On a new installation of OSR 6.0.0, the smbd daemon does not daemonize
itself.

If I run 'mkdev samba', the script never terminates after announcing that
it is Enabling Samba.

If I reboot, /etc/rc2.d/messages/prc_sync records that S99smbd timed
out.

And if I type '/usr/sbin/smbd -D -s /etc/samba/smb.conf' from a root
prompt, it does not daemonize itself and I do not get another shell
prompt.

/var/adm/log/samba/log.smbd contains only three linesa:

[2005/09/20 13:18:41, 0] smbd/server.c798)
smbd version 3.0.13-2sco-SCO started.
Copyright Andrew Tridgell and the Samba Team 1992-2004


--
JP

Jean-Pierre Radley wrote:
> On a new installation of OSR 6.0.0, the smbd daemon does not daemonize
> itself.
>
> If I run 'mkdev samba', the script never terminates after announcing that
> it is Enabling Samba.
>
> If I reboot, /etc/rc2.d/messages/prc_sync records that S99smbd timed
> out.
>
> And if I type '/usr/sbin/smbd -D -s /etc/samba/smb.conf' from a root
> prompt, it does not daemonize itself and I do not get another shell
> prompt.
>
> /var/adm/log/samba/log.smbd contains only three linesa:
>
> [2005/09/20 13:18:41, 0] smbd/server.c798)
> smbd version 3.0.13-2sco-SCO started.
> Copyright Andrew Tridgell and the Samba Team 1992-2004

No help, but I'm using samba on OSR6 with no problem. You didn't say,
but I'll assume you've installed MP1. Have you tried running under
truss to see where it's hanging?

I get the same three lines in log.smbd (which is in /var/log/samba
here).

Roger Cornelius typed (on Tue, Sep 20, 2005 at 03:38:55PM -0700):
| Jean-Pierre Radley wrote:
| > On a new installation of OSR 6.0.0, the smbd daemon does not daemonize
| > itself.
| >
| > [hangs after running 'mkdev samba']
| > [hangs after running /etc/rc2.d/S99smbd]
| >
| > And if I type '/usr/sbin/smbd -D -s /etc/samba/smb.conf' from a root
| > prompt, it does not daemonize itself and I do not get another shell
| > prompt.
|
| No help, but I'm using samba on OSR6 with no problem. You didn't say,
| but I'll assume you've installed MP1. Have you tried running under
| truss to see where it's hanging?

Sorry, yes, MP1 is applied.

Thanks for the suggestion!
Running 'truss /usr/sbin/smbd -D -s /etc/samba/smb.conf' gets into a loop:

open("/etc/resolv.conf", O_RDONLY, 0666) = 4
ioctl(4, TCGETS, 0x080467A4) Err#25 ENOTTY
fxstat(2, 4, 0x080467E4) = 0
brk(0x082AD8B0) = 0
read(4, " h o s t r e s o r d e r".., 8192) = 184
read(4, 0x082A9510, 8192) = 0
close(4) = 0
xsocket(UW71|XPG4, AF_INET, SOCK_DGRAM, 0) = 4
xconnect(UW71|XPG4, 4, 0xBFD74A04, 16) = 0
xsendmsg(UW71|XPG4, 4, 0x08045F48, 0) = 45
poll(0x08045D3C, 1, 5000) (sleeping...)
poll(0x08045D3C, 1, 5000) = 0
close(4) = 0
xsocket(UW71|XPG4, AF_INET, SOCK_DGRAM, 0) = 4
xsendmsg(UW71|XPG4, 4, 0x08045F40, 0) = 45
poll(0x08045D3C, 1, 5000) (sleeping...)
... und so weiter...


Now why in the world is it checking /etc/resolv.conf? Is my problem
due to the fact that this new machine has a fake hostname and domain,
which do not appear anywhere in real DNS tables, but solely in its own
/etc/hosts file?

That does seem to the problem. I followed some suggestions in

http://aplawrence.com/Bofcusm/713.html

and in

http://www.oreilly.com/catalog/samba...k/ch07_03.html

and inserted these two lines into the smb.conf file:

name resolve order = wins
wins server = [the_local_machine_name]

After rebooting, lsof confirms that nmbd/smbd are listening on ports 137,
138, and 139.

When I told 'mkdev samba' that I did want the Unix machine to be its own
WINS server, why didn't it insert such lines for me?

Thanks again, Roger!

--
JP