samba - Jul 2009 - Difference in Samba and CIFS interms of keeping the deleted files opened

Hi,

We have a CIFS server running on a NetApp server and a Solaris host running
Samba-3.3.2.

When we mount both the filesystems to a Drive on a Windows using the net use
command and then try to run a java program which basically does nothing but
continuosly writes a data chunk to a file. On a side note, these same
filesystems are accessible on a Solaris (unix) host too.

When the java program is run and a file is being generated, I go to my unix
terminal and happen to delete the file generated by the java program.
Interestingly, there is an IOexception caught in the java program running on
the Windows machine, when the file is deleted on the CIFS based filesystem
(available on Solaris as a NFS filesystems) but there is no exception caught
when the filesystem happens to be Samba (available on Solaris as /var , a
regular partition).

I delete the file from Unix as the process demands, but also there is no way
to delete a in-use-file in Windows.

I would like to understand the differences in Samba and CIFS in this context
especially why is that so there is an IOexception for a CIFS based
filesystem but not on the samba filesystem. This is reproducible at will.
What could be wrong? What could be made to make samba filesystem also behave
the same way to throw exceptions (Exceptions are good than that not at all
knowing there is a file that is deleted but being still written onto.)

I would be willing to work and share the procedure to reproduce this
behaviour with anyone aware of the Samba/CIFS protocols.

-- 
Nikhil

On Mon, Jul 20, 2009 at 10:14:49PM +0530, Nikhil wrote:
>  Hi,
> 
> We have a CIFS server running on a NetApp server and a Solaris host running
> Samba-3.3.2.
> 
> When we mount both the filesystems to a Drive on a Windows using the net
use
> command and then try to run a java program which basically does nothing but
> continuosly writes a data chunk to a file. On a side note, these same
> filesystems are accessible on a Solaris (unix) host too.
> 
> When the java program is run and a file is being generated, I go to my unix
> terminal and happen to delete the file generated by the java program.
> Interestingly, there is an IOexception caught in the java program running
on
> the Windows machine, when the file is deleted on the CIFS based filesystem
> (available on Solaris as a NFS filesystems) but there is no exception
caught
> when the filesystem happens to be Samba (available on Solaris as /var , a
> regular partition).
> 
> I delete the file from Unix as the process demands, but also there is no
way
> to delete a in-use-file in Windows.
> 
> I would like to understand the differences in Samba and CIFS in this
context
> especially why is that so there is an IOexception for a CIFS based
> filesystem but not on the samba filesystem. This is reproducible at will.
> What could be wrong? What could be made to make samba filesystem also
behave
> the same way to throw exceptions (Exceptions are good than that not at all
> knowing there is a file that is deleted but being still written onto.)
Ok, I think the reason that you're having this problem is that
you're running Samba on Solaris in this case, sad to say.

I don't believe Sun have exposed kernel level oplock (lease) capability
to user space processes, so Samba on Solaris has no way of knowing
that a unix user deleted the file.

Samba running on Linux, (or SGI Irix) has kernel level oplocks,
so can detect access from the local filesystem. As the NetApp
runs a custom kernel (derived a long time ago from FreeBSD I
believe) then their CIFS implementation (like Samba on Linux)
knows when a NFS user has modified the file.

Jeremy.

On Tue, Jul 21, 2009 at 5:06 PM, Gerald Carter <jerry at plainjoe.org>
wrote:
> Nikhil,
>
> > I looked at the enabling kernel oplocks in the samba configuration
file
> and
> > I have enabled "kernel oplocks = yes" in the [global]
section on both of
> > Linux-2.6.18-92 and Linux-2.4.21-47.0.1.ELsmp kernel machines running
the
> > samba and I notice that the behaviour is not as what is said. I still
do
> not
> > see the exception being NOT caught on Windows java program when I
delete
> the
> > file from a Unix terminal.
> >
> > Any idea, where I could be going wrong?
>
> It sounds more like smbd maintains the opens file descriptor
> as long as the client has an open file handle.  Standard
> POSIX semantics allow you to delete open files with the inode
> finally being released after the last open fd is closed.
> I could speculate that it may be less to do with CIFS and more
> to do with the server OS.  But that is pure speculation.
>
> Have you looked at a network trace to see which SMB op caused
> the IO Exception?  Does truncating rather than removing the
> file change the application behavior?
>
>
>
>
> cheers, jerry
> --
> ====================================================================>
http://www.plainjoe.org/
> "What man is a man who does not make the world better?"     
--Balian
>
>
Hi Jeremy,

Yeah.. I tried oplocking thing (btw, is there any set-direct-go howto on
oplocking configuration parameters to be used smb.conf?) and checked via the
smbstatus that the file is being into the samba lock.
But when I rm the file from the Unix terminal, the windows program does not
detect the file deletion.
>> it may be less to do with CIFS and more to do with the server OS.  But
that is pure speculation.
agreed. I am running Solaris-10 where kernel oplock is not available, so
will rule that out.
I have linux box with 2.6 kernel and kernel oplocks=yes set in smb.conf; but
that still does not help in getting the sync set.
>> Does truncating rather than removing the file change the application
behavior?
Interesting. I do "/bin/cp -f /dev/null /var/tmp/test.log" and I
notice that
ls -l does not show that the file is truncated at all.... even though I do
it multiple times; ls -l still does not show that the file is truncated. Am
I missing something that why is rm command or deleting the file behaviour
would differ?


BTW, this is the output of testparam -s -v | grep -i lock output:

Load smb config files from /var/samba/smb.conf
Processing section "[homes]"
Processing section "[printers]"
Processing section "[tmpdir]"
Loaded services file OK.
Server role: ROLE_STANDALONE
kernel oplocks = Yes
lock spin time = 200
oplock break wait time = 0
lock directory = /var/samba/locks
usershare path = /var/samba/locks/usershares
block size = 1024
veto oplock files blocking locks = Yes
fake oplocks = No
locking = Yes
oplocks = Yes
level2 oplocks = Yes
oplock contention limit = 2
posix locking = Yes
strict locking = Yes
 Can you please let me know if any of the parameters need something
else/need not put?

Thanks again, Jeremy.
-- 
Nikhil