Re: Discovery problems

1999-06-21 23:48:50
Subject: Re: Discovery problems
From: Leslie Clark <lclark AT US.IBM DOT COM>
To: nv-l AT lists.tivoli DOT com
Date: Mon, 21 Jun 1999 23:48:50 -0400
First, if the mask is wrong, you fix it in the IP configuration of the device
and then probably delete it from Netview and rediscover it, since Netview
cannot move things like that. It will report it in an event, though, and
in the log file.

Another thought. Sometimes I see this on AIX:  A subnet appears that you
don't expect. After a while you look in it and see several nodes, some of
which have good masks that should not put them there. You keep looking
and you finally find one node on that subnet that has a bad mask that causes
him to be there. That node caused the subnet to be created. The other nodes
got put there by magnetism, or friendship, or something. Some sort of
IP attraction to that first node. You fix the first problem and the rest of the
stay where they belong. If you were to delete the whole subnet, and check
it the minute it reappears, there would be only one node in it and that would
be the culprit. I cannot explain it, but I fix it by correcting the mask on the
first member, and deleting and rediscovering everything that was in it.


Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking

I checked an NT server with a class C mask and class B network address.
It has been discovered under a class B (144.156) network segment
symbol. Monitor->Network Config->Conf menu shows as
subnet mask (which is correct) so although it can see the right mask
its still putting it under wrong symbol. There are two machines like
that. The other interface shown is the MS loopback and I also selected
'do not discover loopback interface somewhere where it is asked' so
this shouldn't be a problem.

Other NT servers with class C mask and have been discovered in the
proper network sysbol (144.156.199).

How did you fix your bad masks problems??


--- Leslie Clark <lclark AT US.IBM DOT COM> wrote:
> It does create a network symbol for each subnet, if
> it can verify via
> snmp the mask configured on each interface. Use
> Monitor.. Network
> Configuration..Addresses to see what Netview is
> seeing. I uncover a
> lot of bad masks this way.
> Then it creates one segment per protocol within that
> subnet, where
> 'protocol' goes with the type of interface it finds.
> Cordially,
> Leslie A. Clark
> IBM Global Services - Systems Mgmt & Networking
> Why doesn't Netmon create a class C segment for all
> devices that have a
> class C subnet mask ( and have the
> same network address
> (e.g. 144.155.199.*). For some devices it does and
> for others it
> doesn't.
> I am using a seed file in which I give the DNS names
> of all my machine
> that I want discovered.
> Kashif
> _________________________________________________________
> Do You Yahoo!?
> Get your free @yahoo.com address at
> http://mail.yahoo.com

Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com

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