Bonjour,
I'm running samba 2.0.4 on a Sparc Solaris 2.5.1 and I have some error logs
when
I'm trying to connect to a sharing directory from *some* NT4 WS (SP4 or
SP5):
pchp2 (194.xxx.xxx.xxx) connect to service centreAS as user truc (uid=1326,
gid=1320) (pid 22939)
[1999/08/26 15:52:57, 0] lib/util_sock.c:write_data(415)
write_data: write failure. Error = Broken pipe
[1999/08/26 15:52:00, 0] lib/util_sock.c:write_data(415)
write_data: write failure. Error = Broken pipe
[1999/08/26 15:52:02, 0] lib/util_sock.c:write_data(415)
write_data: write failure. Error = Broken pipe
[1999/08/26 15:52:14, 0] lib/util_sock.c:write_data(415)
..
..
..
[1999/08/26 15:53:08, 0] smbd/oplock.c:oplock_break(782)
oplock_break: end of file from client
oplock_break failed for file gest99/table.dbf (dev = 8000c7, inode = 506966).
[1999/08/26 15:53:08, 0] smbd/oplock.c:oplock_break(859)
oplock_break: client failure in break - shutting down this smbd.
[1999/08/26 15:53:08, 1] smbd/service.c:close_cnum(514)
[1999/08/26 11:58:40, 1] smbd/service.c:close_cnum(514)
pchp2 (0.0.0.0) closed connection to service centreAS
This happens on some workstations without any particular reason... I guess.. :-)
However I try to make some connections from others workstations (same user's
login) and all works !
If you have some ideas... just have a look at the last line, there is a strange
IP number (0.0.0.0) associated with pchp2...
Regards
Jean-Marc VINET / IRD (ex ORSTOM)
Admin. Systeme et Reseau / Service Informatique
Tel: +33 (0)1 48 02 56 86 / 32, avenue Henri Varagnat
Fax: +33 (0)1 48 02 56 93 / 93143 Bondy Cedex
vinet@paris.ird.fr / FRANCE
-----------------------------------------------------------------
Intitut de Recherche pour le Developpement
Web : http://www.ird.fr
Have searched through the mailing list archives and have seen this show up a couple of other times, but no answers seem to be forthcoming. I am runing Samba 2.0.6 on a unix hp-ux 10.20 system and keep seeing the following errors in the logs: lib/util_sock.c:write_socket_data(537) write_socket_data: write failure. Error = broken pipe followed by: lib/util_sock.c:writewrite_socket(563) write_socket: Error writing 4 bytes of data to socket 5: ERRNO = Broken pipe What's going on here and how can I fix it. I'm using server level security and have enabled the TCP_NODELAY socket option as described in the documentation. At present, there are about 200 people that use Samba to access Excel spreadsheets . The user loose the file that I have to restore on his client Thanks,
Lorenzo was seeing: | lib/util_sock.c:write_socket_data(537) write_socket_data: write failure. | Error = broken pipe ... | lib/util_sock.c:writewrite_socket(563) write_socket: Error writing 4 bytes | of data to socket 5: ERRNO = Broken pipe This is usually a cry of pain from Samba, who's lost its connection to the client. The previous reports that I remember were about the client timing out and reporting a problem, after which it had to be reconnected and the file re-written. Can you describe to the list what this looks like at the client as well at the server? Is the file large, and is this specific to a small number of machines, or hitting everyone? --dave -- David Collier-Brown in Boston Phone: (781) 442-0734, Room BUR03-3632
Anyone know what this error is? write_socket_data: write failure. Error = Broken pipe I'm not getting complaints from users but it is of concern. please reply to dcanedo@concero.com thanks, -- Dave Canedo Concero / IT Department
On Mon, 11 Sep 2000, David C. Canedo wrote:> Anyone know what this error is? > write_socket_data: write failure. Error = Broken pipe > > I'm not getting complaints from users but it is of concern.This normally means that smbd tried to send some bytes to a network connection that has been broken off unexpectedly by the other party. You will get some responses from people who say that you've probably got network problems or crashing clients, because the error often occurs when a client system "goes away" without correctly closing down its connection for some reason. I know of at least one situation where such messages pop up regularly and are *not* indicative of a problem of any sort, however. If you are using "security = server" then every six minutes or so Samba will attempt to send a NetBIOS keepalive packet to the password server, and will continue to attempt to do this even after the authentication connection has been dropped. If this is the source of your log messages then you can quell them by adding "keepalive = 0" to the [global] section of your smb.conf file. I wrote a report on this problem for the Samba list recently: http://lists.samba.org/pipermail/samba/2000-August/024526.html Regards, -- Neil Hoggarth Departmental Computer Officer <neil.hoggarth@physiol.ox.ac.uk> Laboratory of Physiology http://www.physiol.ox.ac.uk/~njh/ University of Oxford, UK
Hi everybody, I've set up a Samba server. It works as a backup server for win98 and NT 4.0 Workstation. I can copy files from Win to Samba. But sometimes I can not copy them back. It concerns mainly bigger files (aprox. 50 MB or more). In the log file I find: [2000/11/20 10:40:13, 1] smbd/service.c:make_connection(550) ballantines (192.168.1.2) connect to service Exchange as user ballantines (uid=502, gid=500) (pid 22742) [2000/11/20 12:50:45, 0] lib/util_sock.c:write_data(508) write_data: write failure. Error = Broken pipe [2000/11/20 12:50:45, 0] lib/util_sock.c:read_socket_data(477) read_socket_data: recv failure for 4. Error = Connection reset by peer [2000/11/20 12:50:45, 1] smbd/service.c:close_cnum(583) ballantines (192.168.1.2) closed connection to service Exchange What's going on? I have no idea. What do these errors mean at all? Thx for any answer Przemek
Hello, I use samba 2.0.7 on solaris 2.7 and 2.8. Every 6 minutes I have a broken pipe in log.smb on hosts where "security = SERVER" in smb.conf but never on host where "security = USER". ... [2001/03/16 13:59:38, 0] lib/util_sock.c:write_socket_data(540) write_socket_data: write failure. Error = Broken pipe [2001/03/16 14:05:38, 0] lib/util_sock.c:write_socket_data(540) write_socket_data: write failure. Error = Broken pipe ... Thank you for any help -- Guy Roussin
[global]
workgroup = ARBEITSGRUPPE
netbios name = XXX
server string = The whole show
security = SERVER
encrypt passwords = Yes
map to guest = Bad User
password server = XXX
keepalive = 30
os level = 2
kernel oplocks = No