samba - Jul 2005 - Samba, Windows 2003 server and large file copy failure

We have a Windows 2003 Small Business Server and Redhat database server.
Windows 2003 server introduces NTLM2 support and SMB signing. It removed
support for NTLM and hence Windows 9x, Windows NT4 and Linux could not
work with the server nor access file shares.

There is a policy change to turn allot of these new security features
off and we have done so. We can now connect to the server.

Now we introduce Redhat ES4 (Enterprise server 4). We tried connecting
to the Windows 2003 server using the mount command and kept freezing the
Linux box.
We then discovered you had to mount to a windows box a special way. We
did so and all was fine for about 3 weeks.

We used "mount.cifs \\\\10.0.0.10\\icebackup /mnt/backup -o
user=iceserver"

10.0.0.10 is the IP of the Windows server and icebackup is the share.
Iceserver is the user of the windows box with full access to the windows
share and full ntfs access to the files in the folder.

Every night at 10:30 a cron job runs copying a firebird database from
the Linux box to the Windows box for backup. The file is not open by any
processes and the copy can have exclusive access to the file. The file
is 979,000 kb (almost 1 gb). This process has been working for over 3
weeks (every night).

Suddenly it no longer works. It gets to 952,000 kb and then copied no
more. The Windows box says the file is in use by the user iceserver, the
Linux box is still waiting to come back to a prompt and it will sit here
like this forever.

Running ps -e shows the "cp" command is running. Running a Kill PID
against it will not stop the process. The only way to stop the process
is a reboot.
When we reboot, the servers time is that of when the file copy stalled.

If the copy runs at 10:30, stalls at 10:40 then the time on the box the
next day is the previous days date at about 10:41.

We are running :
samba-3.0.10-1.4E
samba-client-3.0.10-1.4E
samba-common-3.0.10-1.4E

The log in /var/log/samba/10.0.0.10.log does not seem to have any issues
in it.

Has anyone seen this before and can anyone recommend a solution or fault
detection method?


Michael Jenkin
I.T. Manager  


COPYWORLD - TOSHIBA


( Ph:  +61  8  8363 1011      

( Fax: +61  8  8362 9086 

: Email: michael.jenkin@copyworld.com.au
<mailto:michael.jenkin@copyworld.com.au>  

* Web: http://www.copyworld.com.au <http://www.copyworld.com.au/> 



COPIERS - FAX - PHONE SYSTEMS - COMPUTER SYSTEMS




==========================================================Please note that
neither Copyworld nor the sender accepts any responsibility
for any viruses that may be contained in this e-mail or its attachments.  It
is therefore your responsibility to ensure that your systems have adequate
protection against virus infection. Copyworld uses the latest email scanning
technologies and updates regularly.

The information contained in this e-mail is intended only for the use of the
intended recipient at the e-mail address to which it has been addressed.  If
the reader of this message is not an intended recipient, you are hereby notified
that you have received this document in error and that any review, dissemination
or copying of the message or associated attachments is strictly prohibited.

If you have received this e-mail in error please contact the sender by return
e-mail or call +618 8363 1011 and ask for the sender and then delete it
immediately
from your system.

Michael Jenkin wrote:
> We have a Windows 2003 Small Business Server and Redhat database
> server. 
> 
> We used "mount.cifs \\\\10.0.0.10\\icebackup /mnt/backup -o
> user=iceserver" 
Mount.cifs is a different mailing list.

> 
> Running ps -e shows the "cp" command is running. Running a Kill
PID
> against it will not stop the process. The only way to stop the
> process is a reboot.  
kill -9 PID should kill anything, but you aren't being nice to the process
doing that.
> 
> 
> 
> The log in /var/log/samba/10.0.0.10.log does not seem to have any
> issues in it. 
>
Mount.cifs is a kernel module and so does no logging to the samba logs.
Look in the kernel related logs.
 
> 
> Has anyone seen this before and can anyone recommend a solution or
> fault detection method? 
> 
You could try going the other way with a scheduled xcopy job from the W2K3
server off the RHAS4 server.  You could take one of the examples from the
first few chapters of the Samba3 by example book online and make up a simple
share.  Something like this should do:

# Global parameters
[global]
        workgroup = COPYWORLD
        netbios name = Firebird
        comment = Firebird database server
        server string = Firebird database server
        security = SHARE
        disable spoolss = Yes
        show add printer wizard = No
        wins server = 10.0.0.2
	hosts allow = 10.0.0.(W2K3IP), 127.
	

[Firebird]
        comment = Firebird database
        path = /data/Firebird
        force user = iceserver
        force group = iceservergroup
        read only = No
        guest ok = Yes
        nt acl support = No

Look at the book for the rest of the configs and procedures.  This may not
be as secure as you would like, Caveat Emptor.........
> 
> 
> Michael Jenkin
> 
> I.T. Manager