Ocfs2 users - Oct 2009 - A free space trouble in a ocfs2 partition

Hi,
i have a trouble with a ocfs2 partition.

The "df -kh" command reports 63 GB of free space but only can use
about
5 GB of this free space !!
Then, the ocfs2 partition is wasting 58 GB of free space !!?

df -kh
S.ficheros          Tama?o Usado  Disp Uso% Montado en
/dev/sda1             300G  238G   63G  80% /home/mail

This are the mount options of the partition:

mount
/dev/sda1 on /home/mail type ocfs2 (rw,_netdev,heartbeat=local)

The 300 GB ocfs2 partition is the mail storage of thousands of mail users.

The block size of the partition is 1K to theorically prevent  fragmentation:

tunefs.ocfs2 -q -Q "BS=%5B\n" /dev/sda1
BS= 1024

There are so many free inodes:

df -i
S.ficheros           Nodos-i NUsados NLibres NUso% Montado en
/dev/sda1            78642183 62356449 16285734   80% /home/mail


So, what is the problem with this partition?

Thanks in advanced

Regards,
Pablo

ocfs2 and kernel version?

My guess is that this is because the orphaned inodes are not
getting deleted. This problem has been fixed in the mainline kernel
(2.6.31).

For enterprise users, the fix was provided with OCFS2 1.4.4 for
(RH)EL. SLES10/11 users should upgrade to the latest kernel errata
to get the fix.

If you are using an older mainline kernel, a temporary fix would be
to umount the volume on all nodes and run "fsck.ocfs2 -f device".
Upgrading the kernel to 2.6.31 will also fix the problem.

Sunil

Juan Pablo Marco Cardona wrote:
> Hi,
> i have a trouble with a ocfs2 partition.
>
> The "df -kh" command reports 63 GB of free space but only can use
about
> 5 GB of this free space !!
> Then, the ocfs2 partition is wasting 58 GB of free space !!?
>
> df -kh
> S.ficheros          Tama?o Usado  Disp Uso% Montado en
> /dev/sda1             300G  238G   63G  80% /home/mail
>
> This are the mount options of the partition:
>
> mount
> /dev/sda1 on /home/mail type ocfs2 (rw,_netdev,heartbeat=local)
>
> The 300 GB ocfs2 partition is the mail storage of thousands of mail users.
>
> The block size of the partition is 1K to theorically prevent 
fragmentation:
>
> tunefs.ocfs2 -q -Q "BS=%5B\n" /dev/sda1
> BS= 1024
>
> There are so many free inodes:
>
> df -i
> S.ficheros           Nodos-i NUsados NLibres NUso% Montado en
> /dev/sda1            78642183 62356449 16285734   80% /home/mail
>
>
> So, what is the problem with this partition?
>
> Thanks in advanced
>
> Regards,
> Pablo
>
>
>  
>
>
> _______________________________________________
> Ocfs2-users mailing list
> Ocfs2-users at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-users

Hi Sunil,

this is the kernel version:

uname -a
Linux srvfcor001.cast-info.es 2.6.18-128.2.1.el5 #1 SMP Tue Jul 14 
06:36:37 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux

and the ocfs2 version:

ocfs2-tools-1.4.2-1.el5
ocfs2-2.6.18-128.2.1.el5-1.4.2-1.el5
ocfs2console-1.4.2-1.el5

modinfo ocfs2 | grep ^version:
version:        1.4.2

We have to upgrade to ocfs2 1.4.4 driver to fix this trouble, right?
We do have to to any aditional steps to fix the trouble? Or with the 
ocfs2 upgrade is just enough ?

Thank you!

Regards,
Pablo






Sunil Mushran escribi?:
> ocfs2 and kernel version?
>
> My guess is that this is because the orphaned inodes are not
> getting deleted. This problem has been fixed in the mainline kernel
> (2.6.31).
>
> For enterprise users, the fix was provided with OCFS2 1.4.4 for
> (RH)EL. SLES10/11 users should upgrade to the latest kernel errata
> to get the fix.
>
> If you are using an older mainline kernel, a temporary fix would be
> to umount the volume on all nodes and run "fsck.ocfs2 -f device".
> Upgrading the kernel to 2.6.31 will also fix the problem.
>
> Sunil
>
> Juan Pablo Marco Cardona wrote:
>> Hi,
>> i have a trouble with a ocfs2 partition.
>>
>> The "df -kh" command reports 63 GB of free space but only can
use
>> about 5 GB of this free space !!
>> Then, the ocfs2 partition is wasting 58 GB of free space !!?
>>
>> df -kh
>> S.ficheros          Tama?o Usado  Disp Uso% Montado en
>> /dev/sda1             300G  238G   63G  80% /home/mail
>>
>> This are the mount options of the partition:
>>
>> mount
>> /dev/sda1 on /home/mail type ocfs2 (rw,_netdev,heartbeat=local)
>>
>> The 300 GB ocfs2 partition is the mail storage of thousands of mail 
>> users.
>>
>> The block size of the partition is 1K to theorically prevent  
>> fragmentation:
>>
>> tunefs.ocfs2 -q -Q "BS=%5B\n" /dev/sda1
>> BS= 1024
>>
>> There are so many free inodes:
>>
>> df -i
>> S.ficheros           Nodos-i NUsados NLibres NUso% Montado en
>> /dev/sda1            78642183 62356449 16285734   80% /home/mail
>>
>>
>> So, what is the problem with this partition?
>>
>> Thanks in advanced
>>
>> Regards,
>> Pablo
>>
>>
>>  
>>
>>
>> _______________________________________________
>> Ocfs2-users mailing list
>> Ocfs2-users at oss.oracle.com
>> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>
-- 
logo Cast-info
Juan Pablo Marco Cardona
Administrador de sistemas y redes/Departamento de Sistemas
jmarco.cardona at cast-info.es
M. 608.016.346
08006 Barcelona. c/Tuset 23, 1?. T. 93 445 68 00. F. 93 414 51 31. 
www.cast-info.es

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://oss.oracle.com/pipermail/ocfs2-users/attachments/20091027/ac35e18e/attachment.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 15a_castinfo.gif
Type: image/gif
Size: 2437 bytes
Desc: not available
Url :
http://oss.oracle.com/pipermail/ocfs2-users/attachments/20091027/ac35e18e/attachment.gif

You only need to upgrade. A rolling upgrade will work.
http://oss.oracle.com/pipermail/ocfs2-announce/2009-September/000030.html

Juan Pablo Marco Cardona wrote:
> Hi Sunil,
>
> this is the kernel version:
>
> uname -a
> Linux srvfcor001.cast-info.es 2.6.18-128.2.1.el5 #1 SMP Tue Jul 14 
> 06:36:37 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux
>
> and the ocfs2 version:
>
> ocfs2-tools-1.4.2-1.el5
> ocfs2-2.6.18-128.2.1.el5-1.4.2-1.el5
> ocfs2console-1.4.2-1.el5
>
> modinfo ocfs2 | grep ^version:
> version:        1.4.2
>
> We have to upgrade to ocfs2 1.4.4 driver to fix this trouble, right?
> We do have to to any aditional steps to fix the trouble? Or with the 
> ocfs2 upgrade is just enough ?
>
> Thank you!
>
> Regards,
> Pablo
>
>
>
>
>
>
> Sunil Mushran escribi?:
>> ocfs2 and kernel version?
>>
>> My guess is that this is because the orphaned inodes are not
>> getting deleted. This problem has been fixed in the mainline kernel
>> (2.6.31).
>>
>> For enterprise users, the fix was provided with OCFS2 1.4.4 for
>> (RH)EL. SLES10/11 users should upgrade to the latest kernel errata
>> to get the fix.
>>
>> If you are using an older mainline kernel, a temporary fix would be
>> to umount the volume on all nodes and run "fsck.ocfs2 -f
device".
>> Upgrading the kernel to 2.6.31 will also fix the problem.
>>
>> Sunil
>>
>> Juan Pablo Marco Cardona wrote:
>>> Hi,
>>> i have a trouble with a ocfs2 partition.
>>>
>>> The "df -kh" command reports 63 GB of free space but only
can use
>>> about 5 GB of this free space !!
>>> Then, the ocfs2 partition is wasting 58 GB of free space !!?
>>>
>>> df -kh
>>> S.ficheros          Tama?o Usado  Disp Uso% Montado en
>>> /dev/sda1             300G  238G   63G  80% /home/mail
>>>
>>> This are the mount options of the partition:
>>>
>>> mount
>>> /dev/sda1 on /home/mail type ocfs2 (rw,_netdev,heartbeat=local)
>>>
>>> The 300 GB ocfs2 partition is the mail storage of thousands of mail
>>> users.
>>>
>>> The block size of the partition is 1K to theorically prevent  
>>> fragmentation:
>>>
>>> tunefs.ocfs2 -q -Q "BS=%5B\n" /dev/sda1
>>> BS= 1024
>>>
>>> There are so many free inodes:
>>>
>>> df -i
>>> S.ficheros           Nodos-i NUsados NLibres NUso% Montado en
>>> /dev/sda1            78642183 62356449 16285734   80% /home/mail
>>>
>>>
>>> So, what is the problem with this partition?
>>>
>>> Thanks in advanced
>>>
>>> Regards,
>>> Pablo