Koozali.org: home of the SME Server

smbd error logs? Failure to read file and directories

pjwallis

smbd error logs? Failure to read file and directories
« on: June 30, 2004, 05:58:33 AM »
I am running esmith 5.6 unsupported in a network environment with XP Pro. There are lots of occasional errors where the files or directories are not available from the client machines.

The SMBD log contains lots of entries some are included below. I have searched the manuals and checked on the net and now would appreciate some help in reading and understanding the Logs for Samba.

[2004/06/23 13:39:52, 0] smbd/oplock.c:request_oplock_break(1008)
  request_oplock_break: no response received to oplock break request to pid 14593 on port 33632 for dev = 303, inode = 16507, file_id = 374
[2004/06/23 13:39:52, 0] smbd/open.c:open_mode_check(630)
  open_mode_check: exlusive oplock left by process 14593 after break ! For file Attachments to Exclusive Agency Agreements/Section 49 Consent.doc, dev = 303, inode = 16507. Deleting it to continue...
[2004/06/23 13:39:52, 0] smbd/open.c:open_mode_check(634)
  open_mode_check: Existent process 14593 left active oplock.
[2004/06/23 13:41:00, 0] smbd/oplock.c:request_oplock_break(1008)
  request_oplock_break: no response received to oplock break request to pid 14593 on port 33632 for dev = 303, inode = 16506, file_id = 377
[2004/06/23 13:41:00, 0] smbd/open.c:open_mode_check(630)
  open_mode_check: exlusive oplock left by process 14593 after break ! For file Attachments to Exclusive Agency Agreements/Section 47 Disclosure.doc, dev = 303, inode = 16506. Deleting it to continue...
[2004/06/23 13:41:00, 0] smbd/open.c:open_mode_check(634)
  open_mode_check: Existent process 14593 left active oplock.
[2004/06/23 13:43:03, 0] smbd/oplock.c:request_oplock_break(1008)
  request_oplock_break: no response received to oplock break request to pid 14593 on port 33632 for dev = 303, inode = 3041576, file_id = 310
[2004/06/23 13:43:03, 0] smbd/open.c:open_mode_check(630)
  open_mode_check: exlusive oplock left by process 14593 after break ! For file 22 June 2004.doc, dev = 303, inode = 3041576. Deleting it to continue...
[2004/06/23 13:43:03, 0] smbd/open.c:open_mode_check(634)
  open_mode_check: Existent process 14593 left active oplock.
[2004/06/24 11:26:49, 0] lib/util_sock.c:read_data(436)
  read_data: read failure for 4. Error = Connection reset by peer
[2004/06/28 14:45:26, 0] smbd/oplock.c:oplock_break(796)
  oplock_break: receive_smb timed out after 30 seconds.
  oplock_break failed for file SOLD/pwrptdemo.exe (dev = 303, inode = 2404127, file_id = 21).
[2004/06/28 14:45:26, 0] smbd/oplock.c:oplock_break(868)
  oplock_break: client failure in oplock break in file SOLD/pwrptdemo.exe
[2004/06/28 14:45:26, 0] smbd/reply.c:reply_lockingX(4489)
  reply_lockingX: Error : oplock break from client for fnum = 9261 and no oplock granted on this file (SOLD/pwrptdemo.exe).
[2004/06/29 08:56:16, 0] lib/util_sock.c:read_data(436)
  read_data: read failure for 4. Error = Connection reset by peer
Bottom of Form

Yours sincerely,
Peter Wallis

Offline Reinhold

  • *
  • 517
  • +0/-0
    • http://127.0.0.1
smbd error logs? Failure to read file and directories
« Reply #1 on: June 30, 2004, 08:17:00 AM »
Peter,

Check your network (network-cabling)
...definitely if you are running 10Mbit Coax or hub not switch...
The errors you see may all be attributed to failure of network communication with the XP machine!

Regards
Reinhold
............

Anonymous

smbd error logs? Failure to read file and directories
« Reply #2 on: June 30, 2004, 11:49:23 AM »
I am using Cat 5 with switches and NO hubs. I have recently installed new switches and cabling into the business and it was a clean install at the time.

Peter Wallis

Offline Reinhold

  • *
  • 517
  • +0/-0
    • http://127.0.0.1
smbd error logs? Failure to read file and directories
« Reply #3 on: July 01, 2004, 09:50:05 AM »
Peter,

Since your problems do come from communication failures,
- check the server network card,
- check each xp-client connection (by data transfer and looking for errors)

I do know too little about your lan but
if the speed difference of the PCs involved is very large you might try (as a kludge) to disable locking:
- ssh into the server
pico /etc/smb.conf
change:
oplocks = true
into
oplocks = false

In any case your XP side (from the server nic on) is not answering in time in the samples you show.

Regards
Reinhold
............

pjwallis

smbd error logs? Failure to read file and directories
« Reply #4 on: July 02, 2004, 01:14:40 AM »
Thanks Reinhold for the suggestion - I will give it a go.

Do you or anyone know where I can find the information to understand the smbd logs?

Thanks for all you help
Yours sincerely,
Peter Wallis

Offline Reinhold

  • *
  • 517
  • +0/-0
    • http://127.0.0.1
smbd error logs? Failure to read file and directories
« Reply #5 on: July 02, 2004, 09:47:15 AM »
Peter,

I can recommend my (older) copy of "Using Samba" from O'Reilly which is quite helpful in tracing Samba problems.
It's probably not that detailed on the log you gave but as most O'Reilly books you may read it (at least in part) online. Here's a quick link:
http://www.oreilly.com/catalog/samba/chapter/book/ch09_01.html


Have you tried to increase die smbd -d  "debug level" ?

Regards
Reinhold
............

k_graham

Re: smbd error logs? Failure to read file and directories
« Reply #6 on: July 07, 2004, 05:29:07 PM »
Quote from: "pjwallis"
I am running esmith 5.6 unsupported in a network environment with XP Pro. There are lots of occasional errors where the files or directories are not available from the client machines.
Peter Wallis


Are you running mirrored drives and is one malfunctioning or unhooked?

k_graham

pjwallis

smbd error logs? Failure to read file and directories
« Reply #7 on: July 08, 2004, 03:08:47 PM »
Hi,
Thanks Reinhold for the tips on removing oplocks. Have done that and no apparent errors yet. Tomorrow will be a real test of the system.
The accounting software this place is running specifically states that oplocks should be turned OFF, and now they are.



I am also starting to get a picture that one of the switches in the network may have a faulty port and it has been removed to see if the high numbers of fragments, collisions, droppkt, and RXbad dissappear.

Thanks for all your help.
Peter Wallis