Why do we treat netbios-name<1D> specially?

[Christopher R. Hertel]

Dave,

> "Alex H. Vandenham" <alex at avantel.com> wrote:
> | the question came up as a result of a problem we were having with 
> | multiple workgroups on multiple subnets across a VPN.

This has to do with errors in Microsoft's implementation of the NBNS
server (aka. WINS server).

See: http://msdn.microsoft.com/library/winresource/dnwinnt/S76B0.HTM
and read on...

According to the RFCs, the NBNS server is supposed to keep track of *all*
registered owners of a group name.  That is, each group name should have a
list of IPs associated with it and, when Machine A does a name query on
group name ZARQUON, it should get the entire list.  This, plus the
datagram service facilitates NetBIOS multicast.

Microsoft chose to store only the address 255.255.255.255 instead of the
list.  This is a nasty kludge.  The result is that NetBIOS group
multicasts are restricted to the local IP subnet. 

Microsoft themselves ran into trouble with this, so they came up with an
exception.  The "Internet group name" is name type 0x1C, and they store up
to 25 IP addresses.  The limit of 25 has to do with the size of a UDP
packet.  (In the RFCs, the size of the list was unlimited and a truncation
bit was set if the packet was too small for the full list.  The client
could then optionally negotiate a TCP connection to get the full list. 
Yes, really.)

Samba's implementation follows the Microsoft mechanism, wrong as it is. 
This causes problems for OS/2 systems, in particular, because they expect
correct behavior.  There is a patch available for Samba and I want to talk
to folks next week about applying it.  (Look in the Samba-Technical
archives for "Adding Netbios Datagram Distributor support to Samba".)  The
major test I want to try is to see what happens to Windows boxen in the
face of the corrected behavior.  The patch author (see the samba-tech
archives) claims no problems with Windows boxen. 

I don't think that the TCP/137 option needs to be supported, but I'm not
completely sure.  I note that Microsoft has TCP/137 and TCP/138 open in
some cases (and yes, I mean TCP not UDP).  Have they, do you think,
decided to use these ports for their own purposes?  Hmmm...

Chris -)-----

> "Alex H. Vandenham" <alex at avantel.com> wrote:
> | the question came up as a result of a problem we were having with 
> | multiple workgroups on multiple subnets across a VPN.
> 
> Alex had two subnets and two workgroups, the main one ("TELAVAN")
> with one Samba server on subnet "A" acting as the browse master 
> and the global WINS server.
> 
> On the other ("B") subnet there was a local browse master for
> the main subnet (if I recollect correctly) and the the 
> browse master for the second workgroup, "NEWGROUP".
> 
> Unlike the same configuration with NT servers, NEWGROUP wasn't 
> visible on the "A" subnet. Both workgroups, of course, were 
> visible on subnet "B".
> 
> Alex noticed that the WINS data was asymmetrical on the
> server on subnet "A":
> | "NEWGROUP#00" 956063217 255.255.255.255 e4R	-- Workstation
> | "NEWGROUP#1e" 956063217 255.255.255.255 e4R	-- Browser Service
> Elections
> | "TELAVAN#00" 956109916 255.255.255.255 c4R	-- Workstation
> | "TELAVAN#1b" 956109916 192.168.4.1 192.168.0.1 44R -- Domain Master
> Browser
> | "TELAVAN#1e" 956109916 255.255.255.255 c4R	-- Browser Service
> Elections
> 
> 	The 1b (domain master browser) didn't show up, 
> and the code around line 730 in nmbd_winsserver.c says
>  
>    /*
>     * Special policy decisions based on MS documentation.
>     * 1). All group names (except names ending in 0x1c) are added as
> 255.255.255.255.
>     * 2). All unique names ending in 0x1d are ignored, although a
> positive response is sent.
>     */
> 
> We both wonder if this is the right thing for the WINS server
> to do, and strongly suspect that not having a master browser
> for TELAVAN is what's making the behavior asymmetrical and
> incompatible
> with NT...
> 
> --dave
> 
> David Collier-Brown,  | Always do right. This will gratify some people
> 185 Ellerslie Ave.,   | and astonish the rest.        -- Mark Twain
> Willowdale, Ontario   | //www.oreilly.com/catalog/samba/author.html
> Work: (905) 415-2849 Home: (416) 223-8968 Email: davecb at canada.sun.com
> 


-- 
Christopher R. Hertel -)-----                   University of Minnesota
crh at nts.umn.edu              Networking and Telecommunications Services

    Ideals are like stars; you will not succeed in touching them
    with your hands...you choose them as your guides, and following
    them you will reach your destiny.  --Carl Schultz