Nabble - Samba - General

Samba 3.x reports "not implemented" when Server 2008 SMB client requests FSCTL_GET_OBJECT_ID

2008-10-07T20:37:39Z

Hi Samba list,

I ran across this really bizarre issue and was hoping somebody would
be able to shed some further light on the issue.

If this is better directed to the samba technical list, please let me
know and I will post there instead.

Background
=========

I'm using CommVault Galaxy 7.0 SP4 for backup and decided to share
it's "IndexCache", which is a collection of files & folders, on a UNC
share from our SuSE Linux server.

However, during file I/O a call to Samba to request the
FSCTL_GET_OBJECT_ID fails which causes CV to fail. We are pursuing
things on the CommVault side but wanted to know if there could be done
from Samba side, so I investigated a bit further via Google, mailing
lists and trying different security settings (security = user .vs.
domain, oplocks on off, etc.), here's what I found:

SMB Server logs (Samba) (smbd.log)
===================

MA2 (1.2.3.4) connect to service galaxy initially as user
administrator (uid=xxxx, gid=xxx) (pid 5022)
[2008/10/08 10:15:13, 0] smbd/nttrans.c:call_nt_transact_ioctl(2463)
call_nt_transact_ioctl(0x940cf): Currently not implemented.
[2008/10/08 10:15:49, 1] smbd/service.c:close_cnum(1230)
MA2 (1.2.3.4) closed connection to service galaxy

SMB Client logs (Process Explorer, Server 2008 SP1 server)
============================================
(of interest is the NOT SUPPORTED lines)

226 10:15:17.4357270
AM archiveIndex.exe 6064 QueryEAFile
\\unix\galaxy\IndexCache\CV_Index\2\180\1223420221\tree.dat
0xC0000052
227 10:15:17.4360329
AM archiveIndex.exe 6064 QueryStreamInformationFile
\\brbackup1\galaxy\IndexCache\CV_Index\2\180\1223420221\tree.dat
SUCCESS 0:
::$DATA
228 10:15:17.4362604
AM archiveIndex.exe 6064 FileSystemControl
\\unix\galaxy\IndexCache\CV_Index\2\180\1223420221\tree.dat NOT
SUPPORTED Control: FSCTL_GET_OBJECT_ID
229 10:15:17.4364787
AM archiveIndex.exe 6064 FileSystemControl
\\unix\galaxy\IndexCache\CV_Index\2\180\1223420221\tree.dat NOT
SUPPORTED Control: 0x94160 (Device:0x9 Function:88 Method: 0)
230 10:15:17.4388151
AM archiveIndex.exe 6064 CloseFile
\\unix\galaxy\IndexCache\CV_Index\collects\66789.txt SUCCESS
231 10:15:23.3183477
AM archiveIndex.exe 6064 CloseFile
\\unix\galaxy\IndexCache\CV_Index\2\180\1223420221\tree.dat
SUCCESS

Further information on FSCTL_GET_OBJECT_ID from the MSDN pages
(http://msdn.microsoft.com/en-us/library/aa364570(VS.85).aspx) shows
that it's about getting the "object identifier" for the specified
file.

If anyone would be able to add some comments or provide some technical
direction, that would be much appreciated.

Regards,
Luke
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Re: smbclient kerberos issue

2008-10-07T19:57:47Z

Nope, it's got a real .com to it. The behavior was the hostname
returned the hostname and hostname -f also returned just the
shortname. If it had returned an error instead of just the hostname, I
think it would have been ok from my quick view of the Samba source.

On Tue, Oct 7, 2008 at 2:51 PM, James Zuelow
<James_Zuelow@...> wrote:

> This is just a guess:
>
> Does your domain end in .local?
>
> If so, avahi would hijack DNS lookups for anything like
> domain_controller.company.local -- and since your DC probably doesn't
> have Bonjour installed on it, it gets no answer and reports back with a
> host not found. Unfortunately that's a valid DNS response, so your
> system does not then fall back to regular DNS.
>
> James Zuelow....................CBJ MIS (907)586-0236
> Network Specialist...Registered Linux User No. 186591
>
>> -----Original Message-----
>> From:
>> samba-bounces+james_zuelow=ci.juneau.ak.us@...
>> [mailto:samba-bounces+james_zuelow=ci.juneau.ak.us@...
> .org] On Behalf Of Ryan Bair
>> Sent: Sunday, 05 October, 2008 10:44
>> To: Gerald (Jerry) Carter
>> Cc: samba@...
>> Subject: Re: [Samba] Re: smbclient kerberos issue
>>
>>
>> It seems like it was a problem avahi which mistakenly made its way
>> into my nsswitch.conf. After removing mdns4_minimal and mdns4, I
>> rejoined to the domain and everything works great. I'm a bit confused
>> as to how this caused the problem, but I'm very happy to have it
>> fixed!
>>
>> Thanks
>

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

opensuse 11 samba 3.2.4 ldap add machine

2008-10-07T16:08:02Z

Hi @ll,
i didnt worked on samba for a while now had setup a new domain contoller
on opensuse 11
i used the repositories from download.opensuse.org samba 3.2.4
i mostly copied stuff from working suse 10 samba ldap pdc
and configured /etc/nsswitch.conf smb.conf
in equal ways , testparm shows no bugs
as well as getent passwd etc does not,
populating ldap worked fine
fixing dbus boot stuff in /etc/ldap.conf by boot_policy soft etc
but i didnt got managed joinig the domain by the root user
with a new installed winxp serv pack3 german client
bug message means "no such user"

it looks like it haves problems finding the root user
adding another user and putting him in the domain admin with the latest
smbldaptools script group doesnt helped either.
samba client logs doesnt report root not to be not found

I just googeld around and found likly problems with opensuse 11
mostly telling to upgrade openldap and samba, but now i am on the latest
upgrade level, so before i just loosing more time , just wanna ask if there
are known problems with samba ldap on opensuse 11?
And if ther are known one ,how were they fixed

--
Best Regards

MfG Robert Schetterer

Germany/Munich/Bavaria
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

samba in wxp environment

2008-10-07T13:44:34Z

hello,

i have a linux machine running samba with these settings:

workgroup = ...
netbios name = ...
interfaces = lo eth1
bind interfaces only = yes
hosts allow = ....
smb ports = 139

local master = no
os level = 35
domain master = yes
preferred master = yes
wins support = yes
dns proxy = no
browseable = yes
....

the wxp computers have dhcp enable, the dhcp server has these settings:

option subnet-mask ...
option broadcast-address ...
option routers ...;
option domain-name-servers ....
option netbios-name-servers eth1_ip
....

After I start the samba server, everything (in terms of network
browsing) works fine for about 30 minutes, then
in my netowrk places there is only one computer, the linux server.
Nothing in the logs. I found a similar problem
here: http://lists.samba.org/archive/samba/2004-October/094102.html but
no answer.
My wxp machines were part of a windows NT domain, I dont know if that
matters.

any ideas?

