[Ocfs2-users] ocfs2_extend_file:754 ERROR: status = -28

[Sunil Mushran]

-28 means ENOSPC. It gets printed when the requested space is larger
than the available space. So 25% full makes no difference. All it
takes is to request 1 byte more than the available space. I am more
puzzled by the fact that the error is not showing up in userspace
(via the app).

Having said that, we should not be printing -28 errors at all. It
needs to be handled by the userspace. We recently pushed a patch to
the mainline for the same.

If you are very curious and want to get to the bottom of this, do:

# debugfs.ocfs2 -l INODE allow
... do activity that throws the error...
# debugfs.ocfs2 -l INODE off

Scan the messages file for the -28 error. Above that you should
see a message saying "clusters_to_add = ...". That will tell us
how much the user wanted to extend the file. The number in paren
is the pid. Use ps to lookup the process.

Caution: This method will work best if you know the exact steps to
generate the -28 error. If not, then you may fill up the messages
file even more.

BTW, 2.6.19 is very old. Upgrade to 2.6.21 atleast.

Luke Schierer wrote:
> On Wed, Nov 28, 2007 at 02:13:39PM -0800, Sunil Mushran wrote:
>   
>> Is the only issue here that the error is being printed, or are you also
>> experiencing "out of space" errors?
>>     
>
> these are the only errors I am seeing.  As best I can tell, the system
> is working just fine, and I'm certainly able to create files on it.
>
>   
>> What's the kernel/ocfs2 version. That should tell us more.
>>     
>
> This is a debian lenny system with a 2.6.19.2 kernel, with ocfs2-tools
> 1.2.4-1.
>
> thanks,
> luke
>
>   
>> Luke Schierer wrote:
>>     
>>> In my logs, I am seeing lines such as 
>>> Nov 28 15:17:53 mail kernel: (3793,0):ocfs2_extend_file:754 ERROR:
>>> status = -28
>>>
>>> happening periodically, the number varies from hour to hour.  They
>>> happen only one one of 4 hosts (thre of which, including this one are
>>> xen guests) mounting the volume.
>>>
>>> Googling on the error seems to indicate that this error happens when
>>> the partition is full, but in this case the partition is only about
>>> 25% full. 
>>> Any help figuring out what is going on here and how I can fix it would
>>> be greatly appreciated.
>>>
>>> thanks!
>>>
>>> luke
>>>
>>>
>>> _______________________________________________
>>> Ocfs2-users mailing list
>>> Ocfs2-users at oss.oracle.com
>>> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>>>   
>>>       
>> _______________________________________________
>> Ocfs2-users mailing list
>> Ocfs2-users at oss.oracle.com
>> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>>
>>     
>
> _______________________________________________
> Ocfs2-users mailing list
> Ocfs2-users at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>