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. Redhat system is
fully patched up to date and so is the gcc we are compiling samba with.
See entries from log.smb below.
Andrew
[2006/08/18 08:04:37, 0] lib/util.c:smb_panic(1592)
PANIC (pid 7536): release_posix_lock: unable find entry to delete !
[2006/08/18 08:04:37, 0] lib/util.c:log_stack_trace(1699)
BACKTRACE: 14 stack frames:
#0 /opt/samba/sbin/smbd(log_stack_trace+0x1a) [0x552acab33c]
#1 /opt/samba/sbin/smbd(smb_panic+0x69) [0x552acab1d1]
#2 /opt/samba/sbin/smbd(release_posix_lock+0x11a) [0x552ac64e11]
#3 /opt/samba/sbin/smbd [0x552ac60f86]
#4 /opt/samba/sbin/smbd(brl_unlock+0x46) [0x552ac61a32]
#5 /opt/samba/sbin/smbd(do_unlock+0x164) [0x552ac5e0b1]
#6 /opt/samba/sbin/smbd(reply_lockingX+0x57c) [0x552ab4431f]
#7 /opt/samba/sbin/smbd [0x552ab72a91]
#8 /opt/samba/sbin/smbd [0x552ab72b51]
#9 /opt/samba/sbin/smbd [0x552ab72eb4]
#10 /opt/samba/sbin/smbd(smbd_process+0x140) [0x552ab73caa]
#11 /opt/samba/sbin/smbd(main+0x894) [0x552ad4045a]
#12 /lib64/tls/libc.so.6(__libc_start_main+0xdb) [0x2a96b883fb]
#13 /opt/samba/sbin/smbd [0x552ab078fa]
[2006/08/18 08:04:37, 0] lib/fault.c:dump_core(168)
unable to change to / to dump core
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

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. 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.

Hello Jeremy,

Thanks for the response. I forgot to add in my original posting that we
did not see this problem with Samba3.0.22 on the same system with the
same smb.conf file.

I notice that in the smb.log file, these errors always occur in pairs -
having exactly the same time, date and content. There is usually a few
minutes between their occurance - sometimes as much as half an hour. I
know this probably isn't much help, but I notice that the errors stop
outside working hours - suggesting (to me) that they are occuring in
response to our users initiating some file/print action on their PC's.
This at least gives me a hint to investigate how to replicate the problem.

Unfortunately I am out of the office for the next week - I will get my
colleauges to gather some more information on the problem and try to
find a way to replicate the errors. Will post again when I have any
further information.

Many thanks for your help

Andrew

Jeremy Allison wrote:

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. 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

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:

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. 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

Tue, Aug 29, 2006 at 09:41:15PM +0100, Andrew Wilson wrote:
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 ?

No it shouldn't - it's a bug. But I think it's something I've
fixed in current SAMBA_3_0 svn. I'll look at the SAMBA_3_0_23
svn (still using the older locking code, about to be shipped
as 3.0.23c).

Thanks for the info !

Jeremy.

Tue, Aug 29, 2006 at 01:47:52PM -0700, Jeremy Allison wrote:

No it shouldn't - it's a bug. But I think it's something I've
fixed in current SAMBA_3_0 svn. I'll look at the SAMBA_3_0_23
svn (still using the older locking code, about to be shipped
as 3.0.23c).

, took a quick look at the 3.0.23c code and I don't see how
that condition can happen, so I think we've fixed it.

Jeremy.