Donald H. Bennett, Jr. wrote:
>Here is a problem which I hope someone can help me out on. We have three
>SCO Unix Systems in our office. Two are running samba but the third keeps
>on giving me core dumps. Here is a breakout of each computer.
[snip...]
>Computers 1 and 2 run samba with no problems at all. However, Computer 3
>gives me a core dump when ever I run the SAMBA startup program. Below is
>what I get from the samba log. The operating system on computer #1 and #3
>are the same. The ifconfig and route and basically the same except for the
>computers own IP number. All three computers are on a 8 port HUB with
>Two NT Servers (one Pent and one 486/50) and two Windows 98 machines
>(both Pents).
>[1999/07/28 13:47:16, 1] nmbd/nmbd.c:(684)
> Netbios namerserver version 2.0.5a started.
> Copyright Andrew Tridgell 1994-1998
>[1999/07/28 13:47:16, 0] lib/util_sock.c:(886)
> bind failed on port 137 socket_addr=0.0.0.0 (Address already in use)
>Can anyone offer suggestions. Am I doing something wrong with Computer #3
>or is there something about SCO 5.0.4 and Acer 9100 computer that SAMBA
>doesn't like. I added gcc to Computer #3 and recompiled samba only to
get
>the very same results.
Hi,
You do not have a hardware problem. It is a setup problem with the SCO on
computer #3.
Samba needs access to the NetBIOS ports 137 to 139 and as the message says
there is already a software that is claming the ports exclusively. There
are two SCO software packages that are doing so. One is the "PC
Connect"
package that comes along with SCO OS and the other is the AFPS (NT File and
Print Server). Both packages are installing a real NetBIOS stack on the
system. Samba can not co-exist with real NetBIOS stacks on a UNIX computer.
So check your SCO installation and remove the packages that you do not
need.
Best regards,
Axel Neumann