[Samba] smb log error-Transport end point/getpeername

david rankin drankin at cox-internet.com
Fri Feb 11 05:48:28 GMT 2005

From: "Gerald (Jerry) Carter" <jerry at samba.org>
> david rankin wrote:
> |
> |> I know that Windows XP clients try to connect to port
> |> 139 and 445 in parallel and drop the connection to port
> |> 139 if the connect to 445 is successful.  This is probably
> |> what you are seeing.
> |>
> |
> | Thanks Jerry,
> |
> |    That makes sense. No doubt an intentional feature from
> | $. Is there any way to get around it? When it occurs, there is
> | a real slowdown in  the connection. It is intermittent so
> | it doesn't last that long, but it is sure annoying when
> | it does.....
> You could set 'smb ports = 139' for force netbios connections.
> Are you sure this is causing a slowdown ?  Other than the
> overhead of forking a new smbd temporarily.

    Well, no, I'm not positive, but when whatever hangs, I get a big 
slowdown in file access, file open, etc.. from XP. When I then go and check 
the logs the transport end point and getpeername is the only thing that 
shows up in the log. I admit I haven't had time to put up a console and 
check in the past few days. It is definately an XP issue. It occurs on both 
Suse 9.0 and Mandrake 7.2 with 3.0.7. It occurred with all the 2.0 and 2.2 
sambas as well. I'll keep an eye on it and report back.

David C. Rankin, J.D., P.E.
510 Ochiltree Street
Nacogdoches, Texas 75961
(936) 715-9333
(936) 715-9339 fax

More information about the samba mailing list