TIA

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Gotten getent to work but uid off by 1

2008-10-07T11:15:56Z

I have gotten getent to work, but now the uid for files are off by one and
some of the gid are off by 4 to 24. So if I copy from one system to another
the file ownerships get changed all up.
Any work around for this?
idmap uid = 10000-20000
idmap gid = 10000-20000
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Another getent problem

2008-10-07T10:49:12Z

Thanks that solved the problem.
I wonder when this got changed. The original smb.conf was copied from
another system and modified for the new server.
So somewhere between
3.0.10
and
3.0.28
The change was made. I never saw this in the latest HOW-TO either.
Thanks maybe later today or tomorrow I would have thrown the kitchen sink at
this and finally found the problem, so thanks for saving me a bunch of time.

On Tue, Oct 7, 2008 at 12:02 PM, Gerald (Jerry) Carter <jerry@...>wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Clark Johnston wrote:
> > I have set up a system to be a member server and installed the samba
> > rpms. I then copied over the samba config file and changed it to
> > reflect the new shares and name change. I ran 'net rpc join -
> > UAdministrator%'secret' and I was able to join the domain.
> > Started up smb and then winbind
> > wbinfo -u
> > and I can see the users in the domain
> > getent passwd
> > shows nothing but the users in /etc/passwd
>
> "By default, "winbind enum users" and "winbind enum groups"
> are both disabled. See the smb.conf (5) man page for details.
>
>
>
>
> cheers, jerry
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.6 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQFI65YdIR7qMdg1EfYRApzXAJ91TZwLOdYDymgBSwlA1LmSKe3nqwCgqVST
> QeGBhOk1090EWWs4HSVL0ns=
> =R/kW
> -----END PGP SIGNATURE-----
>

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Another getent problem

2008-10-07T10:02:21Z

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Clark Johnston wrote:
> I have set up a system to be a member server and installed the samba
> rpms. I then copied over the samba config file and changed it to
> reflect the new shares and name change. I ran 'net rpc join -
> UAdministrator%'secret' and I was able to join the domain.
> Started up smb and then winbind
> wbinfo -u
> and I can see the users in the domain
> getent passwd
> shows nothing but the users in /etc/passwd

"By default, "winbind enum users" and "winbind enum groups"
are both disabled. See the smb.conf (5) man page for details.

cheers, jerry
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFI65YdIR7qMdg1EfYRApzXAJ91TZwLOdYDymgBSwlA1LmSKe3nqwCgqVST
QeGBhOk1090EWWs4HSVL0ns=
=R/kW
-----END PGP SIGNATURE-----
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Another getent problem

2008-10-07T09:58:06Z

I have set up a system to be a member server and installed the samba
rpms. I then copied over the samba config file and changed it to
reflect the new shares and name change. I ran 'net rpc join -
UAdministrator%'secret' and I was able to join the domain.
Started up smb and then winbind
wbinfo -u
and I can see the users in the domain
getent passwd
shows nothing but the users in /etc/passwd
checked nsswitch.conf and the following lines are there
passwd: files winbind
shadow: files
group: files winbind

I have the following links
/lib/libnss_winbind.so
/lib/libnss_winbind.so.2
/lib64/libnss_winbind.so
/lib64/libnss_winbind.so.2
/usr/lib/libnss_winbind.so
/usr/lib/libnss_winbind.so.2
/usr/lib64/libnss_winbind.so
/usr/lib64/libnss_winbind.so.2
/usr/lib64/nss/libnss_winbind.so
/usr/lib64/nss/libnss_winbind.so.2

when running ldconfig
/lib64/libnss_winbind.so.2
is pulled.

Tried deleting the cache files
winbindd_cache.tdb
winbindd_idmap.tdb
and restarting smb and winbind

The only error I'm finding is
Oct 6 18:04:45 localhost winbindd[3914]:
cli_rpc_pipe_open_ntlmssp_internal: cli_rpc_pipe_bind failed with error
NT_STATUS_NETWORK_ACCESS_DENIED

samba-common-3.0.28-0.el5.8
samba-3.0.28-0.el5.8
samba-client-3.0.28-0.el5.8
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Many strange errors in logs

2008-10-07T03:19:50Z

Hello all.

What we have:
samba-3.0.32_1
FreeBSD-6.3

PDC + BDC + LDAP.

All seems work mostly fine, but today I can`t add computer to domain.
Error says what domain not exist or cant be reached (im don`t know how
it says in english windows - im translate from russian)

Im start to read logs and found many errors in differrent logs:

tdb_chainlock_with_timeout_internal: alarm (10) timed out for key DC in
tdb /usr/local/etc/samba/secrets.tdb

log.172.16.1.2: domain_client_validate: unable to validate password for
user ilyin-vy in domain HQ to Domain controller DC. Error was
NT_STATUS_IO_TIMEOUT.
log.172.16.1.2: cli_rpc_pipe_close: cli_close failed on pipe \NETLOGON,
fnum 0x7485 to machine DC. Error was Call timed out: server did not
respond after 10000 milliseconds

repeated million times..

What does it mean?

testparm on PDC:

%testparm
Load smb config files from /usr/local/etc/smb.conf
Processing section "[homes]"
Processing section "[netlogon]"
Processing section "[mail]"
Loaded services file OK.
'winbind separator = +' might cause problems with group membership.
Server role: ROLE_DOMAIN_MEMBER
Press enter to see a dump of your service definitions

[global]
dos charset = cp866
unix charset = koi8-r
display charset = koi8-r
workgroup = HQ
server string = DC Server
security = DOMAIN
passdb backend = ldapsam:ldap://localhost/
log file = /var/log/samba/log.%m
max log size = 500
time server = Yes
add user script = /usr/local/sbin/ldapadduser '%u' users
rename user script = /usr/local/sbin/ldaprenameuser '%uold' '%unew'
delete user script = /usr/local/sbin/ldapdeleteuser '%u'
add group script = /usr/local/sbin/ldapaddgroup '%g'
delete group script = /usr/local/sbin/ldapdeletegroup '%g'
add user to group script = /usr/local/sbin/ldapaddusertogroup
'%u' '%g'
delete user from group script =
/usr/local/sbin/ldapdeleteuserfromgroup '%u' '%g'
set primary group script = /usr/local/sbin/ldapsetprimarygroup
'%u' '%g'
add machine script = /usr/local/sbin/ldapaddmachine '%u' computers
logon path =
os level = 64
preferred master = Yes
domain master = Yes
dns proxy = No
wins support = Yes
ldap admin dn = "cn=root,dc=fxclub,dc=org"
ldap group suffix = ou=Groups
ldap machine suffix = ou=Computers
ldap suffix = ou=Samba,dc=fxclub,dc=org
ldap ssl = no
ldap user suffix = ou=Users
winbind separator = +
winbind use default domain = Yes
admin users = admin
inherit acls = Yes
hosts allow = 172.16.1., 192.168.1., 127.
map acl inherit = Yes

[homes]
comment = Home Directories
valid users = %S
read only = No
browseable = No

[netlogon]
comment = Network Logon Service
path = /var/db/samba/netlogon
guest ok = Yes
browseable = No
share modes = No

[mail]
comment = temp
path = /tmp/mail
read only = No
create mask = 0777
guest ok = Yes
browseable = No

--
Best regards,
Proskurin Kirill
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: samba accounts management API

2008-10-07T02:15:21Z

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Русаков Денис wrote:
> Hello, all
> I'd like to create, remove, change samba accounts, groups etc. from my own C program.
> How can I do this, without using samba tools, but some samba API (headers and shared objects).
> Does samba provide some API (headers and shared objects) for accounts creating etc.?
> Thank you

Starting with Samba 3.2 we added a new shared library called libnetapi.
This library is designed very closely to the Windows NetApi equivalent,
and provides functions for all these tasks you are looking for.

