[linux-cifs-client] (no subject)

[Shirish Pargaonkar]

On 7/18/08, Marko Käning <mk362 at mch.osram.de> wrote:
> > The 2.6.22 kernel is pretty old. You might want to try updating it to
> > get the benefits of the last couple of years worth of bugfixes.
>
> Oh, old!?!
> My god, pretty old? It's a SuSE 10.3 installation - the predecessor of the
> current SuSE 11
>
>
> > >  Linux server 2.6.16.53-0.8-smp #1 SMP Fri Aug 31 13:07:27 UTC 2007
> > >  x86_64 x86_64 x86_64 GNU/Linux
>
> The server who doesn't respond quickly enough is much older, isn't it!?!
>
>
> > Not much...mostly these messages are just telling you that the server
> > is being very slow to respond. We should probably tone down this
> > chatter though if that's not already been done in more recent kernels.
>
> Do you mean with "tone down" to "simply shut it up"? :)
>
>
> > The -EBADF errors are a bit more disturbing -- I'm not quite sure what
> > to make of those. Again, you might want to try patching up to a newer
> > kernel and see if the problem just goes away.
>
> Hmmm, can't update my kernel so easily, since I use only standard SuSE
> kernels via their update service.
> _______________________________________________

The EBADF errors occur when client reconnects to the server thinking
server is not responding, so the previous file handles become stale.
If you turn on cifs debugging, (i.e. echo 1 > /proc/fs/cifs/cifsFYI), you
will see reconnect going from client to server (or wireshark trace on
either cifs client machine or samba server machine will show that).

What kind of application are you running?  I have seen these messages
logged when an application such as fsstress is running on client machine.
At some point, nothing is going over socket on clent over the wire to the
server because tcp thinks server has not responded to requests, so
it keeps saying EAGAIN to cifs client (kernel_sendmsg).

> linux-cifs-client mailing list
> linux-cifs-client at lists.samba.org
> https://lists.samba.org/mailman/listinfo/linux-cifs-client
>