Yes and no. In real life, people do this all of the time. In the
development and
support, everyone holds their breath because they know things can go
wrong that are too horrible to correct, so you must be prepared to
rediscover.
I say you should always be prepared to rediscover, so make a seedfile,
print
pictures of the cut/paste job, and document the process.
That said, moving within a subnet, or to a nearby subnet, and moving to
another
state are quite different things. Netview discovers (and organizes its data
base)
with itself at the center of the universe. I would not expect a move to
another
state to go well.
Tivoli Framework has its own dependance on address and name resolution.
There are documented procedures for letting it know about identity changes
before you make them. I've done this and it works, but I hear that it fails
if you
are changing the domain as well as the name.
Netview can accomodate a name or address change if you run
/usr/OV/service/reset_ci. That means netview will know where it is at after
such a move. (Remember that the netview daemons won't come up all
the way after a reboot with a new address, so you might want to disengage
netview from the boot process while you are making these changes).
But that still leaves the map. If the move is too far, you will probably
have
problems, and the problems might not be apparent right away.
I would look at it this way: AIX won't need to be reinstalled. Tivoli
Framework
won't need to be re-installed (well, you might want to avoid using a fully
qualified name if the domain would have to change). Netview won't have to
be reinstalled. You will PROBABLY have to rediscover the network and
re-cut/paste, that's all. Set aside one day for that. It goes fast if you
prepare
for it, and it is the most fun part of playing with Netview!
To prepare for rediscovery, 1) print pictures of any fancy submaps; and
2) dump a list of nodes from the database to act as a seedfile using
nvdbformat.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Hello,
Quick question, a customer would like to build a NetView server in a
current location on the Eastcoast but, intends to move the server to a
permanent location in the future. The location maybe in a different
state on a different subnet. In reading through the TME 10 NetView for
UNIX for Implementer's training guide, it states that Any change that
involves a move of the databases is NOT supported, and likely will not
work. This includes:
* Copying the databases to any other physical box.
* Changing the IP address such that the NetView server is on a new
subnet.
In either case, the databases MUST be completely cleared and the network
rediscovered.
Are the above statements true? Is their a workaround on moving the
databases?
THX in advance
Steyland White
Resident Engineer
Bay/Nortel Networks - "How the world shares ideas"
Steyland@baynetwork.com
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
|