[CentOS] CentOS samba MS Word corruption

Fri Mar 16 13:20:30 UTC 2007
Fred Kienker <fred at at4b.com>

Samba's opportunistic locking and MS Office applications don't get along.
For a small performance hit you can fix the problem by adding:

oplocks = no

to the [global] section and fix your problem.

Just my $.02

============================================
Fred Kienker
AT4B
5261 Sunset Trail
Marietta, GA 30068

Fred at at4b.com <<mailto:Fred at at4b.com>>
770.518.6166 Phone
770.518.6992 Fax

"Advanced Technologies For Business"
============================================
This transmission may contain information that is privileged, confidential
and/or exempt from disclosure under applicable law. If you are not the
intended recipient, you are hereby notified that any disclosure, copying,
distribution, or use of the information contained herein (including any
reliance thereon) is STRICTLY PROHIBITED. If you received this transmission
in error, please immediately contact the sender and destroy the material in
its entirety, whether in electronic or hard copy format. Thank you.



-----Original Message-----
From: centos-bounces at centos.org [mailto:centos-bounces at centos.org]On
Behalf Of Paul Schoonderwoerd
Sent: Friday, March 16, 2007 8:22 AM
To: CentOS mailing list
Subject: [CentOS] CentOS samba MS Word corruption


We're using CentOS samba for about 20 workstations, mostly XP, some 2000 and
two win98. Samba shares a data directory with Word, Excel and Access
documents/db's. Also Exact for Dos is running from this share (executable
and
data).

The users experience a long delay every morning opening the first Word
document (Word 97, Word 2000). This could take 10 minutes or so. If they
close Windows before the document is loaded and open it later on the
document
is corrupted. If they go for a cup of coffee in the meantime everything is
fine. After the first document the rest of them reportedly has no problems
(allthough I doubt this).

Excel, Access and Explorer have no problems !

This is the global section of the smb.conf:
---------
[global]
        log file = /var/log/samba/%m.log
        load printers = yes
        idmap gid = 16777216-33554431
        socket options = TCP_NODELAY SO_SNDBUF=8192 SO_RCVBUF=8192
        wins server = 192.168.9.5
        interfaces = 192.168.9.0/255.255.255.0
        username map = /etc/samba/smbusers
        winbind use default domain = no
        template shell = /bin/false
        dns proxy = no
        cups options = raw
        netbios name = server02
        server string = Samba Server
        idmap uid = 16777216-33554431
        path = /tmp
        workgroup = WORKGROUP
        os level = 20
        printcap name = /etc/printcap
        security = user
        preferred master = no
        max log size = 50
--------
Samba is running as a workgroup server, no domain integration or fancy
stuff.
There is still an older Windows 2000 server running as Wins server which is
to
be replaced by the CentOS server.

I find these entries (below) in /var/log/messages which point to locking
problems. By Googling I found reports from people who think the stock samba
release RedHAt is broken with repect to locking and advice to upgrade to
latest version.

I would love to here your experiences and thoughts on this cause this gives
me
the creeps as you can imagine.

Thanks in advance

Paul
============
Mar 13 11:42:29 server02 smbd[10062]: [2007/03/13 11:42:29, 0]
smbd/oplock.c:oplock_break(839)
Mar 13 11:42:29 server02 smbd[10062]:   oplock_break: receive_smb timed out
after 30 seconds.
Mar 13 11:42:29 server02 smbd[10062]:   oplock_break failed for file
Apps/EXACT70/USR/BEDRYF (dev = fd01, inode = 541383, file_id = 74).
Mar 13 11:42:29 server02 smbd[10062]: [2007/03/13 11:42:29, 0]
smbd/oplock.c:oplock_break(911)
Mar 13 11:42:29 server02 smbd[10062]:   oplock_break: client failure in
oplock
break in file Apps/EXACT70/USR/BEDRYF
Mar 13 11:42:59 server02 smbd[10062]: [2007/03/13 11:42:59, 0]
smbd/oplock.c:oplock_break(839)
Mar 13 11:42:59 server02 smbd[10062]:   oplock_break: receive_smb timed out
after 30 seconds.
Mar 13 11:42:59 server02 smbd[10062]:   oplock_break failed for file
Apps/EXACT70/USR/BEDRYF.IX (dev = fd01, inode = 541384, file_id = 75).
Mar 13 11:42:59 server02 smbd[10062]: [2007/03/13 11:42:59, 0]
smbd/oplock.c:oplock_break(911)
Mar 13 11:42:59 server02 smbd[10062]:   oplock_break: client failure in
oplock
break in file Apps/EXACT70/USR/BEDRYF.IX
Mar 13 11:43:10 server02 smbd[9113]: [2007/03/13 11:43:10, 0]
smbd/oplock.c:request_oplock_break(1055)
Mar 13 11:43:10 server02 smbd[9113]:   request_oplock_break: no response
received to oplock break request to pid 10062 on port 33112 for dev =
 fd01, inode = 28264408, file_id = 68
Mar 13 11:43:10 server02 smbd[9113]: [2007/03/13 11:43:10, 0]
smbd/open.c:open_mode_check(734)
Mar 13 11:43:10 server02 smbd[9113]:   open_mode_check: exlusive oplock left
by process 10062 after break ! For file Apps/EXACT70/EMENU.INI, d
ev = fd01, inode = 28264408. Deleting it to continue...
Mar 13 11:43:10 server02 smbd[9113]: [2007/03/13 11:43:10, 0]
smbd/open.c:open_mode_check(738)
Mar 13 11:43:10 server02 smbd[9113]:   open_mode_check: Existent process
10062
left active oplock.
Mar 13 11:43:19 server02 smbd[9136]: [2007/03/13 11:43:19, 0]
smbd/oplock.c:request_oplock_break(1055)
Mar 13 11:43:19 server02 smbd[9136]:   request_oplock_break: no response
received to oplock break request to pid 10062 on port 33112 for dev =
 fd01, inode = 541384, file_id = 75
Mar 13 11:43:29 server02 smbd[10062]: [2007/03/13 11:43:29, 0]
smbd/oplock.c:oplock_break(839)
Mar 13 11:43:29 server02 smbd[10062]:   oplock_break: receive_smb timed out
after 30 seconds.
Mar 13 11:43:29 server02 smbd[10062]:   oplock_break failed for file
Apps/EXACT70/EMENU.INI (dev = fd01, inode = 28264408, file_id = 68).
Mar 13 11:43:29 server02 smbd[10062]: [2007/03/13 11:43:29, 0]
smbd/oplock.c:oplock_break(911)
Mar 13 11:43:29 server02 smbd[10062]:   oplock_break: client failure in
oplock
break in file Apps/EXACT70/EMENU.INI
Mar 13 11:43:51 server02 smbd[9136]: [2007/03/13 11:43:51, 0]
smbd/oplock.c:request_oplock_break(1055)
Mar 13 11:43:51 server02 smbd[9136]:   request_oplock_break: no response
received to oplock break request to pid 9134 on port 33103 for dev =
fd0
================
--
Paul Schoonderwoerd
Pollux IT - Open Source solutions & Netwerkbeveiliging
www.Pollux-IT.nl
Tel. 0294-283832
_______________________________________________
CentOS mailing list
CentOS at centos.org
http://lists.centos.org/mailman/listinfo/centos