I've been waiting to respond here, hoping somebody would chime in who knows the 
answer for sure to your first question.
I'm glad you've brought it up, 'cause it's a good question, I know there are 
others (including me) who are going to run into the "get a monthly copy" 
requirement.  

I'm thinking it's POSSIBLE it will work, and here's why:  
The doc says you have to create a management class for control information and 
make sure that stays on disk, even if your VE backup pool normally gets 
migrated out to tape.  Before an incremental/CBT backup runs, TSM does a 
restore that looks like this in the activity log:

03/31/2013 20:01:13      ANR0504I Session 37241 for node NNNNNNNN(Userid=),
                          restored or retrieved Backup object: node NNNNNNNN
                          filespace \VMFULL-SASW03, object
                          \SASW03\SNAPSHOT_000003000_20130328201241\Hard Disk
                          1\JOB000003001\MBLK03ee.CTL, version 1 of 1. (SESSION:
                          37241)
If TSM is retrieving the snapshot info that is stored, is that perhaps the way 
it is making the comparison for CBT?  
If that's how it works, then perhaps it could indeed maintain 2 separate 
tracks, if backed up to 2 different "datacenter" nodes.

But I don't really know, I'm just providing this info in hopes of getting 
somebody else to chime in with info.  
Maybe somebody going to Edge could try and find out for sure? 

As for your second question about Oracle, I do know that the way the new VE 
"application aware" backups work for MSSQL is by invoking Microsoft VSS to 
flush the buffers before the VE backup.  That isn't going to happen with any 
non-Microsoft application.

However you back up Oracle now, you should be able to continue doing.  The VE 
backup is run on a normal TSM schedule; I don't see any reason you can't do a 
presched/postsched arrangement to put Oracle in backup mode.  But either for 
Oracle or SQL, the TDP is the only thing that is going to give you log 
management along with the backups.

W



-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of 
Steven Harris
Sent: Thursday, March 28, 2013 10:07 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: [ADSM-L] TSM for VE questions

HI All

Bear with me please, I'm just getting my head around TSM for VE and changed 
block tracking.

I have a government customer that is subject to the Audit Act and so is very 
sensitive about long term data retention.  They are also very keen on 
virtualized Intel environments and have a large number of Windows and Linux 
boxes virtualized.  They basically don't want any new physical hardware in this 
space.

Currently on VSphere 4 we have TSM clients installed on all the VMs.
Now moving to VSphere 5.1 we want to use TSM for VE 6.4, changed block tracking 
and Flashcopy for VMware.

First question.  There is a point-in-time monthly-for-seven-years requirement.  
In the conventional TSM environment we run a daily incremental and once a month 
an incremental to a "monthly" node using -asnode.  Is it possible to do 
something similar with VE? Still maintaining an "incremental" philosophy?

Second.  The customer is standardizing on Oracle under Linux or SQL under 
Windows.  SQL we can handle with VE, but I can't see anything in any of the 
docs about Oracle.  Do we have to install a conventional TSM client and the 
Oracle TDP? I know that Oracle can use Flashcopy for Hardware, but can it 
similarly invoke Flashcopy for VMware?

As always.pointers to the right parts of the manual are appreciated if I've 
missed something.

Regards

Steve

Steven Harris
TSM Admin
Canberra Australia