file copy problems

[Kirk Patton]

Hello,

I seem to be experiencing two problems with Samba lately.

The first is when a user is using MS Outlook and tries to move a mail message from the
server to a personal folder.  The process begins and hangs a while before giving an error
that the folder is not accessible and to close all mail enable apps and restart them.

I set the log level to 5 on the smbd process controlling the connection and noted the
following.

[1999/07/22 10:30:01, 3] smbd/process.c:switch_message(402)
  switch message SMBlockread (pid 13197)
[1999/07/22 10:30:01, 4] smbd/uid.c:become_user(209)
  Skipping become_user - already user
[1999/07/22 10:30:01, 3] lib/util.c:fcntl_lock(2658)
  fcntl lock gave errno 11 (Resource temporarily unavailable)
[1999/07/22 10:30:01, 3] lib/util.c:fcntl_lock(2680)
  lock failed at offset 481 count 1 op 6 type 1 (Resource temporarily unavailable)
[1999/07/22 10:30:01, 3] smbd/blocking.c:push_blocking_lock_request(124)
  push_blocking_lock_request: lock request length=49 blocked with expiry time -1 for fnum
= 4255, name = Mail/keisoon.pst
[1999/07/22 10:30:49, 3] smbd/process.c:timeout_processing(755)
  end of file from client

The other problem is when a user tries to copy a file.  When the copy is made
to the samba share, it takes an unusually long time.  When it completes, the file is
listed as being 4k in size from the windows client.  If the copy operation is retried,
often it is successful.  I set the log level to 5 and observed the following.

[1999/07/22 16:31:07, 5] lib/util.c:show_msg(475)
  smb_bcc=0
[1999/07/22 16:32:16, 0] lib/util_sock.c:read_socket_data(507)
  read_socket_data: recv failure for 11482. Error = Connection reset by peer
[1999/07/22 16:32:16, 3] smbd/process.c:timeout_processing(835)
  receive_smb error (Connection reset by peer) exiting
[1999/07/22 16:32:16, 3] lib/doscalls.c:dos_ChDir(336)


The samba server is running on a RedHat box with kernal 2.2.5-15.  
It is acting as a go between from the main NFS file server and the MS clients.

I noted that the problem does not appear when the samba servers local disk share is used.
Only when the samba server is writing data from a client to a remote NFS mount on the
Samba server.

Has anyone else seen a similiar problem or better yet, know of a fix?

Here is the global section from my smb.conf in case it is revelent.


[global]
  printing=LPRNG
  username map = /usr/local/samba/private/user.map
  netbios name = cmdsmb1
  netbios aliases = ppdsamba
  remote announce = 139.95.53.51
  remote browse sync = 139.95.53.51
  name resolve order = wins bcast lmhosts host 
  server string = "CMD Samba Server"
  workgroup = amdrdsvl
  debuglevel = 1
  security = domain
  password server = ssvlamd16 ssvlamd15
  log file = /usr/local/samba/var/log.smb.%M.%L
  lock dir = /var/samba
  encrypt passwords = yes
  socket options = TCP_NODELAY, SO_SNDBUF=4096, SO_RCVBUF=4096
  deadtime = 360
  admin users = +admin


-- 
Kirk Patton
System Administrator
CA Microprocessor Division
Advanced Micro Devices
Phone (408) 774-7883
Pgr. (408) 870-5461
Fax (408)774-7606