For the upcoming Samba 3.3 release, this library provides around 50
calls and includes example code for at least all account management
functions.

You may want to have a look at:
NetUserAdd,
NetUserDel,
NetUserSetInfo,
and the the NetGroup* functions.

The header file is located under: $SRC/lib/netapi/netapi.h
Example code can found under: $SRC/lib/netapi/examples

Let us know where we can help further.

Hope this helps,
Guenther

- --
Günther Deschner GPG-ID: 8EE11688
Red Hat gdeschner@...
Samba Team gd@...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iEYEARECAAYFAkjrKKkACgkQSOk3aI7hFojZNgCeLrPgVUfGQE/pzHgFpksAKzes
B54An0NQzodllYBnVnSMV8Ww5Jw1aLPj
=85Zy
-----END PGP SIGNATURE-----
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

failed to lookup DC info for domain 'xxx.com' over rpc

2008-10-06T23:42:09Z

all

Any idea why we get such error message when trying to net ads join ?

kinit is working just fine but net ads join is failing

thanks
best regards
Francis
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

user samba

2008-10-06T21:33:00Z

how to make user samba can reset their own password like nt
i try to use smb2www but i can't do it (reset password)
thanx
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

RE: Nmbd is using the wrong ip address as source

2008-10-06T17:28:44Z

On Mon, 2008-10-06 at 09:05 +0300, Teodor Iacob wrote:

> Hello,
>
> Sorry for seeming impatient, I just had some rough night hours because
> of this problem :)
>
> tcpdump shows:
> 08:21:22.920832 IP X.X.X.X.netbios-dgm > 10.10.10.255.netbios-dgm: NBT
> UDP PACKET(138)
> 08:21:22.920986 IP X.X.X.X.netbios-dgm > 10.10.10.255.netbios-dgm: NBT
> UDP PACKET(138)
> 08:21:23.072063 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT
> UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
> 08:21:23.072069 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT
> UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
> 08:21:23.072073 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT
> UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
> 08:21:23.072076 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT
> UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
> 08:21:23.072079 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT
> UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
> 08:21:24.072491 IP X.X.X.X.netbios-dgm > 10.10.10.255.netbios-dgm: NBT
> UDP PACKET(138)
> 08:21:24.072497 IP X.X.X.X.netbios-dgm > 10.10.10.255.netbios-dgm: NBT
> UDP PACKET(138)
> 08:21:24.072656 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT
> UDP PACKET(137): QUERY; REQUEST; BROADCAST
> 08:21:25.073432 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT
> UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
> 08:21:25.073438 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT
> UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
> 08:21:25.073441 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT
> UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
>
> ( where X.X.X.X is the public ip address of the linux server from
> eth0 ), this tcpdump was taken from another linux machine with the
> ip: 10.10.10.2
>
> The requested netstat output:
>
> [root@ns ~]# netstat -nap |grep [sn]mbd
> tcp 0 0 10.10.10.1:139 0.0.0.0:*
> LISTEN 18538/smbd
> tcp 0 0 10.10.10.1:445 0.0.0.0:*
> LISTEN 18538/smbd
> udp 0 0 10.10.10.1:137 0.0.0.0:*
> 18542/nmbd
> udp 0 0 10.10.10.255:137 0.0.0.0:*
> 18542/nmbd
> udp 0 0 10.10.10.1:138 0.0.0.0:*
> 18542/nmbd
> udp 0 0 10.10.10.255:138 0.0.0.0:*
> 18542/nmbd
> unix 2 [ ] DGRAM 194986 18542/nmbd
> unix 2 [ ] DGRAM 194960 18538/smbd
>
> nmblookup from both machines returns this:
> querying server on 10.10.10.255
> name_query failed to find name server
>
> Right now there are no PCs open in the network, but friday when I was
> testing various configurations,
> I was able to see that one of the XP machines reported back as master
> browser ( I could see the share list
> with _MSBROWSE_ as well. )
>
>
> However if I do a unicast request to 10.10.10.1 ( basically it acts as
> a wins server ):
> querying HDV on 10.10.10.1
> 10.10.10.1 HDV<00>
>
> In the nmbd.log I have:
>
> [2008/10/06 08:52:50, 0] nmbd/nmbd.c:main(698)
> Netbios nameserver version 3.0.28-1.el5_2.1 started.
> Copyright Andrew Tridgell and the Samba Team 1992-2007
> [2008/10/06 08:52:51, 0]
> nmbd/nmbd_become_dmb.c:become_domain_master_browser_bcast(290)
> become_domain_master_browser_bcast:
> Attempting to become domain master browser on workgroup HOSTNAME on
> subnet 10.10.10.1
> [2008/10/06 08:52:51, 0]
> nmbd/nmbd_become_dmb.c:become_domain_master_browser_bcast(303)
> become_domain_master_browser_bcast: querying subnet 10.10.10.1 for
> domain master browser on workgroup HOSTNAME
> [2008/10/06 08:53:01, 0]
> nmbd/nmbd_become_dmb.c:become_domain_master_stage2(113)
> *****
>
> Samba server HNS is now a domain master browser for workgroup
> HOSTNAME on subnet 10.10.10.1
>
> *****
> [2008/10/06 08:53:15, 0]
> nmbd/nmbd_become_lmb.c:become_local_master_stage2(396)
> *****
>
> Samba name server HNS is now a local master browser for workgroup
> HOSTNAME on subnet 10.10.10.1
>
> *****
>
> Thank you for your time!
>
> Teodor
>
>
>
>
>
> ______________________________________________________________________
> From: Frank Gruman [mailto:fgatwork@...]
> Sent: Sunday, October 05, 2008 10:28 PM
> To: Teodor Iacob
> Cc: samba@...
> Subject: RE: [Samba] Nmbd is using the wrong ip address as source
>
>
>
>
> On Sun, 2008-10-05 at 18:35 +0300, Teodor Iacob wrote:
>
> > Does anybody have a suggestion for another list or forum to address
> > this issue?
> >
> >
> > > -----Original Message-----
> > > From: Teodor Iacob
> > > Sent: 4 octombrie 2008 09:14
> > > To: samba@...; samba@...
> > > Subject: RE: [Samba] Nmbd is using the wrong ip address as source
> > >
> > >
> > >
> > > > -----Original Message-----
> > > > From: samba-bounces+theo=macsys.eu@...
> > > > [mailto:samba-bounces+theo=macsys.eu@...] On Behalf Of
> > > > samba@...
> > > > Sent: Saturday, October 04, 2008 4:41 AM
> > > > To: samba@...
> > > > Subject: Re: [Samba] Nmbd is using the wrong ip address as source
> > > >
> > > > Teodor Iacob wrote:
> > > > > Hello,
> > > > >
> > > > > I have a linux box with the following S/W versions:
> > > > > CentOS release 5.2 (Final)
> > > > > Linux ns.hostname.ro 2.6.18-92.1.13.el5 #1 SMP Wed Sep 24
> > > > 19:32:05 EDT 2008 x86_64 x86_64 x86_64 GNU/Linux
> > > > > Version 3.0.28-1.el5_2.1
> > > > >
> > > > > My server has 2 IP interfaces, eth0: with the public ip
> > > > address and br0: for the private subnet.
> > > > > Br0 being a bridge ip interface between eth1 and tap0 ( for
> > > > openvpn ).
> > > > >
> > > > > Now my problem is that the nmbd is doing the netbios
> > > > broadcasts using the eth0 ip address instead
> > > > > of the br0 one. I'm only interested in enabling samba on
> > > > the local network (br0). I have checked
> > > > > this by using tcpdump on another linux machine in the
> > > > network and saw the broadcasts with the source
> > > > > address the eth0 ip. Nmbd claims to win the master browsing
> > > > election, but in reality the windows machines
> > > > > ignore the election packets and another machine becomes
> > > > master and the browsing list is messed up..
> > > > > The smbd works just fine, connecting to the shares manually
> > > > works correctly.
> > > > >
> > > > > How can I make nmbd to use the correct ip address? ( btw, I
> > > > removed the bridge config just for testing
> > > > > and only enabled ip on the eth1 and the same problem.. ). I
> > > > saw that in the previous version there was
> > > > > a parameter "-I" to force the source address..
> > > > >
> > > > > For reference this is my smb.conf:
> > > > > [global]
> > > > > workgroup = HOSTNAME
> > > > > netbios name = HNS
> > > > > interfaces = 10.10.10.1 127.0.0.1
> > > >
> > > > ... why not just say "interfaces = br0" so that it binds to your
> > > > private interface (which I presume owns 10.10.10.1)? You
> > > don't need
> > > > to have Samba listen on the loopback address if you can
> > > talk to it on
> > > > another local address.
> > >
> > > I mentioned also br0 in my quest of trying all possibilities
> > > I could think of.. No change.
> > > But bind is only for listening, my problem is when nmbd it
> > > sends packets by itself to make the election. I told you smbd
> > > works fine and I can see from netstat -an at the LISTEN part
> > > that it was only bound to the 10.10.10.1:139 and 10.10.10.1:445
> > >
> > >
> > > >
> > > > > hosts allow = 127.0.0., 10.10.10.
> > > > > bind interfaces only = yes
> > > > [snip]
> > > > > remote announce = 10.10.10.255/HOSTNAME
> > > > > socket address = 10.10.10.1
> > > >
> > > > Don
>
>
> First of all, thank you for your patience on volunteer attempts to
> help you resolve your issue (considering it is still the weekend in
> much of the world).
>
> Now, can you include a portion of the tcpdump you viewed from the
> other machine (particularly the part where the wrong address is
> appearing). Also, I am assuming you have restarted both smbd and nmbd
> after each configuration change.
>
> To verify that your samba settings are being read properly, try to run
> "netstat -nap |grep [sn]mbd" to see which ports are open on your local
> machine. Also, try running "nmblookup -B 10.10.10.255 server" to
> ensure that nmbd is responding to name server queries on the broadcast
> address.
>
> I Googled a moment, and came up with much of the above from her -
> http://samba.org/~tpot/articles/multiple-interfaces.html
>
> Let us know how it goes.
>
> Regards,
> Frank
>

