doing something wrong by not registering our workstations with a DNS?
I have been searching through the documentation today without finding anything
relevant on the subject. When our PC's are configured via DHCP to use our
caching only DNS, will they start looking up machines on our local network using
DNS and forget everything about WINS and broadcast?
It worked just perfectly for nearly a year with no DNS at all. And the problem
first appeared long before we got our internet connection. But the problem
became really anoying at about the same time as we got our internet connection
and the PC's were configured to use a DNS...
Following is just a repetition of all log entries for a failed logon attempt, so
that you have it all in one place. (Is that 10s delay a failed DNS lookup? It
should not be, because the host resolv order is set to "wins bcast".
The machine
also appears to have its entry in wins.dat all right, so I see no reason why the
name lookup should fail?)
log.nmb:
[2000/11/27 10:35:57, 1] nmbd/nmbd_processlogon.c:process_logon_packet(70)
process_logon_packet: Logon from 192.168.1.16: code = 0x0
log.smb:
[2000/11/27 10:36:07, 1] lib/util_sock.c:client_name(1007)
Gethostbyaddr failed for 192.168.1.16
log.pc8:
[2000/11/27 10:36:07, 0] lib/util_sock.c:write_socket_data(540)
write_socket_data: write failure. Error = Broken pipe
[2000/11/27 10:36:07, 0] lib/util_sock.c:write_socket(566)
write_socket: Error writing 4 bytes to socket 6: ERRNO = Broken pipe
[2000/11/27 10:36:07, 0] lib/util_sock.c:send_smb(754)
Error writing 4 bytes to client. -1. Exiting
Svein Roar Nilsen
> the behavior you describe is pretty bad. Could you send me
> you smb.conf file? wrt the 'write errors', check for network
> layer error (such as packet lose, etc...) as well as DNS
> problems (bad reverse name resolution, etc...)
