -o loop whole.img mountpoint
Concatenating the image block would be a costly operation for very large images. I would like to avoid that step.

Any ideas ?

thank you and best regards,
Anastasios



P.S. An explanation of why on earth i would want to do that follows:

The reason why i am asking this is because i am thinking of implementing a backup strategy involving taking an lvm snapshot of a running Ubuntu system and using storeBackup to make incremental snapshots to a network server. StoreBackup has the very useful option of cutting large files into blocks and propagating only the blocks that have changed to the backup target. The bad thing about it is that one ends up with a directory named whole.img containing the individual blocks which i do not know how to mount (say, for individual file recovery) without going through the costly operation of reconstructing the whole first.

And, yes, i could use storeBackup to simply backup the files in the lvm snapshot in the first place but i am a little twitchy about special files (say, sockets, pipes hardlinks and stuff) and about recovering in the case of catastrophic failure (in which case simply dd-ing an image file is a better recovery strategy)

Alternative strategies are also welcome, of course

thanks again,
Anastasios