Thanks for the output.

I can honestly say that I don't know why your public IP would be coming
back in the tcpdump. I will have to leave that one to greater minds
than mine. The other outputs look good except for the nmblookup
failure. I think I didn't provide enough information for you - the
"server" attribute should be the name of a machine on your network (a
node). That should fix the success of that query.

Also, try to add the "-d 8" parameter to add additional debug output to
this command. It should give you a lot more information about how it is
performing lookups. If you even want to test a step further, rather
than setting the "interfaces = <ip_address>" parameter in smb.conf, try
to set it back to the interface name "interfaces = <ifc name, e.g.
eth1>". I recall a posting a long time ago that made references to
this, so all of mine use ifc names rather than IP addresses.

Regards,
Frank

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: processes not closing

2008-10-06T16:49:52Z

On Fri, Oct 03, 2008 at 01:55:59PM -0500, Doug Tucker wrote:

> On Fri, 2008-10-03 at 19:37 +0200, Helmut Hullen wrote:
> > Hallo, Doug,
> >
> > Du (tuckerd) meintest am 03.10.08:
> >
> > > Umm..ok. I can just state the facts, I'm not pretending to
> > > understand them. To make sure I'm communicating properly, If I do:
> >
> > > /etc/init.d/smb restart
> >
> > > the old processes and locked files are gone
> >
> > Sure?
> >
> > ps aux | grep [sn]mbd
> >
> > is more reliable. Ok - it only shows processes, no locked file.
> >
> > Viele Gruesse!
> > Helmut
>
> Oh, you know what, I have provided some bad info, please forgive me.
> I'm thinking one thing and you can't see what I'm doing. At the OS
> level, the process *IS* gone. However, smbstatus still shows the
> process alive, and the user cannot access the file until it no longer
> shows up in smbstatus (such as on a samba restart). Oddly, it even
> shows up with the same pid as what you just killed.

Ok, if you're killing smbd with a kill -9 then it doesnt'
get chance to clean up its share mode records and so you
will see this. However, as subsequent smbd's notice this
pid no longer exists they should self-clean out the dead
open mode records (Samba is designed to be self-repairing
this way).

Jeremy.
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Problems with Share Modes "fill_share_mode_lock failed"

2008-10-06T16:48:10Z

On Fri, Oct 03, 2008 at 08:05:05AM -0300, Ricardo Santos wrote:

> Hi, my details:
>
> Server: 3.2.3
> Client: win2k8
>
> When I try list contents of a shared directory through an ASP script
> (win2k8 is the client), it's take much time and on samba I saw many
> times (4 times for each file) the same error: "fill_share_mode_lock
> failed"
>
> I tried "shares mode = No" but no success
>
> It's the log:
>
> ...:49, 3] locking/locking.c:fetch_share_mode_unlocked(857)
> fill_share_mode_lock failed
> ...:49, 3] smbd/process.c:process_smb(1549)
> Transaction 50769 of length 142 (0 toread)
> ...:49, 3] smbd/process.c:switch_message(1361)
> switch message SMBtrans2 (pid 22032) conn 0xb846b478
> ...:49, 3] smbd/trans2.c:call_trans2qfilepathinfo(3932)
> call_trans2qfilepathinfo: TRANSACT2_QPATHINFO: level = 1004
> ...:49, 3] smbd/vfs.c:check_reduced_name(854)
> reduce_name [web/xml/myfile.xml] [/web_virtual/domain]
> ...:49, 3] smbd/vfs.c:check_reduced_name(963)
> reduce_name: web/xml/myfile.xml reduced to
> /web_virtual/domain/web/xml/myfile.xml
> ...:49, 3] locking/locking.c:fetch_share_mode_unlocked(857)
> fill_share_mode_lock failed
> ...:49, 3] smbd/trans2.c:call_trans2qfilepathinfo(4006)
> call_trans2qfilepathinfo web/xml/myfile.xml (fnum = -1) level=1004
> call=5 total_data=0
> ...:49, 3] smbd/process.c:process_smb(1549)
> Transaction 50770 of length 142 (0 toread)
> ...:49, 3] smbd/process.c:switch_message(1361)
> switch message SMBtrans2 (pid 22032) conn 0xb846b478
> ...:49, 3] smbd/trans2.c:call_trans2qfilepathinfo(3932)
> call_trans2qfilepathinfo: TRANSACT2_QPATHINFO: level = 1005
> ...:49, 3] smbd/vfs.c:check_reduced_name(854)
> reduce_name [web/xml/myfile.xml] [/web_virtual/domain]
> ...:49, 3] smbd/vfs.c:check_reduced_name(963)
> reduce_name: web/xml/myfile.xml reduced to
> /web_virtual/domain/web/xml/myfile.xml
> ...:49, 3] locking/locking.c:fetch_share_mode_unlocked(857)
> fill_share_mode_lock failed
>
> Anyone could help me ?

That shouldn't happen. Please post your smb.conf and
also a debug level 10 log.

Jeremy.
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Cannot get shares to show up

2008-10-06T15:19:26Z

On 10/6/2008, Jesse Stone (jstone1999@...) wrote:
> My "Home" directory comes up as a share in Network Neighborhood but it also
> has browsable=no which is confusing:

You might like the 'hide unreadable' option - it can really make
navigating shares with lots of subfolders much easier, if you only have
access to a few of them...

--

Best regards,

Charles
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Broken links in online docs

2008-10-06T15:19:13Z

All of the links on this page appear to be broken. Of course, I didn't
click on every one, but after discovering thet the link to 'hide
unreadable' was broken, I clicked on a few more, and they all resulted
in Not Found errors...

http://www.samba.org/samba/docs/man/Samba-HOWTO-Collection/AccessControls.html

Scroll down abou halfway to where 'hide unreadable' is found, and start
clicking on links...

Not Found
The requested URL /samba/docs/man/Samba-HOWTO-Collection/smb.conf.5.html
was not found on this server.

--

Best regards,

Charles
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Password server is not connected - "cli not initilised" issue

2008-10-06T15:02:50Z

On Sat, Oct 04, 2008 at 10:30:28PM +1000, Jake Carroll wrote:

> Hi all,
>
> We are currently upgrading some core infrastructure and authentication
> hosts (Mac OS X 10.4.11 Open Directory/OpenLDAP/KDC --> 10.5.5). We've
> run into an interesting snag however.
>
> The Sun host that we use for sharing out files over Samba 3.0.28a (Sun's
> currently shipped, supported version of Samba) worked perfectly with the
> directory master running Mac OS X 10.4.11. When upgraded to 10.5.5 Server
> however, we are given the following errors on the samba/Sun side of
> things:
>
> [2008/10/04 17:40:59, 1, pid=2237] auth/auth_server.c:(247)
> password server is not connected (cli not initilised)
>
> I have done a bit of searching and can't actually find out what this
> *really* means.

We need to see a debug log level 10 to see what is going on here.

Thanks,

Jeremy.
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: CIFS Mounting Issues

2008-10-06T13:45:49Z

usually you get that error because of a bad username/password. instead
of using credentials, try -o username=sgerteisen,password=abc123

Steve Gerteisen wrote:

> Attempting to mount a share from a NetApp FAS using Win2K domain
> authentication. Using a credentials file. For testing purposes,
> actions have been performed as root. System info...
>
> OS: Debian 4.0r4a, kernel 2.6.18-6-amd64.
> mount.cifs version: 1.10-3.0.24
>
> Authentication appears to be occurring, however I am still receiving a
> mount error. Command output below, strace output attached...
>
> Any help would be appreciated.
>
> --begin output--
> workstation1:~# mount -t cifs //192.168.0.25/mysqldb /mnt/fas/ -o
> credentials=/root/.creds
> mount error 13 = Permission denied
> Refer to the mount.cifs(8) manual page (e.g.man mount.cifs)
> workstation1:~# echo $?
> 255
> --end output--
>
> V/R,
> Steve
>

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

CIFS Mounting Issues

2008-10-06T13:34:30Z

Attempting to mount a share from a NetApp FAS using Win2K domain
authentication. Using a credentials file. For testing purposes,
actions have been performed as root. System info...

OS: Debian 4.0r4a, kernel 2.6.18-6-amd64.
mount.cifs version: 1.10-3.0.24

Authentication appears to be occurring, however I am still receiving a
mount error. Command output below, strace output attached...

Any help would be appreciated.

--begin output--
workstation1:~# mount -t cifs //192.168.0.25/mysqldb /mnt/fas/ -o
credentials=/root/.creds
mount error 13 = Permission denied
Refer to the mount.cifs(8) manual page (e.g.man mount.cifs)
workstation1:~# echo $?
255
--end output--

V/R,
Steve

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

setgid bit disappears

2008-10-06T11:30:17Z

I have a samba server (without acls for better performance in ext3) with
shares where users create new directories and put their files.

I need the setgid bit to force the new directories and files to belong to a
shared group of users (not the primary group of the user that creates it).
When users create new directories the setgid bit is inherited and works
fine.

Users can set permissions for their new directories through the Windows
interface. But unfortunately then the setgid bit disappears.
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Cannot get shares to show up

2008-10-06T09:46:09Z

Thanks Ryan,

NOTE: Question at bottom. :)

I guess I need to review the Samba documentation better. I was using the
examples provided in smb.conf and thought that "Browsible" meant that people
other than "Valid Users" could see the stuff inside a share.

For example,

My "Home" directory comes up as a share in Network Neighborhood but it also
has browsable=no which is confusing:

[homes]
comment = Home
valid users = %S
read only = no
browsable = no

I'm not disagreeing with you, just pointing out how I was confused. I will
read through the documentation again on shares to ensure I am setting them
up correctly.

QUESTION: Simply using preexec to mount a drive prior to showing the
contents doesn't sound like it'll be a robust enough solution when you have
multiple people hitting the same share. Is it common practice to point pre
and postexec to a script that determines whether a mount has already been
mounted? Does anyone have any examples of a script that does this?

I am learning Python and have no problems attempting to write this myself as
it seems fairly straight forward. I guess my question is more "Is this the
best approach?"

