Actually, I think the results will be the same whether you use Andy's
syntax, or mine, assuming there have been NOTHING BUT incrementals done
since the image.

Either way, my point is that  there is no "magic" associated with a TSM
image restore followed by a TSM incremental restore.

I don't think it matters whether you start with a hardware image, or a
TSM image, or a system that never died in the first place.  When you
issue the dsmc incremental restore, it doesn't KNOW whether you
previously restored a TSM image or not.

Whatever TSM decides to do, it will do based on information in the TSM
data base.  It reads the data base to decide which files it wants, THEN
goes to position the tape (after first having sorted pointers and stuff
so that it mounts each input tape only once).

I don't remember what type of tape hardware you have, as to whether
positioning will be an issue.  But unless you have an unusual
application that changes the SAME set of files on this host every day,
no matter what you do it will involve positioning, because it is always
possibile you will be skipping over inactive files belonging to this
host to get to the active files on the tape where they reside.

What will be an issue for sure, is the RATE OF CHANGE on this
filesystem, as that will determine how many files you back up per day,
and therefore how many files you have to restore on top of the image.
If it's a low rate of change, this plan should work really well.  If
it's a high rate of change, maybe not...

Besides the rate of change, your restore will be impacted by your tape
hardware, PLUS your disk hardware, and your OS (I don't remember the
beginning of the thread as to whether this was WINdoze or AIX or what.)
So I don't think anybody will be able to give you a definitive answer,
without your setting up a test within your own environment.

Wanda







-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Paul Zarnowski
Sent: Monday, December 13, 2004 5:12 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: image restore with incrementals using flash copy


At 03:24 PM 12/9/2004, Andy Raibeck wrote:
>Wouldn't this work?
>
>    dsmc restore x:\ -subdir=yes -fromdate=mm/dd/yyyy
-fromtime=hh:mm:ss
>-replace=all
>
>where -fromdate and and -fromtime are the date/time that the image was
>taken.

And Wanda Prather wrote:
>Paul,
>
>If you used the flash-copy (hardware) image to do the restore, then did
>a TSM restore with the "ifnewer" parm, how would that be different than
>if you used a TSM image?
>
>W

This might be the answer I was looking for.  But I am concerned that we
get
optimal restore performance.  *If* TSM has to position tape, etc, prior
to
determining whether the file is newer or not, that would cost time.  Do
you
know if this is the case?  This is a filesystem with millions of files.

Thanks again.

..Paul