Re: CIFS mount hangs
Hi,
so I did some testing with a clone of the machine restored to the
state before the upgrade to wheezy.
Squeeze:
- Mount the share
- Disable the share on windows host
- Try to access the share with 'ls' results in 'no such device'
- Enable the share on windows host
- Try to access the share with 'ls' is successful
Wheezy:
- Mount the share
- Disable the share on windows host
- Try to access the share with 'ls' results in 'Server ... has not
responded in 300 seconds' .. and an error from ls after well 300
seconds i would say. Until this the shell hangs and the access can not be interupted with
CTRL-C.
- Enable the share on windows host
- Try to access the share with 'ls' is successful but shows an empty
dir.
- Remount brings the content back
- Connection is lost after some time of inactivity (share is still
available under squeeze)
This is the behavoiur when I manually disable the share. But in 'real
life' the share is not diabled by anyone. It is still visible in the
windows network environment.
I have now tried this:
http://www.mail-archive.com/linux-cifs@vger.kernel.org/msg06161.html
Will see how this will work out.
Thanks so far!
Tobias
> On 21/11/13 22:39, assmann@skygate.de wrote:
>> Hi,
>>
>> I tried your tricks, removing auto, adding _netdev but the problem
>> still exists. I seems to happen when I am not working on the share for
>> a certain amount of time. The share seems to be lost then....this did
>> not happen under wheezy. So I am thinking about, what kind of code to
>> blame? mount.cifs? Maybe try some kind of APT-Pinning to find out?
>> (Never tried this till before).
> I'd try 1x1 (isolation testing) first. Starting with "# mount -a" when
> the network resource is unavailable and looking at any error messages.
> Or try a manual mount (when the network resource is unavailable).
> How does "$ testparm -v" on the new box compare with the same from the
> box that did work?
>>
>> Any ideas to debug and understand are welcome.
> The output of:
> $ testparm -v
> Plus the settings used on the boxes that host the samba resources might
> be informative.
> Do you have any logs from the other boxes?
>>
>> TIA!
>>
>>> On 19/11/13 23:33, assmann@skygate.de wrote:
>>>> Dear list,
>>>>
>>>> since my upgrade to wheezy I have a problem with a cifs mount. It is a
>>>> share of my windows host. The setup worked fine in debian squeeze.
>>>>
> <snipped>
> the _netdev will create an error message, but it still works (allows a
> normal boot if a cifs resource is unavailable).
> None of the wheezy boxes I build have problems with cifs dropping out
> that isn't the fault of the Windoof part of the network, nor do I know
> of any problems similar to yours - so my best guess would be that Debian
> is the wrong end of the problem. I didn't see any relevant bug
> reports.... perhaps if you tell us more about the rest of the network?
> With Squeeze I did have some problems with a mostly Windoof WAN (some
> Apples) due to routing problems but that was solved by using autofs. If
> you can't solve the cause of the problem by looking at the logs on the
> other boxes in your network I'd suggest autofs.
> You could also try the samba mailing list? samba@lists.samba.org
> Kind regards.
Mit freundlichen Grüßen
Tobias Assmann
_______________________________________________________
SkyGate internetworking GmbH
Pfuelstrasse 5, Aufgang VI
D - 10997 Berlin
Handelsreg. Berlin Charlottenburg, HRB 87258
Geschaeftsfuehrer: Stephan Jensen
T: +49- (0)30 - 611038-0
F: +49- (0)30 - 61280465
W: http://www.skygate.de
_______________________________________________________
Reply to: