As I understand the PureDisk product, yes. The copy that goes to tape
will be the full data set. I believe the product is positioned save
space and time from the client and on the equivalent of a DSU. You can
run PureDisk along with NBU, so the stream of data that comes from the
client is inspected for duplicate data (on the media server), and then
space is saved. As a PD client, it inspects data at even smaller
chunks, but does the work at the client and sends the changes to the PD
repository.
I looked at the Commvault 7 documentation, and it was not obvious to me
what the contents of the final tape copy would be.
If I were going to use tape as a final resting place, and use slow cheap
disk as a de-dup location, I would like the tape copy to be the simplest
format possible for reading in the future. And it would be nice if
those archive tapes did not need a complex road map of re-dup chunks to
assemble the data. On a PureDisk server, you can eliminate duplicate
data from all your client images, because at any time you can get that
data back from the single instance which lives on your random access media.
I'm having a hard time conceptualizing the process by which a 100G
client de-duped down to a 20G backup tape would be processed for a
restore. The bits have to come from somewhere, which is why I'm a bit
confused on how Commvault is doing it. I need to read their docs again.
-Jon
> So, ultimately, the data copied to the tape will be the full amount of the
> data and not the deduped data? For example, the original 100GB deduped down
> to 20GB will be the full 100GB and not the 20GB? Or will it be the 20GB, with
> Netbackup referring back to the index and/or base data?
>
> Thanks,
> Michael
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
|