[Samba] Fwd: samba & Oracle ACFS Issues

Nacho del Rey odelreym at gmail.com
Wed Dec 31 01:44:44 MST 2014


Hi Volker

Finally I loaded the following configuration in our test environment

[global]

        deadtime = 60

        keepalive = 10


[dossiers]


        locking = Yes

        ; If yes, turns on byte-range locks.

        strict locking = No

        ; If yes, denies access to an entire file if a byte-range lock
exists in it.

        ; posix locking = Yes

        posix locking = No

        ; If yes, maps oplocks to POSIX locks on the local system.

        oplocks = no

        ; If yes, turns on local caching of files on the client for this
share.

        level2 oplocks = No

        ; If yes, allows oplocks to downgrade to read-only.

        fake oplocks = No

        ; If yes, tells client the lock was obtained, but doesn't actually
lock it.

        blocking locks = Yes

        ; Allows lock requestor to wait for the lock to be granted.

I have been reading a lot about samba & win7 and it is probably we have
another issue when the windows clients enter in hibernating mode, so the
parameters deadtime & keepalive in global section

Yesterday we did a bunch of tests and, for the moment, it seems to run
properly

Thanks for your clue
I will update this issue as soon as I will confirm the solution is permanent

Kind regards and have a happy new year

Nacho.






2014-12-17 8:48 GMT+01:00 Nacho del Rey <odelreym at gmail.com>:

>
>> > >Please remove the SO_RCVBUF and SO_SNDBUF entries.
>> > >Oh, that's ancient. Do you have any possibility to move to
>> > >"security=user"?
>> >                  The smb.conf file was inherited from HP-UX system. Ok,
>> > I'll remove them
>> >
>> > >Did you try "posix locking = no"? That is mostly criticial
>> > >if you are exporting files from a file system with
>> > >unreliable locking like for example NFS.
>> >
>> >           There is no NFS at all over the fs. It is only ACFS + samba.
>> We
>> > tried to disable any interference with posix locks over the ACFS, so the
>> > parameter was set to no. Is it right?
>>
>> In your smb.conf, it was set to yes, that's why I was
>> asking. Setting "posix locking = no" is the right thing to
>> do in this case. I mentioned NFS just as an example of a
>> file system where locking can be problematic. It sounds like
>> ACFS also could have problems here, thus the analogy.
>>
>> Ah, ok. I did not understand you properly. We'll try this change
>
>
>> >
>> > >Can you see what the smbd hosting such a blocked client does?
>> > >If it is in D state (according to ps u), it sits in the
>> > >kernel. If not, you could try stracing the process (strace
>> > >-ttT -p <pid>) and see what it does. gstack <pid> also helps
>> > >often.
>> >
>> >    We got the following strace over the samba process 'locked'
>>
>> The strace looks innocent. Have you been able to see what
>> state the process was in while the client was blocked?
>>
>>
> I will ask for it to my workmates if they can have it one 'top' capture (I
> don't think so)
>
> I'll come back to you as soon as I get the results
>
> Many thanks Volker
>
> Nacho
>
>
>


More information about the samba mailing list