No subject

Robert Dahlem Robert.Dahlem at
Tue Apr 28 13:51:28 GMT 1998

Hi there,

after running somewhat stable for a couple of days (since we started
using 1.9.18p4) tommorow morning nearly all our samba servers in
different subnets doing the remote browse sync thing together were some
kind of unresponsive: 

What I have seen with tcpdump was PCs broadcasting for their server
names but no answers from the servers.

Only one nmbd was running on those machines (we run them as daemons) an
in their log file we found soemthing like

   04/28/1998 08:31:34 netbios nameserver version 1.9.18p4 started
   Copyright Andrew Tridgell 1994-1997
   04/28/1998 08:31:36 become_domain_master_browser_wins: attempting to
become doma
      in master browser on workgroup XXX.DOM, subnet UNICAST_SUBNET.
   become_domain_master_browser_wins: querying WINS server at IP for dom
      ain master browser name XXX.DOM<1b> on workgroup XXX.DOM

   04/28/1998 08:31:36 ***** Samba server XXXBE001 is now a domain
master browser f
      or workgroup XXX.DOM on subnet UNICAST_SUBNET *****

   become_domain_master_browser_bcast: At time 04/28/1998 08:31:36
attempting to be
      come domain master browser on workgroup XXX.DOM on subnet
   become_domain_master_browser_bcast: querying subnet for
domain master
      browser on workgroup XXX.DOM

   04/28/1998 08:31:45 ***** Samba server XXXBE001 is now a domain
master browser f
      or workgroup XXX.DOM on subnet *****

   04/28/1998 08:31:49 *****   Samba name server XXXBE001 is now a
master bro
      wser for workgroup XXX.DOM on subnet *****

   error connecting to (Broken pipe)
   sync_browse_lists: Failed to start browse sync with XXXBA001

After stopping that still living nmbd and restarting it everything was
fine again.

It is perfectly possible that the link between subnet 160.18 and 160.19
was broken at this moment caused by a leased line maintenance done by
our line provider.

What still irritates me that we lost responsiveness of nmbd: a later
test showed that the nmbd process son is stable against SIGPIPE: the
father just goes on and starts another son.

The relevant parts of our smb.conf:

      workgroup = XXX.DOM
      remote announce = xxxdg001 xxxdg002 xxxba001 xxxba002
      remote browse sync = xxxdg001 xxxdg002 xxxba001 xxxxba002
      domain master = yes
      local master = yes
      preferred master = yes
      os level = 35
      name resolve order = host
      wins support = yes
      wins proxy = yes

Has anyone had this situation?  Is there a cure?


Robert.Dahlem at
Radio Bornheim - 2:2461/332 at fidonet +49-69-4930830  (ZyX, V34)
                 2:2461/326 at fidonet +49-69-94414444 (ISDN X.75)

More information about the samba mailing list