ADSM-L

Re: [ADSM-L] NBU user considering switch to TSM

2008-09-30 11:14:56
Subject: Re: [ADSM-L] NBU user considering switch to TSM
From: "Thorneycroft, Doug" <dthorneycroft AT LACSD DOT ORG>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Tue, 30 Sep 2008 08:12:01 -0700
With TSM, it doesn't really make sense to make two copies of a tape, since
things are managed at the object level. Copy Pools are used instead.

-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU]On Behalf Of
Paul Zarnowski
Sent: Tuesday, September 30, 2008 7:56 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: NBU user considering switch to TSM


At 06:58 PM 9/29/2008, steve sorenson wrote:
> >> 1. He could make simultaneous copies during the initial backup, but not
> >> during migration and while copying to the copy pool.
> >>
> >
> > TSM can migrate from random-access disk to serial (tape), simultaneously
> > copying to the copy pool(s).  I don't have first-hand experience with this,
> > but plan to soon.
>
>
>I was referring to the feature of creating two copies of one set of data
>simultaneously.  NBU calls this Inline Tape Copy, where NBU takes one backup
>stream and "twins" it, sending one stream to one device and another stream
>to another device.  It can do this either during the initial backup, or when
>copying backups after the backup is done, like when you're migrating data or
>making copies to go offsite.  The point is to read the data and transfer it
>in memory only one time, but have two copies when you're done.  I was told
>that TSM had a similar feature, but that it could only do this during the
>initial backup, not when copying data.  Do I understand this feature
>correctly?

In TSM, generally you have a disk storage pool in front of tape.  At least,
this is always the way I've used it, as this enables more concurrent backup
sessions and collocation on tape.  With progressive incremental, I don't
really see the need to write to two devices simultaneously as the backup
data is coming in over the network from clients.  The point is, if you want
to ensure that you've always got tape-resident data on at least 2 different
tape volumes, you can do that in TSM.  If you want to be sure that your
data is fully protected on disk, before it gets to tape, I suppose you
could do that with a disk subsystem that implements synchronous mirroring
of some sort.  If you want to have the incoming data be resident on both
disk and a tape synchronously, as it comes in, you can't do that.  I'm not
that familiar with NBU, but if it does not employ progressive incremental,
perhaps the desire to only have the data in memory once is more important?



--
Paul Zarnowski                            Ph: 607-255-4757
Manager, Storage Services                 Fx: 607-255-8521
719 Rhodes Hall, Ithaca, NY 14853-3801    Em: psz1 AT cornell DOT edu