[Samba] Transport endpoint not connected

[Tom Peters]

Well, I WAS puzzled by this persistent error. Then I stumbled onto 
something and I might have fixed it. It looks like with an XP client, XP 
might be trying to renegotiate which port to talk (445 or 139) over at the 
start of every transaction. Whichever one answers first is used for the 
rest of the transfer.

I added this to smb.conf and restarted samba:

smb ports = 139

I can't tell if it's a permanent cure yet, but in 15 minutes of testing, I 
can't make it fail. Usually I can.

Is it true that using port 445 will exclude Win98 machines?

Here's the plea for help I WAS going to send out.

I'm puzzled by a persistent error that seems to be getting reported by smbd 
but is this a TCP/IP issue??

This is Samba 3.0.9-2.3-SUSE. I can read off any share on this machine all 
day/night with no issues. A large part of its reason to exist is serving up 
media files, and it does so without complaint, so long as you're just 
reading them.

At random, attempts to copy files to a Samba share will fail. The failure 
is instantaneous. This can happen even seconds after copying a group of 
files to the same folder successfully. I'm typically using a drag and drop, 
or more precisely, copy-and-paste of the files as listed on screen, by name.

Windows XP clients will report
"Cannot copy <filename>:  The specified network name is no longer 
available. [Ok]"

This results in a zero-length file on the destination. An immediate retry, 
if you answer "yes" to "...exists, overwrite?" always succeeds.

I tried using a compare-and-copy utility that continues even if it 
encounters an error. It does get this error sometimes, on the first file 
copied, then you have to click "ok" but then it goes on and copies all the 
files, successfully.

I used to have a very slow scsi raid array that was the destination of most 
of my file transfers. So I didn't actively pursue this issue too much. Now 
that I have faster SATA drives on a 3Ware card (I benchmarked them- the 
difference is dramatic) and the problem persists, I'm hoping someone can 
give me some help.

Copying a number of files always succeeds if the first one succeeds, 
although I can sometimes make a copy of a number of files fail by 
attempting to start another copy to anywhere on the same share. I haven't 
tested the parameters of that effect yet.

This appears in my messages file whenever this problem occurs:
Mar 13 18:00:30 tolkien smbd[16594]: [2006/03/13 18:00:30, 0] 
lib/util_sock.c:get_peer_addr(1136)
Mar 13 18:00:30 tolkien smbd[16594]:   getpeername failed. Error was 
Transport endpoint is not connected
Mar 13 18:00:30 tolkien smbd[16594]: [2006/03/13 18:00:30, 0] 
lib/util_sock.c:get_peer_addr(1136)
Mar 13 18:00:30 tolkien smbd[16594]:   getpeername failed. Error was 
Transport endpoint is not connected
Mar 13 18:00:35 tolkien smbd[16594]: [2006/03/13 18:00:35, 0] 
lib/util_sock.c:get_peer_addr(1136)
Mar 13 18:00:35 tolkien smbd[16594]:   getpeername failed. Error was 
Transport endpoint is not connected
Mar 13 18:00:35 tolkien smbd[16594]: [2006/03/13 18:00:35, 0] 
lib/util_sock.c:write_socket_data(430)
Mar 13 18:00:35 tolkien smbd[16594]:   write_socket_data: write failure. 
Error = Connection reset by peer
Mar 13 18:00:35 tolkien smbd[16594]: [2006/03/13 18:00:35, 0] 
lib/util_sock.c:write_socket(455)
Mar 13 18:00:35 tolkien smbd[16594]:   write_socket: Error writing 4 bytes 
to socket 5: ERRNO = Connection reset by peer
Mar 13 18:00:35 tolkien smbd[16594]: [2006/03/13 18:00:35, 0] 
lib/util_sock.c:send_smb(647)
Mar 13 18:00:35 tolkien smbd[16594]:   Error writing 4 bytes to client. -1. 
(Connection reset by peer)

Someone please toss me a clue or three?

