[Samba] release_posix_lock: unable to find entry to delete with Samba3.0.23b on Redhat ES4

[Andrew Wilson]

Jeremy,

I have located the source of this error in our config' - we had "posix 
locking=no" in force on the share we are using to host the NT/XP 
profiles (See the definition below). When we remove this (to allow the 
default posix locking) the errors go away - they were occuring during 
log on and log off from Windows XP clients. I probably should have 
spotted this one - so sorry about that. However I do remember having to 
put this setting in place with a much older version of Samba, running 
under Solaris, in order to get the profiles to work reliably. In any 
case I am not sure whether you think this setting should be causing a 
panic in smbd under these circumstances ?

Best Regards

Andrew

[profile]
  comment = My Profile Data
  path = %H/.ntprofile
  profile acls = yes
  read only = no
  browseable = no
  create mask = 0600
  directory mask = 0700
  oplocks = no
  level2 oplocks = no
  posix locking = no


Jeremy Allison wrote:

>On Fri, Aug 18, 2006 at 08:19:16AM +0100, Andrew Wilson wrote:
>  
>
>>We are running Samba3.0.23b on Redhat ES4 and find that we are getting 
>>quite a few of the release_posix_lock errors appearing in the SAMBA log 
>>(See base of this note). There is no obvious impact on the operation of 
>>the system, but I was wondering if anyone else is experiencing the same 
>>and whether we need to worry about this. We have the default "posix 
>>locking=yes" in operation and I note in the man entry that we "should 
>>never need to disable this parameter". However if we do turn it off, the 
>>errors appear to go away. I suspect though that this means that there is 
>>not full lock integrity on our files?
>>
>>We are using exactly the same settings with Samba3.0.23b on both Solaris 
>>8 and Solaris 10 and we do not see this problem. Our Redhat system is 
>>fully patched up to date and so is the gcc we are compiling samba with.
>>    
>>
>
>Strange. You're hitting a deliberate panic due to what smbd believes
>is a logic error in the lock counts. Can you find out more about how
>you reproduce this problem ? Is there anything specific you're doing
>when it happens ?
>
>Jeremy.
>
>________________________________________________________________________
>This e-mail has been scanned for all viruses by Star. The
>service is powered by MessageLabs. For more information on a proactive
>anti-virus service working around the clock, around the globe, visit:
>http://www.star.net.uk
>________________________________________________________________________
>  
>



________________________________________________________________________
This e-mail has been scanned for all viruses by Star. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________