nmbd won't run

[Jamie Stotz]

I am having a hard time getting samba 2.0.4b running on a Sparc20 running Solaris 2.5.1. I have swat
and smbd running, but not nmbd.  I have tried all the stuff in diagnostics.txt and have searched the
archives but am still stuck. I have tried to gather some useful info, in the hopes that some on this
list will be able to figure this out. Thanks in advance.

Here are the results of some of the tests

1. 
The web interface to swat says that smbd is running but that nmbd is not. The button in swat that
starts nmbd doesn't work.

2. 
testparm says all is OK with the config file. There is not much in it anyway. 

3. 
Here are some results from the tests in the diagnostics.txt file.

cobra21#  /opt/samba/bin/smbclient -L cobra21_hme
Added interface ip=192.9.240.2 bcast=192.9.240.255 nmask=255.255.255.0
read_with_timeout: timeout read. read error = Connection reset by peer.
session request to COBRA21_HME failed
session request to *SMBSERVER failed

cobra21# /opt/samba/bin/nmblookup -B cobra21_hme __SAMBA__
Added interface ip=192.9.240.2 bcast=192.9.240.255 nmask=255.255.255.0
Sending queries to 192.9.240.2
name_query failed to find name __SAMBA__

I'm assuming these results are because nmbd isn't running.

4. 
With nmbd enabled in the inetd.conf file, I get this message when I restart inet.d with kill -1 <PID>

cobra21# Dec  3 23:44:40 cobra21 inetd[215]: config: 100083/rpc/tcp still active and was not
reconfigured. 

I don't understand this because the line that mentions 100083/rpc/tcp was already in the inetd.conf
file before I installed samba.

5.
I also get these two messages every few minutes

Dec  4 00:03:27 cobra21 inetd[215]: netbios-ns/udp server failing (looping), service terminated

and

Dec  4 00:02:45 cobra21 inetd[215]: /opt/samba/bin/nmbd: Child Status Changed

The looping message was addressed in a post in the archives, that talked about adding the -r option
to inetd. I made the suggested change:
/usr/sbin/inetd -s -r 300 60
and restarted inetd but it made no difference. I wasn't sure it would, because there are no PC's
trying to remount shares like in the example in the message.

Another post suggested that a udp service was being run as nowait. I checked, and none are. It also
mentioned something similar to the previous message, and suggested running nmbd as a daemon. 

6.
When I tried running nmbd as a daemon I got this message:

cobra21# /opt/samba/bin/nmbd -D
ld.so.1: /opt/samba/bin/nmbd: fatal: relocation error: symbol not found: srandom: referenced in /opt/samba/bin/nmbd
Killed

The man page for srandom implies that it is one of 3 system routines for generating random numbers
but doesn't say what file it is in. I can't find ld.so.1 anywhere on the disk. (See #8)

7.
With nmbd disabled or enabled in the inetd.conf file, smbstatus gives me this:

cobra21# /opt/samba/bin/smbstatus -d
Processing section "[shared_disk]"
using configfile = /opt/samba/lib/smb.conf
lockdir = /var/opt/samba/locks
Opened status file /var/opt/samba/locks/STATUS..LCK

Samba version 2.0.4b
Service      uid      gid      pid     machine
----------------------------------------------

ERROR: Failed to initialise share modes!
Can't initialise shared memory - exiting

8. 
Before I got swat and smbd running I was getting a message similar to #6 about a missing file:
libsec.so.1. I found it on another machine and after i copied it over, swat and smbd ran
successfully. I am wondering if the message in #6 is relating to another missing or damaged piece of
the OS.

9. Logging doesn't seem to be working. Just to be sure, I put a line in the config file to specify
the log level and logfile location, but no log files are being generated.

PS. If you want me to try something, please tell me what to type. I am somewhat new to Unix and
"reinitialize the knibbeling vector" won't mean as much as "reinit -k -v"  :-)


Thanks for any help you can give me. 
Jamie Stotz
EPP Systems Admin.
French Bray
jstotz at frenchbray.com