The GUIDs are intended to help uniquely identify a particular machine (for
reporting purposes), regardless of how many different node names are used
on the machine or which network adapter is used to connect to the TSM
server, or which TSM servers the nodes connect to. So if I use nodes
STORMAN, ANDY, and RAIBECK to connect to a TSM server from my desktop
machine, all three nodes will have the same GUID. Similarly, if STORMAN,
ANDY, and RAIBECK connect to multiple TSM servers, each server will show
the same GUID for those three nodes.

If you use the same node name on multiple machines via the NODENAME
setting, then the server will show the GUID from the machine last used.
The activity log will show when a GUID has been changed via message
ANR1639I, like this:

   ANR1639I Attributes changed for node RAIBECK2: TCP Name
    from STORMAN to AMR-LAPTOP, TCP Address from 9.11.153.65
    to 192.168.0.2, GUID from 39.e6.b6.51.42.a4.11.d7.a9.d0.-
    00.09.6b.12.cc.9a to 51.08.b4.21.81.a9.11.d7.9e.07.00.d0-
    .59.83.d5.82.

If you use the same node name on multiple machines, but specified via
-VIRTUALNODENAME from the secondary machines, then the information for the
node name will not be changed.

As for your question about "consequences", there really aren't any. The
GUID is out there as a possible reporting field, but otherwise you can
safely disregard it. There is no impact on TSM operation.

Regards,

Andy

Andy Raibeck
IBM Software Group
Tivoli Storage Manager Client Development
Internal Notes e-mail: Andrew Raibeck/Tucson/IBM@IBMUS
Internet e-mail: storman AT us.eyebm DOT com (change eye to i to reply)

The only dumb question is the one that goes unasked.
The command line is your friend.
"Good enough" is the enemy of excellence.




David Bronder <david-bronder AT uiowa DOT edu>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
05/12/2003 16:34
Please respond to David Bronder


        To:     ADSM-L AT VM.MARIST DOT EDU
        cc:
        Subject:        Re: Tivoli GUID



Richard Sims wrote:
>
> >This tells me what it is, but now where it's used for...
>
> The manuals don't come out and say what the GUID is for, but one can
> see its purpose...  These days, many hosts respond on multiple IP
addresses
> and network names (Virtual IPs).  Discerning the actual identity of the
> connecting system becomes problematic.  Assigning a unique ID within an
> application allows ready identification.

My question about the GUID has to do with "virtual nodes".  I have lots
of systems with two node names, one for filesystem backups and the
other for API backups (TDP, Backtrack).  I also have nodes used by more
than one physical server (a manual NFS cluster, basically).  Actually,
in the latter case, the servers are also like the former case -- they
use their own name for OS backups and the shared name for the "cluster".

What are the consequences of a GUID being used for multiple nodes, or
for a node's GUID getting flip-flopped depending on which physical
server is connected using that node name (if that's even what would
happen)?  (I'm aware of the little bullet reference to using multiple
physical machines with the same GUID in the client manual, but it gives
no info about the impact of doing so.)

=Dave

--
Hello World.                                    David Bronder - Systems
Admin
Segmentation Fault                                     ITS-SPA, Univ. of
Iowa
Core dumped, disk trashed, quota filled, soda warm.
david-bronder AT uiowa DOT edu