I'm not completely sure if the Celerra compresses the data, but I know
that to play nicely with TSM I had to set up a
server_param server_2 -facility NDMP -modify snapsure -value 1  on the
Celerra's control station.

Which effectively told the Celerra to do a snapshot before sending the
data to TSM.

Now if that is the case, and it truly is a snapshot based off of
changed delta's but not the whole fileset, I could see that each
snapshot that is backed up is not able to deduplicate...

I don't have much experience with the way NetApp's do their backups
with NDMP but I'm wondering if it is the same way, and of course how
others are doing theirs and if it is indeed de-duplicating or not.

Thanks,
Brian

On Dec 3, 2009, at 2:50 PM, Kelly Lipp wrote:

If it's true that the Celera compresses then it probably won't look
like the previous version of the file so it won't deduplicate.  Then
the question becomes: maybe don't compress but rather send all the
bits and let it get deduplicated - will this result in better
storage utilization?

________________________________________
From: ADSM: Dist Stor Manager [ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Wanda Prather [wanda.prather AT JASI DOT COM]
Sent: Thursday, December 03, 2009 10:00 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: [ADSM-L] TSM 6.1 De-dupe + NDMP

Is the NDMP backup going via TCP/IP and into the same storage pool
as your
other backup data?
How can you tell it is not getting deduped?


On Thu, Dec 3, 2009 at 11:54 AM, Brian P. Boyd <boydbria AT duke DOT edu>
wrote:

Hello,
Currently we are using TSM 6.1 to do NDMP backups of our EMC Celerra.
We are getting no de-duplication results from these backups.  I'm
wondering if it is because the celerra is already sending a
compressed
image snapshot file to TSM and TSM just doesn't de-dupe a file like
that.  We're thinking about going to a more traditional backup method
for NDMP backups (going with weekly incrementally, and monthly
fulls...etc) however it would be good to understand why this is
happening.  I find it a bit strange because we still do file-level
restores just fine...

If someone could educate me a bit more that would be great for piece
of mind!

Thanks!
Brian P. Boyd
Sr. SAN Admin
DUKE - OIT
boydbria (at) duke (dot) edu