-Jesse

On Sun, Oct 5, 2008 at 7:23 PM, Ryan Bair <ryandbair@...> wrote:

> Sorry, meant to reply all on that.
>
> On Sun, Oct 5, 2008 at 10:22 PM, Ryan Bair <ryandbair@...> wrote:
> > By showing up I assume you mean showing up in the listing of shares
> > for the computer?
> >
> > The "browsable = no" would be your problem. It makes the share not
> > browsable. :-)
> >
> > --Ryan
> >
> > On Sun, Oct 5, 2008 at 6:50 PM, Jesse Stone <jstone1999@...>
> wrote:
> >> I apologize if this is a really silly question but I've been messing
> with
> >> this for awhile now and cannot see any typos that would prevent these
> shares
> >> from showing up.
> >>
> >> I'm trying two things I've never done before so I want to make sure to
> point
> >> them out:
> >>
> >> 1) I am using groups in the "valid users" section
> >> 2) I am using preexec and postexec on each share
> >>
> >> Example:
> >>
> >> [Media]
> >> comment = Media(Non-kid)!
> >> path = /mnt/media1
> >> read only = yes
> >> browsable = no
> >> valid users = @mediausers
> >> preexec = /bin/mount /mnt/media1
> >> postexec = /bin/umount /mnt/media1
> >>
> >> I've done nothing with the /mnt/media1 folder so I expect my next issue
> will
> >> be permissiones related but I would expect the shares to at least show
> up
> >> for all users within the @mediausers group.
> >>
> >> I'm using preexec ect from the cd-rom example contained within the
> default
> >> smb.conf file.
> >>
> >> -Jesse
> >> --
> >> To unsubscribe from this list go to the following URL and read the
> >> instructions: https://lists.samba.org/mailman/listinfo/samba
> >>
> >
>

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Performance tweaking from Ubuntu to a Macbook vs. Windows through DLink DIR-655

2008-10-06T08:10:49Z

Volker Lendecke wrote:

On Sun, Oct 05, 2008 at 12:32:27PM -0700, Roy Abou Assaly wrote:
> socket options = TCP_NODELAY SO_KEEPALIVE IPTOS_LOWDELAY SO_RCVBUF=65536
> SO_SNDBUF=65536

On modern Linux systems (i.e. from the last 5 years or so),
just dump the buffer parms completely. The Linux kernel is
fine figuring out the buffer sizes itself.

Many thanks Volker. I will be doing just that!

make test fails on Solaris 10 for 3.2.4

2008-10-06T06:51:54Z

Hi,

I'm trying to check the result of my build of 3.2.4 on Solaris 10, but
smbd dies shortly after starting:

Running Test suite
LD_LIBRARY_PATH=/opt/ws/dists/samba-3.2.4/source/bin
CREATE TEST ENVIRONMENT IN './st'...DONE
STARTING NMBD...DONE
STARTING WINBINDD...DONE
STARTING SMBD...DONE
delaying for nbt name registration
smbd failed with status 1!

I looked at the smbd_test.log & I see:

smbd version 3.2.4 started.
Copyright Andrew Tridgell and the Samba Team 1992-2008
standard input is not a socket, assuming -D option
gdb_backtrace: ERROR: No debugger found.
smbd failed with status 1!

Now, I can't tell if smbd is dying because gdb can't be found, or smbd
has died & the test script then tries to run gdb to determine why it
died & can't find it. If I install samba & load smbd by hand, it starts
up without a problem. I haven't found anyone else reporting & a fix, so
I'm hoping someone here could help me.

Thanks.

-John

--
John Center
Villanova University
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

samba accounts management API

2008-10-06T04:10:37Z

Hello, all
I'd like to create, remove, change samba accounts, groups etc. from my own C program.
How can I do this, without using samba tools, but some samba API (headers and shared objects).
Does samba provide some API (headers and shared objects) for accounts creating etc.?
Thank you
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Performance tweaking from Ubuntu to a Macbook vs. Windows through DLink DIR-655

2008-10-06T00:05:35Z

On Sun, Oct 05, 2008 at 12:32:27PM -0700, Roy Abou Assaly wrote:
> socket options = TCP_NODELAY SO_KEEPALIVE IPTOS_LOWDELAY SO_RCVBUF=65536
> SO_SNDBUF=65536

On modern Linux systems (i.e. from the last 5 years or so),
just dump the buffer parms completely. The Linux kernel is
fine figuring out the buffer sizes itself.

Volker

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

attachment0 (196 bytes) Download Attachment

Re: [ANNOUNCE] Samba 3.3.0pre2 Available for Download

2008-10-06T00:04:32Z

On Thu, Oct 02, 2008 at 07:08:17PM -0400, michael@... wrote:
> Regarding "extended cluster support", does that mean CTDB stuff is now
> integrated into this version of Samba?

ctdb as such is still a separate project, but the Samba
integration is all upstream.

Volker

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

attachment0 (196 bytes) Download Attachment

RE: Nmbd is using the wrong ip address as source

2008-10-05T23:05:04Z

Hello,

Sorry for seeming impatient, I just had some rough night hours because of this problem :)

tcpdump shows:
08:21:22.920832 IP X.X.X.X.netbios-dgm > 10.10.10.255.netbios-dgm: NBT UDP PACKET(138)
08:21:22.920986 IP X.X.X.X.netbios-dgm > 10.10.10.255.netbios-dgm: NBT UDP PACKET(138)
08:21:23.072063 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
08:21:23.072069 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
08:21:23.072073 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
08:21:23.072076 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
08:21:23.072079 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
08:21:24.072491 IP X.X.X.X.netbios-dgm > 10.10.10.255.netbios-dgm: NBT UDP PACKET(138)
08:21:24.072497 IP X.X.X.X.netbios-dgm > 10.10.10.255.netbios-dgm: NBT UDP PACKET(138)
08:21:24.072656 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT UDP PACKET(137): QUERY; REQUEST; BROADCAST
08:21:25.073432 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
08:21:25.073438 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST
08:21:25.073441 IP X.X.X.X.netbios-ns > 10.10.10.255.netbios-ns: NBT UDP PACKET(137): REGISTRATION; REQUEST; BROADCAST

( where X.X.X.X is the public ip address of the linux server from eth0 ), this tcpdump was taken from another linux machine with the
ip: 10.10.10.2

The requested netstat output:

[root@ns ~]# netstat -nap |grep [sn]mbd
tcp 0 0 10.10.10.1:139 0.0.0.0:* LISTEN 18538/smbd
tcp 0 0 10.10.10.1:445 0.0.0.0:* LISTEN 18538/smbd
udp 0 0 10.10.10.1:137 0.0.0.0:* 18542/nmbd
udp 0 0 10.10.10.255:137 0.0.0.0:* 18542/nmbd
udp 0 0 10.10.10.1:138 0.0.0.0:* 18542/nmbd
udp 0 0 10.10.10.255:138 0.0.0.0:* 18542/nmbd
unix 2 [ ] DGRAM 194986 18542/nmbd
unix 2 [ ] DGRAM 194960 18538/smbd

nmblookup from both machines returns this:
querying server on 10.10.10.255
name_query failed to find name server

Right now there are no PCs open in the network, but friday when I was testing various configurations,
I was able to see that one of the XP machines reported back as master browser ( I could see the share list
with _MSBROWSE_ as well. )

However if I do a unicast request to 10.10.10.1 ( basically it acts as a wins server ):
querying HDV on 10.10.10.1
10.10.10.1 HDV<00>

In the nmbd.log I have:

