Hi We are using Samba on Debian Sarge (version 3.0.14a-Debian) and have it configured for simple file share access. All services work without issues on the Local network. However we have a number of remote workers who connect in via VPN (Via IPCOP and Linsys IPSEC), these workers are all using Windows XP The problem looks as though it may be related to oplocks or network timeouts / resets. Here are the symptoms Files are written but contain either corrupt data or no data at all. Somtimes the file is written and then deleted again (I suspect often no data has been transferred ) Windows client post errors such as The netowrk path is too deep The network name no longer exists We are consistently getting log messages like these Exhibit A. /var/log/sambs/log.d9rm152j d9rm152j (192.168.2.5) connect to service work initially as user keith (uid=10 04, gid=100) (pid 13006) [2006/07/17 09:01:46, 0] lib/util_sock.c:read_socket_data(384) read_socket_data: recv failure for 4. Error = Connection reset by peer [2006/07/17 09:01:46, 1] smbd/service.c:close_cnum(830) Exhibit B. /var/log/sambs/log.d9rm152j d9rm152j (192.168.2.5) connect to service work initially as user keith (uid=1004, gid=100) (pid 12116) [2006/07/17 08:43:06, 0] lib/util_sock.c:read_socket_with_timeout(321) read_socket_with_timeout: timeout read. read error = Connection reset by peer. [2006/07/17 08:43:06, 0] smbd/oplock.c:oplock_break(841) oplock_break: receive_smb error (Connection reset by peer) oplock_break failed for file Admin/test/DSC01506.JPG (dev = 1641, inode 7356435, file_id = 135). [2006/07/17 08:43:06, 0] smbd/oplock.c:oplock_break(931) oplock_break: client failure in break - shutting down this smbd. [2006/07/17 08:43:06, 1] smbd/service.c:close_cnum(830) --------Exhibits End --------------- smb.conf is as follows [global] log file = /var/log/samba/log.%m passwd chat = *Enter\snew\sUNIX\spassword:* %n\n *Retype\snew\sUNIX\spassword:* %n\n . socket options = TCP_NODELAY obey pam restrictions = yes encrypt passwords = true passdb backend = tdbsam guest passwd program = /usr/bin/passwd %u dns proxy = no netbios name = Emperor netbios aliases = emperor server string = %h server (Samba %v) invalid users = root path = /work default = global workgroup = MSHOME os level = 20 comment = Shared working directory create mode = 775 syslog = 0 panic action = /usr/share/samba/panic-action %d max log size = 1000 directory mode = 775 use sendfile = no [homes] create mask = 0700 comment = Home Directories browseable = no directory mask = 0700 writeable = yes path [printers] comment = All Printers browseable = no path = /tmp printable = yes public = no writable = no create mode = 0700 [print$] comment = Printer Drivers path = /var/lib/samba/printers browseable = yes read only = yes guest ok = no [work] writeable = yes Any help or documentation pointers would be really appreciated -- Thanks and Regards Rick Timmis Adaptive Linux Solutions Please ignore the stupid email address, this is in effort to avoid being avalanched by spam on my main accounts. If you want to get hold of me directly please put URGENT CONTACT in your subject line
Jeremy Allison
2006-Jul-17 15:11 UTC
[Samba] Time outs and data corruption on WAN using VPN
On Mon, Jul 17, 2006 at 11:51:11AM +0100, Rick Timmis wrote:> Hi > > We are using Samba on Debian Sarge (version 3.0.14a-Debian) and have it > configured for simple file share access. > All services work without issues on the Local network. > > However we have a number of remote workers who connect in via VPN (Via IPCOP > and Linsys IPSEC), these workers are all using Windows XP > > The problem looks as though it may be related to oplocks or network timeouts / > resets. > > Here are the symptoms > > Files are written but contain either corrupt data or no data at all. Somtimes > the file is written and then deleted again (I suspect often no data has been > transferred ) > > Windows client post errors such as > > The netowrk path is too deep > The network name no longer exists > > We are consistently getting log messages like these > > Exhibit A. > > /var/log/sambs/log.d9rm152j > > d9rm152j (192.168.2.5) connect to service work initially as user keith > (uid=10 > 04, gid=100) (pid 13006) > [2006/07/17 09:01:46, 0] lib/util_sock.c:read_socket_data(384) > read_socket_data: recv failure for 4. Error = Connection reset by peer > [2006/07/17 09:01:46, 1] smbd/service.c:close_cnum(830) > > Exhibit B. > > /var/log/sambs/log.d9rm152j > > d9rm152j (192.168.2.5) connect to service work initially as user keith > (uid=1004, gid=100) (pid 12116) > [2006/07/17 08:43:06, 0] lib/util_sock.c:read_socket_with_timeout(321) > read_socket_with_timeout: timeout read. read error = Connection reset by > peer. > [2006/07/17 08:43:06, 0] smbd/oplock.c:oplock_break(841) > oplock_break: receive_smb error (Connection reset by peer) > oplock_break failed for file Admin/test/DSC01506.JPG (dev = 1641, inode > 7356435, file_id = 135). > [2006/07/17 08:43:06, 0] smbd/oplock.c:oplock_break(931) > oplock_break: client failure in break - shutting down this smbd. > [2006/07/17 08:43:06, 1] smbd/service.c:close_cnum(830) > > --------Exhibits End ---------------Yep - your network is losing packets. For WAN use with a lossy network you might want to turn off oplocks on a WAN accessed share. Jeremy.