ghost machines, revisited

[Keith Warno]

OK,

Here is what we have, as far as I am aware of.

There are, hmm, a lot of class-C subnets here. :)  I will list them:

.1 - Samba WINS server is here.  Is domain/master/preferred browser
.2 - Samba master/preferred browser here.  There is one other samba box 
that is in its own domain which I use to test the PDC functionality, but 
it is still pointed to the WINS server in .1
.3 - Samba master/preferred browser here
.4 - Nothing here.
.5 - Nothing here except Win2000 Server used for Cisco VoIP.  It is in 
it's own workgroup and does not see any other SMB box, including the 
WINS box in .1
.6 - Mostly win2k clients here.
.7 - Nothing  here.

Now a bunch of nets routed by a another linux box:
.205 thru .215
Samba running here as well.  Preferred/master broswer, points to WINS 
server in .1 too.

So the end result is that, AFAIK:
a) All SMB clients point to one and only one WINS box in .1 ; every 
windoze client configured via DHCP is guranteed to point to the correct 
WINS server, and all of them are configured via dhcp.
b) I have confirmed who is the master browser in each net by way of 
nmblookup.  These are enforced using OS level = <some really high 
number> as has been the case for a year.


Now, if I could find out if there was a stray WINS server somewhere, 
this would help immensely.  Is there a tool I could use to do this?


Thanks for all the feedback thus far.