ADSM-L

Re: [ADSM-L] Ang: [ADSM-L] Backupset restore question - different server OS's

2011-08-23 13:09:19
Subject: Re: [ADSM-L] Ang: [ADSM-L] Backupset restore question - different server OS's
From: "Allen S. Rout" <asr AT UFL DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Tue, 23 Aug 2011 13:06:25 -0400
On 08/23/2011 11:16 AM, Strand, Neil B. wrote:
Yes,  things tend to move quickly around here and I might be attempting to use 
TSM in an unconventional manner.

I have to move about 90 TB of data from data center A to Data center B.  The 
data centers have dissimilar SAN storage and very limited WAN connectivity 
between each other.  Each server at the new data center will be a new system 
and needs to be seeded with data from the currently running system.  Some 
servers will only require a few GB of seed data while others will require 
several hundred GB and a few have several TB of seed data.  I have to have a 
data migration solution ready next week and should begin migration shortly 
after Labor Day.

My Plan in a nutshell:
1.      Generate a backupset on a TSM V5.5 AIX server on encrypted 3592 media 
at data center A
2.      Ship the encrypted media to data center B
3.      Define a backupset and generate a backupsetTOC(optional) from the 
encrypted 3592 media on the TSM v6.2 Linux server at Data center B. Data center 
B has the appropriate encryption keys to decrypt the media.
4.      Restore backupset data to the appropriate client from the TSM V6.2 
Linux server at Data center B
-       No TSM DB Upgrades required or planed.  The TSM V6.2 Linux server is a 
fresh install.


This turns on whether the backupset format is as architecture-agnostic
as the export format.

I would expect the export/import to work across architectures;  for the
backupset, I wouldn't know where to put my money.

I would say that it will either work reliably , or fail immediately.
You might attempt it with something sized less than a tape volume, (or
even just a few GB, and send it over the WAN) and find out.

The encryption et. al. are red herrings in the test process, I'd skip
it. (and then test with non-sensitive data, natch)


- Allen S. Rout

<Prev in Thread] Current Thread [Next in Thread>