[plug] smb client timeout problem - FIXED

[Steve Boak]

Update...

Just in case someone else comes up with this problem, I found the answer by 
combining all the information (more hints actually) that I gleaned from 
google, and then taking a stab in the dark :-)

The simple fix is to use the cifs filesystem instead of smbfs when mounting 
remote samba shares on a Linux box. Apparently cifs has superceded smbfs and 
works much better in this case, but I still need to use smbfs when mounting 
shares from my windows box. Information is rather hard to find on the subject 
of samba/smbfs/cifs.

I had to recompile the kernel to get cifs support.

Not sure of the finer details yet, but at least now I have a working system 
again. Here's the combinations I have found that works:

(Note that all my tests were done with at least kernel 2.6.13, Samba 3.0.20b, 
and Debian Testing on both machines)

Linux client	mount -t cifs ...	shares from Samba server
Linux client	mount -t smbfs ...	shares from Windows-2000
Windows client	uses smbfs?	shares from Samba server

More reading to do...

Steve



On Sat, 12 Nov 2005 09:41 am, Steve Boak wrote:
> Hi all
>
> I'm getting weird results trying to mount smb shares from one linux box to
> the other. Both machines are running Debian Testing with  sambd
> 3.0.20b-Debian, Kernel 2.6.12.2 on the server and 2.6.14.1 on the client.
>
> My 'client' machine is dual-boot, and with windows-2000 I can access all
> the server shares with no problems at all.
>
> When I boot the client machine into Linux, I get the I/O timeouts trying to
> use a mounted smb share on the server.
>
> Since I have samba running on both machines, I tried it the other way - ie
> if I try to mount a share from the client onto the server, I get the same
> problem. So it seems to be something to do with smbfs mount, not the
> server, and probably the 'testing' version, but I am at a loss as to where
> to look next.
>
> Hmm, just tried something else - mounting the same smb share from the
> server locally (on the same machine) - it mounts, but I can't access it
> there either!
>
> Google throws up a few similar errors and a couple of config changes, but
> no conclusive fixes. Using the cifsfs file system instead of smbfs is
> mentioned, but I haven't looked into that yet.
>
> Examples:
>
> shuttle:/home/steve# mount -t smbfs -o username=steve //min/download
> /mnt/tmp Password:
> shuttle:/home/steve#
>
> seems the mount went OK...
>
> shuttle:/home/steve# mount
> [...]
> //min/download on /mnt/tmp type smbfs (rw)
>
> But...
>
> shuttle:/home/steve# ls /mnt/tmp
> ls: /mnt/tmp: Input/output error
> (exactly 30 seconds to timeout - is this significant?)
>
> Various log file contents:
>
> /var/log/samba/log.smbmount on the *client* shows this:
>
> [mount done here]
> [2005/11/11 16:26:13, 0] client/smbmount.c:send_fs_socket(412)
>  mount.smbfs: entering daemon mode for service \\min\download, pid=19000
> ['ls' done here]
> [2005/11/11 16:27:03, 0] tdb/tdbutil.c:tdb_log(767)
>  tdb(/var/run/samba/gencache.tdb): tdb_lock failed on list 80 ltype=0 (Bad
> file descriptor)
> [2005/11/11 16:27:03, 0] tdb/tdbutil.c:tdb_log(767)
>  tdb(/var/run/samba/gencache.tdb): tdb_lock failed on list 80 ltype=1 (Bad
> file descriptor)
> [2005/11/11 16:27:03, 0] tdb/tdbutil.c:tdb_log(767)
>  tdb(/var/run/samba/gencache.tdb): tdb_lock failed on list 48 ltype=0 (Bad
> file descriptor)
> [2005/11/11 16:27:03, 0] tdb/tdbutil.c:tdb_log(767)
>  tdb(/var/run/samba/gencache.tdb): tdb_lock failed on list 80 ltype=1 (Bad
> file descriptor)
>
> /var/log/samba/log.192.168.1.16 on the *server* shows this:
>
> [2005/11/11 16:27:03, 1] smbd/service.c:make_connection_snum(662)
>   192.168.1.16 (192.168.1.16) signed connect to service download initially
> as usr steve (uid=0, gid=1000) (pid 13043)
> [2005/11/11 16:27:03, 0]
> libsmb/smb_signing.c:srv_check_incoming_message(782)
> srv_check_incoming_message: BAD SIG: seq 4 wanted SMB signature of
> [2005/11/11 16:27:03, 0]
> libsmb/smb_signing.c:srv_check_incoming_message(786)
> srv_check_incoming_message: BAD SIG: seq 4 got SMB signature of
> [2005/11/11 16:27:03, 0] libsmb/smb_signing.c:signing_good(240)
>   signing_good: BAD SIG: seq 4
> [2005/11/11 16:27:03, 0] lib/util_sock.c:receive_smb(734)
>   receive_smb: SMB Signature verification failed on incoming packet!
> [2005/11/11 16:27:03, 1] smbd/service.c:close_cnum(835)
>   192.168.1.16 (192.168.1.16) closed connection to service download
>
> /var/log/syslog on the *client* shows this:
>
> Nov 11 16:27:03 localhost kernel: SMB connection re-established (-5)
> Nov 11 16:27:33 localhost kernel: smb_add_request: request [e19a4e80,
> mid=1] timed out!
> Nov 11 16:27:33 localhost kernel: smb_add_request: request [e19a4d80,
> mid=2] timed out!
>
> Any ideas?

-- 
Steve Boak, P.O. Box 240, Nannup, WA 6275
Nannup Telecentre IT support, ph (08) 9756 3022
Free World Dialup (voip) number 454566