Tom

Contents of samba-log.<xpmachineipaddress>
[2006/03/13 18:00:30, 1] lib/util_sock.c:get_peer_name(1095)
   Gethostbyaddr failed for 172.20.0.186
[2006/03/13 18:00:30, 2] lib/access.c:check_access(324)
[2006/03/13 18:00:30, 0] lib/util_sock.c:get_peer_addr(1136)
   getpeername failed. Error was Transport endpoint is not connected
[2006/03/13 18:00:30, 0] lib/util_sock.c:get_peer_addr(1136)
   getpeername failed. Error was Transport endpoint is not connected
[2006/03/13 18:00:31, 1] lib/util_sock.c:get_peer_name(1095)
   Gethostbyaddr failed for 172.20.0.186
[2006/03/13 18:00:31, 2] lib/access.c:check_access(324)
   Allowed connection from 172.20.0.186 (172.20.0.186)
[2006/03/13 18:00:31, 2] smbd/reply.c:reply_special(235)
   netbios connect: name1=TOLKIEN         name2=ELROND
[2006/03/13 18:00:31, 2] smbd/reply.c:reply_special(242)
   netbios connect: local=tolkien remote=elrond, name type = 0
[2006/03/13 18:00:35, 1] lib/util_sock.c:get_peer_name(1095)
   Gethostbyaddr failed for 0.0.0.0
   Allowed connection from 0.0.0.0 (0.0.0.0)
[2006/03/13 18:00:35, 2] smbd/reply.c:reply_special(235)
   netbios connect: name1=TOLKIEN         name2=ELROND
[2006/03/13 18:00:35, 2] smbd/reply.c:reply_special(242)
   netbios connect: local=tolkien remote=elrond, name type = 0

Contents of samba-log.elrond
   [2006/03/13 18:00:01, 2] smbd/close.c:close_normal_file(270)
   tom closed file Science-Tech/Falkirk_bearing.avi (numopen=1)
[2006/03/13 18:00:31, 2] smbd/sesssetup.c:setup_new_vc_session(608)
   setup_new_vc_session: New VC == 0, if NT4.x compatible we would close 
all old resources.
[2006/03/13 18:00:31, 2] smbd/sesssetup.c:setup_new_vc_session(608)
   setup_new_vc_session: New VC == 0, if NT4.x compatible we would close 
all old resources.
[2006/03/13 18:00:31, 2] auth/auth.c:check_ntlm_password(305)
   check_ntlm_password:  authentication for user [tom] -> [tom] -> [tom] 
succeeded
[2006/03/13 18:00:31, 2] lib/access.c:check_access(324)
   Allowed connection from 172.20.0.186 (172.20.0.186)
[2006/03/13 18:00:31, 1] smbd/service.c:make_connection_snum(647)
   elrond (172.20.0.186) signed connect to service movies2 initially as 
user tom (uid=1000, gid=100) (pid 16595)
[2006/03/13 18:00:35, 0] lib/util_sock.c:get_peer_addr(1136)
   getpeername failed. Error was Transport endpoint is not connected
[2006/03/13 18:00:35, 0] lib/util_sock.c:write_socket_data(430)
   write_socket_data: write failure. Error = Connection reset by peer
[2006/03/13 18:00:35, 0] lib/util_sock.c:write_socket(455)
   write_socket: Error writing 4 bytes to socket 5: ERRNO = Connection 
reset by peer
[2006/03/13 18:00:35, 0] lib/util_sock.c:send_smb(647)
   Error writing 4 bytes to client. -1. (Connection reset by peer)
[2006/03/13 18:00:35, 2] smbd/server.c:exit_server(575)
   Closing connections




[Children] A child miseducated is a child lost. --John F. Kennedy
--... ...--  -.. .  -. ----. --.- --.- -...
tpeters at nospam.mixcom.com   (remove "nospam") N9QQB (amateur radio)
"HEY YOU" (loud shouting)  WEB ADDRESS http//www.mixweb.com/tpeters
43° 7' 17.2" N by 88° 6' 28.9" W,  Elevation 815',  Grid Square EN53wc
WAN/LAN/Telcom Analyst, Tech Writer, MCP, CCNA, Registered Linux User 385531