samba - Dec 2010 - Auto-Removal of Straggling File locks due to Ungraceful Client Disconnects

Hi all,

This question has come up many times before in a number of guises.
But I do not believe that it has ever been answered satisfactorily.

This may well reflect fundamental difficulties in the CIFS protocol
However, recent incarnations of Windows Server seem to handle the problem
better
possibly due to tweaks in the underlying TCP/IP stack vis-a-vis Linux.

At the end of the day, a centralised/shared file system should be
immune to client misbehaviour. While file corruption by a misbehaving
client with write access is always a possibility. Surely it is reasonable
that a file should not be left locked indefinitely by a client that has
long since departed the scene.

Suggested solutions usually revolve around:

1. Client configuration tweaks.
2. Samba's "Reset on Zero VC" option in smb.conf.

My response to these is that:

1. A server should not have to depend on well-behaving clients to
maintain basic file system integrity.
2. "Reset on Zero VC" only helps in the case where the disconnected
client actually reconnects.

But what if a client suddenly disconnects due to a crash, power outage
or careless disconnect?
What if the client stays disconnected indefinitely e.g. laptop has
been whipped out and gone off site?

In these situations, a server administrator needs to be on standby.
And straggling locks can be removed manually by:

1. Identifying them using the smbstatus utility
2. Killing the relevant server process.

But this can become extremely tedious in busy ad hoc networks with
mobile users

So can someone please advise me why it is not possible for Samba to:

1. Identify obvious long term client disconnects e.g. by polling every
five minutes.
2. Automatically remove any locks still held by such clients.

Hope I am making some sense here and haven't missed something obvious.

TIA
rvjcallanan

On Thu, Dec 30, 2010 at 03:36:04AM +0000, Vincent Callanan
wrote:
> This question has come up many times before in a number of guises.
> But I do not believe that it has ever been answered satisfactorily.
That's because it can't be anwered really satisfactorily
given that the CIFS protocol does not allow the server to
ping the client if it's still around. All you can do is to
apply heuristics.
> This may well reflect fundamental difficulties in the CIFS protocol
> However, recent incarnations of Windows Server seem to handle the problem
> better
> possibly due to tweaks in the underlying TCP/IP stack vis-a-vis Linux.
If you're on moderately recent Linux, you might want to play
with something like

socket options = SO_KEEPALIVE TCP_KEEPIDLE=300 TCP_KEEPINTVL=10 TCP_KEEPCNT=5

This should kick dead clients after something like 6
minutes. A description of these parameters can be found in
"man tcp".

With best regards,

Volker Lendecke

-- 
SerNet GmbH, Bahnhofsallee 1b, 37081 G?ttingen
phone: +49-551-370000-0, fax: +49-551-370000-9
AG G?ttingen, HRB 2816, GF: Dr. Johannes Loxen