[jcifs] Connection Timeout
tony.thompson at stone-ware.com
Mon Jan 30 14:10:20 GMT 2006
OK, I upgraded Samba to 3.0.20b-3.4-suse and I am still experiencing the
same issue so, I suspect this is something I need to report to the Samba
guys? Something I have noticed is once Samba gets into this state, even
a Windows client takes a long time to get connected but, it does
connect. Can something be done to jCIFS to make it more resilient it
>>> Michael B Allen <mba2000 at ioplex.com> 01/25/06 08:19PM >>>
On Wed, 25 Jan 2006 15:32:43 -0500
"Tony Thompson" <tony.thompson at stone-ware.com> wrote:
> I am using jCIFS 1.2.7 running against Samba (YAST says
> 3.0.14a-0.4-suse) on a SuSE Linux server. I am having an issue that
> appears to be the same as was reported here:
> I only notice it failing after the connection has been idle for a
> (like the Linux server decide my client was done and should go away).
> If I retry the request after it fails, it works. I did a packet
> of the failed attempt and the server is responding to the SYN
> and even part of the conversation starts but then the server doesn't
> appear to respond with an SMB Protocol Response packet. I am
> the packet trace (hopefully it makes it).
> I have never noticed this happening with a Windows server so, I am
> sure if it is a Samba or a jCIFS issue.
Interesting. The capture only has 5 frames in it but indeed it shows
a SYN, to which Samba responds, and then what looks like a legitimate
Negotiate from JCIFS, to which Samba does NOT respond. It wouldn't
hold up in court but if it's true, I think it might be a Samba bug.
Can you temporarily increase the samba log level to see if anything
interesting is logged when Samba doesn't respond to the Negotiate?
Or if you don't feel like wearing your Sherlock Homes hat today you
might just try upgrading Samba to the latest. The Samba guys probably
won't look at the bug seriously unless you try a 3.0.20 version
More information about the jcifs