[2008/10/06 08:52:50, 0] nmbd/nmbd.c:main(698)
Netbios nameserver version 3.0.28-1.el5_2.1 started.
Copyright Andrew Tridgell and the Samba Team 1992-2007
[2008/10/06 08:52:51, 0] nmbd/nmbd_become_dmb.c:become_domain_master_browser_bcast(290)
become_domain_master_browser_bcast:
Attempting to become domain master browser on workgroup HOSTNAME on subnet 10.10.10.1
[2008/10/06 08:52:51, 0] nmbd/nmbd_become_dmb.c:become_domain_master_browser_bcast(303)
become_domain_master_browser_bcast: querying subnet 10.10.10.1 for domain master browser on workgroup HOSTNAME
[2008/10/06 08:53:01, 0] nmbd/nmbd_become_dmb.c:become_domain_master_stage2(113)
*****

Samba server HNS is now a domain master browser for workgroup HOSTNAME on subnet 10.10.10.1

*****
[2008/10/06 08:53:15, 0] nmbd/nmbd_become_lmb.c:become_local_master_stage2(396)
*****

Samba name server HNS is now a local master browser for workgroup HOSTNAME on subnet 10.10.10.1

*****

Thank you for your time!

Teodor

_____

From: Frank Gruman [mailto:fgatwork@...]
Sent: Sunday, October 05, 2008 10:28 PM
To: Teodor Iacob
Cc: samba@...
Subject: RE: [Samba] Nmbd is using the wrong ip address as source

On Sun, 2008-10-05 at 18:35 +0300, Teodor Iacob wrote:

Does anybody have a suggestion for another list or forum to address

this issue?

> -----Original Message-----

> From: Teodor Iacob

> Sent: 4 octombrie 2008 09:14

> To: samba@...; samba@...

> Subject: RE: [Samba] Nmbd is using the wrong ip address as source

>

>

>

> > -----Original Message-----

> > From: samba-bounces+theo=macsys.eu@...

> > [mailto:samba-bounces+theo=macsys.eu@...] On Behalf Of

> > samba@...

> > Sent: Saturday, October 04, 2008 4:41 AM

> > To: samba@...

> > Subject: Re: [Samba] Nmbd is using the wrong ip address as source

> >

> > Teodor Iacob wrote:

> > > Hello,

> > >

> > > I have a linux box with the following S/W versions:

> > > CentOS release 5.2 (Final)

> > > Linux ns.hostname.ro 2.6.18-92.1.13.el5 #1 SMP Wed Sep 24

> > 19:32:05 EDT 2008 x86_64 x86_64 x86_64 GNU/Linux

> > > Version 3.0.28-1.el5_2.1

> > >

> > > My server has 2 IP interfaces, eth0: with the public ip

> > address and br0: for the private subnet.

> > > Br0 being a bridge ip interface between eth1 and tap0 ( for

> > openvpn ).

> > >

> > > Now my problem is that the nmbd is doing the netbios

> > broadcasts using the eth0 ip address instead

> > > of the br0 one. I'm only interested in enabling samba on

> > the local network (br0). I have checked

> > > this by using tcpdump on another linux machine in the

> > network and saw the broadcasts with the source

> > > address the eth0 ip. Nmbd claims to win the master browsing

> > election, but in reality the windows machines

> > > ignore the election packets and another machine becomes

> > master and the browsing list is messed up..

> > > The smbd works just fine, connecting to the shares manually

> > works correctly.

> > >

> > > How can I make nmbd to use the correct ip address? ( btw, I

> > removed the bridge config just for testing

> > > and only enabled ip on the eth1 and the same problem.. ). I

> > saw that in the previous version there was

> > > a parameter "-I" to force the source address..

> > >

> > > For reference this is my smb.conf:

> > > [global]

> > > workgroup = HOSTNAME

> > > netbios name = HNS

> > > interfaces = 10.10.10.1 127.0.0.1

> >

> > ... why not just say "interfaces = br0" so that it binds to your

> > private interface (which I presume owns 10.10.10.1)? You

> don't need

> > to have Samba listen on the loopback address if you can

> talk to it on

> > another local address.

>

> I mentioned also br0 in my quest of trying all possibilities

> I could think of.. No change.

> But bind is only for listening, my problem is when nmbd it

> sends packets by itself to make the election. I told you smbd

> works fine and I can see from netstat -an at the LISTEN part

> that it was only bound to the 10.10.10.1:139 and 10.10.10.1:445

>

>

> >

> > > hosts allow = 127.0.0., 10.10.10.

> > > bind interfaces only = yes

> > [snip]

> > > remote announce = 10.10.10.255/HOSTNAME

> > > socket address = 10.10.10.1

> >

> > Don

First of all, thank you for your patience on volunteer attempts to help you resolve your issue (considering it is still the weekend
in much of the world).

Now, can you include a portion of the tcpdump you viewed from the other machine (particularly the part where the wrong address is
appearing). Also, I am assuming you have restarted both smbd and nmbd after each configuration change.

To verify that your samba settings are being read properly, try to run "netstat -nap |grep [sn]mbd" to see which ports are open on
your local machine. Also, try running "nmblookup -B 10.10.10.255 server" to ensure that nmbd is responding to name server queries
on the broadcast address.

I Googled a moment, and came up with much of the above from her - http://samba.org/~tpot/articles/multiple-interfaces.html

Let us know how it goes.

Regards,
Frank

__________ Information from ESET NOD32 Antivirus, version of virus signature database 3492 (20081003) __________

The message was checked by ESET NOD32 Antivirus.

http://www.eset.com

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Cannot get shares to show up

2008-10-05T19:23:31Z

Sorry, meant to reply all on that.

On Sun, Oct 5, 2008 at 10:22 PM, Ryan Bair <ryandbair@...> wrote:

> By showing up I assume you mean showing up in the listing of shares
> for the computer?
>
> The "browsable = no" would be your problem. It makes the share not
> browsable. :-)
>
> --Ryan
>
> On Sun, Oct 5, 2008 at 6:50 PM, Jesse Stone <jstone1999@...> wrote:
>> I apologize if this is a really silly question but I've been messing with
>> this for awhile now and cannot see any typos that would prevent these shares
>> from showing up.
>>
>> I'm trying two things I've never done before so I want to make sure to point
>> them out:
>>
>> 1) I am using groups in the "valid users" section
>> 2) I am using preexec and postexec on each share
>>
>> Example:
>>
>> [Media]
>> comment = Media(Non-kid)!
>> path = /mnt/media1
>> read only = yes
>> browsable = no
>> valid users = @mediausers
>> preexec = /bin/mount /mnt/media1
>> postexec = /bin/umount /mnt/media1
>>
>> I've done nothing with the /mnt/media1 folder so I expect my next issue will
>> be permissiones related but I would expect the shares to at least show up
>> for all users within the @mediausers group.
>>
>> I'm using preexec ect from the cd-rom example contained within the default
>> smb.conf file.
>>
>> -Jesse
>> --
>> To unsubscribe from this list go to the following URL and read the
>> instructions: https://lists.samba.org/mailman/listinfo/samba
>>
>

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Cannot get shares to show up

2008-10-05T15:50:12Z

I apologize if this is a really silly question but I've been messing with
this for awhile now and cannot see any typos that would prevent these shares
from showing up.

