[Samba] Strange segfaulting bug in Samba, has something to do with logleve.

[webbie]

Hey.

I seem to have the strangest problem with samba.

I have been running samba on a total of four machine, without a problem.
Now with a new machine, it won't work.
UNLESS I set loglevel to 10 in smb.conf, now thats weird.

If I remove the loglevel parameter from the smb.conf (so it reverts to
default, what.. 5 ?) it crashes after just a minute or so of use.

Here is first a log from the server viewing a share, after about five
seconds the client segfaults without any obvious reason.

I tried setting the debug-level of smbclient to 10 by using "smbclient
//gw/north "password" -d10" and it gave me alot of information but NOTHING
about the segfault, it just said seg fault and died on me.

Here is the log anyway,


amos /home/webbie# smbclient //gw/north
added interface ip=192.168.1.1 bcast=192.168.1.255 nmask=255.255.255.0
Password:
Domain=[PHANATIC] OS=[Unix] Server=[Samba 2.2.3a]
smb: \> ls
  .                                   D        0  Tue Sep 17 15:27:58 2002
  ..                                  D        0  Sat Sep 21 12:27:47 2002
  lost+found                          D        0  Sun Sep 15 22:42:47 2002
  asf                                 D        0  Tue Sep 17 16:02:39 2002
  avi                                 D        0  Thu Sep 19 00:31:57 2002
  mpg                                 D        0  Tue Sep 17 16:01:55 2002
  pics                                D        0  Tue Sep 17 15:27:31 2002
Error in dskattr: SUCCESS - 0
smb: \> dir
SUCCESS - 0 listing \*
Error in dskattr: SUCCESS - 0
smb: \> Segmentation fault
amos /home/webbie#


I noticed that smbclient is crashing no matter what loglevel I set, but
the samba daemon does NOT crash when smbclient dies.


And then the log from smbd, doesnt make any sense to me.



[2002/09/22 21:52:51, 1] smbd/service.c:make_connection(615)
  galaxia (192.168.1.31) connect to service north as user webbie (uid=500,
gid=100) (pid 8463)
[2002/09/22 21:52:52, 0] lib/fault.c:fault_report(38)
  ===============================================================
[2002/09/22 21:52:52, 0] lib/fault.c:fault_report(39)
  INTERNAL ERROR: Signal 11 in pid 8463 (2.2.3a)
  Please read the file BUGS.txt in the distribution
[2002/09/22 21:52:52, 0] lib/fault.c:fault_report(41)
  ===============================================================
[2002/09/22 21:52:52, 0] lib/util.c:smb_panic(1064)
  PANIC: internal error
[2002/09/22 21:52:52, 1] smbd/service.c:make_connection(615)
  galaxia (192.168.1.31) connect to service north as user webbie (uid=500,
gid=100) (pid 8464)
[2002/09/22 21:52:53, 1] smbd/service.c:make_connection(615)
  galaxia (192.168.1.31) connect to service webbie as user webbie
(uid=500, gid=100) (pid 8464)
[2002/09/22 21:52:53, 0] lib/fault.c:fault_report(38)
  ===============================================================
[2002/09/22 21:52:53, 0] lib/fault.c:fault_report(39)
  INTERNAL ERROR: Signal 11 in pid 8464 (2.2.3a)
  Please read the file BUGS.txt in the distribution
[2002/09/22 21:52:53, 0] lib/fault.c:fault_report(41)
  ===============================================================
[2002/09/22 21:52:53, 0] lib/util.c:smb_panic(1064)
  PANIC: internal error



For now its running fine with loglevel 10, i guess I could set up a
cronjob to remove the logfile every 30 minutes or so (yikes does samba
spit out alot of info) but Id guess Id let you know.

	  Per Eriksson
	    rambo-style.
	      webbie at phanatic.net