NT 4.0 (WS) refuses to join domain

Gero Treuner gero at faveve.uni-stuttgart.de
Fri Jul 3 00:15:45 GMT 1998


Hi!

I carefully have read the "NT Domain FAQ" and have the impression I have
done the described steps. But, still I get the message that NT doesn't
find the machine account (in german language because of german version).

The configuration is:

           World
             |
             |
   129.69.139.1/24 (muppet.faveve.uni-stuttgart.de)
   Linux 2.0.33 with libc5 (Slackware)
   Samba 1.9.19-prealpha(Snapshot at 1st July 1998)
   129.69.139.129/30 (muppet-if2.faveve.uni-stuttgart.de)
             |
             |
   129.69.139.130/30 (brainbug.faveve.uni-stuttgart.de)
   NT 4.0 Workstation with service pack 3 installed
         
The net for the NT machine (129.69.139.128/30) isn't real because the
routing is done via an static arp entry at muppet, but this should
interfere with connections beyond muppet only and therefore doesn't
affect NetBIOS protocol.

The following things does work:
- Connections of any kind from brainbug to any host in the internet
- Registrating Brainbug in the workgroup for which Samba is responsible
  (named 'NILI'), mounting SMB filesystems with encrypted password from
  Muppet on Brainbug.

But if I try to get Brainbug a member of the NILI domain, I get the
error message that the Brainbug machine account does not exist.
I assume nmbd is working right, because I can see all the things
described in the NT domain FAQ in the paragraph
  'You should get a wonderful message saying "Welcome to the
  SAMBA Domain." If you don't, then please first increase your debug log
  levels ... and you may also see a "Net Server Get Info" being issued on
  the srvsvc pipe. Assuming you got the Welcome message, go through the
  obligatory reboot (the NT box, not the Samba server).'
Hm, everything happens as described in the FAQ, execpt the message
"Welcome to the NILI domain".

There is a hint in the next section of the FAQ that inconsistent
network masks can produce this problem. I double checked that they
match, and tried a full class C subnet (192.168.1.0/24) instead of the 
sub-C subnet 129.69.139.128/30 to eliminate a possible problem with
netmasks not bound to a full byte.

Has anyone an idea what maybe goes wrong? I attach my smb.conf file
to this message, and the log files (log.nmb, log.smb, log.brainbug) are
available at http://muppet.faveve.uni-stuttgart.de/~gero/<filename>.


Gero


P.S.: Please respond with a Cc to me because I am not (yet) a member
      of the mailing list.
-------------- next part --------------
;
; /etc/smb.conf                                                         
;

; Knut und Gero, Tue Jun  9 04:00:12 MET DST 1998

[global]
   log file = /var/log/samba/log.%m
   lock directory = /var/lock/samba
   share modes = yes
; SMB machen wir nur auf dem Brainbug-Segment
   interfaces = 129.69.139.129/30
; durch Komma getrennte Liste von erlaubten Hosts
   allow hosts = brainbug.faveve.uni-stuttgart.de
   guest account = samba

   wins support = yes
   workgroup = NILI
   encrypt passwords = yes
;   announce as = NT

   local master = yes
;   preferred master = true
   domain master = yes
   domain logons = yes

   browseable = yes
   client code page = 437
   time server = true
;   security = user

[homes]
   comment = Home Directories
; Nicht browsable, da ohne Passwoerter ohnehin nur Fehler erzeugt wuerden
   browseable = no
   read only = no
   create mode = 0755


More information about the samba-ntdom mailing list