Segfault in Samba

[Matt LaPlante]

Binary package hint: samba

nmbd has been repeatedly crashing since I upgraded my motherboard and processor hardware (from a 32 bit duron to a 64 bit sempron). This is the only error I've encountered following the upgrade, so I don't think it's the fault of the hardware. The following is the crash output from one such crash:

The Samba 'panic action' script, /usr/share/samba/panic-action,
was called for pid 5988 (/usr/sbin/nmbd).

Below is a backtrace for this process generated with gdb, which shows
the state of the program at the time the error occurred.

If the problem persists, you are encouraged to first install the
samba-dbg package which contains the debugging symbols for samba
binaries. Then, submit the provided information as a bug report to Ubuntu.
For information about the procedure for submitting bug reports, please
see http://www.ubuntulinux.org/support/bugs/document_view

Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1".
[Thread debugging using libthread_db enabled]
[New Thread -1212098288 (LWP 5988)]
0xffffe410 in __kernel_vsyscall ()
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7dae6a3 in waitpid () from /lib/tls/i686/cmov/libc.so.6
#2 0xb7d56281 in strtold_l () from /lib/tls/i686/cmov/libc.so.6
#3 0x080d4b28 in smb_panic2 (why=0x8114018 "internal error",
    decrement_pid_count=1) at lib/util.c:1545
#4 0x080d4c59 in smb_panic (why=0x8114018 "internal error") at lib/util.c:1506
#5 0x080c17ac in sig_fault (sig=11) at lib/fault.c:42
#6 <signal handler called>
#7 find_workgroup_on_subnet (subrec=0x0, name=0x8136458 "WORKGROUP")
    at nmbd/nmbd_workgroupdb.c:177
#8 0x08076327 in write_browse_list (t=0, force_write=1)
    at nmbd/nmbd_serverlistdb.c:350
#9 0x0806261e in msg_reload_nmbd_services (msg_type=3001, src={pid = 0},
    buf=0xbfa72ee0, len=0) at nmbd/nmbd.c:314
#10 0x08062e88 in main (argc=) at nmbd/nmbd.c:597

[didier (did447-deactivatedaccount)]

hi,

 can you add:
- your /etc/samba/smb.conf if you've modified it? At least the first lines where interfaces is set.
- var/log/nmbd.log*, if you can try to find output with the same date than the core it'd be great (maybe around 7 AM ?)

From the source there's actually a bug in nmbd if no workgroups are available but fixing it will only make nmbd exit cleanly.

[Matt LaPlante (cybrmatt)]

Well, I seem to have found the problem in the process of gathering the requested info. The smb.conf had been set to eth0, when in the new motherboard, the NIC comes up as eth1. This still has to be a bug; I dont think a segfault is the desired behavior if an interface becomes unavailable, but suffice it to say fixing the interface config has fixed the crashing.

[didier (did447-deactivatedaccount)]

>The smb.conf had been set to eth0, when in the new motherboard, the NIC comes up as eth1. This still has to be a bug; I >dont think a segfault is the desired behavior if an
cf /etc/iftab why it's eth1 now.

[Chuck Short (zulcss)]

Hello,

I believe this is fixed in hardy. If possible could you please check.

Thanks
chuck

[Matt LaPlante (cybrmatt)]

Acked - it appears to catch the condition gracefully in hardy.

[2008/04/06 20:48:13, 0] nmbd/nmbd.c:main(711)
  Netbios nameserver version 3.0.28a started.
  Copyright Andrew Tridgell and the Samba Team 1992-2008
[2008/04/06 20:48:13, 0] nmbd/nmbd_subnetdb.c:create_subnets(245)
  create_subnets: unable to create any subnet from given interfaces. nmbd is terminating
[2008/04/06 20:48:13, 0] nmbd/nmbd.c:main(785)
  ERROR: Failed when creating subnet lists. Exiting.