I'm trying two things I've never done before so I want to make sure to point
them out:

1) I am using groups in the "valid users" section
2) I am using preexec and postexec on each share

Example:

[Media]
comment = Media(Non-kid)!
path = /mnt/media1
read only = yes
browsable = no
valid users = @mediausers
preexec = /bin/mount /mnt/media1
postexec = /bin/umount /mnt/media1

I've done nothing with the /mnt/media1 folder so I expect my next issue will
be permissiones related but I would expect the shares to at least show up
for all users within the @mediausers group.

I'm using preexec ect from the cd-rom example contained within the default
smb.conf file.

-Jesse
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba

Re: Lost most data on Windows XP machine switching to domain

2008-10-05T15:44:11Z

Thank you for all the responses. Thinking back on how I did it, I bet it
was a permissions issue which prevented the files from migrating to the
Samba server but that still shouldn't have caused them to be deleted. The
files were complete gone. I checked both the old non-domain directory and
the newly created domain directory. I was able to recover 98% of the data
using a special recovery program my IT department provided me.

The option I had to use to find the files using the recovery software was
not "Accidently Deleted" but "Disk was formatted" which makes me think
whatever process distroyed the files did it in a way where they did not go
into the recycle bin but completely distroyed.

-Jesse

On Sat, Oct 4, 2008 at 6:02 PM, Scott Lovenberg
<scott.lovenberg@...>wrote:

> Jesse Stone wrote:
>
>> I'm wondering if anyone has run across that and MUCH more importantly, if
>> the data can be recovered somehow.
>>
>> I'll put as much details as I can at the bottom but here's the gist of the
>> problem:
>>
>> I added my wives computer (which contains 8 years worth of pictures) to
>> the
>> domain. When I logged into the new domain account it was empty and my
>> wives
>> domain users had no access so I did the following:
>>
>> 1) Logged out of the domain account and back into the machine account
>> 2) Added the domain user to the administrative group
>> 3) MOVED (yes, I'm an idiot) everything from my wive's standard profile to
>> the domain profile
>> 4) Logged back in with the domain account
>>
>> Here's what happens:
>> a few random things where in the new domain. For example, 1 bookmark (out
>> of about 50) was in my wive's favorites folder. The "My Pictures" folder
>> contained "Sample Pictures" only.
>>
>> My guess is that 1 of 2 things happened:
>>
>> 1) Samba didn't expect there to be data yet so started out with a fresh
>> new
>> profile. This doesn't explain how some (less than 1% of her data) is
>> available
>>
>> 2) My wive is connecting to the domain via wireless. Somehow, mid-copy
>> the
>> wireless shut off and the data never made it to the roaming profile.
>>
>> Please someone give me good news like "just do this and the data will be
>> recovered!"
>>
>> OK, here's the details (which will show my lack of understanding):
>>
>> I followed the following article when setting up Samba:
>> http://www.howtoforge.com/samba_setup_ubuntu_5.10_p4
>>
>> The only changes I made are that I commented out the following lines
>> (believing this would STOP roaming profiles. I did not actually want
>> roaming profiles and was only planning on setting the "My Documents"
>> folder
>> to use server storage.
>>
>> #logon drive = H: <-- May use later for roaming profiles
>> #logon path = \\%N\profile\%U <-- May use later for roaming profiles
>>
>> (Note, the only thing this did is stop the drive letter from being set.
>> The
>> "profile" directly was still created, only under the /home/%user%/
>> directory
>> instead of /home/samba/profiles/)
>>
>> Here's the entire smb.conf I am using:
>> [global]
>> workgroup = domaintest
>> netbios name = server3200
>> server string = File Server
>> passdb backend = tdbsam
>> security = user
>> username map = /etc/samba/smbusers
>> name resolve order = wins bcast hosts
>> domain logons = yes
>> preferred master = yes
>> wins support = yes
>> # Set CUPS for printing
>> printcap name = CUPS
>> printing = CUPS
>> # Default logon
>> #logon drive = H:
>> #logon script = scripts/logon.bat
>> #logon path = \\%N\profile\%U
>> # Useradd scripts
>> add user script = /usr/sbin/useradd -m %u
>> delete user script = /usr/sbin/userdel -r %u
>> add group script = /usr/sbin/groupadd %g
>> delete group script = /usr/sbin/groupdel %g
>> add user to group script = /usr/sbin/usermod -G %g %u
>> add machine script = /usr/sbin/useradd -s /bin/false/ -d /var/lib/nobody
>> %u
>> idmap uid = 15000-20000
>> idmap gid = 15000-20000
>> # sync smb passwords with linux passwords
>> passwd program = /usr/bin/passwd %u
>> passwd chat = *Enter\snew\sUNIX\spassword:* %n\n
>> *Retype\snew\sUNIX\spassword:* %n\n .
>> passwd chat debug = yes
>> unix password sync = yes
>> # set the loglevel
>> log level = 3
>> [homes]
>> comment = Home
>> valid users = %S
>> read only = no
>> browsable = no
>> [printers]
>> comment = All Printers
>> path = /var/spool/samba
>> printable = yes
>> guest ok = yes
>> browsable = no
>> [netlogon]
>> comment = Network Logon Service
>> path = /home/samba/netlogon
>> admin users = Administrator
>> valid users = %U
>> read only = no
>> [profile]
>> comment = User profiles
>> path = /home/samba/profiles
>> valid users = %U
>> create mode = 0600
>> directory mode = 0700
>> writable = yes
>> browsable = no
>>
>> Please understand that my wife may well divorce me if I can't recover this
>> stuff.
>>
>> -Jesse
>>
>>
> I'm a little mixed up about about the steps that you took... Am I
> interpreting this correctly:
> 1.) You signed on with your wifes domain account, then logged out
> 2.) You then logged in as a local admin and added her domain account to
> the Domain Administrators group
> 3.) Before logging out of the local admin account, you moved all of her
> files to the default domain profile (in the netlogon share) (with
> permissions 0600 as per your profile share configuration)
> 4.) You then logged out of your local admin account and logged back in
> with your wifes domain account
> 5.) Everything is missing at this point.
>
> I'm fairly sure that Windows handles dropped connections during a sign
> on/off with a file that contains successfully transferred files. The fact
> that you have some of her files makes me wonder if you've got a permissions
> issue going on. Are you sure that the files aren't on the domain controller
> with permissions that keep her account from seeing them?
> If I were you, I'd remount that drive read only 60 seconds ago and make a
> copy of it right away. Even if you deleted the files, you can probably get
> a dd_rescue image before you actually blow them away. I've had success with
> that before after fat-fingering an effective "rm -rf /." while logged in as
> root. The Samba team will be happy to know that the Samba server limped
> along while the Windows clients dropped like a lead zeppelin (apologies to
> Keith Moon). :D
>
> I was able to get ~98% of the data back even with a 3ware RAID card,
> reiserfs, and LVM complicating the matter. You've probably got a good shot
> at recovering that data (if it really did go MIA) if you don't write
> anything else to the disk. YMMV, of course.
>

--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba