samba - May 2003 - chrooted smbd crash

When running smbd chrooted it crashes when clients attempt to browse shares 
with the following error (debug level 10):

switch message SMBtrans2 (pid 13594)
setting sec ctx (22078, 505) - sec_ctx_stack_ndx = 0
1 user groups:
505
change_to_user uid=(0,22078) gid=(0,505)
vfs_ChDir to /home/tester
call_trans2qfsinfo: level = 1007
==============================================================INTERNAL ERROR:
Signal 11 in pid 13594 (2.2.7-security-rollup-fix)
Please read the file BUGS.txt in the distribution
==============================================================PANIC: internal
error
Aborted

smbd.conf contains:

[global]
workgroup = HOME
server string = no1v2.com
log file = /var/log/samba/%m.log
max log size = 0
security = user
obey pam restrictions = no
interfaces = lo
bind interfaces only = yes
dns proxy = yes
preserve case = yes
short preserve case = yes
[homes]
comment = Home Directory
browseable = no
writable = yes
valid users = %S
create mode = 0644
directory mode = 0755

I am running the RedHat RPM of Samba 2.2.7 (samba-2.2.7-3.7.2) on Linux 
2.4.9 and am connecting to it through a SSL tunnel from Windows 2000 Pro 
SP3 with sending encrypted SMB passwords disabled to match samba's 
configuration.  The same files and configuration work perfectly on the same 
server when not chrooted (all files were copied from their RPM-installed 
paths to the chroot environment).

The chrooted environment already existed before I added samba to it.  The 
samba-specific files and directories in it are:

/etc/pam.d/samba
/etc/samba
/etc/samba/lmhosts
/etc/samba/smb.conf
/lib/libnss_winbind.so
/lib/libnss_winbind.so.2 -> libnss_winbind.so (symlink)
/lib/libnss_wins.so
/lib/libnss_wins.so.2 -> libnss_wins.so (symlink)
/lib/security/pam_winbind.so
/usr/lib/samba/vfs
/usr/lib/samba/vfs/recycle.so
/usr/sbin/smbd
/usr/share/samba
/usr/share/samba/codepages
/usr/share/samba/codepages/codepage.1125
/usr/share/samba/codepages/codepage.1251
/usr/share/samba/codepages/codepage.437
/usr/share/samba/codepages/codepage.737
/usr/share/samba/codepages/codepage.775
/usr/share/samba/codepages/codepage.850
/usr/share/samba/codepages/codepage.852
/usr/share/samba/codepages/codepage.857
/usr/share/samba/codepages/codepage.861
/usr/share/samba/codepages/codepage.866
/usr/share/samba/codepages/codepage.932
/usr/share/samba/codepages/codepage.936
/usr/share/samba/codepages/codepage.949
/usr/share/samba/codepages/codepage.950
/usr/share/samba/codepages/unicode_map.1125
/usr/share/samba/codepages/unicode_map.1251
/usr/share/samba/codepages/unicode_map.437
/usr/share/samba/codepages/unicode_map.737
/usr/share/samba/codepages/unicode_map.775
/usr/share/samba/codepages/unicode_map.850
/usr/share/samba/codepages/unicode_map.852
/usr/share/samba/codepages/unicode_map.857
/usr/share/samba/codepages/unicode_map.861
/usr/share/samba/codepages/unicode_map.866
/usr/share/samba/codepages/unicode_map.932
/usr/share/samba/codepages/unicode_map.936
/usr/share/samba/codepages/unicode_map.949
/usr/share/samba/codepages/unicode_map.950
/usr/share/samba/codepages/unicode_map.ISO8859-1
/usr/share/samba/codepages/unicode_map.ISO8859-13
/usr/share/samba/codepages/unicode_map.ISO8859-15
/usr/share/samba/codepages/unicode_map.ISO8859-2
/usr/share/samba/codepages/unicode_map.ISO8859-5
/usr/share/samba/codepages/unicode_map.ISO8859-7
/usr/share/samba/codepages/unicode_map.ISO8859-9
/usr/share/samba/codepages/unicode_map.KOI8-R
/usr/share/samba/codepages/unicode_map.KOI8-U
/var/cache/samba
/var/log/samba
/var/run/samba
/var/spool/samba

Am I missing something that's causing the error?  If not, what is causing 
it?

~

Matt Blecker

Btw, how do you relogin to a SMB share as a different user in Windows 2000 
without logging out locally?

On Thu, May 15, 2003 at 12:36:58AM -0400, Matt Blecker
wrote:
> When running smbd chrooted it crashes when clients attempt to browse shares
> with the following error (debug level 10):
> 
> switch message SMBtrans2 (pid 13594)
> setting sec ctx (22078, 505) - sec_ctx_stack_ndx = 0
> 1 user groups:
> 505
> change_to_user uid=(0,22078) gid=(0,505)
> vfs_ChDir to /home/tester
> call_trans2qfsinfo: level = 1007
> ==============================================================> INTERNAL
ERROR: Signal 11 in pid 13594 (2.2.7-security-rollup-fix)
> Please read the file BUGS.txt in the distribution
> ==============================================================> PANIC:
internal error
> Aborted
Can you get me a stack backtrace please ?

Thanks,

	Jeremy.